Prelims.qxd
4/16/2007
10:34 AM
Page i
PSYCHOLOGICAL
CLINICAL SCIENCE
Prelims.qxd
4/16/2007
10:34 AM
Page ii
MODERN PIONEERS IN PSYCHOLOGICAL SCIENCE: AN APS-PSYCHOLOGY PRESS SERIES
This series celebrates the careers and contributions of a generation of pioneers in psychological science. Based on the proceedings of day-long festschrift events at the annual meeting of the Association for Psychological Science, each volume commemorates the research and life of an exceptionally influential scientist. These books document the professional and personal milestones that have shaped the frontiers of progress across a variety of areas, from theoretical discoveries to innovative applications and from experimental psychology to clinical research. The unifying element among the individuals and books in this series is a commitment to science as the key to understanding and improving the human condition.
Prelims.qxd
4/16/2007
10:34 AM
Page iii
PSYCHOLOGICAL
CLINICAL SCIENCE Papers in Honor of Richard M. McFall
Edited by
Teresa A. Treat s Richard R. Bootzin s Timothy B. Baker
aps ASSOCIATION FOR
PSYCHOLOGICAL SCIENCE
Prelims.qxd
4/16/2007
10:34 AM
Page iv
Psychology Press Taylor & Francis Group 270 Madison Avenue New York, NY 10016
Psychology Press Taylor & Francis Group 27 Church Road Hove, East Sussex BN3 2FA
© 2007 by Taylor & Francis Group, LLC Printed in the United States of America on acid-free paper 10 9 8 7 6 5 4 3 2 1 International Standard Book Number-13: 978-0-8058-5561-6 (Hardcover) No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the Psychology Press Web site at http://www.psypress.com
Prelims.qxd
4/16/2007
10:34 AM
Page v
Contents McFall Photo
ix
Foreword
xi
• • • • •
Alan G. Kraut Bruce N. Cuthbert Thomas R. Insel Donald C. Fowles Joseph E. Steinmetz
Preface
• Teresa A. Treat, Richard R. Bootzin, Timothy B. Baker
Acknowledgements
I History and Epistemology of Psychological Clinical Science
xix xxvii
1
1 Psychological Clinical Science: Why and How We Got to Where We Are
3
Richard R. Bootzin
2 The Epistemological and Ethical Dimension of Clinical Science
29
William O’Donohue and Scott O. Lilienfeld
3 The Seduction of Clinical Science in Psychology: Challenges in Psychological and Biological Convergence
53
Gregory A. Miller, Anna S. Engels, and John D. Herrington
II Clinical Science: Topics of Applied Significance
75
4 Clinical Science and the Revolution in Psychological Treatment: The Example of Anxiety Disorders
77
Gabrielle I. Liverant, Jill A. Stoddard, Alicia E. Meuret, and David H. Barlow
v
Prelims.qxd
4/16/2007
10:34 AM
Page vi
vi
CONTENTS
5 Manual-Based Treatment: Evolution and Evaluation
105
G. Terence Wilson
6 The Importance of How: A Call for Mechanistic Research in Tobacco Dependence Treatment Studies
133
Danielle E. McCarthy, Daniel Bolt, and Timothy B. Baker
III Model Clinical Science Research Programs
165
7 Fear, Anxiety, Depression, and the Anxiety Disorder Spectrum: A Psychophysiological Analysis
167
Peter J. Lang, Lisa M. McTeague, and Bruce N. Cuthbert
8 Behavior Therapy for Specific Fears and Phobias: Context Specificity of Fear Extinction
197
Jayson L. Mystkowski and Susan Mineka
9 Assessment of Mental Architecture in Clinical/Cognitive Research
223
James T. Townsend, Mario Fific′, and Richard W. J. Neufeld
10 Using a Simple Associative Learning Procedure to Study Clinical Disorders and Related Brain Function
259
Joseph E. Steinmetz
11 Integrating Clinical and Cognitive Science
289
Teresa A. Treat and Melanie A. Dirks
IV Future Directions for Research, Application, and Training
319
12 Translational Research and the Future of Psychological Clinical Science
321
Bruce N. Cuthbert
13 The Future of the Clinical Science Movement: Challenges, Issues, and Opportunities
349
Robert W. Levenson
V Perspectives on Psychological Clinical Science 14 On Psychological Clinical Science Richard M. McFall
361 363
Prelims.qxd
4/16/2007
10:34 AM
Page vii
CONTENTS
vii
Bibliography of Richard M. McFall
397
About the Editors
405
Author Index
407
Subject Index
427
DVD Interview with Richard McFall
Prelims.qxd
4/16/2007
10:34 AM
Page viii
Prelims.qxd
4/16/2007
10:34 AM
Page ix
Prelims.qxd
4/16/2007
10:34 AM
Page x
Prelims.qxd
4/16/2007
10:34 AM
Page xi
Foreword Foreword: Opening Comments at Festschrift Celebration for Richard M. McFall COMMENTS BY ALAN G. KRAUT Executive Director, Association for Psychological Science (previously American Psychological Society) The May 2004 annual meeting (our 16th!) of the Association for Psychological Science (APS, previously the American Psychological Society), held in Chicago, was the single best-attended APS convention ever. The program included hundreds of research presentations on topics that ran the gamut of psychological subdisciplines, from the most molecular of research in behavioral neuroscience and genetics to the most molar of research in group dynamics, educational policy, and organizational behavior. As the saying goes, a great time was had by all. But a large part of the reason the convention was so well attended and such a resounding success was the event that gave rise to this book: a full day of presentations in honor of Richard M. McFall. The McFall event is part of establishing the APS Convention as the place to honor distinguished psychologists at important points in their careers—distinguished psychologists who have had enormous impact in their own subdisciplines and on the larger field, distinguished psychologists exactly like Dick McFall. To broaden the impact of this important event, we are pleased to join with Psychology Press (PP) to publish the McFall proceedings. The result is this festschrift. Plans are to make this volume the first in an annual festschrift series that PP and APS both believe will become an essential resource for students of psychology.
xi
Prelims.qxd
4/16/2007
10:34 AM
Page xii
xii
FOREWORD
So my main reason in writing these opening remarks is to formally initiate this festschrift series on behalf of the Association for Psychological Science. But I also have a more personal reason for writing this foreword. I have known Dick McFall since my early days in psychology—the early 1980s. That is, from his talks and from his articles, I knew who Richard McFall was long before he knew me. I have always considered him one of my heroes in psychology. You see, I saw in Dick someone who would stand up time and time again and in the face of fads that come and go in psychology, and in the face of an almost machine-like political correctness mantra that has taken over whole areas of our discipline, someone who has the courage to say the things that the rest of us were thinking and would like to have said, the things that we all know are absolutely right. Dick’s message was and is that all psychology, including the practice of clinical psychology, must—must—be based on science. He gives that message without qualification, without apology, and without any concern for those who might be offended by its strength. It was that kind of clear and powerful vision that inspired many of us gathered here today, including me, to go on and try to make our own contribution to the science of psychology. It is the same message that is the foundation of the Association for Psychological Science. So, I want to express my personal gratitude to Dick for being the model of a psychological scientist. Let me also congratulate the organizers of the McFall Festschrift—Teresa Treat, Yale University; Timothy Baker, University of Wisconsin-Madison; Richard Bootzin, University of Arizona; and Robert Levenson, University of California–Berkeley—for assembling an exceptional roster of presenters whose scientific achievements represent the very best of our field. Nothing less would do. COMMENTS BY BRUCE N. CUTHBERT Professor of Psychology, University of Minnesota It has been my good fortune to know Richard McFall in multiple roles for well over three decades—from that of a first-year graduate student to that of a program official at the National Institute of Mental Health (NIMH). It is thus highly gratifying to have this opportunity to offer congratulations to Dick on this salutary occasion, which fittingly reflects the inaugural volume in the festschrift series to be published jointly by the xii
Prelims.qxd
4/16/2007
FOREWORD
10:34 AM
Page xiii
xiii
Association for Psychological Science (previously the American Psychological Society) and Psychology Press. My experiences with Dick over the years have reflected the ways in which he has impacted our field. He taught the first-year graduate assessment course at Wisconsin in the late 1960s with an approach that fit the revolutionary spirit of the times. Rather than presenting a humdrum compendium of assessment techniques, he used the course as an opportunity to open our eyes to a whole new way of considering the nature of personality and psychopathology, to thinking critically about the phenomena and constructs that we were trying to assess, and to evaluating the problems with the traditional instruments that were in current use to diagnose and measure psychopathology. The climactic assignment of this course was to develop a syllabus, along with two or three complete lectures, for a first-semester course entitled “An Introduction to Psychological Clinical Science.” I have carted that now-musty old project along with me ever since—not for nostalgia’s sake, but because it still serves as a touchstone for a comprehensive view of clinical science in psychology. Many years later, Dick and I have had several opportunities to talk at length about developing the science of clinical psychology. My role during these opportunities was as a branch chief for adult psychopathology at NIMH, tasked with leading a work group to develop programs for translational research in behavioral science at the Institute and also serving as a liaison to the Academy of Psychological Clinical Science. It was fascinating to hear Dick’s firsthand description of the way that the department at Indiana had successfully evolved into an entity that largely transcended traditional area-group boundaries to function as a collaborating and cooperative whole. It was especially useful as we at NIMH were trying to develop new models for research and training initiatives, to exchange ideas about how to develop collaborative relationships, and to approach the vital task of training students to function in new academic and health services environments. Many of Dick’s ideas found their way into the research and training announcements that were subsequently published; equally important, however, the energy and vision that he communicated were infectious in promoting enthusiasm for these approaches among NIMH program staff. Among all the well-deserved plaudits for these incalculable contributions to the field, it needs to be said that Dick McFall is as warm, thoughtful, and kind as anyone you would ever hope to meet. I have often wondered, reading over some of the polemical responses to his Manifesto and related works, whether the commentaries would not read
Prelims.qxd
4/16/2007
10:34 AM
Page xiv
xiv
FOREWORD
quite differently if the writers had the good fortune to know Dick personally. In my experience, his willingness to speak out frankly on issues is accompanied by a sincerity and enthusiasm that maintains a clear focus on the issues, and I have long admired his deft touch for the right phrase at the right time. For any young scientist struggling to develop a career in the hurly-burly world of academia, he is a role model in this domain fully as much as for his scientific work. Best wishes on this occasion, Dick, and boundless thanks for all of your contributions to so many individuals and to the entire field. Because it is clear that your retirement is largely nominal, we look forward to your wisdom and your thoughts for years to come. LETTER BY THOMAS R. INSEL Director, National Institute of Mental Health It is a pleasure to extend greetings to Dr. Richard McFall on this happy occasion. Dr. McFall has long been recognized as one of the country's preeminent leaders in developing the science of clinical psychology. Accordingly, it is highly appropriate that his career and accomplishments be celebrated at this festschrift. This occasion comes at a time when science is being challenged to produce significant and enduring outcomes that will improve the health of the nation. For those of us at the National Institute of Mental Health (NIMH), this means ensuring that the exciting breakthroughs in basic science inform the development of superior approaches to the etiology, assessment, and treatment of mental disorders. In a time of constrained budgets, our priorities are driven by the relevance, innovation, and traction these new approaches yield. We must also make certain that as new and effective treatments are developed, they are made widely available to the public. This is essential if we are to accomplish our public health mission. It is a measure of Dick McFall's vision that his career has been directed squarely toward these two areas. First, he has provided stellar leadership in his efforts to ensure that clinical science is informed by the best of basic psychological science. Largely through his initiative, the psychology department at Indiana University is nationally recognized for its unique program in clinical science, in which clinical graduate students develop a full research competency in a basic area of psychological science. Second, Dr. McFall has long been an outspoken champion of the need for evidence-based therapies as the basis for professional practice in
Prelims.qxd
4/16/2007
10:34 AM
Page xv
xv
FOREWORD
clinical psychology. I was fortunate to hear him speak at a recent conference on graduate training that we co-sponsored with the Academy of Psychological Clinical Science, and his thoughtfulness and passion on the topic were palpable. His leadership in the development and dissemination of evidence-based treatments represents a desirable, and indeed ineluctable, trend for the long term. Dr. Richard McFall's accomplishments are clearly evident in the body of his published work, in the work of his students and colleagues, and in the influence that he has had on countless others in the science and practice of psychology and mental disorders. This festschrift is therefore a most appropriate way to mark such a long and distinguished career. Dick, please accept heartfelt congratulations from me and from all of us here at NIMH on this joyous occasion. COMMENTS BY DONALD C. FOWLES Professor of Psychology, University of Iowa As president of the Academy of Psychological Clinical Science (APCS) at the time of this festschrift in honor of Richard M. McFall, I want to take this opportunity to describe and comment on the debt all of the members of APCS owe to Dick. APCS consists of highly research-oriented graduate programs and internships in clinical psychology or, as we now say following Dick’s lead, in psychological clinical science. The need for APCS arises from our unusual interest in training clinical researchers, in contrast to the larger number of programs and internships concerned with training practitioners and the vastly larger numbers of practitioners whose sole concern is delivering clinical services. For various reasons, our goals, our scientific values, and the nature of our training and career pursuits do not mesh well with the dominant practitioner guild forces at the national level. Each program/internship acting alone had little chance to promote our values and training models and to defend them from forces that at times are antithetical. Dick had the vision to imagine an organization such as APCS, but that was the easy part. It was his ability to see how to get there that was truly ingenious. Initially he consulted many colleagues about which graduate programs and which members of those programs embodied a clinical science orientation. He obtained financial support from Indiana University, the Association for Psychological Science (APS, previously the American Psychological Society), and NIMH to hold a conference entitled “Clinical
Prelims.qxd
4/16/2007
xvi
10:34 AM
Page xvi
FOREWORD
Science in the 21st Century” on April 21–24, 1994. He did not, however, contact the programs per se and ask them to send a representative to the conference. Rather, Dick selected one individual from each program, inviting 35. About 25 faculty members attended the conference, along with representatives from APS (Alan Kraut) and NIMH (Jane Steinberg). After 2 days of discussion, the attendees voted to create APCS and elected a six-person Coalition Steering Committee, with Dick as the chair to develop the procedures. The steering committee met several times (with invaluable financial support from APS) during the summer and fall of 1994 and developed criteria for reviewing program applicants for membership in APCS. At this point, there was a fundamental shift from the individuals who were invited to the Indiana Conference to applications from the programs of which they were members. Only the 35 programs associated with the invitations to Indiana were eligible to apply for membership during this bootstraps phase (once APCS was created, any program could apply). Twenty-six programs applied and were admitted. The first meeting of APCS was held on July 1–2, 1995, in connection with the APS meeting in New York City. Dick was elected the first president. There have been many developments since that first meeting. Applications from additional graduate programs followed quickly. In an important development, the first internships applied for membership in 1998. At the time of this writing, we have 44 graduate programs and nine internships as members. The Committee on Accreditation recognizes the clinical science model and attempts to provide site visitors from other academy programs. APCS is routinely invited to participate in national conferences on clinical psychology. Each year, someone from the academy is in charge of the clinical track of the APS program. We have had increasing interactions with NIMH regarding training in clinical science, including a joint NIMH/APCS conference in January 2004. In short, in 10 years Dick took a list of 25 individual clinical scientists and created an organization that is nationally recognized, still growing, and providing national leadership in clinical science training. Throughout these 10 years, Dick has provided leadership for the academy. In addition to organizing the 1994 Indiana Conference, chairing the Steering Committee, and serving as president for the first 3 years, Dick continues to be the person with the clearest vision of who we are and where we should go. As Alan Kraut says in his comments, Dick’s message is that the practice of clinical psychology must be based on science. He
Prelims.qxd
4/16/2007
10:34 AM
Page xvii
xvii
FOREWORD
advocates this basic premise in many contexts—for example, for empirically supported treatments in his famous Manifesto, for the training of clinical graduate students in basic psychological science in his conference on integrative psychological science in April 2002, and for criteria for accreditation in the Alternative Accreditation Steering Committee and other contexts. Regardless of the topic, when we discuss issues in our meetings, or when we represent the academy at other meetings, it turns out that Dick is way ahead of the rest of us: He has the best and most articulate proposals and positions. To say that the academy would not exist without Dick is true, but it is insufficient. We also would not be the kind of academy we are without Dick’s continued leadership and influence. On behalf of all members of APCS, I want to express our appreciation and gratitude for Dick’s inspired leadership and contributions. COMMENTS BY JOSEPH E. STEINMETZ Professor and Chair of Psychology, Indiana University As Chair of Psychology at Indiana University (IU), I am sure I represent all of the IU faculty and students in welcoming all of Dick McFall’s friends and colleagues to this festschrift in his honor. I would like to take this opportunity to talk about two aspects of Dick’s outstanding career that have impacted me personally—his contributions to the field of clinical science in the psychology department at Indiana University, Bloomington, and his contributions to my own academic and research career. Dick has served more than one term as the director of clinical training for IU psychology, a position that has gained him national recognition (and sometimes notoriety) for his unwavering commitment to the advancement of clinical science. Dick’s views on clinical training and the roles of clinical scientists in our field need no elaboration by me; they are a matter of record as exemplified by his famous clinical Manifesto. But Dick practices what he preaches. Our clinical program at IU is a model for translational training and research, and the high level of collaboration between clinical scientists in the department and researchers from areas such as health psychology, neuroscience, and cognitive science provides evidence that this model works well in a department as diverse as ours. Indiana University’s clinical science students are exposed to interdisciplinary, multilevel research strategies and truly appreciate the importance of research that delves into the mechanisms of psychological disorders as well as research that strives to establish empirical support for treatments and therapies.
Prelims.qxd
4/16/2007
xviii
10:34 AM
Page xviii
FOREWORD
Dick’s influence has positively affected other areas of our department. I am proud of the high level of cooperation and collaboration that characterizes the department of psychology at IU. We are a department that is not shy about pushing the envelope of psychological science, as evidenced by recent faculty hires we have made in areas like robotics, computational science, molecular neuroscience, and human brain imaging. This spirit of expanding the horizons of our discipline can be traced to faculty members like Dick McFall, who shunned “area-ism” in their respective fields in favor of a model of cooperating scientists whose interests and efforts cut across disciplinary boundaries. Dick strongly advocated for these developments as an influential senior member of the department. I will always be grateful for this because it made my 10 years as department chair enjoyable, as this is the vision for the department I shared and promoted. Dick McFall has also had a significant impact on my own research and scholarship. For the first 20 years of my career, I concentrated on issues related to the neurobiology of learning and memory, using basic neuroscience techniques and animal models to study how the brain encodes simple associative learning. Dick introduced me to the clinical literature and, through many hours of discussion, helped me to develop an appreciation of how clinical populations could be integrated effectively into my research program. The result has been a series of parallel human and nonhuman studies designed to advance our understanding of clinical science as well as basic science in areas like learning, memory, attention, perception, and information processing. Much of this work has been done in collaboration with Dick. Although I have no formal training in clinical psychology, I now think of myself as a clinical scientist. I suspect that was Dick’s end game. For this I am grateful. On behalf of my colleagues at IU, I congratulate Dick on his career achievements—they are enormous. Although he has nominally retired, I look forward to many more years of interacting with Dick in his roles as an advocate for clinical science, research collaborator, and, most important, friend.
Prelims.qxd
4/16/2007
10:34 AM
Page xix
Preface Richard McFall’s scientific contributions to clinical science—and to psychological science more generally—are widely recognized and appreciated. From the outset of his career, Richard McFall has conducted influential and groundbreaking research and produced citation classics that have shaped the field. His early work examined the effects of selfmonitoring on behavior (e.g., 1970, 1972, 1976, 1977); specified and evaluated the role of rehearsal, modeling, and coaching in social-skills training (e.g., 1971, 1973, 1975); and extended Goldfried and D’Zurilla’s (1969) behavior-analytic model to investigate the link between situation-specific social competence and a wide variety of forms of psychopathology (e.g., 1975, 1978, 1981, 1982). McFall’s abiding interest in clinical assessment and the critical link between theory and assessment also was evident early in his career (e.g., 1977, 1986, 1993, 1998, and 2005). In 1982 (see also 1990), McFall published one of the field’s first social information-processing (SIP) models, in which he distinguished between social skills (cognitive processes such as decoding and decision making) and social competence (behavioral output of the SIP system that is judged for its adequacy), while articulating the role of these processes in clinical phenomena. McFall’s model and extensions have subsequently been applied in almost every area of psychopathology. In the latter half of his career, McFall became an articulate and visionary spokesperson for integrative psychological science (IPS), an approach to the conduct and application of psychological research that draws on the best available theoretical, measurement, and analytical models across areas of psychology and other relevant fields (e.g., 1991, 1995, 1998, 2000, 2006, in press). McFall has advanced IPS not only ideologically, but also by producing pioneering research that has brought the best of contemporary cognitive science and neuroscience to bear on important clinical questions (e.g., 1998, 1999, 2001, 2002, 2006, in press). McFall also envisioned and then instantiated the first and only National Institute of Mental Health (NIMH)-funded IPS training model at Indiana University, in which clinical and nonclinical students have been trained to become legitimate scholars in, and make significant contributions to, clinical science and at least one other allied discipline within the field, such as neural, cognitive, social, or developmental science (see 2006, as well as his chapter in the current volume, for details). This integrative model extends xix
Prelims.qxd
4/16/2007
10:34 AM
Page xx
xx
PREFACE
the National Institutes of Health (NIH) translational model significantly by advocating for the integration of basic and applied expertise within a single individual. Apart from his innovative and influential scientific contributions, McFall also has made an indelible mark on the professional aspects of clinical psychology. Every student in the field now reads his “Manifesto for a Science of Clinical Psychology,” his 1991 presidential address to the Society for a Science of Clinical Psychology (see also 1996 and 2000 for follow-up papers). McFall’s innumerable papers on important professional issues, such as prescription privileges, mental-health care, and training models, also have been quite influential and still are widely read (e.g., 1995, 1998, 2002, in press). In 1994, McFall convened an Association for Psychological Science (APS, previously American Psychological Society)and NIMH-sponsored conference entitled “Clinical Science in the 21st Century.” The Academy of Psychological Clinical Science (APCS), now a thriving organization of 53 leading doctoral and internship training programs, was founded at this meeting. McFall also is the chief architect of the clinical science model of training in clinical psychology (e.g., 2006, as well as his chapter in the current volume), and he remains the model’s most visible spokesperson at the national and international levels. McFall has devoted his entire career to keeping the science in clinical science, and he has done so consistently in both word and deed. THE CURRENT VOLUME It is fitting to honor Richard McFall in the inaugural volume in a new series of festschrifts, published jointly by Psychology Press (PP) and the Association for Psychological Science (APS), that honor the careers and contributions of distinguished psychological scientists. A Festschrift Celebration for McFall was held on May 29–30, 2004, in conjunction with the 2004 APS annual convention in Chicago. A steering committee composed of Teresa Treat, Richard Bootzin, Tim Baker, Bob Levenson, and Alan Kraut planned the festschrift, which was sponsored by APS, LEA, APCS, and the Department of Psychology at Indiana University. Presenters at the festschrift event who contributed to the current volume are Timothy B. Baker, Richard R. Bootzin, Bruce N. Cuthbert, Peter J. Lang, Robert W. Levenson, Gregory A. Miller, Susan Mineka, Richard W. J. Neufeld, Joseph E. Steinmetz, James T. Townsend, Teresa
Prelims.qxd
4/16/2007
PREFACE
10:34 AM
Page xxi
xxi
A. Treat, and G. Terence Wilson. David H. Barlow, Scott O. Lilienfeld, and William O’Donohue were unable to attend the Chicago event, but also generously contributed chapters. In keeping with McFall’s integrative approach to psychological science, contributors include prominent clinical scientists, cognitive scientists, and neuroscientists. Richard R. Bootzin opens Part I—on the historical and epistemological underpinnings of psychological clinical science—with an overview of the history of clinical science since the late 1800s. Throughout his chapter, Bootzin (chap. 1) highlights the increasing integration of clinical science with the rest of psychological science, the ever-present tension between the generation and application of science, and the background for contemporary discussions about the accreditation of clinical programs, differing models of clinical training, and how best to disseminate evidence-based practices to community settings. William O’Donohue and Scott O. Lilienfeld (chap. 2) next use McFall’s Manifesto as a launching pad for an in-depth consideration of the historically troubled relationship between science and practice. O’Donohue and Lilienfeld sketch the epistemological and ethical obligations to which they believe clinical scientists’ research and practice should be subject, including the necessity of incorporating quality-improvement processes into contemporary systems for clinical services delivery. Gregory A. Miller, Anna S. Engels, and John D. Herrington (chap. 3) then raise thought-provoking questions about the proper relationship between psychological and biological characterizations of phenomena of interest to psychological scientists. Miller and colleagues contend that many biologically based explanations for clinical phenomena are dangerously reductionistic and ignore the importance of developing parallel models of psychological mechanisms and, ultimately, integrating the biological and psychological models for a full understanding of the phenomena. Gabrielle I. Liverant, Jill A. Stoddard, Alicia E. Meuret, and David H. Barlow (chap. 4) authored the lead chapter in Part II of the volume, which addresses topics of applied significance in clinical science. Liverant and colleagues illustrate the impact of the clinical science movement on treatment research by tracing the development of evidence-based treatments for anxiety disorders. They describe the learning-theory foundation for contemporary treatments for anxiety disorders; recount how Barlow’s triple vulnerability model has guided the development, refinement, and extension of treatment approaches; and detail more recent efforts to disseminate evidence-based approaches to practice in community settings.
Prelims.qxd
4/16/2007
xxii
10:34 AM
Page xxii
PREFACE
In a rousing chapter, G. Terence Wilson (chap. 5) challenges three commonly held myths about evidence-based, manualized treatments for behavioral health problems. First, Wilson presents evidence indicating that the use of treatment manuals fosters, rather than constrains, therapeutic innovation. Next, he briefly presents the extensive literature that challenges the often-repeated contention that the results of randomized controlled trials (RCTs) are irrelevant to real-world clinical practice. Finally, Wilson soundly refutes the assumption that the use of treatment manuals in clinical practice necessarily precludes the individualization of treatment. Danielle E. McCarthy, Daniel Bolt, and Timothy B. Baker (chap. 6) advocate for increased consideration and evaluation of mechanistic hypotheses about treatment effects on tobacco dependence, given the conceptual and clinical benefits of doing so. After outlining the conceptual, methodological, and statistical criteria necessary to support mediational claims, McCarthy et al. illustrate the use of simple-regression, structural-equation modeling, and stage-sequential approaches to examine the hypothesis that bupropion pharmacotherapy increases the probability of abstinence by reducing smoking cravings. Part III provides an overview of model research programs in clinical science and opens with a fascinating chapter by Peter J. Lang, Lisa M. McTeague, and Bruce N. Cuthbert (chap. 7) on the organization of anxiety-spectrum disorders. Using recently collected data, Lang and colleagues demonstrate that self-reported anxiety increases but physiological reactivity to threatening imagery decreases as the primary diagnosis ranges from specific phobia and social phobia to panic disorder with agoraphobia and generalized anxiety disorder. Futhermore, comorbid depression is associated with attenuated startle reflexes regardless of anxiety disorder diagnosis. These findings provide further evidence for their hypothesized organization of clinical anxiety along negative-affect and physiological-reactivity dimensions. Jayson L. Mystkowski and Susan Mineka (chap. 8) nicely illustrate the potential of capitalizing on decades of basic research in learning science when considering how to enhance the effectiveness and efficiency of behavior therapy for anxiety disorders. Drawing in particular on the animal literature demonstrating the contextualized nature of fear conditioning and extinction, Mystkowski, Mineka, and their collaborators have conducted five studies that document the exquisite context specificity of spider phobics’ “return of fear” after successful brief treatment. Significantly,
Prelims.qxd
4/16/2007
PREFACE
10:34 AM
Page xxiii
xxiii
this program of integrative research bears not only on behavior therapists’ applied efforts, but also on learning theorists’ understanding of context and conditioning. James T. Townsend, Mario Fific′, and Richard W. J. Neufeld (chap. 9) offer an illuminating tutorial on the conceptualization, assessment, and evaluation of clinically relevant individual differences in cognitive architecture (e.g., serial vs. parallel vs. hybrid processing models), which is a critical but understudied aspect of the distinction between automatic and controlled processing. Townsend et al. describe how formal computational models can be used to detect architecture-specific signatures in response-time data obtained in particular behavioral paradigms that also are described. This chapter also illustrates the potential benefits of integrative research—in this case, research that spans clinical and quantitative cognitive science. Joseph E. Steinmetz (chap. 10) offers a captivating overview of his translational research program in clinical neuroscience, conducted in part in collaboration with Richard McFall. Steinmetz and his colleagues have leveraged the wealth of basic research knowledge about the neural circuitry involved in eyeblink conditioning to test hypotheses about the brain–behavior correlates of fetal alcohol syndrome, autism, obsessivecompulsive disorder (OCD), and schizophrenia. Using these different disorders as case studies, Steinmetz reveals how eyeblink conditioning paradigms can be used to evaluate hypotheses about the role of cerebellar pathologies and associative learning processes in psychopathology. In the closing chapter of Part III, Teresa A. Treat and Melanie A. Dirks (chap. 11) highlight the potential utility of applying an integrated class of cognitive scientists’ theoretical, measurement, and analytical models to the study of clinically relevant individual differences in component cognitive processing. Treat and Dirks review recent research conducted in collaboration with Richard McFall that demonstrates that cognitive scientists’ models also generalize well to the study of more complex and socially relevant processing, consistent with a unified approach to the characterization of human cognition that bridges normal and abnormal processing and behavior. Part IV on future directions in clinical science research, application, and training opens with a stimulating and informative chapter by Bruce N. Cuthbert (chap. 12), who has a unique perspective on the issues facing the field given his experience at the National Institute of Mental Health (NIMH). Cuthbert argues that psychological clinical science currently is
Prelims.qxd
4/16/2007
10:34 AM
Page xxiv
xxiv
PREFACE
facing criticism on two fronts: from practitioners regarding their concerns about the empirically supported treatment movement, and from neuroscientists who question the importance of psychological understandings of clinical disorders in the face of exciting advances in clinical neuroscience. Cuthbert sketches the implications of these challenges for future research foci in clinical science, diagnostic systems of psychopathology, and doctoral training in clinical science. Robert W. Levenson (chap. 13) challenges us to broaden our perspectives as we consider future directions in the clinical science movement within the three central areas of diagnosis, treatment, and etiology of clinical phenomena. He highlights increasing interest in transdiagnostic approaches, the controversy surrounding the empirically supported treatment movement, and the rising import of molecular genetic methods for our understanding of psychopathology. He closes the chapter by articulating challenges currently facing the production of clinical science research, the training of clinical scientist researchers, and the stormy relationship between science and practice. In Part V, Richard M. McFall (chap. 14) continues to challenge and inspire us in a stirring closing chapter that provides a fascinating glimpse of his perspective on the past, present, and future of psychological clinical science. Throughout the chapter, McFall emphasizes the importance of adhering to a scientific epistemology in clinical psychology; implementing clinical-science models of doctoral training that are highly individualized, interdisciplinary, and research-focused; and integrating clinical science with the rest of psychological science. Not unlike he did in 1991 with his Manifesto for a Science of Clinical Psychology, McFall once again has offered a rallying cry to psychological clinical scientists. INTERVIEW WITH RICHARD M. MCFALL The DVD included on the back inside cover of this volume contains an 11-track interview with Richard McFall, conducted by Teresa Treat in May 2004, on various issues relevant to McFall’s career and the clinicalscience movement. The interview opens with McFall’s description of how he chose a career in psychology (at the urging of Exner) and the critical role that mentors and significant influences (Kelly, Rotter, and Lang) have played throughout his career. McFall then talks about the background and response to his controversial Manifesto for a Science of Clinical Psychology. The fourth and fifth sections of the interview focus
Prelims.qxd
4/16/2007
PREFACE
10:34 AM
Page xxv
xxv
on the development of and opportunities now facing the Academy of Psychological Clinical Science, as well as the role that the Association for Psychological Science (APS, previously the American Psychological Society) and Alan Kraut, the executive director of APS, have played in clinical science. McFall then defines and articulates the relevance of integrative psychological science to clinical science, including implications for doctoral training programs, the distinction between translational and integrative research, and what clinical science has to offer the rest of psychological science. Professional issues take center stage in the seventh track, as McFall discusses the changing role for clinical psychologists in mental-health care. McFall then reflects on the most gratifying and bittersweet aspects of his career. McFall comments on seven hot topics in clinical psychology in the ninth track: prescription privilges for psychologists, the distinction between the scientist-practitioner and clinical-science models, licensing of psychologists, accreditation of training programs, the mental-health care delivery system, continuing education for psychologists, and the distinction commonly made between efficacy and effectiveness treatment-outcome research. McFall provides an overview of his research career in the penultimate track, focusing on the conceptualization, measurement, and modification of socially relevant cognitive processing in psychopathology, as well as his increasing capitalization on contemporary cognitive science and neuroscience when addressing research questions. Finally, in a marvelous closing segment, McFall provides advice to young scientists. The magnitude and scope of McFall’s impact on psychological clinical science, and on psychological science more generally, are readily apparent throughout this interview.
Prelims.qxd
4/16/2007
10:34 AM
Page xxvi
Prelims.qxd
4/16/2007
10:34 AM
Page xxvii
Acknowledgments Neither McFall’s festschrift celebration nor this ensuing festschrift volume published in his honor would have occurred without the boundless energy and staunch support of Alan Kraut, whom we cannot thank enough for yet another indicator of his tireless efforts on behalf of the advancement of psychological clinical science. The festschrift celebration also would not have been possible without the expert assistance of Alan’s colleagues at the Association for Psychological Science (APS), particularly Sarah Brookhart, Brian Weaver, Rebecca Hachem, Eric Jaffe, Kate Volpe, and Louis Shomette. Many thanks also are due to the financial sponsors of the festschrift celebration: APS, Lawrence Erlbaum Associates, the Academy of Psychological Clinical Science, and the Department of Psychology at Indiana University. Several individuals contributed in particularly noteworthy fashions to the planning and implementation of the festivities in Chicago, including: Bob Levenson, Sue Mineka, Bob Simons, Joe Steinmetz, the McFall family (Kathy, Julie Beth, and Adam), Don Fowles, Bruce Cuthbert, Elizabeth Yeater, Shirley Richardson, the “toasters” at the festschrift dinner, and the presenters of Sunday’s talks. Bill Freeman, Doug Toms, Jerry Forshee, and their colleagues in the Technical Services Group at Indiana University generously invested much time and energy into the production of the DVD interview with McFall that is being distributed with the festschrift volume. Steve Rutter, Nicole Buchmann, and Susan Milmoe of Lawrence Erlbaum Associates, as well as Frank Farach, also provided invaluable advice and assistance as we edited the festschrift volume. We greatly appreciate everyone’s contribution to the festschrift celebration and volume. —T.A.T., R.R.B., T.B.B.
xxvii
Prelims.qxd
4/16/2007
10:34 AM
Page xxviii
Chapter 01.qxd
3/28/2007
12:58 PM
Page 1
I HISTORY AND EPISTEMOLOGY OF PSYCHOLOGICAL CLINICAL SCIENCE
Chapter 01.qxd
3/28/2007
12:58 PM
Page 2
Chapter 01.qxd
3/28/2007
12:58 PM
Page 3
1 Psychological Clinical Science: Why and How We Got to Where We Are Richard R. Bootzin University of Arizona
With apologies to Charles Dickens, these are the best of times and the worst of times for clinical psychology. There have been impressive advances in our understanding of psychopathology, methods for assessing individual characteristics, and the development of treatments based on principles of change for a variety of applications. Many of the most exciting advances have been interdisciplinary and have come at the interfaces between clinical psychology and subdisciplines within psychology, such as cognitive and emotion psychology, health, social, developmental, forensic psychology, and methodology and evaluation, as well as with nonpsychological disciplines such as linguistics, anthropology, sociology, computer science, genetics, and neuroscience. We have new methods, such as neural imaging, behavioral genetics, computational modeling, and graphical data analysis, that expand our understanding and lead to new opportunities for application. At the same time, there are countervailing practical concerns within professional clinical psychology, including licensing, employment opportunities, attempts to expand and articulate the scope of practice, reimbursement parity with physicians, hospitalization privileges, continuing education requirements, and liability insurance, among others that often 3
Chapter 01.qxd
3/28/2007
12:58 PM
Page 4
4
BOOTZIN
create a chasm between clinical scientists and clinical practitioners. This is seen most vividly in the increasingly divergent training models for clinical psychologists and the recurrent battles over accreditation standards. To the nonpsychologist, this strain within the field is hard to understand. Similar distinctions between clinical scientists and clinical practitioners in other fields, such as medicine, nursing, pharmacology, and dentistry, do not appear to produce the same degree of acrimony among members of the same field who work to advance knowledge with those who attempt to apply that knowledge. Is there something different about clinical psychology? There have been dramatic changes in clinical psychology as a profession over the past 100 years. At times it appears that today’s conflicts between science and practice are unique. However, the integration of science and application has waxed and waned since the very beginnings of the separation of psychology from its predecessor disciplines of philosophy and physiology. We explore here the changing roles of research and practice during this history as a means of understanding the choices that present themselves today. In this volume, dedicated to the contributions of Richard M. McFall, it is with considerable pleasure that I acknowledge the substantial influence that Dick McFall has had on me individually, as well as on the field more generally. Dick McFall has played a central role in articulating the importance of the integration of science and practice in clinical psychology and in initiating methods for doing so. His contributions are highlighted throughout the chapter. But first, how did we get to where we are? FROM 1879 TO WORLD WAR I The beginning of the new discipline of psychology is usually identified as 1879 because that was the year when Wilhelm Wundt, a professor of physiology at the University of Leipzig, Germany, established the first laboratory dedicated to the study of psychology (Boring, 1957; Hothersall, 1995). Before Wundt’s laboratory, there were many philosophers, physiologists, physicians, and naturalists who studied psychological phenomena, such as Hermann von Helmholtz (perception), Gustav Fechner (psychophysics), Charles Darwin (evolution), Paul Broca and Carl Wernicke (language and the brain), and Jean Charcot (hysterical disorders).
Chapter 01.qxd
3/28/2007
12:58 PM
Page 5
1. PSYCHOLOGICAL CLINICAL SCIENCE
5
Wundt’s unique contribution was that the focus of his laboratory was on psychological phenomena. He and the many students attracted to work with him labored to found a new discipline. Among those who studied with Wundt were Emil Kraepelin, G. Stanley Hall, James McKeen Cattell, Hugo Münsterberg, Edward Titchener, and Lightner Witmer— all recognizable for their contributions to the founding of psychology as a scientific discipline. Contrary to the idea that early psychologists were narrow experimentalists with little interest in the application of psychological knowledge, many of the founders of psychology in Europe and the United States had broad interests and were interested in application, including in what came to be called clinical psychology. Edward Titchener, who insisted that the new science of psychology be a laboratory science based on the specialized methodology of introspection, was the exception (Hothersall, 1995). Emil Kraepelin, whom Hans Eysenck referred to as the “father of clinical psychology” (Wittling, 1972), established a psychological laboratory dedicated to the study of mental illness and how psychopathology was related to basic psychological processes. In addition to his well-known work on psychiatric diagnosis, Kraepelin was the first to study the effects of alcohol, nicotine, and drugs on human behavior (Wittling, 1972). G. Stanley Hall received his PhD at Harvard with William James and then went to Germany to study with Wundt. When he returned to the United States, Hall established what many consider the first psychology laboratory in the United States in 1881 at Johns Hopkins University. Some list James as having founded the first psychology laboratory at Harvard University in 1874 (e.g., Garvey, 1929). Although he was an influential force within the newly developing psychology, James was not as active an empirical researcher as were Wundt and Hall. Hall founded and was the first president of the American Psychological Association (APA) in 1892. He established a number of journals that reflected his broad interests, including his interest in developmental psychology. The journals included the American Journal of Psychology, the Journal of Genetic Psychology, and the Journal of Applied Psychology. James McKeen Cattell, who was the first American to receive his PhD with Wundt in 1886, continued to work with him on studies of intellectual assessment. These studies focused on cognitive processes as measured by reaction time and other measures of the processes, rather than the
Chapter 01.qxd
6
3/28/2007
12:58 PM
Page 6
BOOTZIN
outcomes of cognition. This method for evaluating intellectual assessment was eventually overshadowed by the assessment of complex intellectual behaviors and outcomes as reflected in the work of Alfred Binet in France. In 1888, Cattell returned to the United States to become the first professor of psychology in the nation at the University of Pennsylvania. He moved to Columbia University in 1891, where among his PhD students were Edward Lee Thorndike, Robert S. Woodworth, and Edward K. Strong, of the Strong Vocational Interest Test (Hothersall, 1995). Cattell cofounded the Psychological Review in 1894 and bought the rights to Science in 1895. Cattell was a vocal advocate for the application of psychology. As early as 1904, he predicted that psychology would be both a science and a profession (Watson, 1978). An example from his own career was that, in 1921, Cattell founded the Psychological Corporation, which continues to this day to market psychological tests and promote the application of psychology. Hugo Münsterberg received his PhD in 1885 with Wundt and then moved to the University of Heidelberg, where he received his MD in 1887. In a statement that foreshadows the scientist-practitioner training model, Münsterberg recommended that getting both a PhD and an MD was ideal preparation for a career in applied psychology (Hothersall, 1995). Münsterberg published an action theory of consciousness in which he stated that muscle sensations were the basis of awareness and consciousness. William James was impressed with Münsterberg’s research and saw similarity to his own theories of emotion. In 1892, James recruited Münsterberg to take over the Harvard psychology laboratory (Hothersall, 1995). James, who was later elected president of the American Philosophical Association as well as president of the APA, expressed interest in spending less time in the laboratory and more time in philosophy. Münsterberg had broad applied interests, including the study and treatment of mental illness. He saw patients in his laboratory at Harvard. In 1909, he wrote a successful popular press book, Psychotherapy, intended to dispel myths about mental illness and the new psychological treatments by Sigmund Freud and others. Münsterberg was an advocate for expanding the boundaries of psychology. In addition to his interest in clinical psychology, he was a pioneer in research in industrial organizational psychology, eyewitness
Chapter 01.qxd
3/28/2007
12:58 PM
Page 7
1. PSYCHOLOGICAL CLINICAL SCIENCE
7
testimony, group decision making, lie detection, and, later in his life, film theory. In each of these areas, descriptions of Münsterberg’s research and conclusions (Hothersall, 1995) sound current. For example, he railed against the use of eyewitness testimony. To illustrate his position, Münsterberg would stage demonstrations of disruptive assaults during classes or lectures. He would have the audience members write detailed descriptions of what occurred (Hothersall, 1995). By pointing out the contradictions that occurred between different witnesses to the same event, he would demonstrate the fragility of memory under stressful conditions. In an interesting parallel to today’s conflicts, Münsterberg’s expansive view of the boundaries of psychology drew criticism. Edward Titchener, who, like Münsterberg, received his PhD with Wundt and then immigrated to the United States, said of Münsterberg, “Dr. Münsterberg has the fatal gift of writing easily—fatal especially in science, and most of all in a young science where accuracy is the one thing most needful” (Titchener, 1891, p. 594; cited by Hothersall, 1995). Over 50 years after Münsterberg’s death, Robert Watson wrote of him, “It is probable that his [Münsterberg’s] present lack of influence can be attributed to the fact that he turned to fields for the application of psychology before they had a research basis on which to operate” (Watson, 1978, p. 410). What is the appropriate balance between psychology’s knowledge base and its applications? Throughout the history of psychology, many have seen policy implications in psychological research knowledge. It is not sufficient to justify policy from basic research alone. Applications and policy must, themselves, be evaluated (Bootzin & Ruggill, 1988; Campbell, 1969; Sechrest & Bootzin, 1996). Münsterberg’s professional activity was one of the many threads that contributed to the development of clinical psychology. Others were interested in the diagnosis and treatment of mental illness. Among the more influential were Pierre Janet, Sigmund Freud, Emil Kraepelin, and Morton Prince. Another thread was the early effort to describe individual differences in intellectual capabilities, including the work of Francis Galton, James McKeen Cattell, and Alfred Binet. Related to this work was the interest in applications from child development for parenting, including the influence of J. Stanley Hall, Lightner Witmer, John Watson, and Mary Cover Jones. There was also a thread involving research on learning, including the works of Ivan Pavlov, Edward Thorndike, John Watson, and, later, B. F. Skinner and Clark Hull.
Chapter 01.qxd
8
3/28/2007
12:58 PM
Page 8
BOOTZIN
The first psychology clinic in the United States was established in 1896 at the University of Pennsylvania by another of Wundt’s Americans, Lightner Witmer. Just as the establishment of Wundt’s psychology laboratory marks the beginning of the science of psychology, the establishment of Witmer’s psychology clinic marks the beginning of clinical psychology in the United States. For Witmer, the clinic was his laboratory. It provided an opportunity to both apply the knowledge of the discipline and generate new knowledge (Kihlstrom & Kihlstrom, 1998; Woody & Robertson, 1997). Witmer performed what he called small experiments on his patients and thus anticipated the use of single-case designs within clinical psychology. In 1908, Witmer founded the journal, The Psychological Clinic, which was published through 1935. For Witmer, science and application were intricately tied together. Witmer was not only the founder of clinical psychology, but also a forefather of today’s emphasis on psychological clinical science. Witmer was not alone in the newly emerging field of clinical psychology. A major figure in the developing field of abnormal psychology at the turn of the 20th century was Morton Prince, a neurologist, who in 1906 founded and was the first editor of the Journal of Abnormal Psychology. Prince, like Charcot and Janet, was interested in dissociative disorders, particularly multiple personality disorders. He provided detailed descriptions of the personalities of normal individuals, as well as of patients. Prince founded the Harvard Psychological Clinic in 1927. Two years later, upon Prince’s death, Henry Murray took over directorship of the clinic and expanded Prince’s work on personality. One of the case studies published by Prince (1912) provided an analysis of a woman with anxiety disorders and panic attacks. As noted by Otlmanns and Mineka (1992), Prince’s description and analysis anticipated current cognitive and social learning explanations of panic disorders with agoraphobia. Prince identified the fear of panic as critical in many phobias and commented as follows on the importance of what have come to be called safety cues: There is no fear, properly speaking, of an open place, or of a closed place, or of a train, or of a theater. The true fear is of having an attack in a situation where, owing to the circumstances of the environment, the patient cannot obtain relief.… These patients with phobias all have anticipatory fear of an attack, and this is particularly intense in anticipation of a situation where help in an attack cannot be expected. None of these phobia cases is afraid to be in these situations provided he is accompanied by a physician, or a person in whom he has confidence. (p. 276)
Chapter 01.qxd
3/28/2007
12:58 PM
Page 9
1. PSYCHOLOGICAL CLINICAL SCIENCE
9
Looking back from today’s perspective, it is surprising that Prince did not have more influence in clinical psychology. As Oltmanns and Mineka (1992) point out, Prince was first and foremost a clinical scientist. Although he was influenced by and expressed respect for the work of those associated with the two movements that were sweeping clinical psychology in the United States, psychoanalysis and behaviorism, he considered himself a member of neither. In a letter to a colleague in England, he wrote, in science there cannot be “schools.” School means dogma. Science means facts of observation and logical inductions.… For myself I consider everything, my own theories and all, as only provisional and someday may be knocked into a cocked hat by new discoveries. After all, “all is opinion and opinion be dammed.” (p. 4; quoted in Oltmanns & Mineka, 1992)
FROM WORLD WAR I TO WORLD WAR II Clinical psychology continued to develop during the first two decades of the 20th century. During World War I, more than 25% of APA members served in World War I in special branches (Griffith, 1922), and many psychology laboratories at universities aided the war effort. The APA set up 12 committees to mobilize and organize help from psychologists in different areas. The work having the most impact included (a) help provided to the air service in research and evaluation of perception in prospective aviators, and in the status of mental states under low oxygen pressure (Griffith, 1922); and (b) the development and use of intelligence and aptitude testing to select and place soldiers and officers throughout the military. The first mass intelligence tests were developed during World War I. A third area, a successful effort to diagnosis and treat “war psychoses,” was the focus of psychologists from England, France, and Germany (Griffith, 1922). Psychology’s successful role in the war effort provided a stimulus for a stronger focus on applied psychology. In 1917, clinical psychologists found that the APA was unwilling to provide assistance in developing standards for clinical practice. Thus, in order to have an organization that would develop standards for clinical practice and fortified by an increasing number of clinical psychologists, the American Association of Clinical Psychologists (AACP) formed independently of the APA. There was considerable acrimony about founding a second professional organization and fear that the existence of AACP would undermine the
Chapter 01.qxd
10
3/28/2007
12:58 PM
Page 10
BOOTZIN
growth and health of APA. As noted by Routh (1994), a prophetic exchange about the future battles involving training was made in the APA Council meeting in 1918. Robert M. Yerkes stated that, “certain educational institutions should specialize in applied psychology and that others should continue with general instruction.” In response, E. L. Thorndike stated that “he believed that in 20 years there would be as many ‘doing’ as teaching psychology, but that both groups must be scientific. He saw no reason why the PhD in psychology should not represent both types” (Psychological Bulletin, 1919; cited in Routh, 1994). Yerkes’ position anticipated the development of professional schools of psychology. In contrast, Thorndike’s position anticipated the scientistpractitioner model and the subsequent development of the psychological clinical science model. The AACP was short-lived. In 1919, just 2 years after its founding, members of AACP joined APA as a separate section after receiving assurances that APA would commit itself to improve the quality of clinical psychology training. The APA leadership promised to develop certification standards to identify those clinical psychologists who had the appropriate training to provide clinical services. At the time, the focus of practice of clinical psychologists was primarily on the diagnosis of mental disorders, and intellectual and personality assessment. Some clinical psychologists also provided psychotherapy, but it was a small, although growing percentage, after World War I. APA committees worked on attempts to provide certification, but in 1927 the effort was ended, 8 years after AACP joined APA. Only 25 clinical psychologists were ever certified (Routh, 1994). State associations, composed mostly of applied psychologists, were growing in number. Clinical and other applied psychologists within the APA, in conjunction with the support of psychologists from the state associations, lobbied the APA to support the profession, not only the science, of psychology, and to provide more representation for applied psychologists in the governance and annual program of the APA. In 1937, when no satisfactory accommodation could be reached, the clinical psychology division of the APA disbanded and members were encouraged to join the newly formed American Association of Applied Psychology (AAAP; Routh, 1994). The AAAP initially had four sections: clinical psychology, consulting psychology, educational psychology, and industrial and business psychology. A section for military psychology was added later.
Chapter 01.qxd
3/28/2007
12:58 PM
Page 11
1. PSYCHOLOGICAL CLINICAL SCIENCE
11
Although there were two separate organizations—APA and AAAP— many of the clinical psychologists continued to belong to both, and annual conventions were held together in all but 1 year so that members could attend both conventions (Routh, 1994). Interest in clinical psychology continued to grow in both academic and applied settings. Despite the separation into its own organization, the proportion of papers in clinical psychology presented at APA continued to increase. At the 1940 APA meeting, clinical psychology papers comprised 25% of all papers and were the largest number of any substantive area (Fernberger, 1943). THE GROWTH OF CLINICAL PSYCHOLOGY FOLLOWING WORLD WAR II The need for appropriately trained clinical psychologists became clear and urgent during World War II. Emotional and mental disorders accounted for more than half of the patients receiving care from the Veterans Administration (VA) in the year following the end of World War II, and there was a shortage of clinical psychologists available to provide care (Benjamin, 2005). In 1944, a joint commission of the APA, the AAAP, the VA, the U.S. Department of Public Health (USDPH), and the newly formed National Institute of Mental Health (NIMH) was held to consider how to provide clinical psychologists to meet the projected mental health needs of the returning veterans. The VA, USDPH, and NIMH agreed to provide funding for the doctoral training of clinical psychologists. Only about a third of clinical psychologists in the community held doctorates, whereas 60% of APA members did (Routh, 1994). The focus on APA as an organization of psychologists with PhDs appears to reflect the beginning of the policy that psychology, including the application of psychology, requires a doctoral degree. The requirement of doctoral training for clinical psychology was not the model accepted in other parts of the world. In many Western countries, such as those in Europe and the British Empire, the PhD was reserved for training in research. Psychological clinical practice was mostly carried out by individuals who had bachelor’s and master’s degrees. The VA agreed to fund doctoral training and employ the newly graduated clinical psychologists. The next year, in 1945, the VA wrote a letter to the APA asking for a list of doctoral programs in clinical psychology so
Chapter 01.qxd
12
3/28/2007
12:58 PM
Page 12
BOOTZIN
that it could offer traineeships to those programs (Sears, 1947). The APA surveyed programs and responded with an initial list of 22 programs. The VA responded quickly, and in the first year of the program there were 200 students from the 22 programs on VA financial support. In subsequent years, the VA increased the number of traineeships, and NIMH began to offer training grants for clinical psychology graduate programs. The prospect of financial support for training from the VA and NIMH, coupled with the commitment by the VA to hire the trained clinical psychologists, served as a major impetus for the growth of the field. At the same time, in 1945, the AAAP rejoined a reorganized APA. To accommodate the interests of clinical and other applied psychologists, each of the AAAP’s sections became an APA division, and the APA’s mission was expanded from advancing the science of psychology to advancing “Psychology as a science, as a profession, and as a means of promoting human welfare” (Wolfle, 1946, p. 3). A new journal to reflect the profession of psychology was started, The American Psychologist. It was subtitled, “The Professional Journal of the American Psychological Association, Inc.” To take advantage of the new funding initiatives for the training of clinical psychologists, there was a need for a model of training. Carl Rogers, elected president of the APA in 1946, appointed David Shakow to chair an APA committee to establish guidelines for training clinical psychologists. The other committee members were Ernest R. Hilgard, E. Lowell Kelly, Bertha Luckey, R. Nevitt Sanford, and Laurance F. Shaffer. The committee’s report (Hilgard et al., 1947) recommended a curriculum involving course work and training in research, assessment, and psychotherapy throughout the PhD graduate program. A full-time internship was recommended to take place in the third year of the 4-year program. Reflecting the focus of clinical psychology in the 1940s, the course work included what many would now see as anachronisms, including course work on projective tests and training in the psychodynamics of personality. The APA started implementing the committee’s recommendations immediately by moving to an accreditation system. Programs filled out questionnaires and were site-visited by members of the committee. All programs that had been identified in the 2 years prior to the VA were considered accredited for the first year. In the first round of accreditation of clinical psychology programs by the APA, 36 programs were continued as accredited, and 7 new programs were added (American Psychological Association, 1948).
Chapter 01.qxd
3/28/2007
12:58 PM
Page 13
1. PSYCHOLOGICAL CLINICAL SCIENCE
13
In August 1949, a conference on graduate education in clinical psychology sponsored by NIMH was held at the University of Colorado at Boulder. The conference endorsed the scientist-practitioner model, which came to be known as the Boulder Model. This was the same model of training recommended by the APA Shakow committee, in which science and practice would be integrated and interdependent. As published in the 1949 report of accreditation guidelines (American Psychological Association, 1949), some of the same complaints about the accreditation process that surfaced later were heard during the first year of accreditation. For example, there was confusion about whether the recommended curriculum was required. Committee members who conducted site visits often had to explain that the recommendations for the clinical psychology curriculum were not required, but were only recommendations. However, each of the curriculum recommendations had been listed separately in the committee’s report (American Psychological Association, 1949), along with other requirements such as having a specialized faculty. Programs received a plus for each curriculum and program recommendation that was already in place. Although site visitors may have considered the recommended curriculum to be advisory, it would be hard to fault the program faculty for concluding that a checklist of courses was the default option. Another familiar reaction is that some program faculty members communicated enthusiastic support about the accreditation process, whereas others communicated that they thought the accreditation committee was usurping power from the university graduate departments by mandating curriculum and insisting that specific program resources, such as practicum facilities, be provided. The familiar challenges of how to provide flexibility and innovation in training and how “voluntary” the voluntary accreditation system really was have been continuing issues from the beginning of APA accreditation. They remain so today. The scientist-practitioner model became so ubiquitous that it is often forgotten that the Boulder conference had contemporary critics, including Hans Eysenck (1949) and Seymore Sarason (1988). Eysenck agreed with identifying training in assessment and research, but not training in psychotherapy, as foundations for clinical psychology training. In anticipation of his article on the ineffectiveness of psychotherapy (Eysenck, 1952) to be published 3 years later, Eysenck argued that there was no scientific basis for psychotherapy, and therefore it had no place in PhD programs based on the science of psychology.
Chapter 01.qxd
14
3/28/2007
12:58 PM
Page 14
BOOTZIN
Before World War II, the scope of practice in clinical psychology had been focused primarily on assessment, but had gradually come to include psychotherapy. At the beginning of World War II, about one third of psychologists included psychotherapy in their practice (Woody & Robertson, 1997). In 1942, Carl Rogers published his book on a nondirective approach to psychotherapy, and it provided an additional impetus for psychologists to include psychotherapy as part of their clinical activities. Although organized psychiatry opposed the expansion of the psychologists’ scope of practice to include psychotherapy, the combination of a national crisis in providing treatment to returning veterans and longterm trends in the expansion in the scope of practice to include treatment led to including training in psychotherapy in the accreditation guidelines. Eysenck’s (1949) critique of the guidelines did not stop training in psychotherapy from becoming one of the foundations of clinical psychology training in the United States. However, a positive, and perhaps ironic, consequence of the debate was that Eysenck’s (1952) analysis of the ineffectiveness of psychotherapy ushered in a revolution in therapy outcome research and the eventual development of empirically supported psychological treatments. Another critic, Seymore Sarason, was one of the attendees at the Boulder Conference. He wrote later (Sarason, 1988) that the conference made a mistake in abandoning the application of clinical psychology with children in schools and other community settings to shift the focus of the field to adult psychiatric disorders. From the beginning of clinical psychology in the United States, with the opening of Witmer’s clinic, the focus of clinical research and practice centered on children. Similar critiques were made in the years following the Boulder conference by Nicholas Hobbs (1964) and George Albee (1969, 2000). Hobbs hoped the community mental health movement of the 1960s would broaden the focus of clinical psychology to improve mental health care in the community and reduce the focus on a disease model of mental illness. Albee also considered the Boulder Model to be overly narrow, based as it was on the psychiatric medical model. Albee (2000) blamed the influence of David Shakow, who had worked mostly in psychiatric settings. Albee argued that clinical psychology lost an opportunity to develop an educational model based on responses to stress, rather than a training model based on diagnostic categories. Many of the criticisms of the Boulder Model resonate even more powerfully today. The current interest in many areas developed despite the
Chapter 01.qxd
3/28/2007
12:58 PM
Page 15
1. PSYCHOLOGICAL CLINICAL SCIENCE
15
curriculum confines of the dominant clinical psychology training model. Included would be interest in positive psychology, family psychology, child and adolescent psychology, health psychology, transdiagnostic reactions to stress, community interventions, and prevention psychology. Despite the qualms about the accreditation process, clinical psychology programs proliferated and were accredited during the next decades. By 1970, there were 81 accredited clinical psychology programs (American Psychological Association, 1970). In the 1960s, applied jobs were plentiful. There were seven positions for every graduate being trained (Wright, 1983). In 1963, the Community Mental Health Act expanded employment opportunities for psychologists by shifting the emphasis of the treatment of mental illness from inpatient hospitals to outpatient care in community clinics. The surplus of positions, compared with the number of clinical psychologists being trained, led to the development of professional schools. The first doctoral program based on a scholar-practitioner, rather than a scientist-practitioner, model was founded by Gordon Derner at Adelphi University in 1951 (Benjamin, 2005; Routh, 1994). In 1968, the University of Illinois established a professional school under the direction of Donald Peterson. A new doctoral degree was awarded, the Doctor of Psychology, abbreviated as PsyD. A year later, the California School of Professional Psychology (CSPP) was established, under the leadership of Nicholas Cummings, who had been a graduate of the Adelphi program, as the first freestanding professional school (Benjamin, 2005; Routh, 1994). Changes were occurring in the science of clinical psychology as well. During the 1940s and 1950s, many faculty members in clinical psychology who followed the scientist-practitioner model engaged in two different and separate enterprises—one on research topics such as personality, psychopathology, assessment, developmental psychology, and learning, and the other on training in clinical practice activities. This led to what McFall (1991) called “the two-headed psychologist,” in whom the activities of science and practice were separate and not well integrated. There were two battles going on in psychology. One was the extent to which training for application belonged in psychology departments. Many psychologists held the conviction that training for application was a burden for those programs interested primarily in advancing knowledge. An elegant statement by Cronbach (1957) pointed out that the clash was
Chapter 01.qxd
16
3/28/2007
12:58 PM
Page 16
BOOTZIN
between two different disciplines of scientific psychology—experimental and individual differences psychology—that used different methodologies and frequently asked different research questions. Cronbach’s statement helped clear the air and provide legitimization for an individual differences psychological science that had been associated with application and had been an important theme from the beginning of scientific psychology in the work of Galton, Kraepelin, James McKeen Cattell, and Binet, among others. The second battle was within clinical psychology. What began to change in the 1950s and 1960s was that an alternative to the two-headed clinician became possible. The goal of developing a science of clinical psychology was seen in the work of many during the 1950s, notably the work of Paul Meehl (1954) in his review of clinical versus statistical prediction and the work at Ohio State University of Julian Rotter (1954) in social learning and George Kelly (1955) on personal constructs. A major training program in clinical psychology developed at Ohio State University. Carl Rogers had been director of the psychology clinic there during World War II, before he moved to the University of Chicago in 1945. Following Rogers at Ohio State were George Kelly and Julian Rotter. During the 1950s and early 1960s, Ohio State was a center of research in clinical psychology. A generation of scientists in clinical psychology was trained there, including Rue Cromwell, Herb Lefcourt, Brendan Maher, Walter Mischel, Lee Sechrest, Mark Stephens, and many others. Dick McFall, who was influenced by both Kelly and Rotter, received his PhD at Ohio State in 1965 and was the last graduate student to get his PhD with Kelly before Kelly left for Brandeis, ending an era. Rotter had moved to the University of Connecticut 2 years earlier. The development of the science of clinical psychology continued in the 1960s with, for example, Walter Mischel’s (1968) book on the role of trait versus situational determinants in assessing personality characteristics and Albert Bandura’s (1961, 1969) watershed reviews of principles of behavior change. In these and other examples, the science of clinical psychology drew on the work of researchers across areas of psychology, but stood on its own in advancing psychological clinical science. The proliferation of training programs, many of which focused more on clinical practice than on knowledge acquisition, led to strains within some departments of psychology and conflicts in accreditation. Some leading private universities (e.g., Harvard University, Stanford University, and
Chapter 01.qxd
3/28/2007
12:58 PM
Page 17
1. PSYCHOLOGICAL CLINICAL SCIENCE
17
Northwestern University) changed their programs from clinical psychology to experimental personality and/or psychopathology to emphasize that their programs provided a focus on training for knowledge acquisition rather than on clinical service. Northwestern University reestablished its clinical psychology program 10 years later, in 1980, and Harvard reestablished its clinical psychology program in 2000. At the same time, there was increasing dissatisfaction expressed with the scientist-practitioner model because some viewed it as unrealistic for clinical psychologists in practice also to be researchers (Woody & Robertson, 1997). An alternative practitioner-training model, in which clinical psychologists would be taught to understand and apply research, but would not learn the skills needed to be a researcher, was proposed. This practitioner professional model, the scholar-practitioner model associated with the awarding of a PsyD, was endorsed along with the scientistpractitioner model, associated with the awarding of a PhD, at a conference held in Vail, Colorado, in 1973. Scholar-practitioner programs have been more commonly found in freestanding professional schools of psychology rather than in universitybased programs. There has been variation, even among professional schools, however, as some award PhDs, and not PsyDs, and some describe their programs as scientist-practitioner, irrespective of the degree awarded. DEVELOPMENT OF THE PSYCHOLOGICAL CLINICAL SCIENCE MODEL Although the vast majority of university-based clinical psychology programs subscribed to the scientist-practitioner model from the 1970s through the 1990s, there was enormous variability in the extent to which different programs emphasized science or practice. As Edward Katkin (1982) observed, “‘Scientist-practitioner’ can be pronounced with a silent ‘scientist’” (p. 9). The pendulum in clinical training has swung to one end of the scientist-practitioner continuum and back again. Although there was a general drift toward clinical service with the development and proliferation of professional schools, there were also some noteworthy landmarks toward a reemphasis on the science of clinical psychology. As early as 1971, Julian Rotter, expressing frustration about an accreditation system that was more concerned about process than outcome, said
Chapter 01.qxd
18
3/28/2007
12:58 PM
Page 18
BOOTZIN
If psychologists are not more active and more explicit in their evaluation of techniques of intervention, they will find themselves restrained from the outside…as a result of their own failure to do what ethical and scientific considerations require. (Rotter, 1971, p. 2)
In 1982, the APA Task Force on the Evaluation of Education, Training, and Service in Psychology released its report that emphasized that, despite the need for accountability in training, there was remarkably little cumulative knowledge regarding training. The report stated, “There is no evidence that any specific educational or training program or experience is related to professional competence” (American Psychological Assocation, 1982, p. 2). It is informative to see a list of the committee members: The chair was Lee Sechrest, and members included Sol Garfield, Ronald Kurz, Neal Miller, Donald Peterson, and Janet Taylor Spence. This was a distinguished group, but its report had little impact on the methods of training, or its evaluation, within clinical psychology programs. Six years later, in a review of training for behavior therapy, Bootzin and Ruggill (1988) noted that the landscape for the evaluation of training had changed little, but the skills of research were beginning to be recognized as important for the development of effective practice. For example, McFall (1985) identified the skills of keen observation, critical thinking, and methodological rigor combined with inventiveness when putting conceptualization to the empirical test and following the lead of empirical evidence as skills that are just as important to the application of knowledge as to its generation. Methods that were commonplace in research were recommended for use by training programs to develop and evaluate clinical skills. Among these methods were the importance of ongoing measurement in assessing change; the development of reliable and valid assessment instruments; the teaching of therapeutic skill through manuals, models, simulations, role playing, and supervision; and verification methods for assessing the extent to which therapeutic procedures were being implemented (Bootzin & Ruggill, 1988; Kazdin, Kratochwill, & VandenBos, 1986). To provide an adequate evaluation of the development of effective treatments or of therapeutic training, it is important to recognize that what is being evaluated is not just a set of treatment interventions. It is also a collection of “small theories” (Lipsey, 1990) about how problem behavior occurs and how treatments affect behavior. It is theories such as these that guide the selection of assessment instruments, the specification
Chapter 01.qxd
3/28/2007
12:58 PM
Page 19
1. PSYCHOLOGICAL CLINICAL SCIENCE
19
of treatment priorities, and the hierarchy of treatment interventions (Bootzin & Ruggill, 1988). The importance of theory in guiding applied research, as opposed to the “costly, inefficient, and limited method of trial and error” (Lewin, 1944/1951, p. 169), has been long recognized. As Lewin (1944/1951) famously said, “there is nothing so practical as a good theory” (p. 169). This point has been expanded and emphasized by Donald T. Campbell (1971/1988) in his writings on the experimenting society and by Sechrest and Bootzin (1996) in discussions of the policy implications of research and its evaluation. Theory, including small theories of treatments, is critically important for advancing knowledge in clinical psychology. In the late 1980s, academic researchers became concerned about the increasing growth of the number of psychologists from professional schools and the increasing focus of the APA on professional issues or what were often described as guild interests. These included issues such as equal reimbursement for psychologists and physicians, improved reimbursement for mental health services, licensing of psychological services, and hospitalization and prescription privileges for psychologists. In 1988, there was an attempt to reorganize APA so that academics and researchers would have increased influence in the governance of APA. The proposed reorganization failed to pass an APA membership vote, and, as a result, about 400 APA members left to form a separate organization to represent psychological science, the American Psychological Society (APS). Charles Kiesler, a former executive director of APA, was the founding president, and Janet Taylor Spence, a former president of APA, was the first elected president. Alan Kraut, then Executive Director for Science at APA, accepted the position of executive director of the newly formed APS. The APS flagship journal, Psychological Science, was edited by William Estes. Over the years, Psychological Science has become one of the leading journals of the field. APS grew quickly and had more than 5,000 members within the first 6 months and grew to more than 16,000 members by 2005. In January 2006, APS changed its name from the American Psychological Society to the Association for Psychological Science to emphasize the primary focus of the association. In a move related to the formation of APS, the American Association of Applied and Preventive Psychology (AAAPP) was formed in 1991 to provide an organizational home for research clinical and prevention psychologists. The AAAPP held its annual meetings at the same place and
Chapter 01.qxd
3/28/2007
12:58 PM
Page 20
20
BOOTZIN
time as APS. The acronym, AAAPP, was purposely chosen to reflect the earlier history of a separate organization for applied psychologists, AAAP. The organizers of the AAAPP included former APA presidents George Albee, Logan Wright, and Bonny Strickland. The AAAPP was an active organization for more than a decade, but by 2004, membership had decreased substantially, and its major continuing asset was its journal, Applied and Preventive Psychology. One likely reason for the lack of long-term success of the AAAPP was that there were other organizations, old and new, committed to the science of clinical psychology, such as Section III of Division 12 of APA and the Academy of Psychological Clinical Science (“the Academy”). The development of the Academy is discussed in a later section. Section III of the clinical psychology division (Division 12) of the APA had been founded as the Section for Clinical Psychology as an Experimental-Behavioral Science in 1966. The organizing committee, under the direction of Leonard Krasner, included prominent clinical psychology researchers Albert Bandura, Cyril Franks, Arnold Goldstein, Fred Kanfer, Peter Lang, Robert Rosenthal, Kurt Salzinger, and Irwin Sarason (Routh, 1994). In 1971, the section started an annual Distinguished Scientist Award, and the first recipient was David Shakow (Routh, 1994), a distinguished researcher in schizophrenia as well as the driving force behind the development of the Boulder Model. Twenty years later, in 1990, during Dick McFall’s presidency, Section III, although keeping ties with APA, became an independent organization, the Society for a Science of Clinical Psychology (SSCP). Membership was open to psychologists regardless of whether they belonged to APA. McFall’s presidential address, “Manifesto for a Science of Clinical Psychology,” was published in the Division 12 newsletter, The Clinical Psychologist (McFall, 1991), and has become the defining statement for the importance of a clinical psychology based on science. THE ACCREDITATION BATTLE In parallel with the substantial advances being made to integrate science and practice, the more science-based clinical psychology programs and their home departments were becoming increasingly frustrated with APA accreditation of clinical psychology programs. In the late 1980s and early 1990s, the accreditation process appeared to be so unfriendly to programs that emphasized clinical science that there were discussions about
Chapter 01.qxd
3/28/2007
12:58 PM
Page 21
1. PSYCHOLOGICAL CLINICAL SCIENCE
21
programs withdrawing from the Committee on Accreditation (CoA) of APA and setting up an alternative accreditation process under the auspices of the APS. In 1992, a Summit on Accreditation, sponsored by the APS and the Council of Graduate Departments of Psychology (COGDOP), with funding from the National Institute of Mental Health (NIMH), was held in Chicago. The steering committee consisted of Marilynn Brewer (chair), Richard Bootzin, Emanuel Donchin, Virginia O’Leary, and Richard Weinberg. At the summit, Dick McFall gave a rousing address on the need to have accreditation reflect the scientific foundation of clinical psychology. As a result of the summit, a steering committee for alternative accreditation was established, and its members were Marilynn Brewer, Emanuel Donchin, Steve Elliot, Don Fowles, Elizabeth Holloway, Richard McFall, and Virginia O’Leary. The committee was often described as working on developing “a lifeboat” in case proposed reforms in accreditation being discussed within the accreditation system were not forthcoming. The CoA accreditation structure was simultaneously restructured as the result of an agreement by the APA, COGDOP, the Council of University Directors of Clinical Psychology (CUDCP), and others. Because there was some overlap in membership between the Brewer committee and the reorganized CoA members who were developing new CoA guidelines, many of the recommendations of the Brewer committee found their way into the operating principles of the restructured CoA that were approved in 1995. Among the important principles in the new accreditation guidelines were that programs would be evaluated according to their own stated training model and that expertise did not need to be demonstrated through a checklist of courses. A danger from accreditation is the addition of new “desirable” requirements that have the inevitable effect of lengthening the training program and reducing time for learning to be a first-rate researcher. If the accreditation process remains sensitive to the needs of programs to develop scientific, as well as clinical, expertise, and supports innovation in training, as opposed to using a course-dominated checklist process of evaluation, then it can be a desirable tool for enhancing training (Bootzin, 2004). Because the new CoA appeared to be adopting standards that allowed science-oriented programs to be treated fairly within the current system, the Brewer committee went into recess, to be called back into session in
Chapter 01.qxd
22
3/28/2007
12:58 PM
Page 22
BOOTZIN
the future if needed. In 2005, the organizations that had negotiated the reorganization of the CoA in the 1990s (the APA, COGDOP, CUDCP, and others) held a conference to restructure again. Because of concerns that the clinical science programs would be disadvantaged, the Brewer committee was called out of recess, and discussions of the possibility of alternative accreditation are again in the air. With the prospect of a negotiated truce in the accreditation battle in the 1990s, clinical science moved forward dramatically on two related fronts: identifying treatments that work, and tackling issues related to the dissemination of treatment. On the first of the two fronts, David Barlow, as president of Division 12, appointed Dianne Chambless to lead a Division 12 task force to develop guidelines for identifying empirically supported treatments in 1993. The first task force report was published in 1995, and additional revisions and refinements followed (Chambless & Ollendick, 2001). The movement for identifying empirically supported treatments continued to develop and evolve in psychology as it has in medicine. The second front focused on dissemination. In any attempt to articulate a model of a science-based clinical psychology, it is important to consider how treatments can be disseminated into community settings. Dick McFall has been an advocate and innovator for training in both the acquisition of knowledge and the application of clinical science. In 1996, McFall proposed that community treatment centers might consider establishing clinics to apply psychological treatments that had been demonstrated to be effective in randomized controlled clinical trials. McFall’s innovation was what he called benchmarking; that is, the published outcomes from the efficacy literature could be used as benchmarks for outcomes in the community. Is it possible that if the community centers adopted the methods employed in research, including the same primary assessment measures, the same manuals used in the studies, and the same type of training and supervision for the therapists, they would produce the same outcomes? Many critics of the effort to identify empirically supported treatments would be likely to assert that such an effort would be bound to fail due to various considerations, including differences in patients due to comorbid problems, differences in therapists in commitment to manualized treatment, and differences in resources needed to free up therapist time to focus on specific clinical problems within a busy community treatment center (e.g., Levant, 2004).
Chapter 01.qxd
3/28/2007
12:58 PM
Page 23
1. PSYCHOLOGICAL CLINICAL SCIENCE
23
Those arguing that empirically supported treatments would not succeed in community settings were wrong, and there has now been a series of demonstrations that the same degree of improvement shown in wellcontrolled efficacy studies can be obtained in effectiveness studies in community settings (e.g., Wade, Treat, & Stuart, 1998, for panic disorder; Franklin, Abramowitz, Kozak, Levitt, & Foa, 2000, for obsessivecompulsive disorder; Tuschen-Caffier, Pook, & Frank, 2001, for bulimia; Lincoln et al., 2003, for social phobia; Merrill, Tolbert, & Wade, 2003, for depression; Weersing, Iyengar, Kolko, Birmaher, & Brent, 2006, for adolescent depression). ACADEMY OF PSYCHOLOGICAL CLINICAL SCIENCE (APCS) Building on the momentum developed by the activity of science-based clinical psychology programs in accreditation, Dick McFall organized the Bloomington Conference, “Clinical Science in the 21st Century,” cosponsored by APS and NIMH in 1994. It was attended by representatives from 25 of 35 invited academic departments. The conference authorized a steering committee (Richard McFall [chair], Richard Bootzin, Don Fowles, Robert Levenson, Beth Meyerowitz, and Gregory Miller) to develop an organization of training programs. The next year, 1995, the Academy of Psychological Clinical Science (APCS; “the Academy”) was formed as a coalition of programs, not as an organization of individuals. The Academy had its first meeting at the APS convention in New York City. Representatives of 21 North American doctoral training programs met to draft a founding mission statement and bylaws for APCS. Richard McFall was elected as the first president. The following five goals were established and are still listed on the Academy web page (http://www.psychclinicalscience.org). 1. Training: To foster the training of students for careers in clinical science research, who skillfully will produce and apply scientific knowledge. 2. Research and theory: To advance the full range of clinical science research and theory and their integration with other relevant sciences. 3. Resources and opportunities: To foster the development of, and access to, resources and opportunities for training, research, funding, and careers in clinical science.
Chapter 01.qxd
24
3/28/2007
12:58 PM
Page 24
BOOTZIN
4. Application: To foster the broad application of clinical science to human problems in responsible and innovative ways. 5. Dissemination: To foster the timely dissemination of clinical science to policymaking groups, psychologists and other scientists, practitioners, and consumers. Today, in 2006, the Academy is a coalition of 44 academic and 9 internship programs that emphasize training, application, and the advancement of knowledge in psychological clinical science. In the decade since its founding, it has had an increasingly important voice on training and accreditation issues. One salient example was a joint conference between the APCS and the NIMH held in January 2004 in Washington, DC, on training in psychological clinical science, chaired by Richard Bootzin from the APCS and Bruce Cuthbert from the NIMH. The themes that were discussed provide an agenda for the future development of psychological clinical science. They included: 1. Objectives and challenges in research training for NIH and NIMH, including (a) being part of and taking leadership in interdisciplinary teams, and (b) conducting clinical science in the context of public health (e.g., developing and disseminating empirically supported treatments for the future development of the field). 2. Exemplars of psychological clinical science training models, including among others the interdisciplinary model described by Dick McFall (2006) and being used at Indiana University, in which students are trained simultaneously in psychological clinical science and an allied basic science area such as cognitive science. 3. Evaluation and outcomes of training: What should we measure and what can we measure? 4. Dissemination and influencing the future, including: (a) Do we have current models to meet the future goals for training? (b) What steps should be done to identify such models? (c) What steps should be done to develop such models? (d) How do we go about disseminating these models? The APCS–NIMH conference provided a glimpse into the future, in which innovations in training are needed to match the exciting changes occurring in clinical science and to be responsive to the need for
Chapter 01.qxd
3/28/2007
12:58 PM
Page 25
1. PSYCHOLOGICAL CLINICAL SCIENCE
25
increased attention to dissemination and policy formulation from the developing knowledge base. CONCLUSION As mentioned at the beginning of the chapter, these are the best and worst of times for clinical psychology. Psychological clinical science is at an important crossroad. Perhaps surprisingly, this is not a new crossroad. Psychologists have struggled with how to both develop and apply psychological knowledge throughout its history. In my view, if there is a lesson to be learned from our history, it is that we can be neither satisfied with the status quo nor simply armchair critics of it. We need to focus on actions that advance knowledge, including knowledge about application. We need to work to provide whatever mechanisms are needed to support those goals. Dick McFall has been an ideal model for how to help advance a field. Time and time again, he has provided cogent analyses and, critically, has been an innovator in helping psychological clinical science advance and secure a firmer base by his efforts on accreditation, founding the Academy of Psychological Clinical Science, providing the impetus for benchmarking as a method of dissemination, and innovating with highquality cross-disciplinary training programs. We are much better prepared to take advantage of the opportunities facing us due to McFall’s many contributions to psychological clinical science. ACKNOWLEDGMENTS I thank Alan Kraut for providing details about the events surrounding the founding of APS and for providing leadership, often out of the spotlight, in the many critically important activities of APS in support of psychological clinical science. My heartfelt thanks also go to Teresa Treat, who, as one of the editors of this volume, made many insightful and helpful suggestions. REFERENCES Albee, G. W. (1969). Who shall be served? My argument with David Shakow. Professional Psychology: Research and Practice, 1, 4–7. Albee, G. W. (2000). The Boulder Model’s fatal flaw. American Psychologist, 55, 247–248.
Chapter 01.qxd
26
3/28/2007
12:58 PM
Page 26
BOOTZIN
American Psychological Association. (1948). Clinical training facilities. American Psychologist, 3, 317–318. American Psychological Association. (1949). Doctoral training programs in clinical psychology: 1949. American Psychologist, 4, 331–341. American Psychological Association. (1970). Doctoral programs in clinical psychology and counseling psychology: 1969. American Psychologist, 25, 1049–1050. American Psychological Association. (1982). Report of the task force on the evaluation of education, training and service in psychology. Washington, DC: Author. Bandura, A. (1961). Psychotherapy as a learning process. Psychological Bulletin, 58, 143–159. Bandura, A. (1969). Principles of behavior modification. New York: Holt, Rinehart, & Winston. Benjamin, L. T., Jr. (2005). A history of clinical psychology as a profession in American (and a glimpse at its future). Annual Review of Clinical Psychology, 1, 1–30. Bootzin, R. R. (2004). Clinical psychologists in academia. In J. M. Darley, M. P. Zanna, & H. L. Roediger III (Eds.), The complete academic (pp. 329–344). Washington, DC: American Psychological Association. Bootzin, R. R., & Ruggill, J. S. (1988). Training issues in behavior therapy. Journal of Consulting and Clinical Psychology, 56, 703–709. Boring, E. G. (1957). A history of experimental psychology (2nd ed.). New York: Appleton-Century-Crofts. Campbell, D. T. (1969). Reforms as experiments. American Psychologist, 24, 409–429. Campbell, D. T. (1988). The experimenting society. In E. S. Overman (Ed.), Methodology and epistemology for social science: Selected papers (pp. 290–314). Chicago: University of Chicago Press. (Original work published 1971) Chambless, D. L., & Ollendick, T. H. (2001). Empirically supported psychological interventions: Controversies and evidence. Annual Review of Psychology, 25, 685–716. Cronbach, L. J. (1957). The two disciplines of psychology. American Psychologist, 12, 671–684. Eysenck, H. J. (1949). Training in clinical psychology: An English point of view. American Psychologist, 4, 173–176. Eysenck, H. J. (1952). The effects of psychotherapy: An evaluation. Journal of Consulting Psychology, 16, 319–324. Fernberger, S. W. (1943). The American Psychological Association—1892–1942. Psychological Review, 50, 33–60. Franklin, M., Abramowitz, J., Kozak, M., Levitt, J., & Foa, E. (2000). Effectiveness of exposure and ritual prevention for obsessive-compulsive disorder: Randomized compared with nonrandomized samples. Journal of Consulting and Clinical Psychology, 68, 594–602. Garvey, C. R. (1929). List of American psychology laboratories. Psychological Bulletin, 26, 652–660. Griffith, C. R. (1922). Contributions to the history of psychology—1916–1921. Psychological Bulletin, 19, 411–428. Hilgard, E. R., Kelly, E. L., Luckey, B., Sanford, R. N., Shaffer, L. F., & Shakow, D. (1947). Recommended graduate training program in clinical psychology. American Psychologist, 2, 539–558.
Chapter 01.qxd
3/28/2007
12:58 PM
Page 27
1. PSYCHOLOGICAL CLINICAL SCIENCE
27
Hobbs, N. (1964). Mental health’s third revolution. American Journal of Orthopsychiatry, 34, 822–833. Hothersall, D. (1995). History of psychology (3rd ed.). New York: McGraw-Hill. Katkin, E. (1982). On reliable knowledge and the proliferation of professional schools of psychology. Clinical Psychologist, 36, 9–11. Kazdin, A. E., Kratochwill, T. R., & VandenBos, G. R. (1986). Beyond clinical trials: Generalizing from research to practice. Professional Psychology: Research and Practice, 17, 391–398. Kelly, G. A. (1955). The psychology of personal constructs. New York: Norton. Kihlstrom, J. F., & Kihlstrom, L. C. (1998). Integrating science and practice in an environment of managed care. In D. K. Routh & R. J. DeRubeis (Eds.), The science of clinical psychology: Accomplishments and future directions (pp. 281–294). Washington, DC: American Psychological Association. Levant, R. F. (2004). The empirically validated treatments movement: A practitioner/educator perspective. Clinical Psychology: Science and Practice, 11, 219–224. Lewin, K. (1951). Problems of research in social psychology. In D. Cartwright (Ed.), Field theory in social science: Selected theoretical papers by Kurt Lewin (pp. 155–169). New York: Harper & Brothers. (Original work published 1944) Lincoln, T., Rief, W., Hahlweg, K., Frank, M., von Witzleben, I., Schroeder, B., et al. (2003). Effectiveness of an empirically supported treatment for social phobia in the field. Behaviour Research and Therapy, 41, 1251–1269. Lipsey, M. W. (1990). Theory as method: Small theories of treatments. In L. Sechrest, E. Perrin, & J. Bunker (Eds.), Research methodology: Strengthening causal interpretations of nonexperimental data (pp. 44–53). Washington, DC: U.S. Department of Health and Human Services. McFall, R. M. (1985). Nonbehavioral training for behavioral clinicians. Behavior Therapist, 8, 27–30. McFall, R. M. (1991). Manifesto for a science of clinical psychology. Clinical Psychologist, 44, 75–88. McFall, R. M. (1996). Consumer satisfaction as a way of evaluating psychotherapy: Ecological validity and all that versus the good old randomized trial (panel discussion). Sixth annual convention of the American Association of Applied and Preventative Psychology, San Francisco. McFall, R. M. (2006). Doctoral training in clinical psychology. Annual Review of Clinical Psychology, 2, 21–49. Meehl, P. E. (1954). Clinical versus statistical prediction: A theoretical analysis and a review of the evidence. Minneapolis: University of Minnesota Press. Merrill, K. A., Tolbert, V. E., & Wade, W. A. (2003). Effectiveness of cognitive therapy for depression in a community mental health center: A benchmarking study. Journal of Consulting and Clinical Psychology, 71, 404–409. Mischel, W. (1968). Personality and assessment. New York: Wiley. Oltmanns, T., & Mineka, S. (1992). Morton Prince on anxiety disorders: Intellectual antecedents of the cognitive approach to panic? Journal of Abnormal Psychology, 101, 607–610. Prince, M. (1912). Report by Morton Prince. In M. Prince & J. J. Putnam (Eds.), A clinical study of a case of phobia: A symposium. Journal of Abnormal Psychology, 7, 259–276.
Chapter 01.qxd
28
3/28/2007
12:58 PM
Page 28
BOOTZIN
Rogers, C. R. (1942). Counseling and psychotherapy. Boston: Houghton Mifflin. Rotter, J. B. (1954). Social learning and clinical psychology. Englewood Cliffs, NJ: Prentice Hall. Rotter, J. B. (1971). On the evaluation of methods of intervening in other people’s lives. Clinical Psychologist, 24, 1–2. Routh, D. K. (1994). Clinical psychology since 1917: Science, practice, and organization. New York: Plenum. Sarason, S. B. (1988). The making of an American psychologist: An autobiography. San Francisco: Jossey-Bass. Sears, R. R. (1947). Clinical training facilities: 1947. American Psychologist, 2, 199–205. Sechrest, L. B., & Bootzin, R. R. (1996). Psychology and inferences about public policy. Psychology, Public Policy, and Law, 2, 377–392. Tuschen-Caffier, B., Pook, M., & Frank, M. (2001). Evaluation of manual-based cognitive-behavioral therapy for bulimia nervosa in a service setting. Behaviour Research and Therapy, 39, 299–308. Wade, W. A., Treat, T. A., & Stuart, G. L. (1998). Transporting an empirically supported treatment for panic disorder to a service clinic setting: A benchmarking strategy. Journal of Consulting and Clinical Psychology, 66, 231–239. Watson, R. I., Sr. (1978). The great psychologists (4th ed.). New York: J. B. Lippincott. Weersing, V. R., Iyengar, S., Kolko, D. J., Birmaher, B., & Brent, D. A. (2006). Effectiveness of cognitive-behavioral therapy for adolescent depression: A benchmarking investigation. Behavior Therapy, 37, 36–48. Wittling, W. (1972). Kraepelin, Emily, B. In H. Eysenck (Ed.), Encyclopedia of psychology (Vol. 2, pp. 172–173). New York: Herder & Herder. Woody, R. H., & Robertson, M. H. (1997). A career in clinical psychology: From training to employment. Madison, CT: International Universities Press. Wolfle, D. (1946). The reorganized APA. American Psychologist, 1, 3–6. Wright, L. (1983). Please don’t tell my mother I’m a clinical psychologist—She still thinks I play piano in a brothel. Clinical Psychologist, 36, 49–51.
Chapter 02.qxd
4/6/2007
2:28 PM
Page 29
2 The Epistemological and Ethical Dimension of Clinical Science William O’Donohue University of Nevada, Reno
Scott O. Lilienfeld Emory University
McFall’s (1991) article, “Manifesto for a Science of Clinical Psychology,” is a classic articulation of an attempted resolution to a problem that has plagued, and continues to plague, clinical psychology and other behavioral health professions: the relationship between science and practice. Essentially, the problem (phrased as a question) is: What is the proper relationship between science and clinical practice? To better understand this problem, we examine alternative forms of the problem statement. In doing so, we also discuss some of the implications and aspects of McFall’s Manifesto, particularly: 1. The problems posed by the lack of clarity in meta-science studies regarding what science is. 2. The problems posed by the fact that clinical psychology’s conception of science has produced slow progress (see Meehl, 1978). 3. The notion that McFall’s Manifesto, in conjoining descriptive (“is”) with ethical (“ought”) statements, creates a blended construct that 29
Chapter 02.qxd
30
4/6/2007
2:28 PM
Page 30
O’DONOHUE AND LILIENFELD
may be thought of as epistemic duties (O’Donohue & Henderson, 1999). 4. The question of whether this discussion has focused excessively on the categories of scientist and practitioner to the neglect of an important third possibility: that of technologist or behavioral engineer. In addition, a critical technical question may be for researchers to focus on developing and evaluating quality improvement and management information systems for delivery systems (Cummings, O’Donohue, & Ferguson, 2002; Fisher & O’Donohue, 2006). Such systems can supplement the historical and contemporary research focus on developing individual assessment devices or therapies. 5. The lack of any coherent statement of an alternative epistemology that could support or provide the epistemic warrant for clinical practice (Stricker & Trierweiler, 1995). 6. An essential compatibility between the roles of science and quality improvement (Walton, 1986), which McFall perceptively discusses. We do not review here the evidence for the problematic relationship between science and practice in clinical psychology because such evidence could easily occupy an entire volume (see Lilienfeld, Lynn, & Lohr, 2003; Singer & Lalich, 1996). There are far too many examples of shoddy, pseudoscientific practices in the mental health field for us to cover here (see e.g., Dawes, 1994; Lilienfeld et al., 2003; O’Donohue & Bradley, 1999; Wood, Nezworski, Lilienfeld, & Garb, 2003). Some of these practices, such as rebirthing (which led to the death of a 10-yearold child in Colorado), can be downright dangerous (see Mercer, Sarner, & Rosa, 2003), whereas others, such as Thought Field Therapy (see Gaudiano & Herbert, 2000), mainly waste time and resources that could be spent in more effective therapies (economists refer to this latter consequence as opportunity cost). Even when harmless by themselves, such treatments can needlessly prolong the pain and dysfunction that the originating clinical condition is causing. To further acquire an understanding of the seriousness and magnitude of this problem, we offer a bold assertion. Specifically, at this point in time, there is no persuasive evidence to disprove the following statement: It is not even clear whether evidence-based practice is more prevalent than pseudoscientific practice. Although most fields and industries have thoroughly embraced quality improvement practices (Walton, 1986), clinical psychology lags woefully behind
Chapter 02.qxd
4/6/2007
2:28 PM
Page 31
2. DIMENSIONS OF SCIENCE
31
and has done little to develop and adopt systematic quality improvement systems. We argue that this should become the top priority of the next decade. THE PROBLEM STATEMENT: THE RELATIONSHIP BETWEEN SCIENCE AND PRACTICE What exactly is the problem? The problem may be better delimited by considering alternative overlapping statements: 1. What should the clinician’s evidence be when he or she practices? For example, when he or she uses a treatment, what evidence should there be of its effectiveness? 2. Does science provide the best evidence—the knowledge and the “know-how” for clinical practice? 3. Is scientific evidence the only kind of evidence that can provide the warrant for clinical practice, or are there other kinds of evidence? 4. Where is the evidence best found? Is it found in experimental cognitive psychology, biological psychiatry, outcome research in clinical psychology, or some combination of these? 5. What exactly is the evidential burden that needs to be met? How much evidence is sufficient to justify a specific clinical decision or practice? 6. Let us imagine granting that science is the best or even the only evidence to support clinical decisions and practice. Even with this stipulation, given that meta-studies of science provide widely different answers to what science actually is, what implications does this equivocality hold for the relationship between science and clinical practice? 7. Are there any legitimate exceptions to the general answers or rules associated with the questions already listed? 8. The APA’s ethical code seems to recognize clinical experience and professional knowledge (e.g., Stricker & Trierweiler, 1995). What exactly are these constructs? Are they in fact legitimate forms of knowledge? Can they pass the philosophical muster of a sophisticated epistemologist? Can they be relevant to the justification of clinical decisions and practice? If so, what is their relationship to scientific knowledge (e.g., what overrides what?)?
Chapter 02.qxd
32
4/6/2007
2:28 PM
Page 32
O’DONOHUE AND LILIENFELD
9. What is the economic value proposition involved in therapy? More specifically, what kinds of evidence should be available to payors (including potential clients) to make rational decisions regarding how much they should be willing to pay? 10. If a clinician is not practicing according to the best evidential standards, what should happen? Is this malpractice, a psychological error, or both? How well is such practice detected and adjudicated? 11. McFall (1991) argued for the importance of quality improvement. A principle of quality improvement (Walton, 1986) is that systems, not individuals, are blamed for mistakes. What can the profession do to create systems that support better evidence-based practice? This problem or set of problems can be regarded as the most fundamental to our field because it defines what constitutes knowledge in our field and what constitutes a legitimate warrant for justifying a clinical decision— indeed for justifying any claim for expertise. The reader should note that this problem (as well as McFall’s Manifesto) is not a scientific problem, but rather a meta-problem and a meta-statement. Some aspects of this problem are strategic, whereas others are ethical. Choices concerning these questions have important consequences. Problematic answers can result when epistemic bars are set too low because this can produce practice that is harmful or less efficient than need be, or by setting standards too high (O’Donohue, 2003), eliminating therapies and tools that could actually help people, but that are not utilized because of a supposed lack of warrant.
McFALL’S RESPONSE TO THE PROBLEM In his classic article, McFall (1991) proposed one cardinal principle and two corollaries: The Cardinal Principle: Scientific clinical psychology is the only legitimate and acceptable form of clinical psychology. (p. 76) First Corollary: Psychological services should not be administered to the public (except under strict experimental control) until they have satisfied these four minimal criteria:
Chapter 02.qxd
4/6/2007
2:28 PM
Page 33
2. DIMENSIONS OF SCIENCE
1. 2. 3. 4.
33
The exact nature of the service must be described clearly. The claimed benefits of the service must be stated explicitly. These claimed benefits must be validated scientifically. Possible negative side effects that might outweigh any benefits must be ruled out empirically. (p. 79)
Second Corollary. The primary and overriding objective of doctoral training programs in clinical psychology must be to produce the most competent clinical scientists possible. (p. 82).
This is perhaps the clearest and most unequivocal statement of a commitment to a scientific epistemology for clinical psychology (in contrast, say, to the murky APA ethics code). McFall (1996, 2000) also refined and extended his basic stance in a series of follow-up articles. In these articles, he did much to clarify and elaborate his position. He shows his affinity for neo-Popperian critical rationalism as the epistemology underlying science. He improved his Criterion 4 in response to a criticism to describe a net benefit (i.e., listing all negative and positive benefits and calculating the net benefit instead of focusing simply on outweighing negative benefits). He strongly asserted the pragmatics of informed consent: We should do no harm by fully informing our clients of what they are buying. The main requirement is: Considering my alternatives, please educate me concerning the evidence regarding how the probabilities change for each course of action available to me. This question must be answered carefully, and no practitioner, under McFall’s epistemic system, is allowed to fudge numbers, slough off the question, or argue (always invalidly) solely from anecdotal experience. With respect to the lattermost “escape hatch,” psychologists must realize that they cannot form valid probability statements in this fashion, given the cognitive biases and heuristics that render anecdotes highly fallible as evidence for truth claims (see Garb, 1996; Meehl, 1997; Tversky & Kahneman, 1974). Even the most cursory overview of the history of medicine reveals the disastrous track record of physiological interventions that turned out to be either useless or harmful, including leeching, blood-letting, blistering, and prefrontal lobotomy. All of these treatments were derived largely or entirely from clinical anecdotes and subjective impressions (Grove & Meehl, 1996). McFall also provides a thoughtful extended discussion of the generalizability question—a point that Quine (1974) viewed as fundamental to epistemology—in that no incident is exactly like past incidents. As a consequence, how does one form reasonable judgments to say that this
Chapter 02.qxd
34
4/6/2007
2:28 PM
Page 34
O’DONOHUE AND LILIENFELD
current situation is sufficiently like past situations that it can be used to form probability statements? The point is that this dilemma is in no way unique to clinical science because it applies to probability statements in all domains of science. A key demand is to be able to pick out features that actually impact probabilities from those that do not. McFall also suggests that we must come to grips with our lack of knowledge in some areas, and that an honest appraisal of this limitation could suggest that at times we should not offer our clinical services because we have no relevant knowledge to offer. We may want to help, but we possess no expertise that would indicate that we can change probabilities in the desired direction. This recognition of our limitations can be emotionally challenging for clinicians to accept—after all, we want to alleviate the client’s pain—but the twin realizations that we have no expertise and that interventions may turn out to be harmful (see Lilienfeld, Fowler, Lohr, & Lynn, in press) support McFall’s simple, but radical, point (see e.g., O’Donohue & Bradley, 1999, for an extension of this point to custody evaluations). The major way the frustration and disappointment of this dilemma can be mitigated is by a thorough commitment to scientific research and quality improvement systems. In this way, knowledge and technology can be acquired to maximize the chances that this state of affairs is temporary. McFall (1996), in response to criticisms, mainly from Peterson (1996), proposed two additional corollaries that, because of their importance and clarity, are quoted here at length: Third Corollary: A scientific epistemology differentiates science from pseudoscience. According to this epistemology:
1. Skepticism is the appropriate and legitimate stance toward all claims about psychological services. 2. The burden of proof regarding the validity of a psychological service rests squarely with the proponents of the services. 3. Skeptics are not required to prove the negative case. The absence of negative evidence is not equivalent to the positive support for the efficacy of a service. 4. Untested services do not deserve special status; the world is full of untested notions. Skeptics must treat untested services as invalid until convinced otherwise by the empirical evidence. 5. Claims about outcomes and theoretical explanations for those outcomes must be tested separately. For example, when evidence shows
Chapter 02.qxd
4/6/2007
2:28 PM
Page 35
2. DIMENSIONS OF SCIENCE
35
that a treatment is beneficial, it is a logical fallacy (“affirming the consequent”) to conclude from this finding that the theoretical explanation for this effect also is correct. 6. Results are specific. Positive results cannot automatically be generalized to untested problems, stimuli, methods, therapists, patients, measures, conditions, and so on. Small changes sometimes produce dramatically different results. 7. Decisions based on nomothetic evidence are more valid, on the whole, than idiographic decisions based on clinical intuition and judgments. (p. 14) Fourth Corollary: The most caring and humane psychological services are those that have been shown empirically to be the most effective, efficient, and safe. Genuine caring requires the highest level of scientific rigor. Anything less, no matter how well intentioned, is likely to be less beneficial for the individuals served.
1. Scientific rigor requires that assessment and treatment protocols be specified in as much detail as possible, validated as specified in the protocol, followed faithfully in clinical applications, and monitored objectively—both in administration and results—in individual cases. 2. The most compassionate procedure for choosing a protocol is one that promotes a fully informed choice based primarily on a careful review of the scientific evidence and secondarily on a conservative appraisal of the local circumstances. 3. The overriding concerns of service providers must be to avoid doing harm or making matters worse. Withholding untested and unproven services usually is the most caring and responsible choice. (p. 14) In these additions, McFall nicely describes most of the substance of what might be called the professional’s epistemic duties (O’Donohue & Henderson, 1999). We next comment on some of the subproblems and clarify some issues associated with McFall’s position in an attempt to make further progress toward McFall’s superordinate goal: consistent improvements in the quality of clinical practice. SUBPROBLEM 1: THERE IS NOT MUCH AGREEMENT REGARDING WHAT CONSTITUTES SCIENCE Philosophy of science and other meta-scientific disciplines have not yielded a consensus about what science is. We review two influential accounts of
Chapter 02.qxd
4/6/2007
36
2:28 PM
Page 36
O’DONOHUE AND LILIENFELD
science to demonstrate this wide divergence: those of Sir Karl Popper and Thomas Kuhn. We encourage the reader to consult summaries of other accounts of science, such as those of Lakatos, Laudan, Feyerabend, and Gross, to acquire an appreciation of the wider diversity of views (see O’Donohue, 1989; O’Donohue, Ferguson, & Naugle, 2003, for introductions). This lack of consensus is of concern because it is not clear what clinical scientists must do to be scientific. Some might contend this issue is not of concern because science, like pornography, is usually easily recognized. We believe this view is mistaken. First, there are well-known disputes in psychology about the scientific status of certain theoretical and research programs. O’Donohue and Halsey (1997) showed that Freud, Rogers, Ellis, and Skinner all thought their programs were first-rate science, but that the programs of their competitors were not. We also witness intense and often acrimonious debates in the literature about the scientific versus pseudoscientific status of certain current therapy movements (see Lilienfeld et al., 2003, for a summary). In addition, we all are familiar with the problems of the unreliability and therefore the lack of validity of journal peer reviews (e.g., Cicchetti, 1991). But most important, we see the “slow progress of soft psychology,” to use Meehl’s (1978) felicitous phrase, and wonder, along with Meehl, whether we as psychologists have a problematic conception of the scientific method. When we turn to meta-scientific studies, we do not obtain a univocal answer. One way out of this conundrum may be that one merely needs to demonstrate that what one is doing is consistent with some account of science. But this superficially appealing solution will prove too broad given the variety of accounts and possible overpermissibility of these arguments by analogy. Popper Because McFall (1991) shows a considerable affinity for the falsificationist account of science of Sir Karl Popper, we briefly describe Popper’s account of science. It is important to bear in mind that, although Popper has probably been the most influential philosopher of science in the 20th century (the other contender being Thomas Kuhn, whom we discuss next), he has been roundly criticized by many philosophers of science, particularly Feyerabend, Lakatos, Kuhn, and Putnam. Popper argued that all science should focus on error elimination. Scientists are free to conjecture, and these conjectures may be influenced by anything, including myths, dreams, speculations, and ideologies. Nevertheless, conjectures
Chapter 02.qxd
4/6/2007
2:28 PM
Page 37
2. DIMENSIONS OF SCIENCE
37
should be bold or risky (i.e., they should generate predictions that rule out many empirical possibilities). Point predictions are bold because they rule out all other values (i.e., 9.8 m/sec2 is bold because it rules out all other numerical values). Because they should not want to retain errors in their ways of believing, good scientists rigorously criticize their theories by determining whether the states of affairs the theories rule out actually occur. In a simple example, if the prediction derived from a scientist’s theory is “Ministers never swear,” the scientist should actively seek out a cussing minister because this empirical state of affairs would falsify the generalization and therefore eliminate an erroneous belief. Better scientists examine ministers on the golf course after a difficult bunker shot and ministers who just hit their thumbs with a hammer, and poorer scientists only observe ministers in the pulpit on Sunday mornings. Popper argued that tests should be severe (i.e., they should look where falsifying instances are most likely to show up). Given falsifying information, scientists revise their account (hopefully by making an even riskier prediction) and then test again, ad infinitum. Popper eventually argued that his philosophy of science was consistent with evolution by natural selection, with bold conjectures being selected out by the environment. According to this evolutionary analogy, survivors are not necessarily true (proven), just good enough to survive for the time being. This evolutionary epistemology shares notable similarities with the epistemologies of other influential philosophers in the 20th century, such as Quine, Skinner, and Campbell (see O’Donohue, Lloyd, & Ferguson, in preparation). Kuhn Kuhn was often seen as Popper’s chief rival and appears to be the philosopher of science whom psychologists know and like best (O’Donohue, 1993). Kuhn, who viewed himself primarily as a historian of science (actually of a few sciences, particularly astronomy and physics), argued that Popper’s account is inconsistent with the actual historical behavior of scientists. Kuhn developed what might be thought of as a sociology of science. In Kuhn’s account, the first stage in the development of a discipline is immature science. According to Kuhn (1970), immature science is characterized by “frequent and deep debates over legitimate methods, problems, and standards of solution, though these serve rather to define schools than to produce agreement” (pp. 47–48). During this stage, there is no consensus, no agreed-on facts or method, little or no agreement on what subject matter is worthy of research (i.e., ontology), and a proliferation
Chapter 02.qxd
38
4/6/2007
2:28 PM
Page 38
O’DONOHUE AND LILIENFELD
of competing schools of thought (Bird, 2000). The second stage in the cyclical process of scientific change is normal science. During normal science, the field demonstrates cumulative progress. Moreover, normal science denotes a consensus in the scientific community: There are agreed-on facts and methods, there is agreement on what subject matter is worthy of research, and what were once competing schools of thought usually settle into a single paradigm. Nevertheless, there are at least 21 different meanings of the term paradigm (see Masterman, 1970), rendering the interpretation of the writings of Kuhn (and especially Kuhn’s followers) obscure. Generally speaking, the term paradigm is used in two distinct ways: as (a) the entire constellation of beliefs, values, techniques, and so on, shared by the members of a given community (or what Kuhn later called the disciplinary matrix); and (b) one element in that constellation, the concrete puzzle solutions, that, employed as models or examples, can replace explicit rules as a basis for the solution of the remaining puzzles of normal science (Kuhn, 1996). This “settling into a single paradigm” usually occurs in the wake of “some notable scientific achievement” (Kuhn, 1974, p. 460). According to Kuhn, this does not mean that the paradigm has to “explain all of the facts [that could confront it]”; a paradigm is required only to explain those facts deemed most important by a given community (Kuhn, 1996). Paradigmatic science is largely a conservative endeavor, consisting of “mopping-up operations” and “puzzle-solving” (Kuhn, 1962, pp. 24, 35–42). Both of these operate to “broaden and deepen the explanatory scope” of a paradigm (Gholson & Barker, 1985). Specifically, “mopping up” and “puzzle solving” involve: (a) striving to bring a paradigm “into closer agreement with nature” (Kuhn, 1963, p. 360); (b) attempts at increasing the accuracy and scope of the paradigm so as to include new phenomena (Kuhn, 1996, p. 25; Losee, 1980); and (c) better articulating the “paradigm theory … resolving some of its residual ambiguities” (Kuhn, 1996, p. 27).1 Normal science proceeds unabated just as long as the paradigm satisfactorily explains the phenomena to which it is applied (Losee, 1980). However, new and unsuspected phenomena are often uncovered by scientific research (Kuhn, 1996). Normal science is almost inevitably confronted with anomalous data (Hoyningen-Huene, 1993). These anomalous data do not necessarily provide refuting counterexamples of the prevailing paradigm. Anomalies 1
Hertz’s refinement of Newton’s Principia Mathematica is one such example (Bird, 2000).
Chapter 02.qxd
4/6/2007
2:28 PM
Page 39
2. DIMENSIONS OF SCIENCE
39
might arise due to instrumental or human error. In fact, when such anomalies initially arise, it is the scientist who is to blame, not the paradigm (Bird, 2000). Kuhn (1962) stated: Normal science … often suppresses fundamental novelties because they are necessarily subversive of its basic commitments … [however], when the profession can no longer evade anomalies that subvert the existing tradition of scientific practice [the paradigm is in crisis]. (pp. 5–6; bracketed additions added)
Crisis. When enough anomalies accumulate, scientists begin questioning whether the dominant paradigm is really appropriate; the prevailing paradigm is said to be in a state of crisis (Laudan, 1977). During a crisis, blame is shifted from scientists to the paradigm, and a “sense of professional insecurity is generated” (Bird, 2000, p. 43). At times of crisis, there is a “blurring of a paradigm and the consequent loosening of the rules for normal research” (Kuhn, 1970, p. 84). When this blurring occurs, it becomes patent that normal science cannot continue as before (Hoyningen-Huene, 1993). The paradigm has “drowned in a sea of anomalies,” and a point is reached when the old paradigm must be discarded, making way for the formulation of a new paradigm (Kuhn, 1996). Contrary to the steady progress of normal science, this replacement of one paradigm for another is a cataclysmic or revolutionary event (Gholson & Barker, 1985). Revolutionary or Extraordinary Science. In the following quotation, Kuhn (1962) defined what he meant by scientific revolution: Scientific revolutions are here taken to be those non-cumulative developmental episodes in which an older paradigm is replaced in whole or in part by an incompatible new one. (p. 91)
By incompatible, Kuhn (1962) suggested that “after a revolution scientists are responding to a different world” (p. 111), rendering competing paradigms largely incommensurable. Kuhn called this psychological phenomenon a Gestalt switch, akin to what observers experience with the famous Necker Cube, which is seen as alternating from one qualitatively different (and mutually exclusive) percept to another. There are then debates between defenders of the old paradigm and promoters of the new. Some of the defenders of the old paradigm remain unpersuaded because they cannot make the Gestalt switch. They remain unconvinced until they retire or die, whereas the proponents of the new paradigm attempt
Chapter 02.qxd
40
4/6/2007
2:28 PM
Page 40
O’DONOHUE AND LILIENFELD
to conduct normal science by using the new paradigm as an exemplar for problem solving. SHOULD CLINICAL SCIENTISTS BE POPPERIANS OR KUHNIANS? We do not discuss this question in any detail here because it actually should be enlarged to read something like: Should clinical scientists be Positivists, Popperians, Kuhnians, Lakatosians, Laudians, Feyerabendians, Grossians, … etc.? Because there are numerous accounts of science that are conflicting in important aspects, to say that clinical psychologists “should be scientific” opens a host of important subproblems. There needs to be more scholarly work on this question. The pragmatic question should be: To what extent does a certain vision of science result in progress for certain kinds of problems? We hypothesize, for example, that the progress associated with behavior analysis was mainly Kuhnian (i.e., behavior analysts attempted to use Skinner’s work as the exemplar of effective problem solving and extended that to other clinical problems). It is also difficult to find many examples of ardent falsification even in what is regarded as good science in clinical psychology. For example, most developers of therapies seem to be protecting their intellectual babies from the arrows of modus tolens, rather than exposing them to falsification attempts (for a striking example, see the ad hoc maneuvers often used by the proponents of eye movement desensitization and reprocessing in response to findings demonstrating that the hypothesized mechanisms of action of this treatment are incorrect; Herbert et al., 2000). This result, from a meta-scientific perspective, would be predicted by Lakatos, but decried by Popper. Nevertheless, McFall is correct that this problem is most egregious in the poorest clinical practice, in which feedback is not sought and virtually all outcomes are seen as confirmations (see Dawes, 1994; Lilienfeld et al., 2003, for a plethora of examples). SUBPROBLEM 2: LEADING DEVELOPERS OF PSYCHOTHERAPIES HAVE NOT PRODUCED A CONSENSUS If the equivocality of what constitutes science were not already sufficient cause for concern, the true proportions of the problem emerge when one
Chapter 02.qxd
4/6/2007
2:28 PM
Page 41
2. DIMENSIONS OF SCIENCE
41
considers that psychologists have produced indigenous accounts of what constitutes science. For example, O’Donohue and Halsey (1997) tried to capture the differing views of science of Skinner, Rogers, and Ellis, all of whom obviously have exerted a significant influence on clinical psychology. For example, Freud claimed that his method was scientific. This assertion may surprise some who think that psychoanalysis can be easily dismissed as unscientific or even pseudoscientific (Cioffi, 1998). Freud characterized his scientific method as deducing useful constructs from careful clinical observations and verifying that these are useful, accurate constructs because they are utilized in therapeutic successes. Freud (1925) stated: The view is often defended that sciences should be built up on clear and sharply defined basal concepts. In actual fact no science, not even the most exact, begins with such definitions. The true beginning of scientific activity consists rather in describing phenomena and in proceeding to group, classify, and correlate them. Even at the stage of description, it is not possible to avoid applying certain abstract ideas to the material in hand, ideas derived from various sources and certainly not the fruit of the new experience only. (p. 60)
Rogers (1968), in contrast, while attempting to synthesize empathic understanding to fully explore the private world of meanings with more conventional accounts of science, stated: Unless I am willing to define these terms operationally, to design a research which will put them to the test, or to encourage others to design such researches; and unless the various extraneous variables are controlled, and the findings support the hypotheses, then we are only in the realm of pattern perception and not of confirmation. (p. 67)
Skinner, having influenced behavioral analysis and behavior therapy, has been criticized by some as being scientistic (i.e., as adhering fetishistically to a scientific methodology even when it is not appropriate), rather than scientific. But his initial claims for scientific status were greeted by some with skepticism because of his single-subject methodology. Skinner’s recommendations for successful scientific practice include a focus on an individual organism, the use of rate of responding as the dependent variable, and an eschewal of inferential statistics, all of which were and are fairly unconventional. We wish to draw four conclusions from this cursory review of the claims for scientific status of these three scholars who have had substantial influence on clinical psychology:
Chapter 02.qxd
42
4/6/2007
2:28 PM
Page 42
O’DONOHUE AND LILIENFELD
1. They do not appear to be influenced directly by any of the classic philosophers of science, instead developing indigenous accounts of the scientific status of their psychologies (Albert Ellis, who cites explicitly the influence of Popper and the neoPopperian W. W. Bartley, is an exception). 2. These accounts of what constitutes science differ widely from one another and widely from the philosophers of science discussed earlier. Feyerabend (and Chairman Mao) would be happy, as a thousand flowers are blooming. 3. These accounts give their psychotherapist adherents some warrant for claiming scientific status for their endeavors. The demarcation line cannot be easily drawn by stating, for example, that X performs psychoanalysis; therefore X is not scientific. 4. More work needs to be conducted on the meta-science of clinical psychology to sort out such questions as: (a) What are legitimate meta-scientific criteria for clinical science? (b) What are the problems with any of the meta-scientific claims of a school of psychotherapy? SUBPROBLEM 3: PSYCHOLOGY’S CONCEPTION OF SCIENCE HAS PRODUCED SLOW PROGRESS Psychology’s adoption of a certain content of science has produced painfully slow progress (Meehl, 1978). There are few solved problems in the soft areas of psychology, including clinical and counseling psychology. This state of affairs may lead some to stray from the fold of a clinical science approach. Meehl (1978) eloquently described some of the possible reasons for this slow progress, including our field’s overreliance on statistical significance testing and tolerance for weak (nonrisky) tests of theoretical predictions. However, there is another possibility. The philosopher of science, Thomas Nickles, suggested that science should be flexible because scientific methodologies need to connect with their subject matter in a manner in which information is gained. The metaphor of a net may be a good one: If the openings of nets are too big, nothing is retained; if they are too small, everything (wheat and chaff) is held. Therefore, the net has to be the right size to capture what is wanted. The slow progress of soft psychology may give rise to the following question: Do we have the wrong view of science (i.e., the wrong-sized net)? A fact that seems to escape many discussions of science among clinical psychologists (many of whom refer to “the” scientific method as a
Chapter 02.qxd
4/6/2007
2:28 PM
Page 43
2. DIMENSIONS OF SCIENCE
43
monolithic truth-gathering device) is that, because various sciences and their subdisciplines use different methodologies, science is different in each of these disciplines. The methodology of evolutionary biology is different than cellular biology, the methodologies of astronomy are different than the methodologies of atomic physics, the methodology of anthropology is different than that of physical chemistry, and so on. Nickles’ point is that methodologies and views of science need to be flexible and creative, and their worth is determined by their problem-solving efficacy. It would be useful to further understand the history of clinical psychology and draw conclusions about what views of science and methodologies lead to what problem-solving efficacy. Nickles’ point would suggest that we may need to expand or revamp our notions of reasonable scientific methodologies (O’Donohue & Lloyd, 2005). SUBPROBLEM 4: IS THERE SOME OTHER KIND OF KNOWLEDGE CALLED PROFESSIONAL KNOWLEDGE? WHAT IS PROFESSIONAL KNOWLEDGE? Needless to say, many clinical psychologists present themselves as experts in both the therapy room (Dawes, 1994) and courts of law (McCann, Shindler, & Hammond, 2003) without legitimate claim to such expertise. As noted earlier, many clinical psychologists draw on their vast reservoir of anecdotes (read: clinical experiences) to make the case for their credentialed knowledge (Meehl, 1997). Perhaps understandably, many laypersons find the basis for this claim to be compelling because they (erroneously) assume that greater clinical experience equals greater expertise. Yet decades of research on clinical judgment and prediction suggest otherwise (Garb, 1998). Moreover, as a wise person once noted, the plural of anecdote is not fact. Multiple anecdotes may often be helpful in the context of discovery and in generating fruitful hypotheses to be tested systematically, but they are rarely helpful in the context of justification (i.e., systematic hypothesis testing; Reichenbach, 1938). The lone—and usually extremely rare— exception to this rule is an existence proof, which falsifies the proposition that a given phenomenon does not exist. Just as it requires only one black swan to demonstrate that the proposition “all swans are white” is false, it requires only one convincingly corroborated case of a recovered memory of early child abuse to demonstrate that the proposition “recovered memories of abuse do not exist” is false (as of this writing, no unambiguous existence proof of recovered memories appears to exist; see McNally, 2003).
Chapter 02.qxd
44
4/6/2007
2:28 PM
Page 44
O’DONOHUE AND LILIENFELD
If anecdotes rarely, if ever, form an adequate foundation for professional knowledge, what is the alternative? Our answer is straightforward: Practitioners must base their actions and probability statements on the best clinical science currently available, and they must be certain not to exceed the boundaries of such science. Moreover, when they do go beyond the existing database, such as when hazarding extensive generalizations from existing research, they must be explicit with their clients when doing so. Of course, a certain degree of generalization in clinical practice is inevitable because no two patients are alike in all respects (McFall, 2002). For example, in applying an empirically supported treatment to a patient, one is by definition generalizing from previous studies based on clients who are not identical to the patient in one’s consulting room. Nevertheless, this fact does not gainsay the point that some forms of generalization are more extensive than others, and that practitioners should always be clear in their communications with clients—and clear in their own heads—about the degree of such generalization. As McFall (1991) noted, “One of the problems facing clinical psychology is that it has oversold itself” (p. 81). Perhaps more than anything else, clinicians who possess professional knowledge are humble (McFall, 1996). They are keenly aware of (a) the strengths and weaknesses of the existing database, and (b) their own inherent fallibility as information processors. Such clinicians also must walk an epistemic tightrope: They must proceed on the basis of extant scientific knowledge while remaining cognizant of the possibility that such knowledge could one day be revised or even overturned. Such humility need not engender insecurity or diffidence. On the contrary, such humility should come with the self-assurance that one is acting as a consummate mental health professional: an individual who brings his or her unique scientific expertise to bear on human suffering, and who is certain not to overstate or oversell the extent of this expertise. EPISTEMIC DUTIES AND THE SELLING OF EXPERTISE O’Donohue and Henderson (1999) argued that clinical psychologists possess epistemic duties. An epistemic duty, in turn, is an obligation to possess some particular body of knowledge. What constitutes knowledge is key here. Knowledge is not intuition, speculation, ideology, narcissism-influenced positive appraisals, or beliefs held due to slothful habits, but rather tested claims that go beyond the store of information of nonspecialists. Otherwise,
Chapter 02.qxd
4/6/2007
2:28 PM
Page 45
2. DIMENSIONS OF SCIENCE
45
specialists are not worth what they charge and may even be dangerous due to the unknown iatrogenic effects of their untested interventions. The public typically assumes that psychologists possess such specialized knowledge, technology, and craft. They believe this because of our degrees, our licensing, and our fees (Why else pay $150 per hour for a psychotherapy session?). That is, we very much lead them to believe these claims. Moreover, these beliefs largely explain why our consumers are willing to go the hassle and expense of consulting a professional. If one of our dogs is limping, we first apply our nonspecialist knowledge to the problem. We may look to see whether it has a thorn or a sore on its foot. If it doesn’t, our lay knowledge is typically exhausted. If the problem persists, we take it to a veterinarian to gain access to increased problemsolving effectiveness. The vet knows how to form and explore other alternatives; he or she knows how to intervene in causal pathways to restore normal functioning. He or she also knows the risks involved in these interventions and can help us understand the best alternatives and their pluses and minuses. For this knowledge, we are quite reasonably willing to pay significant sums of money. The notion of epistemic duty as applied to clinical psychology is that this obligation is foundational to our profession. O’Donohue and Ferguson (2003) criticized the APA ethical code as vague on this duty and problematic in not properly prioritizing it. Furthermore, they criticized the APA in not strongly enforcing such duties, and in tolerating a great deal of practice in which these duties are abrogated (e.g., the use of highly suggestive techniques to recover purported memories of child abuse) and concentrating on such relatively trivial issues as bartering. A priority for the field should be to clearly conceptualize what constitutes a psychological error correlating to medical errors that have been found to result in approximately 90,000 unnecessary deaths per year (Institute of Medicine, 2001). Although some have criticized this exact figure, it is clear that medical errors constitute a significant problem. An interesting but troubling question arises for us: How many deaths and how much morbidity are caused by psychological errors? It is indefensible to suggest that, because we cannot agree on what constitutes acceptable practice, no errors are being made. Nor can one reasonably argue that only in cases marked by egregious consequences, such as a child’s death in rebirthing therapy, do psychological errors occur. This death could have been prevented (as well as future negative consequences) by antecedently showing that all rebirthing therapy constitutes a psychological error and should be banned and punished if practiced.
Chapter 02.qxd
46
4/6/2007
2:28 PM
Page 46
O’DONOHUE AND LILIENFELD
TECHNOLOGY, ENGINEERING, AND BASIC SCIENCE One aspect that the discussion of clinical science has neglected is the distinction between science and technology. We argue (although we suspect that McFall might disagree with us) that it is a mistake to think of all practitioners as scientists and to demand that they behave as scientists, just as it is a mistake to demand that all civil engineers be scientists as they build bridges or roads. It is imperative, however, to demand that they be capable of applying the basic science and the resultant technology aptly and even creatively. We believe that the most acceptable model for practitioners is that of technologists or behavioral engineers. Engineers and technologists rarely operate alone in craft industries. Rather, they function in large organized systems that have components with specialized knowledge and that are to a large degree transportable. Ford learned quality improvement from Toyota, and Caterpillar can transport a manufacturing plant from the Midwest to Mexico. A critical priority is to develop transportable systems in which technologically sound behavioral health delivery systems are embedded in quality improvement systems. This is in contrast with much current research emphasis in clinical science, which concentrates on experimental psychopathology, assessment development and validation, and psychotherapy development and evaluation. Although these topics are certainly important, it is vital that we take the knowledge that we have acquired and develop delivery systems whose technology embodies these best practices. These technologically sophisticated delivery systems in turn can provide much interesting and important data for further scientific and technological advances. Our research efforts have been hampered by the cottage industry approach we have adopted. Other sciences use technology to gain new knowledge; we must similarly develop this attitude and develop these systems. A well-functioning behavioral health delivery system can easily through its quality improvement/management information system provide answers to such questions as: 1. How effective are we at treating problem X? What is the general benchmark for treating this problem (how much do probabilities change)? What changes result in these probabilities with our quality improvement systems?
Chapter 02.qxd
4/6/2007
2:28 PM
Page 47
2. DIMENSIONS OF SCIENCE
47
2. Are there interesting deviations for us to explore? Why do the clients of therapist Y exhibit much more improvement than the clients of other therapists? Can we learn something from Y and incorporate this knowledge into our system? 3. Where are we making mistakes? What do we seem to be bad at? What can we learn from others to improve in these areas? 4. How good is our value proposition? Are we giving payors and consumers what they want? In what ways are they dissatisfied with our services and how can we improve? How can we increase our value proposition? McFall aptly cited the work of Deming (Walton, 1986), whose philosophy of quality improvement centers around three major themes: (a) constancy of purpose, (b) a strong systems orientation, and (c) emphasis on data for decision making. His famous Fourteen Points are: 1. Create constancy of purpose toward improvement of product and service, with the aim of becoming competitive, staying in business, and providing jobs. 2. Adopt the new philosophy. We are in a new economic age. Western management must awaken to the challenge, must learn its responsibilities, and take on leadership for change. 3. Cease dependence on inspection to achieve quality. Eliminate the need for inspection on a mass basis by building quality into the product in the first place. 4. End the practice of awarding business on the basis of price tag. Instead, minimize total cost. Move toward a single supplier for any one item, on a long-term relationship of loyalty and trust. 5. Improve constantly and forever the system of production and service, to improve quality and productivity, and thus constantly decrease costs. 6. Institute training on the job. 7. Institute leadership (see pts 8 and 12). The aim of supervision should be to help people and machines and gadgets do a better job. Supervision of management is in need of overhaul, as well as supervision of production workers. 8. Drive out fear, so that everyone may work effectively for the company. 9. Break down barriers among departments. People in research, design, sales, and production must work as a team to foresee problems of
Chapter 02.qxd
4/6/2007
48
10.
11.
12.
13. 14.
2:28 PM
Page 48
O’DONOHUE AND LILIENFELD
production and in use that may be encountered with the product or service. Eliminate slogans, exhortations, and targets for the work force asking for zero defects and new levels of productivity. Such exhortations only create adversarial relationships, as the bulk of the causes of low quality and low productivity belong to the system and thus lie beyond the power of the work force. Eliminate work standards (quotas) on the factory floor. Substitute leadership. Eliminate management by objective. Eliminate management by numbers, numerical goals. Remove barriers that rob the hourly worker of his or her right to pride of workmanship. The responsibility of supervisors must be changed from sheer numbers to quality. Remove barriers that rob people in management and in engineering of their right to pride of workmanship. This means, inter alia, abolishment of the annual or merit rating and of management by objective. Institute a vigorous program of education and self improvement. Put everybody in the company to work to accomplish the transformation. The transformation is everybody’s job. (pp. 14–16)
We suggest a fifth condition to McFall’s first corollary: Psychological services should not be offered to the public unless they are offered within a sound quality improvement system. In this way, their effects can be continually monitored, deviations from expected outcomes can be immediately detected and investigated, and positive deviations can be explored for possibilities to improve practice. Cummings produced an innovative quality improvement orientation in his American Biodyne (Cummings & Sayama, 1995). His system was transportable, and it was eventually instituted in 34 states. His quality improvement system featured many aspects, including: 1. Fifteen percent of all therapists’ time was spent in quality improvement activities. 2. All management needed to see patients at least one day per week. 3. All therapists were supervised and received feedback on their compliance to protocols. 4. Approved continuing education was paid for. 5. Therapy protocols were developed from systematic reviews of the literature and outcome data were gathered. A critical metric was therapeutic effectiveness, which was a ratio of number of
Chapter 02.qxd
4/6/2007
2:28 PM
Page 49
2. DIMENSIONS OF SCIENCE
49
units of therapy given to number of units of medical services decreased. 6. Weekly case conferences were held in which therapists were rewarded for bringing their biggest failures so all individuals could learn from these failures.
SUMMARY AND CONCLUSIONS Science has caught the attention of many because of its fruits. It has yielded potent knowledge and technologies to resolve key human problems. We fight disease better, we eat better, we travel faster and safer, we feel more comfortable, and we communicate and access information with an ease and proficiency that would not have been possible without the Scientific Revolution. However, there seems to be at best an equivocal commitment to science in the behavioral health professions. McFall, rightly in our estimation, calls for a deeper, more authentic, and unqualified commitment to science as the basis of clinical psychology. He nicely articulates several points related to this commitment. If we take seriously such ethical dictums as “first do no harm,” “treat patients with the most effective, minimally intrusive interventions,” and “always help,” we would do well to read and thoroughly understand his Manifesto. In addition, he most perceptively speaks of the importance not only of a broad commitment to science, but also of the thoroughgoing adoption of quality improvement philosophies and technologies. As we have argued, these two perspectives not only dovetail nicely, but offer the potential to influence the day-to-day activities of the clinician in powerful ways.
REFERENCES Bird, A. (2000). Thomas Kuhn. Princeton, NJ: Princeton University Press. Cicchetti, D. V. (1991). The reliability of peer review for manuscript and grant submissions: A crossdisciplinary investigation. Behavioral and Brain Sciences, 14, 119–135. Cioffi, F. (1998). Freud and the question of pseudoscience. Chicago: Open Court. Cummings, N., & Sayama, M. (1995). Focused psychotherapy: A casebook of brief, intermittent psychotherapy throughout the life cycle. New York: Brunner/Mazel. Cummings, N. A., O’Donohue, W., & Ferguson, K. (2002). The impact of medical cost offset on practice and research: Making it work for you: A report of the first Reno Conference on medical cost offset. Reno, NV: Context Press.
Chapter 02.qxd
50
4/6/2007
2:28 PM
Page 50
O’DONOHUE AND LILIENFELD
Dawes, R. M. (1994). House of cards: Psychology and psychotherapy built on myth. New York: The Free Press. Fisher, J. E., & O’Donohue, W. (Eds.). (2006). Practitioners’ guide to evidence-based psychotherapy. New York: Springer. Freud, S. (1925). Instincts and their vicissitudes. In E. Jones (Ed.), Sigmund Freud, Selected Papers (Vol. 4). London: Basic Books. Garb, H. M. (1996). The representativeness and past-behavior heuristics in clinical judgment. Professional Psychology: Research and Practice, 27(3), 272–277. Garb, H. M. (1998). Studying the clinician: Judgmental research and psychological assessment. Washington, DC: American Psychological Association. Gaudiano, B. A., & Herbert, J. D. (2000). Can we really tap our problems away? A critical analysis of thought field therapy. Skeptical Inquirer, 24, 29–33, 36. Gholson, B., & Barker, P. (1985). Kuhn, Lakatos, and Laudan: Applications in the history of physics and psychology. American Psychologist, 40, 755–769. Grove, W. M., & Meehl, P. E. (1996). Comparative efficiency of informal (subjective, impressionistic) and formal (mechanical, algorithmic) prediction procedures: The clinical-statistical controversy. Psychology, Public Policy, and Law, 2, 293–323. Herbert, J. D., Lilienfeld, S. O., Lohr, J. M., Montgomery, R. W., O’Donohue, W. T., Rosen, G. M., & Tolin, D. F. (2000). Science and pseudoscience in the development of eye movement desensitization and reprocessing: Implications for clinical psychology. Clinical Psychology Review, 20, 945–971. Hoyningen-Huene, P. (1993). Reconstructing scientific revolutions. Chicago: University of Chicago Press. Institute of Medicine. (2001). Crossing the quality chasm: A new health system for the 21st century. Washington, DC: National Academies Press. Kuhn, T. S. (1962). The structure of scientific revolutions. Chicago: University of Chicago Press. Kuhn, T. S. (1963). The function of dogma in scientific research. In A. C. Crombie (Ed.), Scientific change: Historical studies in the intellectual, social and technical conditions for scientific discovery and technical invention, from antiquity to the present (pp. 381–395). London: Heinemann. Kuhn, T. S. (1970). The structure of scientific revolutions (2nd ed.). Chicago: University of Chicago Press. Kuhn, T. S. (1974). Second thoughts on paradigms. In F. Suppe (Ed.), The structure of scientific theories (pp. 459–482). Urbana: University of Illinois Press. Kuhn, T. S. (1996). The structure of scientific revolutions (3rd ed.). Chicago: University of Chicago Press. Laudan, L. (1977). Progress and it problems: Towards a theory of scientific growth. Berkeley: University of California Press. Lilienfeld, S. O., Fowler, K. A., Lohr, J. M., & Lynn, S. J. (in press). Pseudoscience, nonscience, and nonsense in clinical psychology: Dangers and remedies. In R. Wright & N. Cummings (Eds.), Destructive trends in mental health: The well-intentioned path to harm. New York: Brunner-Routledge. Lilienfeld, S. O., Lynn, S. J., & Lohr, J. M. (2003). Science and pseudoscience in clinical psychology. New York: Guilford. Losee, J. (1980). A historical introduction to the philosophy of science (2nd ed.). Oxford, England: Oxford University Press.
Chapter 02.qxd
4/6/2007
2:28 PM
Page 51
2. DIMENSIONS OF SCIENCE
51
Masterman, M. (1970). The nature of a paradigm. In I. Lakatos & A. Musgrave (Eds.), Criticism and the growth of knowledge (pp. 59–89). Cambridge, England: Cambridge University Press. McCann, J. T., Shindler, K. L., & Hammond, T. R. (2003). The science and pseudoscience of expert testimony. In S. O. Lilienfeld, S. J. Lynn, & J. M. Lohr (Eds.), Science and pseudoscience in clinical psychology (pp. 77–108). New York: Guilford. McFall, R. M. (1991). Manifesto for a science of clinical psychology. The Clinical Psychologist, 44, 75–88. McFall, R. M. (1996). Making psychology incorruptible. Applied & Preventive Psychology, 5, 9–15. McFall, R. M. (2000). Elaborate reflections on a simple manifesto. Applied & Preventive Psychology, 9, 5–21. McFall, R. M. (2002). Training for prescriptions vs. prescriptions for training: Where are we now? Where should we be? How do we get there? Journal of Clinical Psychology, 58, 659–676. McNally, R. J. (2003). Remembering trauma. Cambridge, MA: Harvard University Press. Meehl, P. E. (1978). Theoretical risks and tabular asterisks: Sir Karl, Sir Ronald, and the slow progression of soft psychology. Journal of Consulting and Clinical Psychology, 46, 806–844. Meehl, P. E. (1997). Credentialed persons, credentialed knowledge. Clinical Psychology: Science and Practice, 4, 91–98. Mercer, J., Sarner, L., & Rosa, L. (2003). Attachment therapy on trial: The torture and death of Candace Newmaker. Westport, CT: Praeger. O’Donohue, W. (1989). An (even) bolder model: The clinical psychologist as metaphysician-scientist-practitioner. American Psychologist, 44, 1460–1468. O’Donohue, W. (1993). A spell of Kuhn on psychology: An exegetical elixir. Philosophical Psychology, 6, 267–287. O’Donohue, W. (2003). Rational animals: Behavior therapy’s focus on knowledge and the understanding of human behavior. Behavior Therapist, 26, 402–405. O’Donohue, W., & Bradley, A. (1999). Conceptual and empirical issues in child custody evaluations. Clinical Psychology: Science and Practice, 6, 310–322. O’Donohue, W., & Ferguson, K. (Eds.). (2003). Handbook of professional ethics. San Diego: Academic Press. O’Donohue, W. T., Ferguson, K. E., & Naugle, A. E. (2003). The structure of the cognitive revolution: An examination from the philosophy of science. The Behavior Analyst, 26, 85–110. O’Donohue, W., & Halsey, L. (1997). The substance of the scientist-practitioner relation: Freud, Rogers, Skinner and Ellis. New Ideas in Psychology, 15, 35–53. O’Donohue, W., & Henderson, D. (1999). Epistemic and ethical duties in clinical decision-making. Behavior Change, 16, 10–19. O’Donohue, W., & Lloyd, A. (2005). Ethos of contemporary clinical psychology. In R. H. Wright & N. A. Cummings (Eds.), Destructive trends in psychology: The well intentioned road to harm (pp. 1–25). New York: Brunner-Routledge. O’Donohue, W., Lloyd, A., & Ferguson, K. E. (in press). Evolutionary epistemology: An account of knowledge for contemporary psychology. Peterson, D. R. (1996). Making psychology indispensable. Applied & Preventive Psychology, 5, 1–8.
Chapter 02.qxd
52
4/6/2007
2:28 PM
Page 52
O’DONOHUE AND LILIENFELD
Quine, W. V. (1974). Roots of reference. La Salle, IL: Open Court. Reichenbach, H. (1938). Experience and prediction. Chicago: University of Chicago Press. Rogers, C. (1968). Some thoughts regarding the current presupposition of the clinical sciences. In W. R. Coulson & C. Rogers (Eds.), Man and the science of man (pp. 24–38). Columbus, OH: Charles E. Merrill. Singer, M. T., & Lalich, J. (1996). Crazy therapies: What are they? Do they work? San Francisco: Jossey-Bass. Stricker, G., & Trierweiler, S. J. (1995). The local scientist: A bridge between science and practice. American Psychologist, 50, 995–1002. Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185, 1124–1131. Walton, M. (1986). The Deming management method. New York: Perigee. Wood, J. M., Nezworski, M. T., Lilienfeld, S. O., & Garb, H. N. (2003). What’s wrong with the Rorschach? Science confronts the controversial inkblot test. San Francisco: Jossey-Bass.
Chapter 03.qxd
4/6/2007
2:43 PM
Page 53
3 The Seduction of Clinical Science in Psychology: Challenges in Psychological and Biological Convergence Gregory A. Miller, Anna S. Engels, and John D. Herrington University of Illinois at Urbana–Champaign
The evolution of clinical science in psychology has long faced divergent opportunities and conflicting incentives. Meehl (1973), McFall (1991), Dawes (1994), and many others have warned against the powerful and growing temptation that clinical service providers face to act without sufficient grounding in psychological science. Such critiques are often misunderstood as being anticlinician, when in fact they seek to foster the best that clinicians can offer (Miller, 1995, 2004). Such critiques form an important thread in a larger debate in psychology often miscast as questioning the scientific legitimacy of so-called soft psychological research (Meehl, 1978; Miller, 2004). There is no question that normal and abnormal psychological phenomena can be pursued with full scientific rigor. The challenge, as McFall (1991) famously articulated, is for clinical service delivery not simply to find ways to continually integrate science and practice, but to achieve and protect a framing of them in which they are not distinct. McFall argued that it is not enough to bring good science to bear on clinical work. More radically, he argued that the practitioner should be 53
Chapter 03.qxd
54
4/6/2007
2:43 PM
Page 54
MILLER, ENGELS, HERRINGTON
a scientist in every possible sense. Against that goal are arrayed common training models that do not emphasize science and, more important, health care policies, politics, and economics that undermine opportunities for science to be central to practice. That battle continues, but it has the advantage of being widely recognized, at least within clinical science. Another battle has yet to be seriously joined, yet it presents at least as much a challenge to psychological clinical science as the anti-intellectual and economic pressures plaguing clinical service delivery. Rather than selling out scientifically infused clinical work to economic pressures, there is a growing temptation to replace psychological questions with biological questions. Widely noted advances in hemodynamic neuroimaging and molecular genetics and less widely noted but equally promising advances in electromagnetic and optical source image analysis provide powerful pressures for psychological clinical scientists to put aside their focus on psychological phenomena in favor of biological phenomena. These developing technologies offer great promise for addressing questions of normal and abnormal psychological function. Strikingly, however, most of the momentum for the application of these methods to psychopathology research is away from the psychological phenomena that are fundamental to mental disorder in favor of a quite different set of phenomena, logically distinct from psychological events and having an unknown relationship to them. In briefly reviewing recent research that relies on emerging hemodynamic and electromagnetic methods, the present discussion hopes to convey both the excitement and confusion that have overtaken psychopathology research associated with these methods. The primary thesis is that psychological clinical science must guide the evolution of these methods in psychopathology research. A CONCEPTUAL PERSPECTIVE Cognition, emotion, personality, psychopathology—these are all fundamentally psychological concepts. There is no question that in, for example, depression or schizophrenia, there are biological mechanisms gone awry. It is no doubt essential to understand those mechanisms and how to treat them. In turn, biological models can improve psychological models because the psychological and biological stories can mutually guide and constrain each other. Cognition, emotion, personality, and psychopathology are not biological concepts. At present, no theory is available that articulates a means by
Chapter 03.qxd
4/6/2007
2:43 PM
Page 55
3. SEDUCTION OF CLINICAL SCIENCE IN PSYCHOLOGY
55
which biological events cause psychological events, nor a theory that describes the converse. The cardinal feature of schizophrenia is thought disorder, which is clearly a psychological construct. There can be no genetic or biochemical story that is sufficient to account for thought disorder because such an account is not about thought disorder at all; it is about genes or biochemistry. Thus, for example, what has sometimes been called a biochemical theory of schizophrenia cannot be a biochemical theory of schizophrenia because schizophrenia is not a disorder of biochemistry. It is (definitionally) a disorder of thought. Again, there is no question that there are genetic and biochemical stories to be told about schizophrenia. But such stories, necessarily confined to those domains, cannot account for thought disorder, and thus for schizophrenia. A biochemical theory could only be a theory of disrupted biochemistry in schizophrenia—enormously important to develop, but not to be confused with a theory of thought disorder. Even if clinical science advances to the point where, by psychological, biochemical, or genetic means, the abnormal biochemistry of schizophrenia can be effectively treated, this would not provide an account of the inherently psychological symptoms of schizophrenia. Absent a theory that describes how biological events cause psychological events, it is not meaningful to speak of biological events as underlying psychological events. A complete account might, perhaps, provide such mechanisms—or might conversely explain how psychological events underlie or otherwise drive biological events. At present, however, the field is in no position to portray one domain as underlying the other (for an extended discussion of this point, see Miller & Keller, 2000; Miller, 1996; Taitano & Miller, 1998). A consequence of the foregoing view of psychopathology is that brain measures cannot localize psychological phenomena because psychological phenomena do not have a location. This is a logical, rather than an empirical issue (Kosslyn & Koenig, 1992; Marr, 1982). For example, fear is often inferred on the basis of observables, such as avoidance behavior, autonomic changes, and self-report, but the construct of fear is not fully captured by or reducible to those measures (Kozak & Miller, 1982). Instead, fear is often defined functionally, and functions do not have locations (Fodor, 1968; Miller, 1996; Miller & Keller, 2000; Miller & Kozak, 1993). Fear is not in the amygdala (Lang, Davis, & Öhman, 2000; LeDoux, 1995), pleasure is not in the orbital frontal cortex (Davidson, 1998), and trust is not in the dorsal striatum (King-Casas et al., 2005). This logical issue is distinct from
Chapter 03.qxd
56
4/6/2007
2:43 PM
Page 56
MILLER, ENGELS, HERRINGTON
the certainty that when specific psychological events occur in specific organisms, specific brain events occur. But there is no necessary one-to-one mapping of a class of psychological events to a class of brain events. Indeed, in the present state of the field, there is no account available of the mechanisms by which brain events (electrical, chemical, hemodynamic, or even structural changes) either affect psychological phenomena or are affected by them. The present point is not about whether such effects occur, only about our inability, to date, to account for them. Faced with the present inability to articulate causal mechanisms crossing the boundary between biology and psychology, the literature often falls prey to dualism or reductionism, and reductionism is often implicit in discussions of functional brain localization in normative cognition and emotion in psychopathology. But reductionism is not tenable in cognitive, affective, or clinical neuroscience. Mental events are not the same thing as neural activity; phenomenological experience cannot be described in terms of ion flows, synaptic connections, and so forth.…The mind is what the brain does: a description of mental events is a description of brain function [not brain tissue].…[Mental events] and brain events are members of different [logical] categories, and one cannot be replaced by the other. (Kosslyn & Koenig, 1992, pp. 4, 432; italics added)
This meta-conceptual perspective is helpful in evaluating what localizationoriented brain research methods can and cannot offer the field, at least at present. Brief reviews of recent results from two ongoing research programs serve to highlight some of the promise of hemodynamic and electromagnetic measures of brain activity, but also serve to illustrate the present limitations of those methods. SENSORY GATING AND OVERLOAD IN SCHIZOPHRENIA Informal clinical impressions and systematic studies have described reports from patients with schizophrenia of significant experiences of sensory overload (Hetrick & Smith, under review; Kisley, Noecker, & Guinther, 2004; McGhie & Chapman, 1961). Such reports of subjective experience are clearly about psychological phenomena. As argued earlier, it would be inappropriate to suppose that neural evidence of dysfunction in stimulus gating could somehow suffice to explain this subjective experience, but it is nonetheless tempting to speculate about relevant neural
Chapter 03.qxd
4/6/2007
2:43 PM
Page 57
3. SEDUCTION OF CLINICAL SCIENCE IN PSYCHOLOGY
57
abnormalities. A simple paradigm providing considerable control of stimulus parameters involves presentation of pairs of stimuli (Graham, 1975, 1979) and has been used in many variants and with a number of different physiological measures, including heart rate (Cook & Turpin, 1997), startle eyeblink (Dawson, Schell, Swerdlow, & Filion, 1997; Lang, Bradley, & Cuthbert, 1997), and event-related brain potentials (ERPs; Adler et al., 1982; Perlstein, Fiorito, Graham, & Simons, 1993). Adler et al. (1982) pioneered a version of this paradigm in which two brief clicks were presented in succession, with the amplitude of the P50 component of the ERP markedly reduced in nonpatients, which they termed sensory gating. They found that, under some circumstances, individuals diagnosed with schizophrenia showed little or no gating of P50 amplitude. This gating deficit proved difficult to replicate at first, and concerns were raised about the psychometric properties of the ratio metric used (Smith, Boutros, & Schwarkopf, 1994). However, methods were refined and better described, and numerous replications were eventually reported, with a recent review paper concluding that the evidence for a P50 gating deficit in schizophrenia is compelling, showing an impressive 1.56 standard deviation effect size (Bramon, Rabe-Hesketh, Sham, Murray, & Frangou, 2004). Bramon et al. noted that this effect size matches or exceeds the most robust brain morphometry and neuropsychological findings available in schizophrenia. Adler et al. and subsequent authors have interpreted the gating failure as highly relevant to the subjective stimulus overload reported in schizophrenia. Converging evidence has come from a variety of clinical, neuropsychological, and genetic studies (Freedman et al., 1997; Freedman, Adler, & Leonard, 1999; Huang et al., 2003; Kisley et al., 2004; Leonard et al., 2002; Thoma et al., 2003, 2005; Yee, Nuechterlein, Morris, & White, 1998). In a series of recent studies, we have undertaken to identify the neural circuitry driving the P50 gating deficit (Edgar et al., 2003, 2005; Hanlon et al., 2005; Huang et al., 2003; Thoma et al., 2003, 2004, 2005). Densearray magnetoencephalography (MEG) coupled with structural magnetic resonance imaging (sMRI) has been called magnetic source imaging (Lewine & Orrison, 1995). For biophysical reasons having to do with the primary generators of P50 being in bilateral superior temporal gyrus (STG), MEG is particularly well suited for localization of P50 generators, although there are substantial individual differences in the orientation of the generators (Edgar et al., 2003). Huang et al. (2003) found that in control subjects, 97% of the variance in P50 recorded at the Cz
Chapter 03.qxd
58
4/6/2007
2:43 PM
Page 58
MILLER, ENGELS, HERRINGTON
electrode site, on which most P50 gating studies rely, could be accounted for by activity from sources located in STG. In a matched patient group, the STG sources accounted for significantly less (86%) of the Cz P50 variance. These findings confirm the central role of STG sources in P50 generation and illustrate some of the promise of electroencephalography (EEG)/MEG source localization methods. Although it is often assumed that scalp EEG is both sufficiently distant from neural activity and sufficiently complex to be difficult to decompose in terms of specific sources, appropriate combinations of paradigms and methods show this to be unfounded. In the Huang et al. (2003) data, one might suppose that the nontrivial residual variance in the patient group reflected greater noise in some sense—a problem that plagues much clinical research. In fact, however, the residual was highly organized, suggesting a rather pure 40-Hz generator elsewhere in the brain that contributes to patient (but not control group) EEG around P50. Work is underway to identify such a source, with candidates in prefrontal cortex and hippocampus under consideration (Hanlon et al., 2003, 2005, 2006). These studies demonstrate the effectiveness of MEG source localization even deep in the brain, counter to common assumptions that MEG’s impressive spatial resolution is confined to superficial cortex. In fact, MEG source localization resolution can be on the order of a few millimeters (Pantev et al., 1995; Romani, 1986, in the auditory modality; Huang et al., 2000; Kawamura et al., 1996, in the somatosensory modality; Aine, Huang, Stephen, & Christner, 2000; Stephen et al., 2002, in the visual modality). MEG’s resolution is presently limited only by head movement, registration error between MEG and anatomic images, and the ballistocardiogram. MEG is at its best with sources in superficial cortex, where its spatial localization capability is often superior to that of functional MRI (fMRI) or other hemodynamic neuroimaging methods and vastly superior in terms of temporal resolution. In skull phantom work, Leahy, Mosher, Spencer, Huang, and Lewine (1998) demonstrated an average spatial localization error of 3 mm with 122-channel MEG (61 locations, 2 orthogonal channels at each location) and 7–8 mm with up to 64-channel EEG across 32 dipoles. The 3-mm figure was only approximately the 2.26mm error associated with registration of sensor location, indicating that the fundamental MEG methodology can do well indeed. Newer MEG machines provide considerably denser arrays. Studies of the relationship between sensor density and source localization accuracy have suggested
Chapter 03.qxd
4/6/2007
2:43 PM
Page 59
3. SEDUCTION OF CLINICAL SCIENCE IN PSYCHOLOGY
59
that benefits accrue as density increases as high as 128 or even 256 channels positioned roughly over the superior-posterior half of the head (Srinivasan, Tucker, & Murias, 1998), although in some cases lower channel counts do well (e.g., Babiloni et al., 2004; Michel et al., 2004). Often more important than total number of sensors is distribution, including coverage of inferior head surfaces (Tucker, Luu, Frishkoff, Quiring, & Poulsen, 2003). Furthermore, in many cases, absolute localization accuracy is not particularly important. Rather, identification of the number of distinct sources, their time courses, and their functional roles is of interest (Scherg, Ille, Bornfleth, & Berg, 2002). Thus, winnertake-all comparisons among neuroimaging methods’ sensor montages or measurement methods in terms of millimeter accuracy warrant less attention than they are often given. WHAT IS BEING LOCALIZED? PART I The starting point for this line of research on P50 was clinical reports of sensory overload in schizophrenia. A considerable body of literature evolved, marked by gradual substantive progress in the face of replication issues, methodology disputes, technological innovations, and other features from the usual array of process challenges an area of science faces. The present goal is not a definitive review of the substantive findings in the schizophrenia/P50 gating literature (see Bramon et al., 2004), but an examination of what this literature reveals about the state of the art in the relationship between psychological and biological phenomena. Granting compelling evidence of a P50 gating deficit, the identification of its immediate sources in STG, and clear directions for further localization of more extended brain circuits, what does this research say about mechanisms of patients’ experience of sensory overload? It may be argued that explaining subjective experience should not be among the goals of scientific endeavor. There are at least two responses to this view. Minimally, it remains the case that the sizable P50 gating deficit literature arose from an attempt to understand reports of just such experiences in the hopes of learning critical things about schizophrenia. More generally, research focused on biological phenomena has often been inspired by reports of psychological phenomena. The biological research may lead to important insights about disrupted sensory control systems in schizophrenia and the relationship of that disruption to biological phenomena implicated in other symptoms of schizophrenia. But explanations
Chapter 03.qxd
60
4/6/2007
2:43 PM
Page 60
MILLER, ENGELS, HERRINGTON
in terms of biological phenomena are explanations only of biological phenomena, not of psychological phenomena. A richer story is needed that encompasses psychological and biological phenomena and causal relationships between them, if any. A second response to skepticism about whether subjective experience belongs in science is that subjective experience can be treated like any other mentalistic psychological phenomena, such as memory, attention, motivation, language comprehension, or emotion. Phenomena that are not directly observable are routinely treated as legitimate and even central in psychology and other areas of science, as long as there are observable phenomena understood as having some specified relationship to them (Kozak & Miller, 1982; Meehl, 1978). That is, science inevitably makes inferences about unobserved events on the basis of observed events. The present stance is not that research on subjective experience should be a primary goal of science, but simply that there is no barrier to including it. The research on P50 gating may mature to the point where psychological or biological interventions are developed that fully normalize gating in schizophrenia. Evidence suggests that some newer antipsychotic medications improve gating (Adler et al., 2004, 2005; Freedman, Adler, Waldo, Pachtman, & Franks, 1983; Light, Geyer, Clementz, Cadenhead, & Braff, 2000; Nagamoto et al., 1996). Given normalized P50 gating, a key question would be whether patients’ reports of sensory overload also subside. Absent an articulated model explaining how experienced sensory overload drives P50 gating, or how P50 gating drives experienced sensory overload, a fundamental issue arises. It is not clear what the logical or causal relationship is between sensory overload (an inherently psychological event) and P50 gating (an inherently biological event). More generally, it is not clear how interventions in one domain alter the other domain. At issue here is not merely whether there may be a correlation between an intervention in one domain and a change in another domain. It is not difficult to design studies that observe correlations, and careful designs may even provide a basis for inferring causation. But the goal is bolder: the identification of mechanisms— explanations—at play when our self-regulation strategies fail and when our interventions succeed. A reductionist assumption that whatever brain mechanisms explain the P50 gating failure in schizophrenia also explain sensory overload in schizophrenia—the view that explaining the brain explains the mind—makes the logical error discussed earlier for
Chapter 03.qxd
4/6/2007
2:43 PM
Page 61
3. SEDUCTION OF CLINICAL SCIENCE IN PSYCHOLOGY
61
biochemical theories of schizophrenia. Brain mechanisms simply are not mechanisms of psychological events. EMOTION REGULATION IN PSYCHOPATHOLOGY A second example of the promise of hemodynamic and electromagnetic measures of brain activity comes from a growing research interest in normal and disturbed emotion processes that has encouraged studies of diverse brain mechanisms. Inevitably, much of this work has sought to identify and localize brain regions with specific roles in emotion. Numerous concepts and experimental paradigms have been brought to bear, often focused on proposed dichotomies such as left versus right anterior brain activity (Davidson, 1992; Harmon-Jones & Allen, 1998), appetitive versus aversive motivation (Bradley, Codispoti, Cuthbert, & Lang, 2001; Schupp et al., 2000), approach versus avoidance aspects of emotion (Davidson, Jackson, & Kalin, 2000; Davidson, Pizzagalli, Nitschke, & Putnam, 2002; Harmon-Jones & Allen, 1998), approach versus inhibition (Sutton & Davidson, 1997), or valence versus arousal dimensions (Davidson, 1992; Heller, Koven, & Miller, 2003; Lang et al., 1997). Paradigms have employed personalized emotional imagery (Miller et al., 1987; Shin et al., 2004) and standardized pictures, sounds, or words (Bradley & Lang, 1999a, 1999b; Lang, Bradley, & Cuthbert, 1999) under a variety of task conditions. This research area is exploding, as the methods of cognitive psychophysiology/cognitive neuroscience come to be routinely applied to emotion research in what is sometimes called affective neuroscience (Davidson & Sutton, 1995; Panksepp, 1998). Lang et al. (1997) framed much adaptive and maladaptive emotiondriven behavior in terms of a distinction between appetitive and defense motivation. Tucker, Derryberry, and Luu (2000) discussed depression and anxiety in terms of neural control processes involving elaborate connections across vertically integrated brain regions. Davidson (2003; Davidson et al., 2002) outlined emotion-regulation mechanisms as manifestations of diatheses of affective style that drive emotional behavior along approach and withdrawal dimensions. Our recent fMRI work has cast the relationship of cognition and emotion in terms of self-regulation of attention mechanisms operating in the face of emotional distractors whose failure may foster psychopathology (Compton et al., 2003; Engels et al., 2004; Herrington et al., 2005; Miller et al., 2004; Mohanty et al., 2005, in press). These approaches to the characterization of psychopathology
Chapter 03.qxd
4/6/2007
2:43 PM
62
Page 62
MILLER, ENGELS, HERRINGTON
Working model of current context
Dorsolateral Prefrontal Cortex
Hippocampus
Amygdala Context learning, context recognition Emotional alerting FIGURE 3.1 Proposed modulation of dorsolateral prefrontal cortex (DLPFC) by hippocampus and amygdala. Negatively valenced stimuli prompt abnormal limbic and striatal activity, which in turn exacerbates DLPFC-related cognitive disturbance in schizophrenia.
reflect a long-standing but underappreciated perspective on big-picture neural function, which views the goal of overt behavior as the selfmodulation of external input and internal homeostasis (e.g., Montague, Hyman, & Cohen, 2004; Powers, 1973). Psychopathology can thus be conceptualized not simply as a failure of self-regulation, but potentially as the operation of intact control systems receiving distorted input, intact control systems with control parameters set incorrectly, or control systems in which important segments are essentially broken. Such distinctions potentially have fundamental implications for prevention and intervention. One of the challenges in identifying brain circuits involved in psychopathology is determining the causal relations between brain regions.
Chapter 03.qxd
4/6/2007
2:43 PM
Page 63
3. SEDUCTION OF CLINICAL SCIENCE IN PSYCHOLOGY
63
FIGURE 3.2 In an emotional Stroop paradigm during fMRI recording, 16 control participants showed more left dorsolateral prefrontal cortex (DLPFC) activation than did 16 schizotypal participants (left panel) in response to negatively valenced words, beyond that obtained for neutral words. This group difference was reversed in right DLFPC (middle panel) and in bilateral amygdala (right panel). (Following radiological convention, right side of the brain is on left side of panel in axial and coronal slices. Based on data reported in Mohanty et al. [2005]. Color version available on Psychology Press Web site: http://www.psypress.com/brainscans-etc) Whether the abnormal DLPFC lateralization drives or is driven by the abnormal amygdala activation in this paradigm awaits further research, but the Grace and Moore (2000) model suggests the latter.
Grace (2000; Grace & Moore, 1998) proposed a modulation of dorsolateral prefrontal cortex (DLPFC) by hippocampus and amygdala (see Fig. 3.1), with distinct regulatory roles for the latter two structures. Via the nucleus accumbens, hippocampal and amygdalar activity gates activity of neurons in PFC, with hippocampus gating input in a way that influences context and amygdala providing an emotional interrupt that overrides input from ongoing context. In schizophrenia, Grace proposed that negatively valenced stimuli prompt abnormal limbic and striatal activity, which in turn exacerbates DLPFC-related cognitive disturbance. Mohanty et al. (2005) found support for this model in nonpatients having high or moderate levels of schizotypy. In an emotional Stroop paradigm, subjects judged the ink color of words whose meaning was pleasant, neutral, or unpleasant during hemodynamic measurement of regional brain activity. Control participants showed more left DLPFC activation than did schizotypal participants in response to negatively valenced words, beyond that obtained for neutral words (Fig. 3.2, left panel). This group difference was reversed in right DLFPC (middle panel) and in bilateral amygdala (right panel). Whether the abnormal
Chapter 03.qxd
64
4/6/2007
2:43 PM
Page 64
MILLER, ENGELS, HERRINGTON
FIGURE 3.3 In a color-word Stroop paradigm, 17 control participants performed a color-word Stroop test during fMRI recording (left panels) and during dense-array EEG recording (right panel). Five brain regions (Talairach coordinates are listed) showing enhanced activity for incongruent stimuli (such as the word red in blue ink) were used to constrain EEG source dipole locations. Source analysis was then undertaken with the EEG data using those five regions to constrain the dipole locations. Because the BESA source localization software (Megis Software, Gräfelfing, Germany) used difference waveforms (as was also done during the fMRI analysis) from the event-related brain potential (ERP), no sources reflected early ERP components. The ovals indicate periods of peak activation in the five regions, with bilateral dorsolateral prefrontal cortex (DLPFC) preceding parietal and anterior cingulate cortex activation. From Miller et al. (2004). Color version available on Psychology Press http://www.psypress.com/brainscans-etc.
DLPFC lateralization drives or is driven by the abnormal amygdala activation in this paradigm awaits further research, but the Grace and Moore (1998) model suggests the latter. The temporal resolution of these types of hemodynamic measures, however, limits the ability to infer causal sequence among brain regions when processing unfolds in hundreds of milliseconds or less. Miller et al. (2004) reported work underway recording dense-array EEG and fMRI in the same subjects doing the same task in different sessions. This work
Chapter 03.qxd
4/6/2007
2:43 PM
Page 65
3. SEDUCTION OF CLINICAL SCIENCE IN PSYCHOLOGY
65
involves both emotional Stroop and traditional color-word Stroop task variants. Figure 3.3 illustrates how fMRI-guided EEG source localization may be able to resolve functional connectivity questions in such paradigms. These preliminary results strongly support the model of Banich et al. (2001; Milham, Banich, Claus, & Cohen, 2003), which suggests that DLPFC serves an executive function, directing a parietal attentional system to attend to an abstract, task-relevant attribute (ink color) and undertake motor planning. Parietal cortex then performs a mapping of stimulus attributes to response preparation (e.g., red stimulus mapped to middle button). Anterior cingulate cortex (ACC) then proceeds to select a specific response (press the button for red) if there is no conflict between ink color and word meaning. If DLPFC is unsuccessful in setting up the rest of the system for the task, ACC becomes more active to deal with response selection conflict. Accordingly, for incongruent stimuli in Figure 3.3, activity in DLPFC onsets first and activity in ACC onsets last. WHAT IS BEING LOCALIZED? PART II The work just reviewed, combining sMRI, fMRI, and dense-array EEG with clinical symptom measures, attentional control tasks, and overt behavioral performance assessment, illustrates some of the considerable potential of multimethod integration in understanding control-system disruption in psychopathology. But what do these results say about localization of function in the brain? Schizotypes’ greater amygdala activation to negative words in Figure 3.2 suggests a mechanism by which negative affect disrupts cognition in schizotypy and schizophrenia. But does that mean that the disruption is in the amygdala? In the data, what is in the amygdala is altered blood oxygenation and/or blood flow. Assuming that this change reflects local neural changes, this is clear evidence of abnormal functional biology in schizotypy. However, the disruption in question is in the domain of psychology (disrupted cognition), not biology. Is the cognitive disruption nothing more than the biological disruption? As discussed earlier, an affirmative answer would be a classic mistake of reductionism. By cognitive disruption we do not mean anything at all about the biology of the individual showing the disruption. Surely there must be something different in the schizotype’s brain. Studying that biological disruption may lead to critical insights into—but not a substitution for—the psychological mechanisms that produce the cognitive disruption. Cognitive disruption cannot have a geographic address in the
Chapter 03.qxd
66
4/6/2007
2:43 PM
Page 66
MILLER, ENGELS, HERRINGTON
brain (or anywhere else). It may be a consequence of brain dysfunction if one wants to understand psychological events as emergent properties of brain function. But cognition does not have a location, so the psychological disruption cannot be in the amygdala. Similarly, in Figure 3.3, do we understand something about Stroop interference simply because we see activity in DLPFC preceding activity elsewhere? We might say that at least some frontal-lobe activity precedes and potentially drives parietal-lobe systems. It is an observation about two brain regions. It is an inference about the relationship between two psychological functional modules, two attentional systems. It is not an observation of the (psychological) attentional systems. We do not have a theory adequate to account for (psychological) attentional events in terms of biological events. Specifically here, we do not have a theory adequate to account for (psychological) Stroop interference in terms of brain events. With luck, we can build parallel psychological and biological models of functional psychological modules and functional biological , modules each of which accounts for interference as assessed within its own domain. Success in building one model may profoundly facilitate building a model in the other domain. But a model in one domain is not an adequate account of—or even a thorough model of—phenomena in the other domain. Unfortunately, this distinction is often lost in public discourse, illustrated in a statement by Leshner (1997) when director of the National Institute of Drug Abuse: “That addiction is tied to changes in brain structure and function is what makes it, fundamentally, a brain disease” (p. 46). If being “tied to” brain structure or function is sufficient to make something a brain disease, it would seem that everything is a “brain disease,” making the characterization meaningless. More recently, a prominent review paper (Harrison & Weinberger, 2005) provided a variety of characterizations of the logical status of schizophrenia, without addressing the evident inconsistency: “schizophrenia is beyond doubt a brain disease” (p. 41); “schizophrenia is predominantly a genetic disorder” (p. 43); reference to a view of “schizophrenia as a disorder of synaptic signalling” (p. 56); and, in summary, “The evidence…is consistent with the view that the disorder is fundamentally one of abnormal information processing at the highest level” (p. 57). Only the latter is viable on logical grounds, given the historical meaning of the term schizophrenia as a psychological disorder. Harris and Weinberger marshal compelling evidence that there are a brain disorder, a genetic disorder, and a synaptic disorder
Chapter 03.qxd
4/6/2007
2:43 PM
Page 67
3. SEDUCTION OF CLINICAL SCIENCE IN PSYCHOLOGY
67
in schizophrenia, all of which need to be understood and integrated for a complete understanding of relevant phenomena, but their existence does not reduce schizophrenia into a disorder of any of those types. A similar point can be made for nonbiological approaches. Work on so-called neural network modeling (which is actually algebraic, not neural; Montague et al., 2004) of Stroop interference (Cohen & ServanSchreiber, 1992) has provided important insights into how relevant psychological and biological mechanisms unfold during Stroop interference, with potential applications in psychopathology. But we would not mistake the computer model of a neural network for either the cognitive or neural functions that unfold. The computer model is not intended as a complete model of the psychology or the biology. It is an impressive metaphor and heuristic, modeling portions of the function of the natural phenomena. Marr (1982), discussing psychological and biological approaches to vision, wrote: The explication of each level involves issues that are rather independent of the other…some phenomena may be explained at only one or two [levels].… In attempts to relate psychophysical problems to physiology, too often there is confusion about the level at which problems should be addressed. (p. 25)
Although it can be enormously valuable, localization of brain activity is insufficient. We should not confuse implementation with concept. Again quoting Marr (1982): An algorithm is likely to be understood more readily by understanding the nature of the problem being solved than by examining the mechanism (and the hardware) in which it is embedded…trying to understand [a psychological construct such as] perception by studying only neurons is like trying to understand bird flight by studying only feathers. (p. 27)
Accordingly, terms such as localization of function and functional connectivity can be misleading. The goal should be to localize brain activity believed to support or implement (psychological) function, but in so doing we are not actually localizing psychological functions. Again, a psychological construct such as fear, belligerence, or affection does not have a location in the brain. Localization must be understood as a means, not the goal, of psychological research on brain activity. Advances in psychological clinical science depend as much on careful analysis of the concepts that drive our research and service as on careful attention to methodology in our experiments and interventions. Recent
Chapter 03.qxd
68
4/6/2007
2:43 PM
Page 68
MILLER, ENGELS, HERRINGTON
work reviewed here illustrates both the promise of emerging neuroimaging methods and the temptation to overinterpret what they can deliver. In our efforts to understand psychopathology, we must not confuse psychological questions, which concern the fundamental phenomena, with biological questions, which surely have a central role in advancing psychological clinical science. ACKNOWLEDGMENTS This research was supported by NIDA R21 DA14111, NIMH R01 MH61358, Carle Clinic, and the Beckman Institute. The research was conducted while Anna Engels was a predoctoral trainee in the Cognitive Psychophysiology training program of the Department of Psychology, University of Illinois at Urbana–Champaign, NIMH T32 MH19554, and while John Herrington was a predoctoral trainee in the Quantitative Methods program of the Department of Psychology, University of Illinois at Urbana–Champaign, NIMH T32 MH14257. The authors thank Lawrence Adler, Marie Banich, Joseph Barkmeier, Jose Cañive, J. Christopher Edgar, Joscelyn Fisher, Faith Hanlon, Wendy Heller, Mingxiong Huang, Rebecca Levin, Michael Milham, Aprajita Mohanty, Sarah Sass, Jennifer Stewart, Bradley Sutton, Robert Thoma, Andrew Webb, Michael Weisend, and Tracey Wszalek for their contributions to the work reported here. REFERENCES Adler, L. E., Cawthra, E. M., Donovan, K. A., Harris, J. G., Nagamoto, H. T., Olincy, A., et al. (2005). Improved P50 auditory gating with ondansetron in medicated schizophrenia patients. American Journal of Psychiatry, 162, 386–388. Adler, L. E., Olincy, A., Cawthra, E. M., McRae, K. A., Harris, J. G., Nagamoto, H. T., et al. (2004). Varied effects of atypical neuroleptics on P50 auditory gating in schizophrenia patients. American Journal of Psychiatry, 161, 1822–1828. Adler, L. E., Pachtman, E., Franks, R. D., Pecevich, M., Waldo, M. C., & Freedman, R. (1982). Neurophysiological evidence for a defect in neuronal mechanisms involved in sensory gating in schizophrenia. Biological Psychiatry, 17, 639–654. Aine, C., Huang, M., Stephen, J., & Christner, R. (2000). Multi-start algorithms for MEG empirical data analysis reliably characterize locations and time-courses of multiple sources. NeuroImage 12, 159–172. Babiloni, F., Babiloni, C., Carducci, F., Romani, G. L., Rossini, P. M., Angelone, L. M., et al. (2004). Multimodal integration of EEG and MEG data: A simulation study with variable signal-to-noise ratio and number of sensors. Human Brain Mapping, 22, 52–62.
Chapter 03.qxd
4/6/2007
2:43 PM
Page 69
3. SEDUCTION OF CLINICAL SCIENCE IN PSYCHOLOGY
69
Banich, M. T., Milham, M. P., Jacobson, B. L., Webb, A., Wszalek, T., Cohen, N. J., et al. (2001). Attentional selection and the processing of task-irrelevant information: Insights from fMRI examinations of the Stroop task. Progress in Brain Research, 134, 459–470. Bradley, M. M., Codispoti, M., Cuthbert, B. N., & Lang, P. J. (2001). Emotion and motivation: I. Defensive and appetitive reactions in picture processing. Emotion, 1, 276–298. Bradley, M. M., & Lang, P. J. (1999a). Affective norms for English words (ANEW): Instruction manual and affective ratings (Tech. Rep. C-1). Gainesville: Center for Research in Psychophysiology, University of Florida. Bradley, M. M., & Lang, P. J. (1999b). International affective digitized sounds (IADS): Stimuli, instruction manual and affective ratings (Tech. Rep. B-2). Gainesville: Center for Research in Psychophysiology, University of Florida. Bramon, E., Rabe-Hesketh, S., Sham, P., Murray, R. M., & Frangou, S. (2004). Metaanalysis of the P300 and P50 waveforms in schizophrenia. Schizophrenia Research, 70, 315–329. Cohen, J. D., & Servan-Schreiber, D. (1992). Context, cortex, and dopamine: A connectionist approach to behavior and biology in schizophrenia. Psychological Review, 99, 45–77. Compton, R. J., Banich, M. T., Mohanty, A., Milham, M. P., Herrington, J., Miller, G. A., et al. (2003). Paying attention to emotion: An fMRI investigation of cognitive and emotional Stroop tasks. Cognitive, Affective & Behavioral Neuroscience, 3, 81–96. Cook, W. III, & Turpin, G. (1997). Differentiating orienting, startle, and defense responses: The role of affect and its implications for psychopathology. In P. J. Lang, R. F. Simons, & M. T. Balaban (Eds.), Attention and orienting: Sensory and motivational processes (pp. 137–164). Mahwah, NJ: Lawrence Erlbaum Associates. Davidson, R. J. (1992). Anterior cerebral asymmetry and the nature of emotion. Brain and Cognition, 20, 125–151. Davidson, R. J. (1998). Affective style and affective disorders: Perspectives from affective neuroscience. Cognition and Emotion, 12, 307–330. Davidson, R. J. (2003). Presidential address: Affective neuroscience and psychophysiology: Toward a synthesis. Psychophysiology, 40, 655–665. Davidson, R. J., Jackson, D. C., & Kalin, N. H. (2000). Emotion, plasticity, context and regulation: Perspectives from affective neuroscience. Psychological Bulletin, 126, 890–906. Davidson, R. J., Pizzagalli, D., Nitschke, J. B., & Putnam, K. M. (2002). Depression: Perspectives from affective neuroscience. Annual Review of Psychology, 53, 545–574. Davidson, R. J., & Sutton, S. K. (1995). Affective neuroscience: The emergence of a discipline. Special Cognitive Neuroscience Issue for Current Opinion in Neurobiology, 5, 217–224. Dawes, R. (1994). House of cards: Psychology and psychotherapy built on myth. New York: The Free Press. Dawson, M. E., Schell, A. M., Swerdlow, N. R., & Filion, D. L. (1997). Cognitive, clinical, and neurophysiological implications of startle modification. In
Chapter 03.qxd
70
4/6/2007
2:43 PM
Page 70
MILLER, ENGELS, HERRINGTON
P. J. Lang, R. F. Simons, & M. T. Balaban (Eds.), Attention and orienting: Sensory and motivational processes (pp. 257–279). Mahwah, NJ: Lawrence Erlbaum Associates. Edgar, J. C., Huang, M., Weisend, M. P., Sherwood, A., Miller, G. A., Adler, L. E., et al. (2003). Interpreting abnormality in event-related brain potentials: A case study of P50 sensory gating. Biological Psychology, 65, 1–21. Edgar, J. C., Miller, G. A., Moses, S. N., Thoma, R. J., Huang, M. X., Hanlon, F. M., et al. (2005). Cross-modal generality of the gating deficit. Psychophysiology, 42, 318–327. Engels, A. S., Mohanty, A., Herrington, J. D., Ho, M. R., Wenzel, E. A., Fisher, J. E., et al. (2004, October). Regional specificity of anxiety subtypes during emotion processing. Paper presented at the annual meeting of the Society for Research in Psychopathology, St. Louis, MO. Fodor, J. A. (1968). Psychological explanation. New York: Random House. Freedman, R., Adler, L. E., & Leonard, S. (1999). Alternative phenotypes for the complex genetics of schizophrenia. Biological Psychiatry, 45, 551–558. Freedman, R., Adler, L. E., Waldo, M. C., Pachtman, E., & Franks, R. D. (1983). Neurophysiological evidence for a defect in inhibitory pathways in schizophrenia: Comparison of medicated and drug-free patients. Biological Psychiatry, 18, 537–551. Freedman, R., Coon, H., Myles-Worsley, M., Orr-Urtreger, A., Olincy, A., Davis, A., et al. (1997). Linkage of a neurophysiological deficit in schizophrenia to a chromosome 15 locus. Proceedings of the National Academy of Sciences of the United States of America, 94, 587–592. Grace, A. A. (2000). Gating of information flow within the limbic system and the pathophysiology of schizophrenia. Brain Research–Brain Research Reviews, 31, 330–341. Grace, A. A., & Moore, H. (1998). Regulation of information flow in the nucleus accumbens: A model for the pathophysiology of schizophrenia. In F. Lenzenweger & R. H. Dworkin (Eds.), Origins and development of schizophrenia: Advances in experimental psychopathology (pp. 123–157). Washington, DC: American Psychological Association. Graham, F. K. (1975). The more or less startling effects of weak prestimulation. Psychophysiology, 12, 238–248. Graham, F. K. (1979). Distinguishing among orienting, defense, and startle reflexes. In H. D. Kimmel, E. H van Olst, & J. F. Orlebeke (Eds.), The orienting reflex in humans (pp. 137–167). Hillsdale, NJ: Lawrence Erlbaum Associates. Hanlon, F. M., Miller, G. A., Thoma, R. J., Irwin, J., Moses, S. N., Huang, M., et al. (2005). Distinct M50 and M100 auditory gating deficits in schizophrenia. Psychophysiology, 42, 417–427. Hanlon, F. M., Weisend, M. P., Hamilton, D. A., Jones, A. P., Thoma, R. J., Huang, M., et al. (2006). The virtual Morris water task identifies a hippocampal deficit in schizophrenia. Schizophrenia Research, 87, 67–80. Hanlon, F. M., Weisend, M. P., Huang, M., Lee, R. R., Moses, S. N., Paulson, K. M., et al. (2003). A noninvasive method for observing hippocampal function. NeuroReport, 14, 1957–1960.
Chapter 03.qxd
4/6/2007
2:43 PM
Page 71
3. SEDUCTION OF CLINICAL SCIENCE IN PSYCHOLOGY
71
Hanlon, F. M., Weisend, M. P., Yeo, R. A., Huang, M., Thoma, R. J., Lee, R. R., et al. (2005). A specific test of hippocampal deficit in schizophrenia. Behavioral Neuroscience, 119, 863–875. Harmon-Jones, E., & Allen, J. J. B. (1998). Anger and frontal brain activity: EEG asymmetry consistent with approach motivation despite negative affective valence. Journal of Personality and Social Psychology, 74, 1310–1316. Harrison, P. J., & Weinberger, D. R. (2005). Schizophrenia genes, gene expression, and neuropathology: On the matter of their convergence. Molecular Psychiatry, 10, 40–68. Heller, W., Koven, N. S., & Miller, G. A. (2003). Regional brain activity in anxiety and depression, cognition/emotion interaction, and emotion regulation. In K. Hugdahl & R. J. Davidson (Eds.), The asymmetrical brain (pp. 533–564). Cambridge, MA: MIT Press. Herrington, J. D., Mohanty, A., Koven, N. S., Fisher, J. E., Stewart, J. L., Banich, M. T., et al. (2005). Emotion-modulated performance and activity in left dorsolateral prefrontal cortex. Emotion, 5, 200–207. Hetrick, W. P., & Smith, D. A. (under review). Exploring sensory gating phenomenology through scale development. Huang, M., Aine, C., Davis, L., Butman, J., Christner, R., Weisend, W., et al. (2000). Sources on the anterior and posterior banks of the central sulcus identified from magnetic somatosensory evoked responses using Multi-Start Spatio-Temporal Localization. Human Brain Mapping, 11, 59–76. Huang, M. X., Edgar, J. C., Thoma, R. J., Hanlon, F. M., Moses, S. N., Lee, R. R., et al. (2003). Predicting EEG responses using MEG sources in superior temporal gyrus reveals source anomaly in patients with schizophrenia. Clinical Neurophysiology, 114, 835–850. Kawamura, T., Nakasato, N., Seki, K., Kanno, A., Fujita, S., Fujiwara, S., et al. (1996). Neuromagnetic evidence of pre- and post-central cortical sources of somatosensory evoked responses. Electroencephalography and Clinical Neurophysiology, 100, 44–50. King-Casas, B., Tomlin, D., Anen, C., Camerer, C. F., Quartz, S. R., & Montague, P. R. (2005). Getting to know you: Reputation and trust in a two-person economic exchange. Science, 308, 78–83. Kisley, M. A., Noecker, T. L., & Guinther, P. M. (2004). Comparison of sensory gating to mismatch negativity and self-reported perceptual phenomena in healthy adults. Psychophysiology, 41, 604–612. Kosslyn, S. M., & Koenig, O. (1992). Wet mind: The new cognitive neuroscience. New York: The Free Press. Kozak, M. J., & Miller, G. A. (1982). Hypothetical constructs versus intervening variables: A re-appraisal of the three-systems model of anxiety assessment. Behavioral Assessment, 14, 347–358. Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1997). Motivated attention: Affect, activitation, and action. In P. J. Lang, R. F. Simons, & M. T. Balaban (Eds.), Attention and orienting: Sensory and motivational processes (pp. 97–135). Mahwah, NJ: Lawrence Erlbaum Associates. Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1999). International affective picture system (IAPS): Instruction manual and affective ratings (Tech. Rep. A-4). Gainesville: Center for Research in Psychophysiology, University of Florida.
Chapter 03.qxd
72
4/6/2007
2:43 PM
Page 72
MILLER, ENGELS, HERRINGTON
Lang, P. J., Davis, M., & Öhman, A. (2000). Fear and anxiety: Animal models and human cognitive psychophysiology. Journal of Affective Disorders, 61, 137–159. Leahy, R. M., Mosher, J. C., Spencer, M. E., Huang, M. X., & Lewine, J. D. (1998). A study of dipole localization accuracy for MEG and EEG using a human skull phantom (Tech. Rep. LA-UR-98-1442). Los Alamos, NM: Los Alamos National Laboratory. LeDoux, J. E. (1995). Emotion: Clues from the brain. Annual Review of Psychology, 46, 209–235. Leonard, S., Gault, J., Hopkins, J., Logel, J., Vianzon, R., Short, M., et al. (2002). Association of promoter variants in the alpha 7 nicotinic acetylcholine receptor subunit gene with an inhibitory deficit found in schizophrenia. Archives of General Psychiatry, 59, 1085–1096. Leshner, A. (1997). Addiction is a brain disease, and it matters. Science, 278, 45–47. Lewine, J. D., & Orrison, W. W., Jr. (1995). Magnetoencephalography and magnetic source imaging. In W. W. Orrison, Jr., J. D. Lewine, J. A. Sanders, & M. F. Hartshorne (Eds.), Functional brain imaging (pp. 369–471). St. Louis, MO: Mosby. Light, G. A., Geyer, M. A., Clementz, B. A., Cadenhead, K. S., & Braff, D. L. (2000). Normal P50 suppression in schizophrenia patients treated with atypical antipsychotic medications. American Journal of Psychiatry, 157, 767–771. Marr, D. (1982). Vision: A computational investigation into the human representation and processing of visual information. New York: Freeman. McFall, R. M. (1991). Manifesto for a science of clinical psychology. Clinical Psychologist, 44, 75–88. McGhie, A., & Chapman, J. (1961). Disorders of attention and perception in early schizophrenia. British Journal of Medical Psychology, 34, 103–116. Meehl, P. (1973). Why I do not attend case conferences. In P. E. Meehl (Ed.), Psychodiagnosis: Selected papers (pp. 225–301). Minneapolis: University of Minnesota Press. Meehl, P. E. (1978). Theoretical risks and tabular asterisks: Sir Karl, Sir Ronald, and the slow progress of soft psychology. Journal of Consulting and Clinical Psychology, 46, 806–834. Michel, C. M., Murray, M. M., Lantz, G., Gonzalez, S., Spinelli, L., & Grave de Peralta, R. (2004). EEG source imaging. Electroencephalography and Clinical Neurophysiology, 115, 2195–2222. Milham, M. P., Banich, M. T., Claus, E. D., & Cohen, N. J. (2003). Practice-related effects demonstrate complementary roles of anterior cingulate and prefrontal cortices in attentional control. Neuroimage, 18, 483–493. Miller, G. A. (1995). Strong medicine. [Review of Robyn M. Dawes, House of cards: Psychology and psychotherapy built on myth.] Psychological Science, 6, 129–132. Miller, G. A. (1996). Presidential address: How we think about cognition, emotion, and biology in psychopathology. Psychophysiology, 33, 615–628. Miller, G. A. (2004). Another quasi-thirty years of slow progress. Applied and Preventive Psychology: Current Scientific Perspectives, 11, 61–64. Miller, G. A., Herrington, J. D., Mohanty, A., Engels, A. S., Edgar, J. C., Banich, M. T., et al. (2004, October). Frontal-cortex circuity in emotional self-regulation. Paper presented in symposium, Frontal Lobes and Psychopathology Revisited, at the annual meeting of the Society for Research in Psychopathology, St. Louis, MO.
Chapter 03.qxd
4/6/2007
2:43 PM
Page 73
3. SEDUCTION OF CLINICAL SCIENCE IN PSYCHOLOGY
73
Miller, G. A., & Keller, J. (2000). Psychology and neuroscience: Making peace. Current Directions in Psychological Science, 9, 212–215. Miller, G. A., & Kozak, M. J. (1993). A philosophy for the study of emotion: Threesystems theory. In N. Birbaumer & A. Öhman (Eds.), The structure of emotion: Physiological, cognitive and clinical aspects (pp. 31–47). Seattle, WA: Hogrefe & Huber. Miller, G. A., Levin, D. N., Kozak, M. J., Cook, E. W., McLean, A., & Lang, P. J. (1987). Individual differences in imagery and the psychophysiology of emotion. Cognition and Emotion, 1, 367–390. Mohanty, A., Engels, A. S., Herrington, J. D., Heller, W., Ringo Ho, M.- H. R., Banich, M. T., et al. (in press). Differential engagement of anterior cingulate cortex subdivisions for cognitive and emotional function. Psychophysiology. Mohanty, A., Herrington, J. D., Koven, N. S., Wenzel, E. A., Webb, A. G., Heller, W., et al. (2005). Neural mechanisms of affective interference in schizotypy. Journal of Abnormal Psychology, 114, 16–27. Montague, P. R., Hyman, S. E., & Cohen, J. D. (2004). Computational roles for dopamine in behavioural control. Nature, 431, 760–767. Nagamoto, H. T., Adler, L. E., Hea, R. A., Griffith, J. M., McRae, K. A., & Freedman, R. (1996). Gating of auditory P50 in schizophrenics: Unique effects of clozapine. Biological Psychiatry, 40, 181–188. Panksepp, J. (1998). Affective neuroscience: The foundations of human and animal emotions. New York: Oxford University Press. Pantev, C., Bertrand, O., Eulitz, C., Verkindt, C., Hampson, S., Schuierer, G., Elbert, T. (1995). Specific tonotopic organizations of different areas of the human auditory cortex revealed by simultaneous magnetic and electric recordings. Electroencephalography and Clinical Neurophysiology, 94, 26–40. Perlstein, W., Fiorito, E., Graham, F. K., & Simons, R. F. (1993). Lead stimulation effects on reflex blink, exogenous brain potentials, and loudness judgments. Psychophysioloqy, 30, 347–358. Powers, W. T. (1973). Behavior: The control of perception. New York: Aldine DeGruyter. Romani, G. L. (1986). Tonotopic organization of the human auditory cortex revealed by steady state neuromagnetic measurements. Acta Otolaryngology, Supplement 432, 33–34. Scherg, M., Ille, N., Bornfleth, H., & Berg, P. (2002). Advanced tools for digital EEG review: Virtual source montages, whole-head mapping, correlation, and phase analysis. Journal of Clinical Neurophysiology, 19, 91–112. Schupp, H. T., Cuthbert, B. N., Bradley, M. M., Cacioppo, J. T., Ito, T., & Lang, P. J. (2000). Affective picture processing: The late positive potential is modulated by motivational relevance. Psychophysiology, 37, 257–261. Shin, L. M., Orr, S. P., Carson, M. A., Rauch, S. L., Macklin, M. L., Lasko, N. B., et al. (2004). Regional cerebral blood flow in amygdala and medial prefrontal cortex during traumatic imagery in male and female Vietnam veterans with PTSD. Archives of General Psychiatry, 61, 168–176. Smith, D. A., Boutros, N. N., & Schwarzkopf, S. B. (1994). Reliability of P50 auditory event-related potential indices of sensory gating. Psychophysiology, 31, 495–502. Srinivasan, R., Tucker, D. M., & Murias, M. (1998). Estimating the spatial Nyquist of the human EEG. Behavior Research Methods, Instruments, and Computers, 30, 8–19.
Chapter 03.qxd
74
4/6/2007
2:43 PM
Page 74
MILLER, ENGELS, HERRINGTON
Stephen, J. M., Aine, C. J., Christner, R. F., Ranken, D., Huang, M., & Best, E. (2002). Central versus peripheral visual field stimulation results in timing differences in dorsal stream sources as measured with MEG. Vision Research, 42, 3059–3074. Sutton, S. K., & Davidson, R. J. (1997). Prefrontal brain asymmetry: A biological substrate of the behavioral approach and inhibition systems. Psychological Science, 8, 204–210. Taitano, K., & Miller, G. A. (1998). Neuroscience perspectives on emotion in psychopathology. In W. Flack & J. Laird (Eds.), Emotion in psychopathology: Theory and research (pp. 20–44). New York: Oxford. Thoma, R. J., Hanlon, F. M., Moses, S. N., Edgar, J. C., Huang, M. X., Weisend, M. P., et al. (2003). Lateralization of auditory sensory gating and neuropsychological dysfunction in schizophrenia. American Journal of Psychiatry, 160, 1595–1605. Thoma, R. J., Hanlon, F. M., Moses, S. N., Ricker, D., Huang, M. X., Edgar, C., Irwin, J., Torres, F., Weisend, M. P., Adler, L. E., Miller, G. A., & Cañive, J. M. (2005). M50 sensory gating predicts negative symptoms in schizophrenia. Schizophrenia Research, 73, 311–318. Thoma, R. J., Hanlon, F. M., Sanchez, N., Weisend, M. P., Huang, M., Jones, A., et al. (2004). Auditory sensory gating deficit and cortical thickness in schizophrenia. Neurology and Clinical Neurophysiology, 62. Retrieved December 6, 2006, from www.ncnpjournal.com. Tucker, D. M., Derryberry, D., & Luu, P. (2000). Anatomy and physiology of human emotion: Vertical integration of brainstem, limbic, and cortical systems. In J. Borod (Ed.), Handbook of the neuropsychology of emotion (pp. 56–79). New York: Oxford University Press. Tucker, D. M., Luu, P., Frishkoff, G., Quiring, J., & Poulsen, C. (2003). Frontolimbic response to negative feedback in clinical depression. Journal of Abnormal Psychology, 112, 667–678. Yee, C. M., Nuechterlein, K. H., Morris, S. E., & White, P. M. (1998). P50 suppression in recent-onset schizophrenia: Clinical correlates and risperidone effects. Journal of Abnormal Psychology, 107, 691–698.
Chapter 04.qxd
3/28/2007
12:58 PM
Page 75
II CLINICAL SCIENCE: TOPICS OF APPLIED SIGNIFICANCE
Chapter 04.qxd
3/28/2007
12:58 PM
Page 76
Chapter 04.qxd
3/28/2007
12:58 PM
Page 77
4 Clinical Science and the Revolution in Psychological Treatment: The Example of Anxiety Disorders Gabrielle I. Liverant, Jill A. Stoddard, Alicia E. Meuret, and David H. Barlow Center for Anxiety and Related Disorders, Boston
It is no coincidence that the development of a scientifically based clinical psychology began with and spans the career of Dick McFall. From its beginnings in the 1960s, clinical psychologists, other mental health professionals, and mental health policymakers have become convinced that a scientifically based approach to clinical psychology fundamentally rooted in evidence greatly advances our field to the benefit of both practitioners and consumers. Dick McFall, over the course of his career, has had an enormous influence on this process. This influence is evident in his creation of the Academy of Psychological Clinical Science to guide training programs in ways to become more deeply and broadly scientifically based, and in his steadfast belief that an approach based on science is the only path to progress. It is also evident in the specific example of his research. In this chapter, we describe one example of the progression of a scientific base in our field—the development of knowledge on the nature and treatment of anxiety and its disorders. We begin with a brief history of the 77
Chapter 04.qxd
78
3/28/2007
12:58 PM
Page 78
LIVERANT ET AL.
development of scientifically based interventions for anxiety disorders and continue with an explication of advances in our understanding of the psychopathology of anxiety. The reciprocal influence of increased understanding of psychopathology and treatment advances is underscored because one of the fundamental tenets of a scientifically based process of psychological interventions is that these interventions are closely targeted to the psychopathology on which they focus. We then proceed with a further description of the process of evaluation of psychological treatments for anxiety disorders and recent cutting-edge developments based on new science or technological advances. Finally, we describe the beginnings of important progress in disseminating these procedures to the front lines of clinical practice. EARLY DEVELOPMENT OF SCIENTIFICALLY BASED INTERVENTIONS Over the past century, advances in conceptualizations of human behavior and emotional disorders have evolved through a sequence of stages that have shaped the face of psychological treatment for anxiety disorders today. Early scientists, such as John B. Watson, Rosalie Raynor (Watson & Raynor, 1920), and Mary Cover Jones (1924), endeavored to make the application of psychology to human problems a more objective science. Although this earliest phase of behavioral thought did not translate to an immediate or robust applied science, it laid the groundwork for a number of important innovations in psychological practice. By the mid-20th century, behavioral theory had blossomed into a number of specific, applied principles marking the beginnings of behavior therapy. At that time, psychological treatments began to target specific problems (e.g., specific phobias) in lieu of vague, hypothetical underlying personality dispositions. Early psychological treatments for anxiety and fear arose out of stimulus–response learning theory and related experimental studies of fear acquisition. Watson and Raynor (1920) conducted one of the earliest and best-known experimental investigations of fear acquisition. They demonstrated that repeated pairings of a loud noise (unconditioned stimulus) with the presentation of a white rat (conditioned stimulus) resulted in a child’s conditioned fear of rats (that generalized across similar stimuli). Findings from this and similar investigations led to a general belief that all phobias were acquired through classical conditioning. Thus, the fundamental basis of early behavioral treatments for fear and anxiety hinged on extinction
Chapter 04.qxd
3/28/2007
12:58 PM
Page 79
4. REVOLUTION IN PSYCHOLOGICAL TREATMENT
79
of conditioned fears when the conditioned stimulus (feared object) was presented repeatedly in the absence of the unconditioned stimulus. Mowrer’s (1939) two-factor model supplemented the classical fear conditioning paradigm by asserting that anxiety and fear persisted (i.e., did not undergo extinction) because phobic individuals learned to avoid the feared object, thus preventing the opportunity for extinction to occur. Additionally, avoidance behavior was negatively reinforced because it resulted in reduced anxiety. The actual establishment of behavior therapy for anxiety and fear is probably best depicted by the pivotal work of Wolpe (1958). Wolpe developed systematic desensitization (SD) to reduce pathological fears in humans, citing a process he referred to as reciprocal inhibition. In his laboratory research, Wolpe found that conditioned anxiety inhibited eating. Thus, he reasoned that eating might reciprocally inhibit anxiety. According to Wolpe (1958), “if a response antagonistic to anxiety can be made to occur in the presence of anxiety-evoking stimuli so that it is accompanied by a complete or partial suppression of the anxiety responses, the bond between these stimuli and the anxiety responses will be weakened” (p. 71). Although the reciprocal inhibition theory has long been discredited, it was the basis for one of the first scientifically based psychological interventions, SD. SD is characterized by gradual, imaginal exposure to a hierarchy of feared stimuli in conjunction with relaxation, a hypothetically fearincompatible response. SD was among the first operationally defined and empirically testable therapeutic approaches. Initially, SD was found to be efficacious in reducing mild specific fears (Hasselt & Hersen, 1994). However, the intervention demonstrated limited efficacy when applied to clinical populations, a wider variety of feared stimuli, and more complex anxiety conditions such as agoraphobia (Barlow, 1988, 1994). Because of these limitations, SD began to fall out of favor as a treatment strategy of choice. Nevertheless, the development of SD marked an important transition in the history of psychological treatment for anxiety by allowing for the translation of basic behavioral science into the practice of psychotherapy. Moreover, it ignited a movement in psychotherapy research to expand its focus from largely process-oriented inquiries to those that included psychotherapy outcomes (Barlow & Hersen, 1984). With the promise of SD and translational research in general came a substantial increase in psychotherapy investigations, especially exposure-based interventions for anxiety and fear. Agras, Leitenberg, and Barlow (1968)
Chapter 04.qxd
80
3/28/2007
12:58 PM
Page 80
LIVERANT ET AL.
and Marks (1971) demonstrated that in vivo exposure to feared situations without relaxation resulted in effective and rapid reductions in fear and avoidance in patients with panic disorder and agoraphobia. Stampfl and Levis (1967) described decreases in anxiety and obsessive-compulsive disorder (OCD) symptoms with implosion therapy or flooding. These were considered novel approaches because traditional beliefs and the zeitgeist at the time held that experiencing high levels of anxiety could result in harm to the patient. However, research demonstrated that exposure in the absence of relaxation was efficacious in reducing anxiety and fear and did not cause harm to patients. In vivo exposure remains an important ingredient in behavioral treatment for anxiety disorders today. Within a decade of Wolpe’s groundbreaking innovations, anxiety interventions based on pure stimulus–response learning theory were seen as insufficient, primarily due to the inability of strictly associationistic principles to address the matter of cognitive processing of emotional stimuli (Follette & Hayes, 2000). Researchers began to investigate more integrative theories of behavior change, and subsequent mediational models of psychopathology materialized. Principles from social learning theory (Bandura, 1977), emotional processing theory (e.g., Foa & Kozac, 1986; Lang, 1977; Rachman, 1980), cognitive science (e.g., Beck, Emery, & Greenberg, 1985), and other related writings (e.g., Staats, 1975) led to a more sophisticated understanding of psychopathology and the mechanisms of action underlying psychotherapeutic techniques. Out of this understanding emerged cognitive-behavioral approaches to treatment. Our deepening understanding of the nature of emotional disorders allowed for an elaboration of old stimulus–response models of anxiety and fear acquisition and reduction. For example, consequent research established that, beyond stimulus–response associations, individuals exhibiting clinical fear states commit evaluative errors (Foa & Kozac, 1986; Rachman, 1980). In other words, phobic individuals (a) believe anxiety will endure unless the feared stimulus is escaped, (b) overestimate the likelihood of harm resulting from fear/anxiety, and (c) regard the consequences of fear to be more aversive than do nonphobic individuals. Thus, research investigating the mechanisms of action in exposure therapy indicated that exposure to a feared stimulus facilitated extinction and physiological habituation to anxiety and also modified dysfunctional associations and interpretations (Barlow, 2002). Notably, although catastrophic cognitions have been shown to occur with disorders, it is unclear whether they are causal or consequent in the actual etiology of these disorders
Chapter 04.qxd
3/28/2007
12:58 PM
Page 81
4. REVOLUTION IN PSYCHOLOGICAL TREATMENT
81
(Bouton, Mineka, & Barlow, 2001). Nevertheless, these findings have led directly to the creation of new approaches to the treatment of anxiety. Currently, cognitive-behavioral techniques aimed at modifying distorted beliefs have become a mainstay in anxiety treatment (e.g., Beck et al., 1985; Clark, 1986). An equally important development was the recognition of the strong relationship between anxiety and physiological or somatic sensations in panic and related disorders (Barlow, 1988; Barlow, Craske, Cerny, & Klosko, 1989). That is, individuals with anxiety disorders are not simply reacting to and avoiding external stimuli or cognitions. Internal somatic sensations play an important role in the genesis and maintenance of anxiety and its disorders. For example, research in Russia in the 1960s demonstrated that fear could be conditioned to internal, somatic stimuli. In this research, dogs exposed to electric shock (unconditioned stimulus) repeatedly paired with mild colon stimulation (conditioned stimulus) then experienced significant conditioned anxiety during subsequent natural bowel movements. This conditioned anxiety response to internal sensations came to be known as interoceptive conditioning (Razran, 1961) and was an important discovery for etiological theories and treatment of panic disorder. Panic attacks have been posited to act as conditioning trials during which an individual associates early somatic sensations at the onset of an attack with sensations occurring throughout the resolution of the attack. As a result, somatic cues subsequently act as conditioned stimuli that can elicit a full-blown panic attack (Barlow, 1988; Bouton et al., 2001; Goldstein & Chambless, 1978). Numerous empirical investigations have provided substantial support for the role of interoceptive conditioning in the development of panic disorder (see Bouton et al., 2001). Thus, we developed treatment techniques based on principles of interoceptive conditioning called interoceptive exposure (Barlow, 1988). Interoceptive exposure involves elicitation of, and exposure to, physical symptoms that resemble those of a panic attack. Through repeated exposure to interoceptive cues (elicited during exercises such as spinning in a chair or running in place and adapted to the patient’s profile), individuals habituate to the physiological sensations of anxiety and dysfunctional attributions regarding threat and the probability of harm arising from these stimuli are reinterpreted (Barlow & Craske, 2000). Contemporary, efficacious treatments for panic disorder and agoraphobia typically incorporate components of situational/in vivo exposure, interoceptive exposure, and cognitive restructuring (e.g., Barlow & Craske, 2000).
Chapter 04.qxd
3/28/2007
12:58 PM
Page 82
82
LIVERANT ET AL.
Psychology’s expanding focus on psychotherapy outcomes has necessitated a growing number of specific and reproducible treatment protocols (Hayes, Barlow, & Nelson-Gray, 1999). Early evidence for the efficacy of these protocols has stimulated a growing trend toward large, often multisite, randomized, controlled clinical trials. This research has resulted in our current climate of disorder-specific, manualized therapeutic techniques focused on the alleviation of particular anxiety symptoms (Lambert, Okiishi, Finch, & Johnson, 1998). Advances in psychological treatment have paralleled and contributed to growing knowledge of factors involved in the development and maintenance of anxiety and related disordered emotional states. A comprehensive body of research examining the etiology of emotional disorders has provided a heightened understanding of the underpinnings of psychopathology. This deeper understanding, in turn, has reciprocally influenced the nature and development of treatment for anxiety and related emotional disorders. This reciprocal influence between psychopathology and interventions is a unique characteristic of the development of scientifically based treatments when compared with schools of psychotherapy that are often not scientifically based. THE ORIGINS OF ANXIOUS APPREHENSION AND CLINICAL DISORDERS Over the last half century, research has advanced our comprehension of the factors that contribute to the development of anxious apprehension, anxiety disorders, and related emotional disorders. Barlow (1988, 2002) proposed the triple vulnerability model of anxiety disorders, which describes three sets of synergistic vulnerabilities that contribute to the development of specific disorders. These vulnerabilities include: (a) a generalized biological vulnerability, (b) a generalized psychological vulnerability resulting from early life experiences, and (c) a specific psychological vulnerability, which results in the development of specific disorders. Research findings relevant to each of these areas are reviewed next. Generalized Biological Vulnerability Family and genetic studies have provided considerable evidence that being emotional or nervous runs in families and is heritable (Barlow, 2002). Recent research has also demonstrated that personality constructs such as neuroticism, negative affect, behavioral inhibition, positive affect,
Chapter 04.qxd
3/28/2007
12:58 PM
Page 83
4. REVOLUTION IN PSYCHOLOGICAL TREATMENT
83
and extraversion are heritable traits with differential relationships to mood and anxiety disorders (e.g., Clark, Watson, & Mineka, 1994; Jang, Livesley, & Vernon, 1996; Smoller et al., 2003). Notably, correlational and longitudinal research has shown that the personality traits of neuroticism, negative affect, behavioral inhibition, positive affect, extraversion, and behavioral activation are related to the development, occurrence, and course of mood and anxiety disorders (e.g., Brown, Chorpita, & Barlow, 1998; Gershuny & Sher, 1998; Kasch, Rottenberg, Arnow, & Gotlib, 2002). Although the exact relationships between these related constructs have not been fully explicated, it is likely that neuroticism, negative affect, and behavioral inhibition are related manifestations of a generalized biological vulnerability to develop emotional disorders (Barlow, 2002). Additionally, research examining genetic contributions to anxiety and mood disorders generally supports the existence of a common genetic diathesis for anxiety and related disorders and suggests that environmental factors may be responsible for the differences in the development of specific disorders (Andrews, Stewart, Allen, & Henderson, 1990; Kendler, 1996). However, some preliminary research has suggested unique genetic contributions for some specific disorders—namely, specific phobias (e.g., Fyer, Mannuzza, Chapman, Martin, & Klein, 1995; Kendler, Karkowski, & Prescott, 1999). In addition, evidence supports the existence of a specific genetic contribution for the experience of panic attacks and related acute defensive reactions (Kendler et al., 1995; Marks, 1986; Martin, Jardine, Andrews, & Heath, 1988). Overall, the existence of a specific anxious gene has not been supported despite considerable investigation (Barlow, 2002). Instead, a polygenetic model appears to contribute to a generalized biological vulnerability to experience anxiety, which may be expressed as heritable personality constructs such as neuroticism and negative affect (Plomin, DeFries, McClearn, & Rutter, 1997). Generalized Psychological Vulnerability Barlow (2002) described individuals with anxiety and related disorders as being characterized by a sense of uncontrollability when confronted with threatening situations. This sense of uncontrollability is associated with the production of negative emotions. Significantly, findings from both the animal and human literatures offer support for the presence of a generalized psychological vulnerability, a sense of uncontrollability or unpredictability
Chapter 04.qxd
84
3/28/2007
12:58 PM
Page 84
LIVERANT ET AL.
over significant events or resources, and highlight the role of this vulnerability in the development of anxiety and related disorders. For example, animal models of anxious apprehension are based on the production of experimental neurosis, a state characterized by behavioral and physiological signs similar to those observed during the experience of anxiety in humans. Experimental neurosis has been produced in laboratory settings through a variety of different procedures (e.g., punishment of appetitive responses, long periods of restraint). In a comprehensive review, Mineka and Kihlstrom (1978) posited that there is a common causal factor, uncontrollability and unpredictability of environmental events of vital importance to the animal, which is shared by all of these laboratory paradigms. A large amount of experimental research has manipulated the controllability and predictability of stimuli in different species and found a strong relationship between lack of control over events and the development of states that resemble anxiety and depression (e.g., Overmier & Seligman, 1967; Weiss, 1971). More specifically, findings from studies with primates have demonstrated that developmental experiences with unpredictable and uncontrollable negative events produce long-term increases in primate anxiety (e.g., Mineka, Gunnar, & Champoux, 1986). In these types of experimental paradigms, the initial stressful event, such as separation from a primary caregiver or introduction to strangers, produces a panic-like reaction in immature primates, which is followed by a chronic pattern of anxiety (Barlow, 2002). Additionally, Sapolsky’s extensive research examining the physiology of stress among primates in the wild suggests that the most important factor in determining responsivity in the hypothalamo-pituatary-adrenal (HPA) axis is the predictability of resources and sources of social support (Sapolsky, Alberts, & Altman, 1997). Other primate researchers have also proposed that early life stressful experiences in combination with a genetic predisposition result in a neurobiological diathesis composed of chronic changes in the HPA axis and other biological systems seemingly associated with anxiety (Coplan et al., 1998). These animal findings are also in accordance with research with humans that has implicated family influences, attachment, and a developing sense of control in neuroendocrine functioning in children (e.g., Granger, Weisz, & Kauneckis, 1994; Gunnar, Larson, Hertsgaard, Harris, & Broderson, 1992). Taken as a whole, these findings from animal research are largely consistent with work examining psychological correlates of anxiety in
Chapter 04.qxd
3/28/2007
12:58 PM
Page 85
4. REVOLUTION IN PSYCHOLOGICAL TREATMENT
85
humans. For example, Julian Rotter (1954) hypothesized that an internal versus external locus of control mediates the reinforcing properties of an event. Later research with psychometrically validated self-report measures has found an association between external locus of control and anxiety and depression (e.g., McCauley, Mitchell, Burke, & Moss, 1988). In particular, evidence suggests that a more specific definition of control, perceived controllability over threat and reactions to threat, is associated with anxiety among adults (Rapee, Craske, Brown, & Barlow, 1996). Theory and research with the related construct of attributional style has suggested that internal, stable, and global attributions moderate the relationship between the experience of negative events and the development of psychopathology—namely, depression (Abramson, Seligman, & Teasdale, 1978). Building on this work, Abramson, Metalsky, and Alloy (1989) hypothesized that attributional style contributes to a sense of hopelessness, which is related to depression. Longitudinal findings have generally supported a relationship between attributional style and depression (e.g., Nolen-Hoeksema, Girgus, & Seligman, 1986). However, this focus on depression in the literature may obscure the relationship between control-related cognitions and anxiety and the relationship between controlrelated cognitions and negative affect. For example, Luten, Ralph, and Mineka (1997) found that pessimistic attributional style was associated with anxiety, negative affect, and depression. Additionally, longitudinal research examining the temporal relationship between anxiety and depression suggests anxiety in children may be predictive of later depression (Cole, Peeke, Martin, Truglio, & Seroczynski, 1998). Taken together, these findings suggest that attributional style or control-related cognitions may be a nonspecific diathesis for negative affect, which is related to the experience of both anxiety and depression (Brown et al., 1998). In addition, a large body of work has developed examining the influence of parenting styles on the development of control-related cognitions as well as the development of anxiety and depression. In a comprehensive review, Chorpita and Barlow (1998) identified two parenting dimensions that either facilitate or impede the development of a sense of control. These dimensions are: (a) warmth or sensitivity, consistency, and contingency; and (b) encouragement of autonomy and absence of intrusion or overcontrolling style. Additionally, studies using a variety of correlational methodologies have demonstrated a relationship between these two parenting dimensions or similar constructs and the experience of anxiety and depression (e.g., Reiss et al., 1995; Silove, Parker, Hadzi-Pavlovic,
Chapter 04.qxd
3/28/2007
12:58 PM
Page 86
86
LIVERANT ET AL.
Manicavasagar, & Blaszczynski, 1991). Interestingly, findings from the field of attachment are also largely consistent with this work examining the development of a sense of control and its relationship to anxiety and depression (e.g., Bowlby, 1980; Chorpita & Barlow, 1998). Building on research examining the development of control-related cognitions and psychopathology, several studies have attempted to examine the nature of the relationship between this generalized psychological vulnerability, life events, and emotional disorders. Specifically, research has examined the question of whether a sense of uncontrollability acts as a moderator or mediator between events and anxiety and depression during childhood development (Chorpita, Brown, & Barlow, 1998; Cole & Turner, 1993; Turner & Cole, 1994). In early development, results demonstrate that a generalized psychological sense of uncontrollability serves as a mediator between events and psychopathology (Chorpita et al., 1998; Cole & Turner, 1993). In contrast, in later childhood and adulthood, this psychological vulnerability appears to function as a moderator between events and anxiety and depression (Nolen-Hoeksema, Girgus, & Seligman, 1992; Turner & Cole, 1994). When combined, these findings may suggest a significant developmental sequence in the creation of this psychological vulnerability where early life events contribute to the formation of a cognitive pattern in early development (i.e., mediational model) and then, in later development, the existing cognitive template serves to enhance the relationship between life events and anxiety and depression (i.e., moderational model; Barlow, 2002). In summary, the combination of a genetically based and heritable biological vulnerability with the presence of a generalized psychological vulnerability based on early experience sets the occasion for the development of clinical disorders such as generalized anxiety and depressive disorders when triggered by stressful life events (Barlow, 2002). It is possible that panic attacks (false alarms) may occur in response to high baseline anxiety, perhaps as a function of a somewhat different genetic diathesis as well as the experience of stressful events. However, the exact nature of the clinical presentation or disorder is not defined without the additional influence of a third set of vulnerabilities: specific psychological vulnerabilities. Specific Psychological Vulnerabilities Specific psychological vulnerabilities, which are also developed from early learning experience, direct the focus of an individual’s anxiety onto
Chapter 04.qxd
3/28/2007
12:58 PM
Page 87
4. REVOLUTION IN PSYCHOLOGICAL TREATMENT
87
a specific object or event (Barlow, 2002). For example, in panic disorder, the focus of anxious apprehension is somatic or interoceptive cues, which may signal the next panic attack. In specific phobia, specific psychological vulnerabilities shape the particular object or situation that becomes feared. In social phobia, early learning experiences related to the danger of social evaluation have been implicated in the development of the disorder. In OCD, obsessional thoughts, images, and impulses become the focus of anxious apprehension. The influence of specific psychological vulnerabilities is demonstrated by studies investigating the relationship between early life experience and the development of panic disorder. For example, relative to patients with other anxiety disorders and control subjects, individuals with panic symptoms reported more frequent experience of chronic illness in their households during childhood, more frequent observation of their parents’ experience of paniclike symptoms, and the receipt of more parental encouragement for sick-role behavior during their experience of panic-like symptoms (Ehlers, 1993). Additionally, research suggests that a focus on specific somatic sensations, such as suffocation cues, may also be influenced by specific learning histories (e.g., Craske, Poulton, Tsao, & Plotkin, 2001). With respect to social phobia, research has demonstrated that parents of children with social anxiety discuss the threatening aspects of social interactions with their children and reinforce social avoidance in their children (Barrett, Rapee, Dadds, & Ryan, 1996). Additionally, relative to the parents of controls, parents of socially anxious individuals are described by their children as significantly more fearful and avoidant in social situations and more concerned with social evaluation (Bruch & Heimberg, 1994). Overall, the development of specific anxiety disorders seems to emerge from these specific psychological vulnerabilities, which are based on vicarious learning during development. Specific psychological vulnerabilities focus anxious apprehension on specific objects or situations, producing symptom clusters consistent with our current classification of mental disorders. However, these specific psychological vulnerabilities alone are not sufficient to produce a clinical disorder. Instead, it is the synergistic interaction of a heritable biological vulnerability, a generalized psychological vulnerability (a sense of uncontrollability or unpredictability over important events or resources), and these specific psychological vulnerabilities that produces clinical disorders (Barlow, 2002; see Fig. 4.1).
Chapter 04.qxd
3/28/2007
12:58 PM
Page 88
88
LIVERANT ET AL.
Diatheses
Synergistic Vulnerabilities
Biological Vulnerabilities
Generalized Psychological Vulnerability
Stress
False Alarms (Panic)
Generalized Anxiety
Depression FIGURE 4.1 The triple vulnerability model of anxiety disorders. From Barlow (2002), with permission.
Triple Vulnerabilities and Current Treatments Stemming from these advances in our understanding of the origins of anxious apprehension and specific anxiety and mood disorders, specialized treatments for individual disorders have been developed, which target the combination of vulnerabilities implicated in the experience of unique emotional disorders. Notably, psychological treatments for a wide range of emotional disorders share treatment elements that target common, generalized vulnerabilities. For example, the majority of current treatments for emotional disorders address a sense of uncontrollability and unpredictability over significant events, as well as a perception of lack of control over emotions. Overall, the primary processes that treatments use to address this generalized psychological vulnerability can be grouped into three broad categories: antecedent cognitive reappraisal, emotional exposure, and changing action tendencies (Barlow, Allen, & Choate, 2004). For instance, cognitive-behavioral treatments for a variety of emotional disorders
Chapter 04.qxd
3/28/2007
12:58 PM
Page 89
4. REVOLUTION IN PSYCHOLOGICAL TREATMENT
89
modify negative appraisals of external events as well as strengthen an individual’s sense of self-efficacy in coping with these events (e.g., Beck, 1972; Steketee, 1993). Moreover, cognitive-behavioral treatments focus on changing emotional action tendencies through the reduction of both emotional and situational avoidance patterns and the development of new, more adaptive behaviors (e.g., Jacobson, Martell, & Dimidjian, 2001). In addition to addressing more generalized biological and psychological vulnerabilities, current empirically supported treatments precisely tailor these broad therapeutic techniques to target the unique objects or events relevant to individual disorders (Chambless & Hollon, 1998). Thus, our emerging scientific knowledge from experimental psychopathology has translated into empirically supported treatments. FURTHER DEVELOPMENT OF SPECIFIC EMPIRICALLY SUPPORTED TREATMENTS: THE CASE OF PANIC DISORDER In this section, the latest findings from empirical studies of psychological treatments for panic disorder with and without agoraphobia (PDA) are presented, along with cutting-edge evaluation of three new areas based on scientific or technological advances. These new areas include: brief treatment, new developments in breathing retraining, and virtual reality. As noted earlier, by the early 1980s, numerous studies showed that patients with agoraphobia who underwent exposure-based procedures experienced substantial clinical improvement in their symptomatology (e.g., Agras et al., 1968; Marks, 1971). Nevertheless, only 10% to 20% approached levels of normal functioning, whereas the majority continued to suffer from anxiety, panic, and residual avoidance (Jansson & Öst, 1982). By the mid-1980s, comprehensive treatment approaches for PDA were designed to modify and attenuate catastrophic thoughts, dysregulated emotions, and exaggerated avoidance behavior through communication of the therapy rationale, cognitive restructuring, and interoceptive exposure. As mentioned previously, disorder-specific manuals were developed to facilitate dissemination of empirically supported treatments into clinical settings. One widely used treatment for PDA developed in our center is referred to as panic control treatment (PCT; Barlow & Craske, 2000; Craske, Barlow, & Meadows, 2000). Overall, cognitive-behavioral treatments (CBT) consistently have demonstrated efficacy in the treatment of PDA. In a meta-analysis, Gould, Otto, and Pollack (1995) evaluated 43 controlled studies and found that CBT showed the largest effect sizes and the
Chapter 04.qxd
90
3/28/2007
12:58 PM
Page 90
LIVERANT ET AL.
smallest rate of attrition when compared with drug treatment or combined psychological and drug treatment. Studies using interoceptive exposure not only provided the largest effect sizes, but showed an incremental increase in quality of life for patients (e.g., Telch, Schmidt, Jaimez, Jacquin, & Harrington, 1995). In a recent multicenter study (Barlow, Gorman, Shear, & Woods, 2000), representing the largest combined treatment study for PDA to date (n = 312 panic patients), patients with mild to moderate agoraphobia were randomly assigned to receive imipramine, CBT, CBT plus placebo, pill placebo only, or CBT plus imipramine. All four active treatments resulted by the end of treatment in marked improvements that were significantly greater than for the pill placebo condition. No differences were observed between imipramine and CBT, yet the combination treatment outperformed CBT on several measures, but was not superior to CBT plus placebo. These data suggest that any increased efficacy of adding imipramine to CBT by the end of treatment was accounted for by the nonspecific effects associated with taking a pill. Furthermore, the outcome of combined CBT and drug tended to become less optimal once medication was withdrawn, relative to CBT alone (also see Marks et al., 1993), suggesting the importance of CBT for long-term maintenance of symptom improvement. Given the efficacy of CBT, investigators focused their efforts on developing briefer, more cost-effective versions of CBT for PDA. Promising results were presented by Craske, Maidenberg, and Bystritsky (1995) and Clark and colleagues (1999). Craske et al. (1995) found that a brief foursession PCT protocol was superior to nondirective supportive therapy in reducing panic attack frequency, worry, and phobic fears (avoidance). Clark and colleagues (1999) found similar results when they compared their standard 12-session cognitive therapy to a condensed 5-session protocol in terms of high-end-state functioning. Another promising treatment approach is an 8-day intensive treatment called sensation-focused intensive treatment (SFIT; Heinrichs, Spiegel, & Hofmann, 2002). Preliminary data for SFIT suggest significant improvements in self-report and clinical measures at posttreatment and 5-month follow-up assessments (Spiegel & Barlow, 2000). Overall, CBT for PDA has been shown to be efficacious, and many of the fundamental strategies used in these treatments are currently undergoing further refinement. However, we still know relatively little about the active ingredients of these cognitive-behavioral approaches. Marks and Dar
Chapter 04.qxd
3/28/2007
12:58 PM
Page 91
4. REVOLUTION IN PSYCHOLOGICAL TREATMENT
91
(2000), therefore, called for more research to identify therapy components, which, alone or in combination, reduce each facet of fear (cognitive, behavioral, and physiological). Breathing training (BT) is one example of a component in CBT that has become controversial. Researchers favoring the cognitive model of panic treatment have argued that patients can abuse breathing techniques by using them as safety aids (Salkovskis, Clark, & Gelder, 1996). Other researchers have criticized BT on the grounds that it is a rational placebo (Garssen, de Ruiter, & Van Dyck, 1992), leads to limited improvement (Schmidt et al., 2000), and/or is actively antitherapeutic by acting as a false safety aid (Craske, Rowe, Lewin, & Noriega-Dimitri, 1997). Although theoretically intriguing, no convincing empirical evidence for rejecting breathing training has been demonstrated as of yet (Meuret, Wilhelm, Ritz, & Roth, 2003). In addition, these relatively negative views about safety aids are rather new; previous interpretations of learning theory held that an association of a breathing training-induced low-anxiety state with the anxiety-producing situation would lead to faster fear extinction and therapeutic benefits (Roth, Wilhelm, Pettit, & Meuret, 2005; Wolpe & Rowan, 1988). In fact, breathing training as previously practiced in CBT was probably suboptimal. Traditional breathing exercises or instructions such as “take deep breaths” or “breathe slowly and regularly” risk producing hypocapnia and increasing symptoms (e.g., shortness of breath) by decreasing pCO2. When respiratory rate (RR) is being reduced, compensatory increases in tidal volume can be expected, particularly in patients who are afraid they may not get enough air (Ley, 1991; Meuret et al., 2003). This may explain why conventional BT did not improve outcome when it was added to a CBT package (Schmidt et al., 2000). More recently, mechanisms of change in behavior, cognition, and physiology were evaluated during a behavioral intervention specifically targeting pCO2 levels (Meuret, Wilhelm, & Roth, 2001, 2004). The capnometry-assisted breathing therapy consisted of five 1-hour individual breathing retraining sessions over 4 weeks. Other cognitive-behavioral treatment components were excluded from the breathing therapy, such as systematic cognitive restructuring or exposure to situations or physiological sensations other than those produced during hyperventilation. Patients were taught to increase their end-tidal pCO2 and reduce their RR and tidal volume irregularity using a handheld capnometer for feedback. Overall, results show that this procedure was successful in relieving the symptoms of PDA. At posttreatment, the treated group, but not the wait-list group,
Chapter 04.qxd
92
3/28/2007
12:58 PM
Page 92
LIVERANT ET AL.
improved significantly in PDA severity, anxiety sensitivity, disability, and respiratory measures. Psychological and physiological measures remained improved or improved further at 2- and 12-month reassessments. At 12-month follow-up, 68% of the treated patients no longer experienced panic attacks, and 64% met the conjoint criteria of highend-state functioning. The percentage reaching panic-free status and high-end-state functioning was comparable to findings for other CBT treatments for PDA. Compliance was high for both session attendance and completion of homework exercises. Before treatment, patient pCO2 levels were on average in a hypocapnic range (lower than normal levels of pCO2), but at posttreatment, pCO2 was normalized, an improvement that was maintained 12 months later. Improvement in respiratory and psychological measures was correlated with more normal levels of pCO2 and RR associated with lowered levels of panic severity and disability. These results lend some support to the therapy’s rationale that hypocapnia can cause and maintain anxiety (Ley, 1985) and that patients can learn compensatory skills that allow them to remain normocapnic in anxietyinducing situations. However, alternative interpretations must also be considered. For example, improvement could have been a result of desensitization of an overly sensitive suffocation-alarm system (Klein, 1993), reduction of catastrophic thinking by providing patients with a sense of control, or fear reduction through interoceptive exposure (the latter two interpretations follow a CBT rationale). In this study, Meuret et al. (2004) attributed the high treatment compliance and low attrition to the twofold feedback provided, which included direct capnometer and therapist-delivered capnometer feedback. As with other technology-based treatment approaches, capnometry feedback has a cost. Prices for capnometers range from approximately $1,195 to $2,500. The cost for standard 13-session CBT is approximately $1,300. The fivesession BT treatment would cost $500 plus $1,195 for the capnometer, totaling $1,695 for the first patient ($395 more than CBT). However, the cost would reduce to below standard CBT for subsequent patients because the device can be used multiple times. Integrating technology with psychological treatments such as biofeedback in ambulatory physiological monitoring, computer-administered therapy, or adjunctive palmtop computer therapy allows the use of powerful tools with the potential to enhance treatment compliance, effectiveness, and dissemination (e.g., Kenardy et al., 2003; Newman, Kenardy, Herman, & Taylor, 1997). A good example is virtual reality exposure therapy (VRET).
Chapter 04.qxd
3/28/2007
12:58 PM
Page 93
4. REVOLUTION IN PSYCHOLOGICAL TREATMENT
93
VRET uses technology to create immersive, anxiety-eliciting environments to activate the fear structure, which, according to Foa and Kozak’s (1986) emotional processing theory, is mandatory for habituation and extinction. Research on the effectiveness of VRET for anxiety disorders has been conducted on claustrophobia, fear of flying, fear of driving, acrophobia, spider phobia, fear of public speaking, PDA, and posttraumatic stress disorder (PTSD; Krijn, Emmelkamp, Olafsson, & Biemond, 2004). Promising results show that VRET is as effective as situational exposure in treating fear of heights (e.g., Emmelkamp et al., 2002) and fear of flying (e.g., Rothbaum, Hodges, Smith, Lee, & Price, 2002). However, findings with other anxiety disorders are not yet conclusive due to lack of randomized clinical trials. Overall, VRET may offer advantages over situational exposure such as cost-effectiveness, the ability to adapt exposures to individual needs, the possibility of generating more gradual assignments, and numerous opportunities for repeated exposure. Nevertheless, current drawbacks of VRET, including high attrition and the inability of some patients to involve themselves in the virtual world, await further investigation and improvement (Krijn et al., 2004). DISSEMINATION Although the development of psychological treatments with proven efficacy emerging from the principles of psychological science is an exciting development, equally important are the substantial advances in solving the thorny problem of disseminating evidence-based interventions into practice. Once again, as in so many other areas, Dick McFall has been a leader in this endeavor. Under his sponsorship and guidance, a series of studies was conducted at the Center for Mental Health, a community mental health center in Indiana, examining the transportability of evidence-based practices to front-line clinical settings. For example, Wade, Treat, and Stuart (1998) examined the effectiveness of PCT (Barlow & Craske, 2000; Barlow et al., 2000) for the treatment of all Center for Mental Health patients presenting with PDA. Having carefully trained clinicians in diagnostic procedures and the treatment, the investigators utilized a benchmarking strategy in which they compared their treatment outcome data with those from controlled trials of Barlow et al. (1989) and Telch et al. (1993). There were no exclusionary criteria for participation, and the treatment was delivered by current staff at the community mental health center. Interestingly, outcomes were remarkably similar to those from the controlled trials. Patients in the
Chapter 04.qxd
94
3/28/2007
12:58 PM
Page 94
LIVERANT ET AL.
community mental health center improved on nearly all measures of functioning, and 87.2% of the patients were panic free at the end of the treatment, compared with 84.6% in the Barlow et al. (1989) trial and 85.3% in the Telch et al. (1993) trial. In a later study, these investigators demonstrated that the improvements were durable. Specifically, at a 1-year followup, 70% of the original patients (89% of those contacted) were still panic free (Stuart, Treat, & Wade, 2000). More recently, investigators from this center have extended this analysis to patients presenting with depression with similar outcomes (Merrill, Tolbert, & Wade, 2003). Indeed, other evidence is accumulating for the effectiveness of scientifically based psychological interventions in the context of evidencebased practice. For example, in clinical trials for treatment for social phobia and OCD, patients excluded because they did not meet all inclusion criteria were treated with the interventions used in the controlled trials. The results indicate that rates of improvement were comparable to rates from those who were originally included in the trials (Franklin, Abramowitz, Kozak, Levitt, & Foa, 2000; Juster, Heimberg, & Engelberg, 1995). These studies are important because it is often assumed that patients who are excluded for not meeting all inclusion criteria, particularly from early clinical trials, would not do as well as those included (Westen, Novonty, & Thomson-Brenner, 2004). In fact, there is little evidence to support that assumption, although it is also clear that clinical trials must make every effort to be as inclusive as possible to facilitate analysis of generalizability of results. In addition, greater efforts at organization, communication, and collaboration between research facilities and clinical service settings are taking place. Perhaps the best example is found in practice research networks such as the Pennsylvania Practice Research Network, which has created an infrastructure involving the collaboration of clinical researchers and full-time practitioners to examine the disseminability of newly developed interventions (e.g., Borkovec, Echaemendia, Ragusea, & Ruiz, 2001). There is every reason to expect that more of these practice research networks will emerge in coming years in view of the increasing recognition of the need to advance knowledge of the generalizability of efficacious interventions and the importance of communication between research and service settings. In another notable recent example of efforts to facilitate research on dissemination, the Substance Abuse and Mental Health Services Administration (SAMHSA) has established a National Child Traumatic
Chapter 04.qxd
3/28/2007
12:58 PM
Page 95
4. REVOLUTION IN PSYCHOLOGICAL TREATMENT
95
Stress Network with funding of over $30,000,000. The network’s mission is to develop and disseminate empirically supported interventions to ameliorate trauma-related stress and impairment in children and their families (Substance Abuse and Mental Health Services Administration, 2003). This initiative is particularly noteworthy due to the balance with which it attends to both efficacy and effectiveness research. To accomplish this aim, it funds two principal types of centers: one for treatment development and evaluation, and the other for treatment delivery feedback and adaptation. Funding is specifically allocated to facilitate collaboration and communication among the centers so that all treatment development centers are in continual communication and partnership with treatment delivery centers. All treatment delivery centers have extensive access to research expertise and are encouraged to provide feedback to treatment development centers to influence treatment adaptations and refinements. This initiative provides an important example of the best methods to link research and practice and to disseminate evidencebased practice directly to the front lines of patient care. Similarly, the National Institute on Drug Abuse has funded a large-scale effort to study dissemination of evidence-based practice in the form of the Clinical Trials Network Initiative. Once again, the purpose of this initiative is to evaluate promising new treatments as they are delivered on the front lines of care. Finally, various state governments have been active in establishing large-scale efforts to study and promote dissemination of scientifically based practice. Perhaps the best example occurred in Hawaii as a consequence of the “Felix Consent Decree,” in which it was mandated by the court that all children suffering from psychological disorders that would impair their ability to learn in the public school system receive remedial treatment. Efforts to comply with this consent decree involved statewide committees to ascertain empirically supported interventions, most of them psychological, and additional organized efforts to disseminate these procedures across the state (Chorpita et al., 2002). Of course, this is just the beginning of this type of research, but the momentum is clearly there to expand efforts in this important area in years to come. Perhaps nowhere is the promise of a scientifically based clinical psychology more evident than in the United Kingdom with the recent revolutionary transformations in the National Health Service (NHS). As of 2004, over 5,000 clinical psychologists were working in this system, and policy developments promise a great expansion of positions for psychologists. Specifically, in 1996, the NHS Executive Review described
Chapter 04.qxd
96
3/28/2007
12:58 PM
Page 96
LIVERANT ET AL.
and evaluated evidence for a variety of psychological therapies used to treat children and adults in the NHS. This group concluded that these approaches were clearly effective based on the best evidence available at that time and advised commissioners, policymakers, employers, and providers about how to drive forward the agenda to make these interventions more readily available (The British Psychological Society, 2004). In 2001, when the NHS underwent a substantial reorganization, the government decided to invest considerable monies to promote its agenda for improving health care, and scientifically based psychological interventions were clearly recognized as an important component of these changes. In fact, the NHS projected a sizable gap in the demand for psychologists and addressed this problem in collaboration with the British Psychological Society by projecting that annual growth in the number of clinical psychologists should be increased to 15% each year for the foreseeable future. As a result of the enhanced visibility and demand for clinical psychologists in the NHS, salaries are also increasing substantially, in many instances bringing them on par with physicians. SUMMARY AND CONCLUSIONS Over the course of Dick McFall’s illustrious career, governments and health care policymakers around the world have been persuaded of the value of a scientifically based clinical psychology to the point where it is becoming a prominent part of health care delivery systems. Although the results from our efforts are promising, it is also clear that we have a long way to go in fully understanding the nature of psychopathology, as well as psychological contributions to physical pathology, and in ascertaining the most effective interventions and preventive procedures to address these problems. Facilitating a true integration of discoveries from basic research with clinical application has been a consistent theme of Dick McFall’s work, and the values of translational research are now widely recognized by policymakers around the world. It is also clear that full elucidation of clinical utility or effectiveness depends on close working relationships between clinical scientists developing these interventions and practitioners utilizing them in the community. Practitioners play a major role in this process, and in so doing become full partners and fulfill the role of scientistpractitioner so often envisioned as an ideal in clinical psychology (Barlow, Hayes, & Nelson, 1984; Hayes et al., 1999). For these remarkable advancements, no small amount of credit is due to the efforts of Dick McFall.
Chapter 04.qxd
3/28/2007
12:58 PM
Page 97
4. REVOLUTION IN PSYCHOLOGICAL TREATMENT
97
REFERENCES Abramson, L. Y., Metalsky, G. I., & Alloy, L. B. (1989). Hopelessness and depression: A theory based subtype of depression. Psychological Review, 96, 358–392. Abramson, L. Y., Seligman, M. E. P., & Teasdale, J. D. (1978). Learned helplessness in humans: Critique and reformulation. Journal of Abnormal Psychology, 87, 49–74. Agras, W. S., Leitenberg, H., & Barlow, D. H. (1968). Social reinforcement in the modification of agoraphobia. Archives of General Psychiatry, 19, 423–427. Andrews, G., Stewart, G., Allen, R., & Henderson, A. S. (1990). The genetics of six neurotic disorders: A twin study. Journal of Affective Disorders, 19, 23–29. Bandura, A. (1977). Social learning theory. Upper Saddle River, NJ: Prentice-Hall. Barlow, D. H. (1988). Anxiety and its disorders: The nature and treatment of anxiety and panic. New York: Guilford. Barlow, D. H. (1994). Effectiveness of behavior treatment for panic disorder with and without agoraphobia. In B. E. Wolfe & J. D. Maser (Eds.), Treatment of anxiety and panic: A consensus development conference (pp. 105–120). Washington, DC: American Psychiatric Press. Barlow, D. H. (2002). Anxiety and its disorders: The nature and treatment of anxiety and panic (2nd ed.). New York: Guilford. Barlow, D. H., Allen, L. B., & Choate, M. L. (2004). Toward a unified treatment for emotional disorders. Behavior Therapy, 35, 205–230. Barlow, D. H., & Craske, M. G. (2000). Mastery of your anxiety and panic (3rd ed.). New York: Oxford University Press. Barlow, D. H., Craske, M. G., Cerny, J. A., & Klosko, J. S. (1989). Behavioral treatment of panic disorder. Behavior Therapy, 20, 261–282. Barlow, D. H., Gorman, J. M., Shear, M. K., & Woods, S. W. (2000). Cognitivebehavioral therapy, imipramine, or their combination for panic disorder: A randomized controlled trial. Journal of the American Medical Association, 283, 2529–2536. Barlow, D. H., Hayes, S., & Nelson, R. (1984). The scientist-practitioner. Elmsford, NY: Pergamon. Barlow, D. H., & Hersen, M. (1984). Single case experimental designs: Strategies for studying behavior change (2nd ed.). New York: Pergamon. Barrett, P. M., Rapee, R. M., Dadds, M. M., & Ryan, S. M. (1996). Family enhancement of cognitive style in anxious and aggressive children. Journal of Abnormal Child Psychology, 24, 187–203. Beck, A. T. (1972). Depression: Causes and treatment. Philadelphia: University of Pennsylvania Press. Beck, A. T., Emery, G., & Greenberg, R. L. (1985). Anxiety disorders and phobias: A cognitive perspective. New York: Basic Books. Borkovec, T. D., Echemendia, R. J., Ragusea, S. A., & Ruiz, M. (2001). The Pennsylvania Practice Research Network and future possibilities for clinically meaningful and scientifically rigorous psychotherapy effectiveness research. Clinical Psychology: Science and Practice, 8, 155–167. Bouton, M. E., Mineka, S., & Barlow, D. H. (2001). A modern learning theory perspective on the etiology of panic disorder. Psychological Review, 108, 4–32.
Chapter 04.qxd
98
3/28/2007
12:58 PM
Page 98
LIVERANT ET AL.
Bowlby, J. (1980). Attachment and loss: Vol. 3. Loss: Sadness and depression. New York: Basic Books. Brown, T. A., Chorpita, B. F., & Barlow, D. H. (1998). Structural relationships among dimensions of the DSM–IV anxiety and mood disorders and dimensions of negative affect, positive affect, and autonomic arousal. Journal of Abnormal Psychology, 110, 49–58. Bruch, M. A., & Heimberg, R. G. (1994). Differences in perceptions of parental and personal characteristics between generalized and nongeneralized social phobics. Journal of Anxiety Disorders, 8, 155–168. Chambless, D. L., & Hollon, S. D. (1998). Defining empirically supported therapies. Journal of Consulting and Clinical Psychology, 66, 7–18. Chorpita, B. F., & Barlow, D. H. (1998). The development of anxiety: The role of control in the early environment. Psychological Bulletin, 124, 3–21. Chorpita, B. F., Brown, T. A., & Barlow, D. H. (1998). Perceived control as a mediator of family environment in etiological models of childhood anxiety. Behavior Therapy, 29, 457–476. Chorpita, B. F., Yim, L. M., Donkervoet, J. C., Arensdorf, A., Amundsen, M. J., McGee, C., Servano, A., Yates, A., Burns, J. A., & Morelli, P. (2002). Toward large-scale implementation of empirically supported treatments for children: A review and observations by the Hawaii Empirical Basis to Services Task Force. Clinical Psychology: Science and Practice, 9, 165–190. Clark, D. M. (1986). A cognitive approach to panic. Behavior Research and Therapy, 24, 461–470. Clark, D. M., Salkovkis, P. M., Hackman, A., Wells, A., Ludgate, J., & Gelder, M. (1999). Brief cognitive therapy for panic disorder: A randomized controlled trial. Journal of Consulting and Clinical Psychology, 67, 583–589. Clark, L. A., Watson, D., & Mineka, S. (1994). Temperament, personality, and the mood and anxiety disorders. Journal of Abnormal Psychology, 103, 103–116. Cole, D. A., Peeke, L. G., Martin, J. M., Truglio, R., & Seroczynski, A. D. (1998). A longitudinal look at the relation between depression and anxiety in children and adolescents. Journal of Consulting and Clinical Psychology, 66, 451–460. Cole, D. A., & Turner, J. E. (1993). Models of cognitive mediation and moderation in child depression. Journal of Abnormal Psychology, 102, 271–281. Coplan, J. D., Trost, R. C., Owens, M. J., Cooper, T. B., Gorman, J. M., Nemeroff, C. B., Rosenblum, L. A. (1998). Cerebrospinal fluid concentrations of somatostatin and biogenic amines in grown primates reared by mothers exposed to manipulated foraging conditions. Archives of General Psychiatry, 55, 473–477. Craske, M. G., Barlow, D. H., & Meadows, E. A. (2000). Mastery of your anxiety and panic (3rd ed.). New York: Oxford University Press. Craske, M. G., Maidenberg, E., & Bystritsky, A. (1995). Brief cognitive-behavioral versus non directed therapy for panic disorder. Journal of Behavior Therapy and Experimental Psychiatry, 26, 113–120. Craske, M. G., Poulton, R., Tsao, J. C. I., & Plotkin, D. (2001). Paths to panicagoraphobia: An exploratory analysis from age 3 to 21 in an unselected birth cohort. American Journal of Child and Adolescent Psychiatry, 40, 556–563. Craske, M. G., Rowe, M., Lewin, M., & Noriega-Dimitri, R. (1997). Interoceptive exposure versus breathing retraining within cognitive-behavioural therapy for
Chapter 04.qxd
3/28/2007
12:58 PM
Page 99
4. REVOLUTION IN PSYCHOLOGICAL TREATMENT
99
panic disorder with agoraphobia. British Journal of Clinical Psychology, 36, 85–99. Ehlers, A. (1993). Somatic symptoms and panic attacks: A retrospective study of learning experiences. Behaviour Research and Therapy, 31, 269–278. Emmelkamp, P. M., Krijn, M., Hulsbosch, A. M., de Vries, S., Schuemie, M. J., & van der Mast, C. A. (2002). Virtual reality treatment versus exposure in vivo: A comparative evaluation in acrophobia. Behaviour Research and Therapy, 40, 509–516. Foa, E. B., & Kozak, M. J. (1986). Emotional processing of fear: Exposure to corrective information. Psychological Bulletin, 99, 20–35. Follette, W. C., & Hayes, S. C. (2000). Contemporary behavior therapy. In C. R. Snyder & R. E. Ingram (Eds.), Handbook of psychological change: Psychotherapy processes & practices for the 21st century (pp. 381–408). New York: Wiley. Franklin, M. E., Abramowitz, J. S., Kozak, M. J., Levitt, J. T., & Foa, E. B. (2000). Effectiveness of exposure and ritual prevention for obsessive-compulsive disorder: Randomized compared with nonrandomized samples. Journal of Consulting and Clinical Psychology, 68, 594–602. Fyer, A. J., Mannuzza, S., Chapman, T. F., Martin, L. Y., & Klein, D. F. (1995). Specificity in familial aggregation of phobic disorders. Archives of General Psychiatry, 52, 564–573. Garssen, B., de Ruiter, C., & Van Dyck, R. (1992). Breathing retraining: A rational placebo? Clinical Psychology Review, 12, 141–153. Gershuny, B. S., & Sher, K. J. (1998). The relation between personality and anxiety: Findings from a 3-year prospective study. Journal of Abnormal Psychology, 107, 252–262. Goldstein, A. J., & Chambless, D. L. (1978). A reanalysis of agoraphobia. Behavior Therapy, 9, 47–59. Gould, R. A., Otto, M. W., & Pollack, M. H. (1995). A meta-analysis of treatment outcome in panic disorder. Clinical Psychology Review, 8, 819–844. Granger, D. A., Weisz, J. R., & Kauneckis, D. (1994). Neuroendocrine reactivity, internalizing behavior problems, and control-related cognitions in clinic-referred children and adolescents. Journal of Abnormal Psychology, 103, 267–276. Gunnar, M. R., Larson, M., Hertsgaard, L., Harris, L. M., & Broderson, L. (1992). The stressfulness of separation among 9-month-old infants: Effects of social context variables and infant temperament. Child Development, 63, 290–303. Hasselt, V. B., & Hersen, M. (Eds.). (1993). Advanced abnormal psychology. New York: Plenum. Hayes, S. C., Barlow, D. H., & Nelson-Gray, R. O. (1999). The scientist practitioner: Research and accountability in the age of managed care (2nd ed.). Boston: Allyn & Bacon. Heinrichs, N., Spiegel, D. A., & Hofmann, S. G. (2002). Panic disorder and agoraphobia. In F. Bond & W. Dryden (Eds.), Handbook of brief cognitive behaviour therapy (pp. 55–76). Chichester, England: Wiley. Jacobson, N. S., Martell, C. R., & Dimidjian, S. (2001). Behavioral activation treatment for depression: Returning to contextual roots. Clinical Psychology: Science and Practice, 8, 255–270.
Chapter 04.qxd
100
3/28/2007
12:58 PM
Page 100
LIVERANT ET AL.
Jang, K. L., Livesley, W. J., & Vernon, P. A. (1996). Heritability of the big five personality dimensions and their facets: A twin study. Journal of Personality, 64, 577–591. Jansson, L., & Ost, L. (1982). Behavioral treatments for agoraphobia: An evaluative review. Clinical Psychology Review, 2, 311–336. Jones, M. C. (1924). The elimination of children’s fears. Journal of Experimental Psychology, 7, 383–390. Juster, H. R., Heimberg, R. G., & Engelberg, B. (1995). Self-selection and sample selection in a treatment study of social phobia. Behaviour Research and Therapy, 33, 321–324. Kasch, K. L., Rottenberg, J., Arnow, B. A., & Gotlib, I. H. (2002). Behavioral activation and inhibition systems and the severity and course of depression. Journal of Abnormal Psychology, 111, 589–597. Kenardy, J. A., Dow, M. G. T., Johnston, D. W., Newman, M. G., Thomson, A., & Taylor, C. B. (2003). A comparison of delivery methods of cognitive-behavioral therapy for panic disorder: An international multicenter trial. Journal of Consulting and Clinical Psychology, 71, 1068–1075. Kendler, K. S. (1996). Major depression and generalized anxiety disorder: Same genes, (partly) different environments-revisited. British Journal of Psychiatry, 168(Suppl. 30), 68–75. Kendler, K. S., Karkowski, L. M., & Prescott, C. A. (1999). Fear and phobias: Reliability and heritability. Psychological Medicine, 29, 539–553. Kendler, K. S., Walters, E. E., Neale, M. C., Kessler, R. C., Heath, A. C., & Eaves, L. J. (1995). The structure of genetic and environmental risk factors for six major psychiatric disorders in women: Phobia, generalized anxiety disorder, panic disorder, bulimia, major depression, and alcoholism. Archives of General Psychiatry, 52, 374–382. Klein, D. F. (1993). False suffocation alarms, spontaneous panics, and related conditions. An integrative hypothesis. Archives of General Psychiatry, 50, 306–317. Krijn, M., Emmelkamp, P. M. G., Olafsson, R. P., & Biemond, R. (2004). Virtual reality exposure therapy and anxiety disorders: A review. Clinical Psychology Review, 24, 259–281. Lambert, M. J., Okiishi, J. C., Finch, A. E., & Johnson, L. D. (1998). Outcome assessment: From conceptualization to implementation. Professional Psychology–Research & Practice, 29, 63–70. Lang, P. J. (1977). Imagery in therapy: An information processing analysis of fear. Behavior Therapy, 8, 862–886. Ley, R. (1985). Blood, breath, and fears: A hyperventilation theory of panic attacks and agoraphobia. Clinical Psychology Review, 5, 271–285. Ley, R. (1991). The efficacy of breathing retraining and the centrality of hyperventilation in panic disorder: A reinterpretation of experimental findings. Behaviour Research and Therapy, 29, 301–304. Luten, A., Ralph, J. A., & Mineka, S. (1997). Pessimistic attributional style: Is it specific to depression versus anxiety versus negative affect? Behaviour Research and Therapy, 35, 703–719. Marks, I. M. (1971). Phobic disorders four years after treatment: A prospective follow-up. British Journal of Psychiatry, 129, 362–371.
Chapter 04.qxd
3/28/2007
12:58 PM
Page 101
4. REVOLUTION IN PSYCHOLOGICAL TREATMENT
101
Marks, I. M. (1986). Genetics of fear and anxiety disorders. British Journal of Psychiatry, 149, 406–418. Marks, I., & Dar, R. (2000). Fear reduction by psychotherapies. Recent findings, future directions. British Journal of Psychiatry, 176, 507–511. Marks, I. M., Swinson, R. P., Basoglu, M., Kuch, K., Noshirvani, H., O’Sullivan, G., Lelliott, P. T., Kirby, M., McNamee, G., Sengun, S., & Wickwire, K. (1993). Alprazolam and exposure alone and combined in panic disorder with agoraphobia. A controlled study in London and Toronto. British Journal of Psychiatry, 162, 776–787. Martin, N. G., Jardine, R., Andrews, G., & Heath, A. C. (1988). Anxiety disorders and neuroticism: Are there genetic factors specific to panic? Acta Psychiatrica Scandinavica, 77, 698–706. McCauley, E., Mitchell, J. R., Burke, P., & Moss, S. (1988). Cognitive attributes of depression in children and adolescents. Journal of Consulting and Clinical Psychology, 56, 903–908. Merrill, K. A., Tolbert, V. E., & Wade, W. A. (2003). Effectiveness of cognitive therapy for depression in a community mental health center: A benchmarking study. Journal of Consulting and Clinical Psychology, 71, 404–409. Meuret, A. E., Wilhelm, F. H., Ritz, T., & Roth, W. T. (2003). Breathing training in panic disorder treatment: Useful intervention or impediment to therapy? Behavior Modification, 27, 731–754. Meuret, A. E., Wilhelm, F. H., & Roth, W. T. (2001). Respiratory biofeedbackassisted therapy in panic disorder. Behavior Modification, 25, 584–605. Meuret, A. E., Wilhelm, F. H., & Roth, W. T. (2004). Respiratory feedback for treating panic disorder. Journal of Clinical Psychology, 60, 197–207. Mineka, S., Gunnar, M., & Champoux, M. (1986). Control and early socioemotional development: Infant rhesus monkeys reared in controllable versus uncontrollable environments. Child Development, 57, 1241–1256. Mineka, S., & Kihlstrom, J. (1978). Unpredictable and uncontrollable aversive events. Journal of Abnormal Psychology, 87, 256–271. Mowrer, O. H. (1939). A stimulus-response analysis of anxiety and its role as a reinforcing agent. Psychological Review, 46, 553–565. Newman, M. G., Kenardy, J., Herman, S., & Taylor, C. B. (1997). Comparison of palmtop-computer-assisted brief cognitive-behavioral treatment to cognitivebehavioral treatment for panic disorder. Journal of Consulting and Clinical Psychology, 65, 178–183. Nolen-Hoeksema, S., Girgus, J. S., & Seligman, M. E. P. (1986). Learned helplessness in children: A longitudinal study of depression, achievement, and attributional style. Journal of Personality and Social Psychology, 51, 435–442. Nolen-Hoeksema, S., Girgus, J. S., & Seligman, M. E. P. (1992). Predictors and consequences of childhood depressive symptoms: A 5-year longitudinal study. Journal of Abnormal Psychology, 101, 405–422. Overmier, J. B., & Seligman, M. E. P. (1967). Effects of inescapable shock upon subsequent escape and avoidance behavior. Journal of Comparative and Physiological Psychology, 63, 23–33. Plomin, R., DeFries, J. C., McClearn, G. E., & Rutter, M. (1997). Behavioral genetics: A primer (3rd ed.). New York: Freeman.
Chapter 04.qxd
102
3/28/2007
12:58 PM
Page 102
LIVERANT ET AL.
Rachman, S. J. (1980). Emotional processing. Behaviour Research and Therapy, 18, 51–60. Rapee, R. M., Craske, M. G., Brown, T. A., & Barlow, D. H. (1996). Measurement of perceived control over anxiety-related events. Behavior Therapy, 27, 279–293. Razran, G. (1961). The observable unconscious and the inferable conscious in current Soviet psychophysiology: Interoceptive conditioning, semantic conditioning, and the orienting reflex. Psychological Review, 68, 81–147. Reiss, D., Hetherington, E. M., Plomin, R., Howe, G. W., Simmens, S. J., Henderson, S. H., O’Connor, T. J., Bussell, D. A., Anderson, E. R., & Law, T. (1995). Genetic questions for environmental studies: Differential parenting and psychopathology in adolescence. Archives of General Psychiatry, 52, 925–936. Roth, W. T., Wilhelm, F. H., Pettit, D., & Meuret, A. E. (2005). Rescuing the hyperventilation theory of panic: Reply to Ley (2005). Psychological Bulletin, 13, 199–201. Rothbaum, B. O., Hodges, L., Smith, S., Lee, J. H., & Price, L. (2002). A controlled study of virtual reality exposure therapy for the fear of flying. Journal of Consulting and Clinical Psychology, 68, 1020–1026. Rotter, J. B. (1954). Social learning and clinical psychology. Englewood Cliffs, NJ: Prentice-Hall. Salkovskis, P. M., Clark, D. M., & Gelder, M. G. (1996). Cognition–behaviour links in the persistence of panic. Behaviour Research and Therapy, 34, 453–458. Sapolsky, R. M., Alberts, S. C., & Altman, J. (1997). Hypercortisolism associated with social subordinance or social isolation among wild baboons. Archives of General Psychiatry, 54, 1137–1143. Schmidt, N. B., Woolaway-Bickel, K., Trakowski, J., Santiago, H., Storey, J., Koselka, M., & Cook, J. (2000). Dismantling cognitive-behavioral treatment for panic disorder: Questioning the utility of breathing retraining. Journal of Consulting and Clinical Psychology, 68, 417–424. Silove, D., Parker, G., Hadzi-Pavlovic, D., Manicavasgar, V., & Blaszczynski, A. (1991). Parental representations of patients with panic disorder and centralized anxiety. British Journal of Psychiatry, 159, 835–841. Smoller, J. W., Rosenbaum, J. F., Biederman, J. K., Dai, D., Racette, S. R., Laird, N. M., et al. (2003). Association of a genetic marker at the corticotrophin-releasing hormone locus with behavioral inhibition. Biological Psychiatry, 54, 1376–1381. Spiegel, D. A., & Barlow, D. H. (2000, November). Eight-day treatment of panic disorder with moderate to severe agoraphobia: Preliminary outcome data. Poster presented at the 34th annual convention of the Association of Advancement of Behavior Therapy, New Orleans, LA. Staats, A. W. (1975). Social behaviorism. Homewood, IL: Dorsey. Stampfl, T. G., & Levis, D. J. (1967). Essentials of implosive therapy. A learningtheory based psychodynamic behavioral therapy. Journal of Abnormal Psychology, 72, 496–503. Steketee, G. (1993). Treatment of obsessive compulsive disorder. New York: Guilford. Stuart, G. L., Treat, T. A., & Wade, W. A. (2000). Effectiveness of an empirically based treatment for panic disorder delivered in a service clinic setting: 1-year follow-up. Journal of Consulting and Clinical Psychology, 68, 506–512. Substance Abuse and Mental Health Administration. (2003). SAMSHA agency overview. Retrieved January 9, 2004, from www.samsha.gov/about/about.html.
Chapter 04.qxd
3/28/2007
12:58 PM
Page 103
4. REVOLUTION IN PSYCHOLOGICAL TREATMENT
103
Telch, M. J., Lucas, J. A., Schmidt, N. B., Hanna, H. H., Jaimcz, T. S., & Lucas, R. A. (1993). Group cognitive behavioral treatment of panic disorder. Behaviour Research and Therapy, 31, 279–287. Telch, M. J., Schmidt, N. B., Jaimez, T. L., Jaquin, K. M., & Harrington, P. J. (1995). Impact of cognitive-behavioral treatment on quality of life in panic disorder patients. Journal of Consulting and Clinical Psychology, 63, 823–830. The British Psychological Society. (2004). English survey of applied psychologists in health & social care and in the probation and prison service. Leicester, England: Author. Turner, J. E., & Cole, D. A. (1994). Developmental differences in cognitive diatheses for child depression. Journal of Abnormal Child Psychology, 22, 15–32. Wade, W. A., Treat, T. A., & Stuart, G. L. (1998). Transporting an empirically supported treatment for panic disorder to a service clinic setting: A benchmarking strategy. Journal of Consulting and Clinical Psychology, 66, 231–239. Watson, J. B., & Raynor, R. (1920). Conditioned emotional reactions. Journal of Experimental Psychology, 3, 1–14. Weiss, J. M. (1971). Effects of coping behavior in different warning conditions on stress pathology in rats. Journal of Comparative and Physiological Psychology, 77, 1–13. Westen, D., Novonty, C. M., & Thompson-Brenner, H. (2004). Empirical status of empirically supported psychotherapies: Assumptions, findings, and reporting in controlled clinical trials. Psychological Bulletin, 130, 631–663. Wolpe, J. (1958). Psychotherapy by reciprocal inhibition. Stanford, CA: Stanford University Press. Wolpe, J., & Rowan, V. C. (1988). Panic disorder: A product of classical conditioning. Behavior Research and Therapy, 26, 441–450.
Chapter 04.qxd
3/28/2007
12:58 PM
Page 104
Chapter 05.qxd
3/28/2007
12:58 PM
Page 105
5 Manual-Based Treatment: Evolution and Evaluation G. Terence Wilson Rutgers University
The development of manual-based psychological treatments for a wide range of clinical disorders has had a significant impact not only on clinical research, but also on clinical practice. Theory-driven, manual-based treatments have become a defining feature of evidence-based treatments for specific clinical disorders that have been evaluated in numerous randomized controlled trials (RCTs). The advantages of manual-based treatment include well-documented efficacy, less reliance on intuitive clinical judgment, and greater ease in training and supervising therapists in specific clinical strategies and techniques (Wilson, 1998a). Another nontrivial benefit has been the development of various selfhelp interventions derived from manual-based protocols (Fairburn & Carter, 1997). Nonetheless, the advent of manual-based treatment has generated considerable controversy (Addis, Wade, & Hatgis, 1999; Garfield, 1996; Wilson, 1998a). Different criticisms that have been leveled against manual-based treatment have focused not only on the use of standardized protocols (manuals; e.g., Strupp & Anderson, 1997), but also on the more general concept of empirically supported or evidence-based treatment (e.g., Garfield, 1996; Westen, Novotny, & Thompson-Brenner, 2004), and in some instances on the use of RCTs as a research methodology for
105
Chapter 05.qxd
3/28/2007
12:58 PM
Page 106
106
WILSON
evaluating the efficacy of psychological treatments (Seligman, 1995; Westen et al., 2004).1 The present chapter is limited to addressing three specific objections to manual-based treatment and the use of RCTs for documenting their efficacy and effectiveness. I focus on these particular criticisms because they ignore or misrepresent ongoing and evolving research that promises to enhance our clinical effectiveness and theoretical understanding of mechanisms of change. It is a personal pleasure and privilege to participate in a festschrift honoring Dick McFall, a friend and distinguished colleague whose work I have long admired. An influential educator and researcher, Dick’s unwavering commitment to the highest standards of clinical science in the study of clinical psychology serves as a model for all who seek to develop effective, evidence-based treatments. MANUAL-BASED TREATMENT AND THERAPEUTIC INNOVATION One of the most puzzling misconceptions about the development of manual-based psychological therapies has been the contention that it would hinder theoretical and clinical innovation. Gaston and Gagnon (1996), for example, predicted that it would result in a set of “stagnant, codified accepted treatments” (p. 17). This view is misguided. Indeed, the opposite has been the case. Manual-based treatment has significantly spurred therapeutic innovation, as illustrated in the context of the following three examples. Exposure Treatment for Anxiety Disorders A little history is informative. Consider what happened to systematic desensitization, which at one point in the early stages of behavior therapy was arguably the best-known, most widely used, and empirically supported treatment for phobic and other anxiety disorders (Lazarus, 1961;
1
The choice of the term manual to describe structured, evidence-based treatment was unfortunate. It conjures up images of a more-or-less mechanical approach to therapy. The connotative links are to “maintenance manuals” for lawn mowers and Toyota Camrys. The term invites clever—albeit misleading—references to “manual labor” (Parloff, 1998) in depicting therapists adopting this approach as mere technicians, compared with purportedly clinically sophisticated therapists who are in no need of such evidence-based protocols.
Chapter 05.qxd
3/28/2007
12:58 PM
Page 107
5. MANUAL-BASED TREATMENT
107
Wolpe, 1958). One of Wolpe’s (1958) great contributions (he was primarily a practicing clinician) was that he spelled out in testable, operational detail the procedural elements of systematic desensitization. As part of his well-known doctoral dissertation, Paul (1966) developed what he called a systematic desensitization treatment manual. To my knowledge, this is the earliest formal use of the term treatment manual in the behavior therapy literature. The availability of a detailed and therapist-friendly treatment protocol (manual) for Wolpe’s clinically inspired treatment enabled theoretically sophisticated and methodologically expert clinical researchers to conduct controlled studies of the treatment in both the laboratory (Bandura, 1969; Lang, 1969) and the field (Paul, 1969). Innovative dismantling studies, as they came to be known, showed that the therapeutic efficacy of the treatment was not due to the therapist–patient relationship, therapeutic expectancies, or other so-called nonspecific influences. Similarly, some of the components that Wolpe (1958) believed to be vital to behavior change, such as progressive relaxation training or the invariable use of a hierarchical presentation of phobic stimuli, were shown to be nonessential. The collective outcome of these lines of research was the conclusion that exposure to relevant anxiety-eliciting cues was a necessary and sufficient condition for therapeutic success. Wolpe’s (1958) theory of reciprocal inhibition was promptly discarded (Wilson & Davison, 1971; much to his displeasure), and more powerful and flexible forms of exposure treatment were increasingly adapted to the treatment of the full spectrum of anxiety disorders. Today there is little question that exposure-based therapy is the treatment of choice (Barlow, 2002). Only first- or secondgeneration behavior therapists would know much about systematic desensitization. Cognitive Therapy for Depression A more contemporary example is provided by Beck’s cognitive therapy (CT), now well established as an effective treatment for depression. Recent evidence from well-controlled RCTs has shown that CT appears to be as effective as antidepressant medication even with severely depressed patients (DeRubeis, Gelfand, Tang, & Simons, 1999; DeRubeis et al., 2005; Hollon et al., 2005). As with Wolpe, one of Beck’s many contributions was to spell out clearly in a manual how treatment was administered (Beck, Rush, Shaw, & Emery, 1979). Hollon (1999) succinctly deconstructed CT into the following list of overlapping and sequential
Chapter 05.qxd
108
3/28/2007
12:58 PM
Page 108
WILSON
elements: rationale for treatment; systematic self-observation; behavioral activation (BA); monitoring thoughts; challenging the accuracy of thoughts; exploring core underlying beliefs; and relapse prevention. As with systematic desensitization, this detailed description of the therapy encouraged researchers to subject the treatment to rigorous experimental scrutiny.
Behavioral Activation. Jacobson and his colleagues (1996) carried out a component analysis (dismantling study) of CT for depression. They showed that the early phase of CT alone, which emphasizes behavioral activation (BA), was as effective as the complete treatment protocol both at the end of treatment and, most tellingly, at a 2-year follow-up (Gortner, Gollan, Dobson, & Jacobson, 1998). Most recently, Dimidjian et al. (2004) extended this finding in showing that BA was as effective as antidepressant medication and more effective than CT in the treatment of severe depression. Full analysis of the implications of these important—and, to many, surprising—findings is beyond the scope of this chapter. Suffice it to say that they challenge the necessity of some of the defining cognitive components of CT and possibly call into question the current cognitive theory behind CT. BA has been further refined into a distinctive therapy for depression—a functional analytic treatment that has been detailed in a treatment manual (Martell, Addis, & Jacobson, 2001). Hollon (2001) suggested that BA may be easier to learn than CT. Given our difficulties in disseminating evidencebased treatments (discussed later), the efficacy of BA is an encouraging development. Mindfulness-Based Cognitive Therapy. Another manual-based innovation directly influenced by CT is mindfulness-based cognitive therapy (MBCT; Segal, Teasdale, & Williams, 2004). CT focuses explicitly on the content or validity of patients’ dysfunctional beliefs. Teasdale et al. (2002) have argued that “this focus leads, implicitly, to changes in relationships to negative thoughts and feelings and to increased metacognitive awareness” (p. 275). The latter is defined as a cognitive set in which “negative thoughts and feelings are seen as passing events in the mind rather than as inherent aspects of self or as necessarily valid reflections of reality” (p. 285). Enhanced metacognitive awareness, rather than change in the content of beliefs, is posited to be responsible for the longterm efficacy of cognitive therapy. MBCT is designed to promote metacognitive awareness as a means of reducing the risk of relapse in
Chapter 05.qxd
3/28/2007
12:58 PM
Page 109
5. MANUAL-BASED TREATMENT
109
the face of future stress in patients who have recovered from a depressive episode. Preliminary findings suggest that MBCT is effective in reducing relapse in recovered recurrently depressed patients compared with treatment as usual (Ma & Teasdale, 2004; Teasdale et al., 2002). Moreover, the results are consistent with the hypothesis that increased metacognitive awareness mediated the therapeutic effect.
Early Response to Treatment. Ilardi and Craighead (1994) first pointed out that CT for depression produces much of its therapeutic benefit early in treatment. According to their analysis, approximately 60% to 80% of total reduction in depression assessed at posttreatment occurred within the first 4 weeks of therapy. Subsequent research has indicated that this is a robust finding that applies to other manual-based CBT treatments for different disorders (Wilson, 1999). The finding, directly attributable to the well-defined structure and sequencing of manual-based CT, has wide-ranging theoretical, methodological, and clinical implications. In terms of theory, the finding raises serious questions about the mechanisms of action of CBT (Ilardi & Craighead, 1994). Existing theories did not predict this finding, and it has spurred constructive theoretical debate. The methodological implications are clear-cut—the study of the mechanisms of action of CBT (and perhaps of any psychological treatment?) requires targeted and repeated assessment of the hypothesized mechanisms from the onset of treatment (Kraemer, Wilson, Fairburn, & Agras, 2002). From the practical, clinical perspective, the early response finding meshes with the evidence of the efficacy of BA (Jacobson et al., 1996). Simply put, the initial treatment procedures in CT for depression basically comprise behavioral activation. Is this sufficient for lasting therapeutic improvement? What then is the role of the more cognitive procedures that unfold later in the sequence of CT? Are they necessary? The early response phenomenon, combined with the findings on BA, challenges what is purported to be the distinctive added value of schemafocused therapy (SFT; Young, Beck, & Weinberger, 2001). This approach has proved popular with clinicians. SFT is aimed at underlying cognitive vulnerabilities, as opposed to a focus on symptom reduction. Presumably the focus of behavioral activation is on the latter, whereas SFT emphasizes a focus on core underlying beliefs or early maladaptive schemas. But if the heavily cognitive component of CT (the focus on core beliefs) apparently does not add to the efficacy of BA, why would a much-expanded
Chapter 05.qxd
3/28/2007
12:58 PM
Page 110
110
WILSON
concern with hypothesized cognitive content of schemas be required? At present, there is no evidence of the specific efficacy of SFT, let alone data indicating that it might be superior to current CBT in the treatment of any clinical disorder. Of particular practical significance is the finding that early response to treatment has emerged as a robust predictor of subsequent treatment outcome—not only in depression, but also in other disorders (Wilson, 1999; see the example of eating disorders discussed later in this chapter). Likely nonresponders to manual-based CBT can be identified more efficiently, and treatment can be modified or switched to enhance the chances of successful outcome. In short, the brief history of Beck’s manual-based CT for depression has seen the development of new and different treatments, and novel research on mechanisms of action of CT. This is hardly the stuff of which stagnation is made! Cognitive Behavior Therapy for Eating Disorders Theory-driven, manual-based CBT for eating disorders (Fairburn, Marcus, & Wilson, 1993) is now well documented as the current treatment of choice for bulimia nervosa (BN) and binge eating disorder (BED; National Institute for Clinical Excellence [NICE], 2004; Wilson & Shafran, 2005). As proponents of this approach were quick to note (Wilson, 1996a), this manual-based CBT still has limited efficacy and does not help a significant number of patients. Far from leading to stagnation or complacency, however, the treatment has been the target of theoretical and clinical analyses designed to develop an improved second-generation manual that is more effective and applicable to a wider range of eating disorders (including anorexia nervosa and eating disorders not otherwise specified [EDNOS]; Fairburn, Cooper, & Shafran, 2003; Wilson, 2005; also see later discussion). Conclusion Any development that enhances accountability and increases our ability to critically test the efficacy of specific treatments and their presumed mechanisms will lead to research and likely innovation. Manual-based treatment represents such a development and has clearly led to important innovations in psychological therapy. There is every indication that it will continue to do so.
Chapter 05.qxd
3/28/2007
12:58 PM
Page 111
5. MANUAL-BASED TREATMENT
111
THE DISSEMINATION AND CLINICAL UTILITY OF MANUAL-BASED TREATMENTS The efficacy of CBT as a treatment for many clinical disorders is well established (Nathan & Gorman, 2002). Yet its lack of dissemination to routine clinical practice has been repeatedly documented (e.g., Barlow, Levitt, & Bufka, 1999; Mussell et al., 2000; Persons, 1997). One of the reasons for this unsatisfactory state of affairs, I would argue, is the misconception that the findings of RCTs evaluating manual-based treatments are of little if any relevance to routine clinical practice. Exclusion and Inclusion Criteria in RCTs of Manual-Based Treatments The misconception is based, in part, on the false assumption that RCTs typically exclude difficult patients—patients with multiple comorbidities— in a limited focus on a sample of patients with a single problem and a good prognosis. This charge has been analyzed in detail and found wanting (e.g., Barlow et al., 1999; Crits-Christoph, Wilson, & Hollon, 2005; Stirman, DeRubeis, Crits-Christoph, & Brody, 2003; Stirman, DeRubeis, Crits-Christoph, & Rothman, 2005; Weisz, Weersing, & Henggeler, 2005). Of course, some studies have broader exclusion criteria than others and have included patients with limited problems. Yet RCTs have increasingly included patients with severe psychopathology, high rates of psychiatric comorbidity, and frequent histories of previously failed therapy. As several commentators have noted, the most common reason for excluding individuals from RCTs is that their problems are not severe enough to meet the inclusion criteria (e.g., Crits-Christoph et al., 2005; Jacobson & Christensen, 1996). Not surprisingly, patient samples in some RCTs might have greater severity of the target disorder and more comorbidity than some unselected clinical samples in routine practice (e.g., Hirsch, Jolley, & Williams, 2000; Merrill, Tolbert, & Wade, 2003; Westbrook & Kirk, 2005). As always, it depends on the nature of the specific RCT and clinical samples in question. In his commentary on evidence-based treatment and the individual patient, Summerskill (2005) had the following to say: It can be tempting to consider the application of trial data in rigid terms: “Could my patient have been randomized in this trial? If so the results are applicable; if not, they may not be.” A more matter-of-fact approach to clinical complexity lies at the heart of Sackett, Richardson, Rosenberg, and Haynes’
Chapter 05.qxd
3/28/2007
12:58 PM
Page 112
112
WILSON
(1997) message: “Is my patient so different from those in the trial that its results cannot help me make my treatment decision?” It is always easy to find reasons why a patient is different from trial participants. This is one reason that the more family practitioners feel they know their patients, the less likely they are to apply external evidence to guide management (Summerskill & Pope, 2002). But are paternalistic assumptions in patients’ best interests? (p. 13)
Prognostic Effects of Comorbid Clinical Disorders It is commonly assumed by critics that the comorbid disorders that are allegedly the basis for exclusion from RCTs (e.g., personality disorders) are known to result in a worse treatment outcome (Westen et al., 2004). In reality, whether psychiatric comorbidity influences the clinical effectiveness of manual-based treatments is a function of the specific clinical disorder, the nature of the comorbidity, and the particular treatment in question. There are well-documented instances in which neither Axis I nor Axis II comorbidity has a discernible impact on outcome (e.g., Barlow et al., 1999; Wilson, 1998b). Therefore, RCTs do not necessarily inflate treatment outcome. Consider the following illustration of this general point. Westen and his colleagues (2004) contended that RCTs evaluating CBT for BN have excluded potential patients with Axis II psychopathology such as borderline personality disorder. Leaving aside the data showing that this is an inaccurate assertion (see e.g., Agras, Walsh, Fairburn, Wilson, & Kraemer, 2000), what do we know of the prognostic significance of comorbid borderline personality disorder in patients with BN? Scholarly analyses of the evidence have shown that it is premature to conclude that co-occurring borderline personality disorder predicts a worse outcome (Grilo, 2002; NICE, 2004). Moreover, the natural course of BN is not influenced by personality disorders (Grilo et al., 2003). More controlled research is needed to determine the specific relationship between personality disorders and treatment outcome in BN and other eating disorders. Generalizability of Findings of RCTs of Manual-Based Treatment Studies As with any experiment, the issue arises about the generalizability of the findings of RCTs to conditions other than those of the particular study— the question of external validity. Concerns about the external validity of
Chapter 05.qxd
3/28/2007
12:58 PM
Page 113
5. MANUAL-BASED TREATMENT
113
findings are hardly specific to research on psychological treatment. The issues involved in the generalizability from RCTs in medical research in general were the focus of a recent series of reviews in The Lancet (Rothwell, 2005). The generalizability of the findings of efficacy studies to diverse clinical samples across different clinical settings must be evaluated directly in clinical effectiveness research. In the ultimate analysis, the applicability of the findings of RCTs to clinical practice depends on the design of the individual study, the patient sample, and the clinical setting to which the results are to be generalized (Chambless & Hollon, 1998; Kazdin & Wilson, 1978). The critical dimensions along which generalizability must be assessed are the patient characteristics, clinical setting, therapist training and expertise, and specific treatment methods. In marked contrast to such a systematic scientific approach, the clinical literature is replete with warnings that the findings of efficacy studies (RCTs) either do not—or, more cautiously, may not—apply to real patients treated in real-world settings (e.g., Goldfried & Wolfe, 1996; Havik & VandenBos, 1996). One does not need to be a cognitive therapist to identify the all-or-nothing thinking implicit in this common refrain. Miklowitz and Clarkin (1999) made this point some years ago: “We run the danger of dichotomous thinking in which RCTs are viewed as irrelevant to community health care whereas studies done in mental health clinics, however poorly designed, take greater precedence” (p. 2). Imagine two patients being treated for BN. One is a high-functioning young woman, attending a prestigious Ivy League university, who was referred to a therapist in independent practice in upper middle-class suburbia. The other is a young Hispanic woman, from a single-mother home in the inner city, who responded to a public announcement of free treatment as part of a National Institute of Mental Health (NIMH)-funded study of BN at a major urban medical school. She could not otherwise have afforded treatment. Who is the real patient here? Who has the better prognosis? What is the real world here? This example highlights the failings of drawing a simplistic dichotomy between a research study and routine clinical practice. The reality is that we must address the needs of a wide range of different patients drawn from a diverse spectrum of real worlds. Innovative research that explicitly investigates the degree to which different treatments generalize to conditions other than those of controlled efficacy studies is a priority. As summarized later in this chapter, much
Chapter 05.qxd
114
3/28/2007
12:58 PM
Page 114
WILSON
progress has already been made with encouraging results. Assuming, implicitly or otherwise, that the findings of RCTs that evaluate manualbased treatments do not generalize to routine clinical setting is premature, if not wrong. History, again, is instructive. The 1970s were marked by controversy over the value of analogue research in behavior therapy. The methodology was designed to identify the necessary and sufficient conditions of behavior change and to test hypotheses about mechanisms of change under tightly controlled laboratory conditions. The participants were often research volunteers with a single problem behavior, rather than treatment-seeking patients with multiple problems. As noted earlier, exposure was shown to be the critical element in systematic desensitization (Lang, 1969). Subsequent research established the clinical applicability and efficacy of exposure-based treatments to a variety of anxiety disorders in real patients in RCTs in clinical settings. Exposure is now widely accepted as an effective treatment for anxiety disorders (Barlow, 2002). Similarly, in the early 1970s, McFall and his students pioneered laboratory-based evaluation of assertion training. An innovative feature of the research was an evaluation of how well the intervention’s effects generalized to a real-life setting (McFall & Twentyman, 1973). Assertion training has since been widely incorporated into clinical practice (Alberti & Emmons, 2001).
Randomized Controlled Trials. As several commentators pointed out, RCTs need not be restricted to studies of treatment efficacy—they can also be used for evaluating the generalizability of treatment effects (Chambless & Hollon, 1998; Jacobson & Christensen, 1996). For example, Fairburn (2004) described an ongoing RCT of the treatment of eating disorders that has no exclusion criteria. All patients seeking treatment at two community psychiatric centers offering specialty treatment for eating disorders are randomly assigned either to current manual-based CBT (Fairburn, Marcus, & Wilson, 1993) or an enhanced version of the same basic approach (Fairburn et al., 2003). Any clinical eating disorder merits inclusion; the sample is not limited to any specific DSM–IV-defined diagnosis (e.g., BN). A major advantage of this innovative study of unselected patients is that it includes individuals with EDNOS who comprise the majority of patients in routine clinical settings, but who have previously been excluded from efficacy research (Wilson, 2005). The patients in this study exemplify a clinically representative and relevant sample. The therapists, however, are specifically trained and
Chapter 05.qxd
3/28/2007
12:58 PM
Page 115
5. MANUAL-BASED TREATMENT
115
supervised in the administration of the treatments. On the therapist and treatment dimensions, therefore, the study does not meet criteria for clinical representativeness (Shadish et al., 1997).
Quasi-Experimental and Nonexperimental Strategies. Comprehensive analysis of the generalizability of treatment effects requires a range of methodological strategies ranging from RCTs to nonexperimental and uncontrolled studies of outcome across diverse patients and clinical service settings. Different methodologies can be ordered along a continuum ranging from efficacy studies, on the one hand, to an uncontrolled, fully clinically representative approach, on the other hand. A complete review of the growing literature on this subject is beyond the scope of this chapter. Suffice it here to provide some illustrative examples. In a quasi-experimental design, Juster, Heimberg, and Engelberg (1995) compared three groups of patients seeking treatment for social phobia. The first group comprised those who were included in the RCT, the second group were those excluded primarily for medical or diagnostic reasons, and the third group were those individuals who declined random assignment to treatment. The innovative feature of this study was that Groups 2 and 3 were treated with the same CBT used in the RCT, and their response was compared with that of the patients formally included in the study. The results show that all three groups of patients demonstrated comparable improvement. This study in essence manipulated patient characteristics while holding constant the clinical setting, therapists, and specific treatment. Perhaps the most flexible yet informative research strategy for evaluating generalizability is benchmarking, a strategy first described by McFall (1996). In a benchmarking study, treatments of established efficacy in RCTs are administered in clinical service settings with unselected patients. The outcome in the service setting is then compared with that from RCTs completed in research clinics. The prototype of this research strategy is the Wade, Treat, and Stuart (1998) study conducted at the Center for Behavioral Health (CBH) in Bloomington, Indiana.2 In this study, the Barlow and Craske (1989) treatment manual for panic disorder, which has been shown to be effective in RCTs (Barlow, 2002), was 2 The investigators were past (Wade) and then current (Treat, Stuart) doctoral students from Indiana University; the inspiration behind the research was that of faculty member and clinical scientist Dick McFall.
Chapter 05.qxd
116
3/28/2007
12:58 PM
Page 116
WILSON
implemented by therapists of varying levels of experience and training at the CBH in the treatment of unselected patients with panic disorder. The therapists were trained by Wendy Wade, who had learned the treatment during a visit to Barlow’s research clinic. The results reveal that the CBH therapists achieved a success rate comparable to that of CBT evaluated in RCTs. Impressively, these results were maintained at a 1-year follow-up (Stuart, Treat, & Wade, 2000). The basic finding of the Wade et al. (1998) benchmarking study—that manual-based CBT produces outcomes in clinical service settings comparable to those of RCTs—has been replicated several times by different investigators treating a range of clinical disorders. Examples include obsessive-compulsive disorder (OCD; Franklin, Abramowitz, Kozak, Levitt, & Foa, 2000), social phobia (Lincoln et al., 2003), PTSD (Gillespie, Duggy, Hackmann, & Clark, 2002), and depression (Merrill et al., 2003). Shadish et al. (1997) developed a set of criteria to define the clinical representativeness of treatment outcome research. The most clinically representative end of the generalizability spectrum includes a nonuniversity setting—patients with heterogeneous problems, rather than one focal disorder, and who are clinically referred, and therapists who are professionals with regular caseloads. Treatment conditions are uncontrolled— that is, no use of a formal therapy manual, no specific training or supervision of therapists for the purposes of the study, and no treatment integrity checks. Benchmarking studies such as Franklin et al. (2000) and Wade et al. (1998) would fall closer to the controlled research end of the continuum given these criteria. Other studies of CBT, however, have met the most stringent Shadish et al. (1997) criteria for clinical relevance (e.g., Hirsch et al., 2000; Persons, Bostrom, & Bertagnolli, 1999; Westbrook & Kirk, 2005). In the largest study of its kind, Westbrook and Kirk (in press) analyzed the outcome of 1,276 patients (ages 18–65 years) treated by the specialized CBT service within the National Health Service in the United Kingdom. The authors reported effect sizes (ES) and clinical significance statistics on two standardized measures. The ES for the Beck Depression Inventory (BDI), for example, was 1.15, with the proportion of patients meeting criteria for reliable clinical change comparable to the findings of the Persons et al. (1999) clinical sample and the Elkin et al. (1989) RCT. Westbrook and Kirk (2005) concluded that their findings “suggest that CBT in this context is an effective treatment, albeit with probably not quite such good results as it achieves in research trials.”
Chapter 05.qxd
3/28/2007
12:58 PM
Page 117
5. MANUAL-BASED TREATMENT
117
Clinically representative analyses of this sort are inevitably flawed methodologically in several respects, such as lack of controls, missing data, and uncontrolled pharmacological treatment. I agree with Westbrook and Kirk (2005), who argue that the clinical relevance of research findings are ultimately determined by a range of methodologies varying in internal and external validity. Conclusion Evidence-based treatments for several disorders have been shown to be effective across clinical service settings, unselected patients with or without concurrent pharmacotherapy, and therapists with varying levels of training. Critics often ignore the evidence just summarized here or try to discount the methodological adequacy of studies of generalizability (see the Weisz et al., 2005, commentary on this issue). Ironically, calls for the evaluation of psychotherapy as it is practiced in clinical service settings typically ignore existing research showing that customary treatment in the child and adolescent treatment literature, which is not supported by controlled clinical research, appears to be ineffective, with ESs averaging about zero (Weisz et al., 2005). Some preliminary findings from this early stage of research on generalizability that warrant further investigation are the following. Patients in uncontrolled studies have experienced comparable effects to patients in RCTs despite receiving fewer sessions of treatment in service settings (e.g., Merrill et al., 2003; Roy-Byrne et al., 2005). Contrary to the claims of advocates of longer term psychotherapy (Seligman, 1995; Westen et al., 2004), more is not always better either in RCTs or uncontrolled clinical practice. As in RCTs, psychiatric comorbidity appears to be a negative predictor of outcome in some instances (e.g., Merrill et al., 2003), but not others (e.g., Roy-Byrne et al., 2005). Future studies need to identify what focal problems are influenced by what psychiatric comorbidity. Ideally, this research might pinpoint moderators rather than simple predictors of outcome, thereby allowing more rational treatment planning. Necessary and sufficient levels of therapist training and expertise remain unclear. We know that within RCTs, therapists effects are usually nonsignificant (Crits-Christoph & Mintz, 1991; Loeb et al., 2005; Wilson, 1998a). This is attributable to the selection in efficacy studies of competent therapists who are then carefully trained and closely supervised.
Chapter 05.qxd
118
3/28/2007
12:58 PM
Page 118
WILSON
As a result, the therapists acquire the technical expertise and have the interpersonal skills to administer manual-based treatments in a clinically sophisticated manner. The same is probably true for benchmarking studies, in which clinic therapists received specific instruction in the use of a specific treatment. Therapists in routine clinical service settings do not have this training or monitoring, and therapist effects are more likely under these uncontrolled conditions (Crits-Christoph & Mintz, 1991). The level of therapists’ raining has varied considerably in studies that have evaluated the effects of evidence-based CBT in clinical service settings. Some benchmarking studies have provided intensive training and continuing supervision of doctoral- and master’s-level therapists (e.g., Merrill et al., 2003). In the Roy-Byrne et al. (2005) study of the treatment of panic disorder in a primary care setting, CBT was administered by “a CBT naive, midlevel behavioral health specialist” (p. 290). Gillespie et al. (2002) trained five clinicians from a range of professional backgrounds (including nursing and social work) who were working in routine clinical positions. Training mainly comprised a 2-day workshop in CBT for PTSD, followed by monthly videoconferencing case supervision by experts in CBT thereafter. The therapists in Westbrook and Kirk’s (2005) uncontrolled study were professionals as well as trainees. Consistent with other research (e.g., Bickman, 1999), effectiveness studies have shown that degree of therapist experience was unrelated to outcome (e.g., Hahlweg, Fiegenbaum, Frank, Schroeder, & von Witzleben, 2001; Lincoln et al., 2003). Nevertheless, specific expertise in using an evidence-based treatment such as CBT makes a difference. Howard (1999) found that, among doctoral-level therapists with the same level of experience, those with training in CBT for anxiety disorders were more effective in treating patients with those problems. In summary, mental health providers with relatively little experience and less than a doctoral degree can be trained to deliver effective treatment for some problems in routine care settings. Nevertheless, therapist expertise in the principles and practice of CBT in general, aside from mastery of a specific treatment manual, is vital in complex and treatmentresistant cases. It is also important in the implementation of comprehensive and flexible protocols, which necessarily require more therapist judgment than more highly standardized or limited manuals (Wilson, 1998a). The recent treatment of adolescents with major depression (TADS) study has been described as a bridge between efficacy and effectiveness
Chapter 05.qxd
3/28/2007
12:58 PM
Page 119
5. MANUAL-BASED TREATMENT
119
research (Curry & Wells, 2005). Conducted across 13 different sites, it included patients who were representative of adolescents treated by clinicians in routine clinical practice. The short-term results show that antidepressant medication was significantly superior to pill placebo. Medication plus CBT was most effective overall in reducing depression and suicidal risk. But the effects of CBT were poor—no better than placebo—and less successful than in previous studies of adolescents (TADS Team, 2004). Hollon, Garber, and Shelton (2005) have attributed the relative ineffectiveness of CBT in this study to the type of CBT used and how it was implemented. Many of the therapists were inexperienced, with minimal training in CBT. Many of the on-site trainers/supervisors had less than optimal experience in treating adolescent depression with CBT. Hollon et al. (2005) argue that the investigators opted for a manual that “seemed overly comprehensive and far too structured” (p. 150). As a result, experienced therapists may have been constrained “from implementing CBT in an individuated fashion, resulting in an intervention that did not fully represent the best or even typical clinical practice” (p. 151). Moreover, “CBT therapists had so many things to do that they did not have enough time to do anything as well as they would have liked” (p. 150). Whether the Hollon et al. (2005) interpretation of the poor showing of CBT is valid is debatable; additional analyses of the TADS data might provide answers to the questions they raised. What is important in the current context is the acknowledgment that the efficacy of complex manual-based CBT treatments, both in controlled RCTs and in studies of their generalizability, is contingent on therapist expertise. This point has been made repeatedly in the CBT literature (e.g., Franks & Wilson, 1973; Jacobson & Hollon, 1996; Wolpe & Lazarus, 1966). Undoubtedly the same holds true for other psychological therapies. INDIVIDUALIZING TREATMENT: THE EVOLUTION OF MANUAL-BASED THERAPY Manual-based CBT requires that the therapist individualize treatment in several different ways. These include formulating a treatment plan for the individual patient within the overall treatment model; actively engaging patients in treatment within the collaborative framework of CBT; ongoing session-to-session assessment based on self-monitoring that helps determine the timing and nature of treatment; identification of specific
Chapter 05.qxd
120
3/28/2007
12:58 PM
Page 120
WILSON
dysfunctional beliefs and specific triggers for problem behaviors; use of multiple techniques, some of which may be better suited to particular patients than others; and addressing comorbid disorders when necessary (Wilson, 1996b). Specific therapist skills, including the ability to develop a good therapeutic alliance and balance a focus on treatment structure with flexibility, are essential (Wilson, 1998a). In effective manual-based treatments, a highly positive therapeutic alliance is strongly correlated with adherence to the treatment protocol (Addis et al., 1999; Loeb et al., 2005). Despite the many accounts of manual-based treatment—not to mention the content of actual manuals—over the past several years, some misconceptions persist. For example, Weisz et al. (2005) point out that the Westen et al. (2004) critique portrays manuals as “rigidly structured documents that minimize the patient’s active involvement in the treatment process, prevent therapists from using clinical judgment, reduce the therapist to a ‘research assistant’ whose job is to ‘run subjects’ … and are incompatible with an emphasis on broad principles of change” (p. 422). Even a cursory review of the relevant literature would reveal that this is a gross misrepresentation of competently conducted manual-based treatment. More simply, however, we have only to ask how manual-based CBT could possibly be effective—as has been conclusively shown in efficacy and effectiveness studies—if this sort of criticism were valid? Although manual-based treatment thus far has hardly ignored individualization, nor been ineffective in treating patients with multiple problems, much more can be done in developing manual-based therapies that address the specific needs of individual patients. One problem has been that the application of manual-based treatment thus far has been determined, in large part, by categorical DSM–IV diagnoses. Heterogeneity exists across individuals within DSM–IV diagnostic categories. The mechanisms that maintain the specific disorder vary across individuals, and therefore the same treatment is not equally effective for all members of a diagnostic category. Matching interventions to DSM–IV diagnoses as the sole basis for treatment selection is fundamentally at odds with the functional analysis of the individual patient that has been a core conceptual and clinical feature of behavior therapy from its earliest days. We need to move beyond the atheoretical, heterogeneous categories of DSM–IV to more refined matches of specific treatments with particular problems in individual patients guided by detailed functional analyses of the variables that maintain the problem behaviors in question. There is
Chapter 05.qxd
3/28/2007
12:58 PM
Page 121
5. MANUAL-BASED TREATMENT
121
nothing inherent in manual-based treatments that links them to DSM diagnoses, and there is no reason that they cannot be flexibly used in treatment driven by the functional analysis of behavior. Indeed, this has been the case in more flexible and comprehensive manuals, and in part accounts for the success of these interventions. The challenge in emphasizing a greater individualization of treatment is to balance this clinically appealing flexibility with the welldocumented strengths of the structured focus of manual-based treatment. The problems of intuitive clinical judgment have been amply documented (Dawes, 1994). Empirically supported, manual-based treatments are prescriptive in the same sense that the NICE treatment guidelines are. But they do not ignore clinical judgment: “Guidelines are not a substitute for professional knowledge and clinical judgement…there will always be some people and situations for which clinical guideline recommendations are not readily applicable. The NICE guidance does not, therefore, override the individual responsibility of health-care professionals to make appropriate decisions” (NICE, 2004, p.10). What is distinctive about this approach to treatment is the balance between research and clinical judgment. As Wilson and Shafran (2005) argued: Clinical judgment is decisive when evidence is lacking on what treatment to use. It is essential when an evidence-based treatment needs to be adapted to the niceties of an individual or when an alternative approach is needed. On the other hand, where sufficient evidence exists to allow general recommendation…the best practice must be to implement the treatment that enjoys the most empirical support rather than invoke subjective judgment. (p. 81)
Enhancing Manual-Based Treatment: The Example of Eating Disorders Fairburn et al. (2003) developed an innovative and enhanced manualbased treatment for the full range of eating disorders. Ultimately, valid matching of specific treatments to particular patients hinges on an improved understanding of (a) the mechanisms that maintain the clinical disorder in question, and (b) the mechanisms whereby specific treatments work. Accordingly, Fairburn et al. (2003) broadened the cognitive-behavioral model of the mechanisms that maintain BN, from which the first generation of manual-based therapy was derived (Fairburn et al., 1993), and extended it to all eating disorders. The expanded model
Chapter 05.qxd
122
3/28/2007
12:58 PM
Page 122
WILSON
has solid theoretical and empirical foundations. A major goal of the enhanced treatment is to identify specific patient profiles so that treatment can be tailored to them using specific modules that target the expanded range of maintaining mechanisms. In their emphasis on a psychological analysis of presenting problems, Fairburn et al. (2003) state that, “diagnosis is not of relevance to treatment” (p. 522). Instead, they propose a transdiagnostic theory and treatment of all eating disorders. Their fundamental rationale is that all the eating disorders share common maintaining mechanisms. Furthermore, Fairburn et al. (2003) underscore the necessarily idiographic nature of personalized treatment formulations in implementing this new framework. The latter emphasis, of course, harks back to the functional analysis that has been a seminal part of behavior therapy. This refined transdiagnostic treatment approach for manual-based treatment also addresses another common criticism of manual-based treatment. It is often argued that clinical practice in the real world is self-correcting—if one method is unsuccessful, another is adopted (Seligman, 1995). In contrast, it is alleged that manual-based treatment proceeds in an unchanging, lock-step fashion. There is, however, little evidence to indicate that routine clinical practice is self-correcting. The meager data that exist suggest that therapists tend to stick with the treatment they started, regardless of outcome (Wilson, 1998a). Fairburn et al. (2003) build a self-correcting feature into their treatment. Stage 1 involves eight sessions of core CBT treatment with a primary focus on behavioral change. The next few sessions focus on formally evaluating progress. In the case of problems, the focus is on identifying barriers to change and assessing the role of additional maintaining mechanisms, with a view to formulating a revised, personalized treatment plan. This taking stock of initial progress fits with the evidence that manualbased CBT is marked by an early response to treatment that is the most robust predictor of outcome at posttreatment and longer term follow-up (Agras, Crow, Halmi, Mitchell, Wilson, & Kraemer, 2000; Fairburn, Walsh, Agras, Wilson, & Stice, 2004). Absent sufficient improvement at this early stage, treatment needs to be modified or switched to another modality (e.g., antidepressant medication). Fairburn (2004) reported encouraging initial results from this enhanced CBT treatment. A preliminary investigation by Ghaderi (in press) also suggested the superiority of a broader, more individualized CBT approach over a more focused, standardized CBT treatment for BN.
Chapter 05.qxd
3/28/2007
12:58 PM
Page 123
5. MANUAL-BASED TREATMENT
123
This transdiagnostic approach could be applied to other groups of related clinical disorders. A related development is Barlow, Allen, and Choate’s (2004) proposal for the unified treatment of a negative affect syndrome featuring the anxiety disorders and depression. The core treatment consists of three fundamental components: antecedent cognitive reappraisal, overcoming emotional avoidance, and modifying emotional action tendencies. These core components could be modified or enhanced through additional strategies to accommodate different patient profiles across this spectrum of psychopathology. PRINCIPLE-DRIVEN INDIVIDUALIZATION OF TREATMENT Alternative proposals for individualizing treatment, while remaining responsive to clinical science, have emphasized principle-driven approaches (e.g., Beutler, 2000; Salkovskis, 2002). There is much to recommend this strategy, which overlaps heavily with manual-based treatment. Theory-driven, empirically based principles have been the life blood of CBT (Bandura, 1969). These principles are vital to therapeutic innovation and development (e.g., Clark, 2004), and they guide the flexible and scientifically informed implementation of manual-based treatment. As Stirman et al. (2005) observe, because “some of the treatments studied in RCTs are modifications of the same modality for different diagnoses, clinicians may find that they can apply the principles of those treatments to more than one diagnosis. With training in the treatments tested in RCTs, clinicians will be able to conceptualize the interactions between co-occurring diagnoses and use the concepts of these therapies to treat their patients” (p. 133). In arguing against manual-based treatment in general clinical practice, Beutler (2000) cautioned that therapists would have to learn too many different manuals, some of which may undermine “clinicians’ general therapeutic skill” (p. 6). The latter need not be the case, as noted earlier, but the former is a legitimate practical concern. As an alternative, Beutler recommended that clinicians flexibly apply a “refined list of empirically supported principles of treatment” that allow the use of their “favorite procedures” (p. 8)—thereby integrating science with clinical judgment. One limitation of relying only on principles is that it may miss the rich clinical content and context of treatment manuals. Moreover, in a
Chapter 05.qxd
124
3/28/2007
12:58 PM
Page 124
WILSON
related fashion, it might also miss specific maintaining mechanisms of different clinical disorders. This would be a problem especially for the mental health provider with less training in the treatment or lacking specialization in the target disorder. For example, Beutler (2000) advocates using the principle of “exposure and extinction”—and for good reason. This core principle of numerous CBT treatment strategies enjoys impressive empirical support (Barlow, 2002); it cuts across varied treatment manuals for different disorders. One of these disorders is BN. Consider, then, the treatment of BN based on the principle of exposure. A critically important component of manual-based CBT for BN is the early intervention to reduce dysfunctional dietary restraint and restore more normal patterns of healthy eating (Fairburn et al., 1993). Helping a BN patient who skips meals and avoids entire classes of “forbidden foods” to resume eating three meals a day plus planned snacks, on the one hand, and to systematically incorporate previously forbidden foods into her meals, on the other hand, is classic exposure therapy (Wilson, Fairburn, & Agras, 1997). This procedure reduces to overcoming fear and avoidance of potentially gaining weight. But the principle of exposure alone, however, would not instruct therapists in how to overcome dysfunctional dietary restraint in an efficient or optimal manner. Missing from the manual derived in part from the principle, for instance, would be advice on the sequencing of specific interventions—about targeting dietary restraint early in therapy, and about focusing first on establishing a regular pattern of eating before attempting to address forbidden foods. We know that manual-based CBT effects change early in treatment, and that this change mediates subsequent outcome (Kraemer et al., 2002). Numerous other examples could be cited illustrating how evidence-based treatment manuals put flesh on the theoretical skeleton of fundamental principles of behavioral change and, in so doing, can offer invaluable practical guidance to clinicians. Another limitation of purely principle-driven treatment is its reliance on the clinical judgment of the therapist. Empirically supported, manualbased treatment is, in part, prescriptive, as are the evidence-based NICE guidelines. Beutler (2000), among others, overlooks the evidence that therapists given free reign to select their preferred techniques will not necessarily choose the most effective methods. The best illustration of this problem, and one that makes the case for selective prescription, is Schulte, Kunzel, Pepping, and Schulte-Bahrenberg’s (1992) study of behavior therapists treating phobic disorders. To summarize, therapists who used
Chapter 05.qxd
3/28/2007
12:58 PM
Page 125
5. MANUAL-BASED TREATMENT
125
standardized in vivo exposure achieved significantly superior results compared with those who were free to select whatever techniques they wished. The difference was attributable to therapists in the latter condition neglecting to use exposure treatment. As I noted previously, this finding that therapists rejected the empirically validated therapy in favor of their own personal predilections underscores the problems with personalistic case formulations. If behavior therapists can persuade themselves to ignore exposure treatment for specific phobia in favor of other methods based on their clinical judgment, there would seem to be no end to the possibilities with more heterogeneous disorders and other less empirically based theoretical approaches. It should come as little surprise, therefore, that in clinical practice empirically validated methods are routinely ignored in favor of intuition and personal experience (Wilson, 1996b). Finally, the appeal to principle-driven treatment implicitly assumes that therapists will be doctoral-level clinical psychologists who combine clinical skill with knowledge of relevant scientific research. The reality is that psychological therapy increasingly is being provided by master’slevel counselors from a wide range of disciplines with less than optimal backgrounds in the scientific foundations and principles of behavior change. These mental health providers, in particular, can benefit from the more specific guidance provided by evidence-based treatment manuals. CONCLUDING COMMENTS The fundamentally important questions in analyzing the outcome of psychological therapy are still what treatments work, for whom, and why. Research on manual-based treatments has begun to provide answers to these questions and will continue to play an important role in advancing future knowledge. Manual-based treatments specify therapeutic procedures and identify mechanisms that maintain the target disorder. By enhancing accountability, manuals have already spurred the development of new treatment methods and raised intriguing theoretical questions about mechanisms of action. Manual-based treatments provide a critical means of increasing dissemination of effective psychological therapies and facilitate broadening the range of mental health providers who can provide effective treatment. Our best treatments currently are good but not good enough. The challenge is to make them more effective and for a broader range of clinical disorders. There is no more important goal than understanding the
Chapter 05.qxd
3/28/2007
12:58 PM
Page 126
126
WILSON
mechanisms of change of effective treatments. Without the operational specification of treatment manuals, we will have trouble isolating the necessary and sufficient conditions of effective treatments—a precursor to unraveling mechanisms of action. Of course manual-based therapies can be implemented badly. But any therapy—manualized or other—can be implemented poorly. Obviously, therapy can be effective without following a manual. Nonetheless, manuals are an important means within a science-based approach of improving efficacy and understanding how treatments work. We need better and bolder visions of the future, rather than advocacy of the status quo in the training of clinical psychologists. Westen et al. (2004), for example, proposed a model of empirically informed treatments as an alternative to empirically supported, manual-based treatments tested in RCTs. As Crits-Christoph et al. (2005) pointed out, this proposal is more likely to prevent than promote the adoption of evidence-based practice: While such a model might seem entirely reasonable, it begs the question. What evidence will inform whom, and how will it be evaluated? Actually, such a model already exists in the clinical psychology accreditation criteria of the American Psychological Association. To be accredited, doctoral programs in clinical psychology are required to expose students to the scientific underpinnings of psychology, but it is left to individual programs to adopt any philosophy of clinical training they wish provided they articulate it in a coherent manner. We can do better. Groups such as the Academy of Psychological Clinical Science have proposed a different view of a connection between science and practice from that of Westen et al. (2004) that emphasizes the importance of training in and dissemination of evidence-based treatment. (p. 415)
Dick McFall was the guiding force behind the establishment of the Academy of Psychological Clinical Science, as this volume makes clear. ACKNOWLEDGEMENTS I am grateful to Tanya Schlam and Robyn Sysko for their helpful comments on this chapter. REFERENCES Addis, M. E., Wade, W. A., & Hatgis, C. (1999). Barriers to dissemination of evidence-based practices: Addressing practitioners’ concerns about manual-based psychotherapies. Clinical Psychology: Science and Practice, 6, 430–441.
Chapter 05.qxd
3/28/2007
12:58 PM
Page 127
5. MANUAL-BASED TREATMENT
127
Agras, W. S., Crow, S. J., Halmi, K. A., Mitchell, J. E., Wilson, G. T., & Kraemer, H. (2000). Outcome predictors for the cognitive-behavioral treatment of bulimia nervosa: Data from a multisite study. American Journal of Psychiatry, 157, 1302–1308. Agras, W. S., Walsh, B. T., Fairburn, C. G., Wilson, G. T., & Kraemer, H. C. (2000). A multicenter comparison of cognitive-behavioral therapy and interpersonal psychotherapy for bulimia nervosa. Archives of General Psychiatry, 57, 459–466. Alberti, R. E. & Emmons, M. L. (2001). Your perfect right. San Luis Obispo, CA: Impact. Bandura, A. (1969). Principles of behavior modification. New York: Holt, Rinehart & Winston. Barlow, D. H. (2002). Anxiety and its disorders (2nd ed.). New York: Guilford. Barlow, D. H., Allen, L. B., & Choate, M. L. (2004). Toward a unified treatment for emotional disorders. Behavior Therapy, 35, 205–230. Barlow, D. H., & Craske, M. G. (1989). Mastery of your anxiety and panic. Albany, NY: Graywind. Barlow, D. H., Levitt, J. T., & Bufka, L. F. (1999). The dissemination of empirically supported treatments: A view to the future. Behaviour Research and Therapy, 37, S147–S162. Beck, A., Rush, J., Shaw, B., & Emery, G. (1979). Cognitive therapy of depression. New York: Guilford. Beutler, L. E. (2000, September 1). Empirically based decision making in clinical practice. Prevention & Treatment, 3, Article 27. Bickman, L. (1999). Practice makes perfect and other myths about mental health services. American Psychologist, 54, 965–978. Chambless, D. L., & Hollon, S. D. (1998). Defining empirically supported therapies. Journal of Consulting and Clinical Psychology, 66, 7–18. Clark, D. M. (2004). Developing new treatments: On the interplay between theories, experimental science and clinical innovation. Behaviour Research and Therapy, 42, 1089–1104. Crits-Christoph, P., & Mintz, J. (1991). Implications of therapist effects for the design and analysis of comparative studies of psychotherapies. Journal of Consulting and Clinical Psychology, 59, 20–26. Crits-Christoph, P., Wilson, G. T., & Hollon, S. (2005). Empirically supported psychotherapies: Commentary on Westen et al. Psychological Bulletin, 131, 412–417. Curry, J. F., & Wells, K. (2005). Striving for effectiveness in the treatment of adolescent depression: Cognitive behavior therapy for multisite community intervention. Cognitive and Behavioral Practice, 12, 177–185. Dawes, R. (1994). House of cards. New York: The Free Press. DeRubeis, R. J., Gelfand, L. A., Tang, T. Z., & Simons, A. D. (1999). Medications versus cognitive behavior therapy for severely depressed outpatients: Mega-analysis of four randomized comparisons. American Journal of Psychiatry, 156, 1007–1013. DeRubeis, R. J., Hollon, S. D., Amsterdam, J. D., Shelton, R. C., Young, P. R., Salomon, R. M., O’Reardon, J. P., Lovett, M. L., Gladis, M. M., Brown, L. L., & Gallop, R. (2005). Cognitive therapy vs. medications in the treatment of moderate to severe depression. Archives of General Psychiatry, 62, 409–416.
Chapter 05.qxd
128
3/28/2007
12:58 PM
Page 128
WILSON
Dimidjian, S., Hollon, S., Dobson, K., Schmaling, K., Kohlenberg, R., McGlinchey, J., Markley, D., Atkins, D., Addis, M., & Dunner, D. (2004, May). Behavioral activation, cognitive therapy, and anti-depressant medication in the treatment of major depression: Methods and acute phase outcomes. Paper presented at the annual meeting of the American Psychiatric Association, New York. Elkin, I., Shea, M. T., Watkins, J. T., Imber, S. D., Sotsky, S. M., Collins, J. F., D. Glass, P. Pilkonis, W. Leber, J. P. Docherty, et al. (1989). National Institute of Mental Health Treatment of Depression Collaborative Research Program: General effectiveness of treatments. Archives of General Psychiatry, 46, 971–982. Fairburn, C. G. (2004, April). The relationship between treatment research and clinical practice. Paper presented at the International Conference on Eating Disorders of the Academy of Eating Disorders, Orlando, FL. Fairburn, C. G., & Carter, J. C. (1997). Self-help and guided self-help for bingeeating problems. In D. M. Garner & P. E. Garfinkel (Eds.), Handbook of treatment for eating disorders (2nd ed., pp. 494–499). New York: Guilford. Fairburn, C. G., Cooper, Z., & Shafran, R. (2003). Cognitive behaviour therapy for eating disorders: A “transdiagnostic” theory and treatment. Behaviour Research and Therapy, 41, 509–529. Fairburn, C. G., Marcus, M. D., & Wilson. G. T. (1993). Cognitive behaviour therapy for binge eating and bulimia nervosa: A comprehensive treatment manual. In C. G. Fairburn & G. T. Wilson (Eds.), Binge eating: Nature, assessment and treatment (pp. 361–404). New York: Guilford. Fairburn, C. G., Walsh, B. T., Agras, W. S., Wilson, G. T., & Stice, E. (2004). Early change in treatment predicts outcome in bulimia nervosa. American Journal of Psychiatry, 161, 2322–2324. Franklin, M. E., Abramowitz, J. S., Kozak, M. J., Levitt, J. T., & Foa, E. B. (2000). Effectiveness of exposure and ritual prevention for obsessive-compulsive disorder: Randomized compared with nonrandomized samples. Journal of Consulting and Clinical Psychology, 68, 594–602. Franks, C. M., & Wilson, G. T. (Eds.). (1973). Annual review of behavior therapy (Vol. 1). New York: Brunner/Mazel. Garfield, S. L. (1996). Some problems associated with “validated” forms of psychotherapy. Clinical Psychology: Science and Practice, 3, 241–244. Gaston, L., & Gagnon, R. (1996). The role of process research in manual development. Clinical Psychology: Science and Practice, 3, 13–24. Ghaderi, A. (in press). Does individualization matter? A randomized trial of standardized (focused) versus individualized (broader) cognitive behavior therapy for bulimia nervosa. Behaviour Research and Therapy. Gillespie, K., Duffy, M., Hackmann, A., & Clark, D. M. (2002). Community based cognitive therapy in the treatment of post-traumatic stress disorder following the Omagh bomb. Behaviour Research and Therapy, 40, 345–357. Goldfried, M. R., & Wolfe, B. E. (1996). Psychotherapy practice and research: Repairing a strained alliance. American Psychologist, 51, 1007–1016. Gortner, E. T., Gollan, J. K., Dobson, K. S., & Jacobson, N. S. (1998). Cognitivebehavioral treatment for depression: Relapse prevention. Journal of Consulting and Clinical Psychology, 66, 377–384.
Chapter 05.qxd
3/28/2007
12:58 PM
Page 129
5. MANUAL-BASED TREATMENT
129
Grilo, C. M. (2002). Recent research of relationships among eating disorders and personality disorders. Current Psychiatry Reports, 4, 18–24. Grilo, C. M., Sanislow, C. A., Shea, M. T., Skodol, A. E., Stout, R. L., Pagano, M. E., Yen, S., & McGlashan, T. H. (2003). The natural course of bulimia nervosa and eating disorder not otherwise classified is not influenced by personality disorders. International Journal of Eating Disorders, 34, 319–330. Hahlweg, K., Fiegenbaum, W., Frank, M., Schroeder, B., & von Witzleben, I. (2001). Short- and long-term effectiveness of an empirically supported treatment for agoraphobia. Journal of Consulting and Clinical Psychology, 69, 375–382. Havik , O. D., & VandenBos, G. R. (1996). Limitations of manualized psychotherapy for everyday clinical practice. Clinical Psychologist: Science and Practice, 3, 264–267. Hirsch, C., Jolley, S., & Williams, R. (2000). A study of outcome in a clinical psychology service and preliminary evaluation of cognitive-behavioural therapy in real practice. Journal of Mental Health, 9(5), 537–549. Hollon, S. D. (1999). Rapid early response in cognitive behavior therapy: A commentary. Clinical Psychology: Science and Practice, 6, 305–309. Hollon, S. D. (2001). Behavioral activation treatment for depression: A commentary. Clinical Psychology: Science and Practice, 8, 271–274. Hollon, S. D., DeRubeis, R. J., Shelton, R. C., Amsterdam, J. D., Salomon, R. M., O’Reardon, J. P., Lovett, M. L., Young, P. R., Haman, K. L., Freeman, B. B., & Gallop, R. (2005). Prevention of relapse following cognitive therapy versus medications in moderate to severe depression. Archives of General Psychiatry, 62, 417–422. Hollon, S. D., Garber, J., & Shelton, R.C. (2005). Treatment of depression in adolescents with cognitive behavior therapy and medications: A commentary on the TADS project. Cognitive and Behavioral Practice, 12, 149–155. Howard, R. C. (1999). Treatment of anxiety disorders: Does specialty training help? Professional Psychology: Research and Practice, 30, 470–473. Ilardi, S. S., & Craighead, W. E. (1994). The role of nonspecific factors in cognitivebehavior therapy for depression. Clinical Psychology: Science & Practice, 1, 138–156. Jacobson, N. S., & Christensen, A. (1996). Studying the effectiveness of psychotherapy: How well can clinical trials do the job? American Psychologist, 51, 1031–1039. Jacobson, N. S., Dobson, K. S., Truax, P. A., Addis, M. E., Koerner, K., Gollan, J. K., Gortner, E., & Prince, S. E. (1996). A component analysis of cognitive-behavioral treatment for depression. Journal of Consulting and Clinical Psychology, 64, 295–304. Jacobson, N. S., & Hollon, S. D. (1996). Cognitive-behavior therapy versus pharmacotherapy: Now that the jury’s returned its verdict, it’s time to present the rest of the evidence. Journal of Consulting and Clinical Psychology, 64, 74–80. Juster, H. R., Heimberg, R. G., & Engelberg, B. (1995). Self selection and sample selection in a treatment study of social phobia. Behaviour Research and Therapy, 33, 321–324. Kazdin, A. E., & Wilson, G. T. (1978). Criteria for evaluating psychotherapy. Archives of General Psychiatry, 35, 407–4l8. Kraemer, H. C., Wilson, G. T., Fairburn, C. G., & Agras, W. S. (2002). Mediators and moderators of treatment effects in randomized clinical trials. Archives of General Psychiatry, 59, 877–883.
Chapter 05.qxd
130
3/28/2007
12:58 PM
Page 130
WILSON
Lang, P. J. (1969). The mechanics of desensitization and the laboratory study of human fear. In C. M. Franks (Ed.), Behavior therapy: Appraisal and status (pp. 160–191). New York: McGraw-Hill. Lazarus, A. A. (1961). Group therapy of phobic disorders by systematic desensitization. Journal of Abnormal and Social Psychology, 63, 504–510. Lincoln, T. M., Rief, W., Hahleg, K., Frank., M., von Witzleben, I., Schroeder, B., & Fiegenbaum, W. (2003). Effectiveness of an empirically supported treatment for social phobia in the field. Behaviour Research and Therapy, 41, 1251–1269. Loeb, K. L., Wilson, G. T., Labouvie, E., Pratt, E. M., Hayaki, J., Walsh, B. T., Agras, W. S., & Fairburn, C. G. (2005). Therapeutic alliance and treatment adherence in two interventions for bulimia nervosa: A study of process and outcome. Journal of Consulting and Clinical Psychology, 73, 1097–1106. Ma, S. H., & Teasdale, J. D. (2004). Mindfulness-based cognitive therapy for depression: Replication and exploration of differential relapse prevention effects. Journal of Consulting and Clinical Psychology, 72, 31–40. Martell, C. R., Addis, M. E., & Jacobson, N. S. (2001). Depression in context. New York: Norton. McFall, R. M. (1996, July). Consumer satisfaction as a way of evaluating psychotherapy: Ecological validity and all that versus the good old randomized trial. Panel discussion at the sixth annual convention of the American Association of Applied and Preventative Psychology, San Francisco, CA. McFall, R. M., & Twentyman, C. T. (1973). Four experiments on the relative contributions of rehearsal, modeling, and coaching to assertion training. Journal of Abnormal Psychology, 81, 199–218. Merrill, K. A., Tolbert, V. E., & Wade, W. A. (2003). Effectiveness of cognitive therapy for depression in a community mental health center: A benchmarking study. Journal of Consulting and Clinical Psychology, 71, 404–409. Miklowitz, D. J., & Clarkin, J. F. (1999). Balancing internal and external validity. Prevention & Treatment, 2, Article 0004c. Mussell, M. P., Crosby, R. D., Crow, S. J., Knopke, A. J., Peterson, C. B., Wonderlich, S. A., & Mitchell, J. E. (2000). Utilization of empirically supported psychotherapy treatments for individuals with eating disorders. International Journal of Eating Disorders, 27, 230–237. Nathan, P. E., & Gorman, J. M. (Eds.). (2002). A guide to treatments that work (2nd ed.). New York: Oxford University Press. National Institute for Clinical Excellence. (2004). Eating disorders—Core interventions in the treatment and management of anorexia nervosa, bulimia, nervosa and related eating disorders (NICE clinical guideline no. 9). London: NICE. accessed November 9, 2004, from www.nice.org.uk/pdf/cgoo9. Parloff, M. B. (1998). Is psychotherapy more than manual labor? Clinical Psychology: Science & Practice, 4, 376–381. Paul, G. L. (1966). Insight vs. desensization in psychotherapy. Stanford, CA: Stanford University Press. Paul, G. L. (1969). Outcome of systematic desensization: II. Controlled investigations of individual treatment, technique variations, and current status. In C. M. Franks (Ed.), Behavior therapy: Appraisal and status (pp. 105–159). New York: McGraw-Hill.
Chapter 05.qxd
3/28/2007
12:58 PM
Page 131
5. MANUAL-BASED TREATMENT
131
Persons, J. B. (1997). Dissemination of effective methods: Behavior therapy’s next challenge. Behavior Therapy, 28, 465–472. Persons, J. B., Bostrom, A., & Bertagnolli, A. (1999). Results of randomized controlled trials of cognitive therapy for depression generalize to private practice. Cognitive Therapy and Research, 23, 535–548. Rothwell, P. M. (2005). External validity of randomized controlled trials: “To whom do the results of this trial apply?” Lancet, 365, 82–93. Roy-Byrne, P. P., Craske, M. G., Stein, M. B., Sullivan, G., Bystritsky, A., Katon, W., Golinelli, D., & Sherbourne, C. D. (2005). A randomized effectiveness trial of cognitive-behavioral therapy and medication for primary care panic disorder. Archives of General Psychiatry, 62, 290–298. Sackett, D. L., Richardson, W. S., Rosenberg, W., & Haynes, R. D. (1997). Evidencebased medicine: How to practice and teach EBM. London: Churchill Livingstone. Salkovskis, P. M. (2002). Empirically grounded clinical interventions: Cognitivebehavioural therapy progresses through a multi-dimensional approach to clinical science. Behavioural and Cognitive Psychotherapy, 30, 3–9. Schulte, D., Kunzel, R., Pepping, G., & Schulte-Bahrenberg, D. (1992). Tailor-made versus standardized therapy of phobic patients. Advances in Behaviour Research and Therapy, 13, 67–92. Segal, Z. V., Teasdale, J. D., & Williams, M. (2004). Mindfulness-based cognitive therapy: Theoretical rationale and empirical status. In S. C. Hayes, V. M. Follette, & M. M. Linehan (Eds.), Mindfulness and acceptance: Expanding the cognitive-behavioral tradition (pp. 45–65). New York: Guilford. Seligman, M. E. P. (1995). The effectiveness of psychotherapy. American Psychologist, 50, 965–974. Shadish, W. R., Matt, G. E., Navarro, A. M., Siegle, G., Crits-Christoph, P., Hazelrigg, M. D., Jorm, A. F., Lyons, L. C., Nietzel, M. T., Prout, H. T., Robinson, L., Smith, M. L., Svartberg, M., & Weiss, B. (1997). Evidence that therapy works in clinically representative conditions. Journal of Consulting and Clinical Psychology, 65, 355–365. Stirman, S. W., DeRubeis, R. J., Crits-Christoph, P., & Brody, P. E. (2003). Are samples in randomized controlled trials of psychotherapy representative of community outpatients? Journal of Consulting and Clinical Psychology, 71, 963–972. Stirman, S. W., DeRubeis, R. J., Crits-Christoph, P., & Rothman, A. (2005). Can the randomized controlled trial literature generalize to nonrandomized patients? Journal of Consulting and Clinical Psychology, 73, 127–135. Strupp, H. H., & Anderson, T. (1997). On the limitations of therapy manuals. Clinical Psychology: Science and Practice, 4, 76–82. Stuart, G. L., Treat, T. A., & Wade, W. A. (2000). Effectiveness of an empirically based treatment of panic disorder delivered in a service clinic setting: 1-year follow-up. Journal of Consulting and Clinical Psychology, 68, 506–512. Summerskill, W. (2005). Evidence-based practice and the individual. Lancet, 365, 13–14. Summerskill, W. S. M., & Pope, C. (2002). “I saw the panic rise in her eyes, and evidence-based medicine went out of the door.” An exploratory qualitative study of the barriers to secondary prevention in the management of coronary heart disease. Family Practice, 19, 605–610.
Chapter 05.qxd
132
3/28/2007
12:58 PM
Page 132
WILSON
Teasdale, J. D., Moore, R. G., Hayhurst, H., Pope, M., Williams, S., & Segal, Z. V. (2002). Metacognitive awareness and prevention of relapse in depression: Empirical evidence. Journal of Consulting and Clinical Psychology, 70, 275–287. Treatment for Adolescents with Depression Study (TADS) Team. (2004). Fluoxetine, cognitive-behavioral therapy, and their combination for adolescents with depression: Treatment for Adolescents With Depression Study (TADS) randomized controlled trial. Journal of the American Medical Association, 292, 807–820. Wade, W. A., Treat, T. A., & Stuart, G. L. (1998). Transporting an empirically supported treatment for panic disorder to a service clinic setting: A benchmarking strategy. Journal of Consulting and Clinical Psychology, 66, 231–239. Weisz, J. R., Weersing, V. R., & Henggeler, S. W. (2005). Jousting with straw men: Comment on Westen, Novotny, and Thompson-Brenner (2004). Psychological Bulletin, 131, 418–426. Westbrook, D., & Kirk, J. (2005). The clinical effectiveness of cognitive behaviour therapy: Outcome for a large sample of adults treated in routine practice. Behaviour Research and Therapy, 43, 1243–1261. Westen, D., Novotny, C. M., & Thompson-Brenner, H. (2004). The empirical status of empirically supported psychotherapies: Assumptions, findings, and reporting in controlled clinical trials. Psychological Bulletin, 130, 631–663. Wilson, G. T. (1996a). Treatment of bulimia nervosa: When CBT fails. Behaviour Research and Therapy, 34, 197–212. Wilson, G. T. (1996b). Manual-based treatments: The clinical application of research findings. Behaviour Research and Therapy, 34, 295–315. Wilson, G. T. (1998a). Manual-based treatment and clinical practice. Clinical Psychology: Science and Practice, 5, 363–375. Wilson, G. T. (1998b). The clinical utility of randomized controlled trials. International Journal of Eating Disorders, 24, 13–29. Wilson, G. T. (1999). Cognitive behavior therapy for eating disorders: Progress and problems. Behaviour Research and Therapy, 37, 579–596. Wilson, G. T. (2005). Psychological treatment of eating disorders. In S. NolenHoeksema (Ed.), Annual review of clinical psychology (Vol. 1, pp. 439–466). Palo Alto, CA: Annual Reviews. Wilson, G. T., & Davison, G. C. (1971). Processes of fear reduction in systematic desensitization: Animal studies. Psychological Bulletin , 76, l–l4. Wilson, G. T., & Shafran, R. (2005). Eating disorders guidelines from NICE. Lancet, 365, 79–81. Wilson, G. T., Fairburn, C. G., & Agras, W. S. (1997). Cognitive behavioral therapy for bulimia nervosa. In D. M. Garner & P. Garfinkel (Eds.), Handbook of treatment for eating disorders (pp. 67–93). New York: Guilford. Wolpe, J. (1958). Psychotherapy by reciprocal inhibition. Stanford, CA: Stanford University Press. Wolpe, J., & Lazarus, A.A. (1966). Practice of behavior therapy. New York: Pergamon. Young, J. E., Beck, A. T., & Weinberger, A. (2001). Depression. In D. H. Barlow (Ed.), Clinical handbook of psychological disorders (3rd ed., pp. 264–308). New York: Guilford.
Chapter 06.qxd
3/28/2007
6:25 PM
Page 133
6 The Importance of How: A Call for Mechanistic Research in Tobacco Dependence Treatment Studies Danielle E. McCarthy, Daniel Bolt, and Timothy B. Baker University of Wisconsin–Madison
Roughly a quarter century ago, Richard McFall called on his fellow smoking-cessation researchers to strive to bootstrap theories of smoking behavior from empirical findings and to use these theories to generate treatments (McFall, 1978). He touted theoretical research as “the most advanced research approach” (p. 710), and he lamented the scarcity of good theories of smoking behavior to guide treatment development (McFall, 1978). McFall pointed out that diverse treatments, derived from distinct theories, produced uncannily equivalent treatment effects, and he identified the critical role played by nonspecific factors such as motivation, structure, and self-monitoring in such effects (McFall, 1970; McFall & Hammen, 1971). He outlined the value of adopting constructive or dismantling research strategies, and he used such designs in his own research (McFall, 1978; McFall & Lillesand, 1971). He also highlighted the importance of collecting process measures during the treatment phase and of looking for treatment-specific effects on process measures (Marston & McFall, 1971; McFall, 1978). In summary, McFall pointed 133
Chapter 06.qxd
134
3/28/2007
6:25 PM
Page 134
MCCARTHY, BOLT, BAKER
out how little was known about treating smoking, and he outlined a rational path, based on a synthesis of theory and methods, to foster the future development of tobacco treatments. Despite copious research on tobacco-dependence treatments over the past few decades, many of McFall’s laments and calls for action remain unanswered. Tobacco researchers are still enamored with the horse race technique of pitting treatments against one another or a straw-man control condition to see which ultimately achieves the highest abstinence rate, without determining how or why one treatment bests another. In this chapter, we renew and elaborate McFall’s prescient and neglected call for mechanistic research in tobacco-dependence treatment development and evaluation. We first review the potential conceptual and clinical yield of mechanistic research. We then review conceptual criteria for testing mechanistic or mediational hypotheses (hypotheses that assert that a treatment exerts an effect on a target outcome through a specific, intervening variable called a mediator). Finally, we provide examples of mechanistic research methods using data from a smokingcessation clinical trial. VALUE OF MECHANISTIC RESEARCH Conceptual Benefits of Mechanistic Research Mechanisms of action of pharmacological and psychosocial tobaccodependence treatments can be studied using mediational analyses at the physiological, psychological, or behavioral level of analysis. In mediational analyses, investigators examine relations among an independent or initial variable (e.g., treatment), a putative process variable or mediator (e.g., coping skill mastery), and a target outcome (e.g., 6-month abstinence; Kenny, Kashy, & Bolger, 1998). Mediational inferences require (at minimum) that the initial variable influences the mediator as predicted, and that the mediator and the outcome are related as predicted (Kenny et al., 1998; see Fig. 6.1A). In a cessation counseling program, for example, one might expect individual counseling to lead to increased skills for coping with stress or with urges to smoke, which in turn would lead to increased probability of abstinence. Mechanistic research can yield a deeper level of understanding, and richer theories, of tobacco dependence and treatments than can simple outcome research. Mediational hypotheses are both causal and specific, two important characteristics of well-developed theories. In fact, inferences
Chapter 06.qxd
3/28/2007
6:25 PM
Page 135
6. THE IMPORTANCE OF HOW
135
FIGURE 6.1 (A) Relations between variables that must be demonstrated to support claims of mediation (e.g., Baron & Kenny, 1986). The treatment variable (X) must be significantly related to the mediator (M) via Path a. The mediator must be significantly related to the outcome variable (Y) via Path b. In the strictest interpretation of mediation, path c linking the treatment variable to the outcome must also be significant. (B) Path c’ should be reduced to 0 in the case of complete mediation. The numbers below the arrows depict the standardized regression coefficients among combination medication and counseling treatment versus all other treatments (X), the difference between craving scores on Day 6 versus Day 0 of a quit attempt divided by 7 (M), and CO-confirmed 7-day point-prevalence smoking status 6 weeks post-quit date (Y). Higher values on the mediator indicate increases in craving over the first week post-quit. Combination treatment is associated with lower slopes in craving, which are in turn related to decreased risk of relapse. Significant coefficients (at alpha .05) are noted with an asterisk.
about the effects of treatment, even in randomly controlled clinical trials, require some consideration of possible mechanisms of treatment actions. Demonstration of causality requires that (a) a relationship exist between the putative cause and effect, (b) the cause precede the effect, (c) alternative explanations of the cause–effect relationship have been ruled out, and (d) the putative mechanism linking the cause and effect be plausible, given extant knowledge about the phenomena of interest (Haynes, 1992; Kazdin, 1999). These criteria for the demonstration of causality state that mechanism must be considered, if not tested, before drawing causal inferences. We argue that in treatment research the mechanisms of action deemed plausible should be explicitly stated and tested whenever resources permit. We, like McFall (1978), advocate this admittedly difficult and resource-intensive practice because causation is the key target of experimental research and because examining the mechanisms whereby treatments achieve their effects could improve our understanding of both tobacco dependence and tobacco-cessation processes.
Chapter 06.qxd
136
3/28/2007
6:25 PM
Page 136
MCCARTHY, BOLT, BAKER
Mechanistic research studies are also more powerful than simple outcome intervention studies because they test multiple theories simultaneously. Theories regarding treatment effects link theories of change and theories about the factors that cause or maintain a target behavior or condition (Eddy, Dishion, & Stoolmiller, 1998). Others have argued that an important part of treatment development is the articulation of a small theory (Lipsey, 1993) to explain the means by which a treatment program affects an outcome. This theory of treatment effects can be parsed into two parts: an action (or program) theory that specifies the manner in which the treatment should affect the mediator, and a conceptual (or psychosocial) theory that specifies the relation between the mediator and the outcome of interest (i.e., why changing the mediator should affect outcome; Chen, 1990, Kenny et al., 1998; MacKinnon, Taborga, & Morgan-Lopez, 2002). When Marston and McFall (1971) stressed the importance of collecting process measures to be able to detect whether “different treatments do, in fact, produce discriminably different response curves during the treatment period” (p. 154), they were highlighting the importance of testing the action theory using process measures. Because mechanistic research tests both the action and conceptual theories simultaneously, mechanistic research can yield information regarding “the genesis of the outcome variables of interest” and allow researchers to “build and test a theory regarding more general causal mechanisms responsible for the outcome behavior” (Judd & Kenny, 1981, p. 603). In this way, mechanistic research is efficient in that it tests models of a target behavior and models of change simultaneously. For example, testing the relations among counseling treatment, coping skills, and abstinence can tell us whether our action theory is refuted or retained (i.e., whether counseling leads to enhanced coping) and whether our conceptual theory is refuted or retained (i.e., whether coping is associated with increased abstinence likelihood; Collins, Graham, & Flaherty, 1998; MacKinnon, Taborga, & Morgan-Lopez, 2002; Weersing & Weisz, 2002). If the action theory is refuted, this suggests that the treatment was ineffective in changing the target mediator and that our intervention model may need revision. If the conceptual theory is refuted, however, our understanding of the factors that promote abstinence may be flawed, and we may need to select another intervention target. In this way, mechanistic research may help make sense of inconsistent results. McFall (1978) argued that, without process information, “it is difficult to rise above one’s failures and to design
Chapter 06.qxd
3/28/2007
6:25 PM
Page 137
6. THE IMPORTANCE OF HOW
137
better treatments” (p. 708). To help us rise above our failures to replicate treatment effects, we can now use mediational meta-analysis procedures to investigate the support for various action and psychosocial theories across treatment studies (Shadish & Sweeney, 1991). Mechanistic research is also preferable to simple outcome research because the former is less prone to some of the well-known pitfalls of null hypothesis significance testing. A good theory about treatment must make testable predictions about how treatments achieve their effects, rather than simply conducting a horse race among treatments and placebo conditions (McFall, 1978). Making multiple, precise predictions about how variables are related, rather than just proposing relations among a subset, makes tests riskier and reduces the risk of chance findings. This is so because the combination of multiple relations is always less likely than the occurrence of a single relation, and demonstration of mediation requires the co-occurrence of multiple relations. In this way, mediational tests are more than are tests of simple direct treatment effects. Elaborate Treatment models that are not exposed to heightened risk of refutation are less compelling (and less scientific) than are models that have passed such tests (Meehl, 1978). Clinical Benefits of Mechanistic Research The study of mechanisms of treatment effects is an important endeavor for clinical as well as theoretical reasons. Knowledge of the mechanisms of action of specific agents or treatment components may suggest new treatment combinations (Morgenstern & Longabaugh, 2000). For example, if we knew that nicotine replacement and antidepressant medication therapies exerted effects on smoking behavior through distinct mechanisms, we could rationally expect these effects to be additive, even if outcomes are similar for either treatment used alone (e.g., bupropion SR and the nicotine nasal spray; Fiore et al., 2000). If two treatments were found to work through nonspecific mechanisms (e.g., enhanced abstinence self-efficacy) or through similar mechanisms (e.g., withdrawal suppression), we would not expect combining such treatments to improve abstinence rates substantially (although they may do so in a dose-related manner). As such, it is important to uncover the mechanisms of action of our extant treatments because such understanding may suggest a rational basis for combining treatments.
Chapter 06.qxd
138
3/28/2007
6:25 PM
Page 138
MCCARTHY, BOLT, BAKER
In addition, if we understood how a treatment affects a given risk factor, we could match individuals with that risk factor to the treatment in question. Thus, if we knew that cognitive behavioral therapy for depression (or smoking cessation) altered negative schemata, we could select individuals for CBT based on an assessment of their negative cognitive belief structure (Morgenstern & Longabaugh, 2000). In this way, improved understanding of treatment mechanisms could lead to better treatment matching. Mechanistic knowledge could also lead to better understanding of moderation effects more generally. It may be that the many individual difference and contextual variables that have been found to moderate treatment effects do so because treatments work through different mechanisms in different people or situations. For example, people with a history of depression have been found to benefit less from nicotine replacement therapies than do people without a history of depression (Smith et al., 2003). Depression-vulnerable individuals may respond less because nicotine replacement therapies ameliorate withdrawal-related distress, but do not reduce the coping-skill deficits that may foster tobacco use among these affectively vulnerable individuals. Mechanistic research has the potential to identify mediating variables that may contribute to the emergence of important interactions between individual differences and treatment. Identifying treatment processes that may account for person-by-treatment interactions may enable us to develop treatments that help treatment-refractory individuals by activating different critical processes or activating them in a different way (e.g., teaching different urge-reduction strategies to men and women). If we know what works in a treatment and how it works, we can generate hypotheses about how to amplify treatment effects on a rational basis, rather than merely extending or intensifying treatment or offering booster sessions based on the premise that more (of some poorly understood entity) is better (Kazdin, 2001). A better understanding of treatment mechanisms could also enhance the dissemination and delivery of treatments. In general, discovering the potent mechanisms of treatment can lead to more efficient and transportable (Kazdin, 1999; MacKinnon, Taborga, & Morgan-Lopez 2002; Morgenstern & Longabaugh, 2000) treatment delivery. Although many treatments are manualized to facilitate dissemination, these manuals do not identify the critical components or processes of treatment (Kazdin, 2001). If we identified the critical ingredients in treatment and
Chapter 06.qxd
3/28/2007
6:25 PM
Page 139
6. THE IMPORTANCE OF HOW
139
the active processes that result in positive change, we could ensure that these critical components of efficacious treatments are emphasized in treatment manuals and optimized in treatment delivery in diverse settings (Kazdin & Kendall, 1998; MacKinnon & Lockwood, 2003). Thus, mechanistic research has the capacity to separate the wheat from the chaff in multicomponential treatments (Judd & Kenny, 1981). Such knowledge could ensure that the potent components of a treatment are undiluted by unnecessary or iatrogenic components (Judd & Kenny, 1981; MacKinnon, Taborga, & Morgan-Lopez, 2002). Mechanistic research by itself cannot identify active treatment ingredients, however. The inclusion of appropriate control conditions is essential. In well-designed placebo-controlled, constructive, factorial, fractional, or dismantling studies (Kenny et al., 1998; McFall, 1978), mediational analyses can reveal much about how and why placebo or active control conditions differ from an index treatment. For example, we might find that all treatments, even inert placebo or attentional control conditions, have positive effects through a nonspecific mechanism of action, such as the instillation of hope (e.g., Howard, Lueger, Maling, & Martinovich, 1993; McFall & Hammen, 1971). Mechanistic research, when coupled with appropriate study designs, can help identify qualitative and quantitative differences in the effects of comparison treatments and control conditions, even when they have equivalent impacts on ultimate outcomes such as abstinence measures. Finally, mechanistic research can help us identify when a treatment is not working. If we know by what mechanism a treatment results in an ultimate desired outcome (and the time frame in which treatment processes unfold), we can assess treatment response early in treatment and modify treatment for people who are not changing in the desired fashion. Psychotherapy researchers have strongly advocated the use of process measures to inform decision making in therapy, but tend to focus on intermediate outcomes, active ingredients, or rate of change rather than mediators (e.g., DeRubeis & Feeley, 1990; Feeley, DeRubeis, & Gelfand, 1999; Goldfried, Greenberg, & Marmar, 1990; Tang & DeRubeis, 1999). In theory, treatment may be titrated on an ongoing basis for participants based on their standing on a mediator, rather than waiting for the treatment failure to culminate. In this and other ways described earlier, mechanistic research could improve the efficiency of clinical interventions.
Chapter 06.qxd
3/28/2007
6:25 PM
Page 140
140
MCCARTHY, BOLT, BAKER
CONCEPTUAL CONCERNS AND CRITERIA FOR MEDIATION Design and Assessment Concerns
Experimental Design. First, let us consider when it is appropriate to conduct mediational analyses. Some scholars (West & Aiken, 1997) have argued that merely comparing a treatment package to a control condition is not sufficient for mediational analysis. These authors argue that factorial, fractional, or dismantling designs in which one treatment component is hypothesized to influence a single mediator are better suited to mediational analyses. Other authors merely require that at least two treatments be compared (Holland, 1988; Rubin, 1974) to support causal inferences. Comparisons of two active treatments may be suitable for mediational analyses, even when no significant difference in outcome between treatments is detected (Morgenstern & Longabaugh, 2000). Scholars have argued that the best mediational design (barring direct manipulation of the mediator) is a study comparing “several programs based on different theories of tobacco use and a control group” (MacKinnon, Taborga, & Morgan-Lopez, 2002, p. 578). Such designs allow investigators to test simultaneously multiple theories about treatment effects and more general theories of the determinants of tobacco use. In addition, McFall argued that the minimal treatment condition is a better basis for comparison than a no-treatment control condition due to the apparent effects of nonspecific factors in treatment responses (Marston & McFall, 1971). Assessment Schedule. The assessment battery and timing of procedures are critical to the study of mediation. Logically, the treatment manipulation must precede the mediator, which in turn must precede the outcome (Holland, 1988; Kazdin, 2000). As others have pointed out, it is not enough merely to assess the constructs in the appropriate order; the constructs must exert effects in the proper order (Cole & Maxwell, 2003). For instance, although the assessment of a mediator may occur after treatment manipulation, variance in the mediator may reflect temporally remote events. For example, beneficial increases in self-efficacy may occur immediately on enrollment in a cessation study, be unrelated to treatment, and yet influence the outcome. Similarly, treatment and mediational processes may overlap in time (i.e., mediational processes may begin while treatment is ongoing). If the temporal ordering of the
Chapter 06.qxd
3/28/2007
6:25 PM
Page 141
6. THE IMPORTANCE OF HOW
141
treatment and mediator cannot be established, the baseline level of the mediating variable should be included in analyses as a control variable (Cinciripini et al., 2003; Cole & Maxwell, 2003). This allows one to infer that change in the mediator during or after treatment is predictive of outcome and accounts for the treatment effect. Timing, not just temporal ordering, of the assessments is critically important as well. The magnitude of the effects of a treatment on a mediator, and a mediator on an outcome, naturally varies based on when the relations are assessed. In general, one would want to use research and theory to estimate the time courses of relevant events in the mediational model (e.g., the temporal patterning of both treatment effects and mediator effects). Moreover, one would want to consider what levels of change, and durations of change, would be needed to exert desired effects. For instance, if enhanced coping with major stressors were the mediator, the conceptual model would involve assumptions about the occurrence and timing of stressors in people’s lives. Kenny et al. (1998) noted that, to detect mediation, one ideally wants the relation between the treatment and the mediator, and the relation between a mediator and outcome, to be large. This can be difficult to achieve, however, given that these relations tend to be complementary (i.e., as one increases, the other decreases) because their combined effect cannot exceed the overall relation between the treatment and outcome (Kenny et al., 1998). Ideally, then, the mediator would be assessed at the midpoint between the independent variable (treatment) and outcome (Kenny et al., 1998). At first blush, it may seem best to assess the mediator when it is most tightly related to treatment. Such an approach would create a strong test of the action theory guiding the treatment. A strong correlation between the treatment and mediator creates a high level of collinearity, however, and this interferes with estimation of mediator–outcome relations. In this way, strengthening the test of the action theory may undermine the test of the conceptual theory linking the mediator and outcome. A complementary cost is incurred if the timing of assessments is modified to maximize the association between the mediator and outcome (thus favoring the conceptual theory over the action theory). In tobacco-cessation research, the preferred outcomes for treatment studies are quite distal from the initiation of treatment (e.g., abstinence rates 6–12 months after the target quit date). From a public health perspective, such distal endpoints are attractive because they capture a socially meaningful outcome (Wiggins, 1973). From a mechanistic perspective,
Chapter 06.qxd
142
3/28/2007
6:25 PM
Page 142
MCCARTHY, BOLT, BAKER
however, more proximal outcome measures may be preferable. Examining hypothesized mediational relations over a shorter timeframe may reduce the influence of the myriad processes and factors that likely influence long-term abstinence, but that are unrelated to treatment (e.g., having a spouse who uses tobacco). Evidence shows that smoking-cessation treatments tend to affect survival (prevent relapse) most in the first days or weeks of treatment (McFall & Hammen, 1971; Piasecki, Fiore, McCarthy, & Baker, 2002). Thus, examination of short-term treatment effects and outcomes may be especially sensitive and appropriate for testing treatment and mediator effects.
Assessment Battery. The nature of the assessment of the mediator also has important implications for the conduct and interpretation of mediational analyses. The grave and insidious impact of error in the measurement of a putative mediator is well documented (Baron & Kenny, 1986; Judd & Kenny, 1981; Kenny et al., 1998; MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002; West & Aiken, 1997). Typically, error in the measurement of the mediator leads to underestimation of the mediated effect and overestimation of the direct (unmediated) effect, thus increasing the likelihood of retaining the null hypothesis that the mediated effect is not significant. For this reason, researchers recommend examination of effect sizes and confidence intervals instead of reliance on significance testing in mediational research (Baron & Kenny, 1986; MacKinnon, Taborga, & Morgan-Lopez, 2002; MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002). In multiple mediator models, error in the measurement of one mediator can lead to bias in the estimation of other mediators as well (West & Aiken, 1997). As such, constructs including the mediator should ideally be assessed using multiple indicators (Cole & Maxwell, 2003; Hoyle & Smith, 1994), preferably those that are maximally dissimilar from one another to reduce the likelihood of retaining method variance in the latent mediator construct (Cole & Maxwell, 2003). Latent variable approaches can remove error variance from the target construct (i.e., mediator) if multiple indicators are used, particularly if the indicators are diverse. For example, self-report Likert-type items tapping withdrawal symptoms could be supplemented with observer ratings of irritability and objective measures of negative affect (e.g., eyeblink startle responses). Although this more complicated approach adds to the assessment burden for participants and the analytical complexity of the data, the latent variable approach is deemed the best approach to mediational analyses (Hoyle & Smith, 1994; Kenny et al., 1998) because it can isolate error in the measurement of the mediator.
Chapter 06.qxd
3/28/2007
6:25 PM
Page 143
6. THE IMPORTANCE OF HOW
143
Nuisance Variables. It is essential to assess and control for variables other than treatment that might influence both a mediator and outcome. This issue is critical to the internal validity of a mechanistic study. Omitted variables have the potential both to inflate estimates of the path between the mediator and outcome and increase the estimate of the direct effect (Herting, 2002). In the real world, in which multicausality is the rule rather than the exception (Cole & Maxwell, 2003), it may not be possible to identify and assess all the possible confounding variables that could threaten the internal validity of a mechanistic study. In light of this, it may be best to adopt statistical strategies, such as controlling for earlier levels of the mediator and outcome when estimating later causal relations (assuming that an unestimated fourth variable has already exerted its effects; Cole & Maxwell, 2003; Hoyle & Smith, 1994). Sample Size. A final design consideration is sample size. For a variety of reasons, mediational analyses are frequently underpowered (Baron & Kenny, 1986; MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002). As such, medium to large samples are often required to test mediational hypotheses sensitively. Multiple mediator models or models containing complex causal chains require especially large sample sizes (West & Aiken, 1997). Interestingly, when there are strong relations between treatment and the mediator, larger sample sizes [N(1 – rxm2)] are required to detect mediation (Kenny et al., 1998). This occurs because a strong relation between the treatment variable and mediator results in collinearity when examining these variables’ unique relations with outcome (Baron & Kenny, 1986). This suggests that more potent treatments require larger sample sizes to detect mediation, contrary to what one would expect. As such, caution should be applied to interpretation of negative results in even moderate or large samples. For this reason, some have advocated that authors report effect sizes or confidence intervals in addition to significance tests (Baron & Kenny, 1986; MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002). Conceptual Criteria for Mediation Scholars have discussed the criteria necessary for the demonstration of mediation for at least the past five decades (MacCorquodale & Meehl, 1948; Rozeboom, 1956). Today, there is one prevailing set of criteria adopted by most researchers. Essentially, mediation is established using a multiple correlation approach (Judd & Kenny, 1981; West & Aiken, 1997) to parse the direct and indirect effects of an initial (Kenny et al.,
Chapter 06.qxd
144
3/28/2007
6:25 PM
Page 144
MCCARTHY, BOLT, BAKER
1998) variable (e.g., treatment) on an outcome, where the indirect effect passes through a mediator. The typical goal of mediational analyses is to identify the variable(s) that account for a treatment effect on an outcome.
Core Criteria. In a seminal, oft-cited article, Baron and Kenny (1986) defined the term mediator and established clear criteria for the demonstration of mediation. As they define the term and as we use it here, a mediator is a variable that “accounts for the relation between the predictor and criterion” (p. 1176). In our case, a mediator is a variable that accounts for the success of a tobacco-dependence treatment in preventing relapse. How does a mediator account for treatment effects? In a simple, single-mediator model (see Fig. 6.1), a mediator must meet the following criteria, according to Baron and Kenny (1986): (a) A treatment condition significantly accounts for variance in the mediator, (b) the mediator accounts for variance in relapse outcome, and (c) controlling for mediator relations with treatment and outcome eliminates or reduces the relation between the treatment and outcome. These relations can be tested optimally with structural equation modeling (SEM) techniques using a series of nested models (Hoyle & Smith, 1994; Kenny et al., 1998). SEM is optimal because it permits use of a latent variable approach in which error variance can be removed from the mediator and outcome. Some SEM programs can now handle dichotomous outcome variables (e.g., MPlus), thus permitting use of SEM in situations that violate the assumption of multivariate normality. For manifest variables, multiple or logistic regression strategies can be used to test the significance of individual and partialed paths (Kenny et al., 1998). A generic mediation model is depicted in Figure 6.1A. To infer mediation and corroborate a process model, one must show that “Each variable in the causal chain affects the variable that follows it in the chain, when all variables prior to it, including the treatment, are controlled” (Judd & Kenny, 1981, p. 605). The essential steps for demonstrating mediation are showing that the treatment and the mediator are associated (i.e., Path a is significant) and that the mediator and the outcome are associated (i.e., Path b is significant) when treatment is statistically controlled (Kenny et al., 1998). In addition, if a significant association is found between the treatment and the outcome (Path c is significant), one expects to find that this association is reduced or eliminated when the mediator is statistically controlled in analyses (Path c′ is reduced or nonsignificant). Complete mediation requires that the direct path from treatment to outcome be
Chapter 06.qxd
3/28/2007
6:25 PM
Page 145
6. THE IMPORTANCE OF HOW
145
reduced to zero when the mediator is included in the model (as shown in Fig. 6.1B; Baron & Kenny, 1986; Judd & Kenny, 1981). Over the years, scholars have softened on the issue of complete mediation and grown more accepting of partial mediation (e.g., Kenny et al., 1998, vs. Judd & Kenny, 1981), which is the most likely case in most models. Thus, the original criterion articulated by Judd and Kenny (1981)—that only one predictor (viz., the most proximal cause in the causal chain) should be significant in each of the three regression equations tested has been relaxed considerably, in apparent recognition of the fact that there are few mediating variables that are sufficient to explain all of a treatment’s effect on an outcome. The first criterion for mediation articulated by Baron and Kenny (1986) and others (a significant relation between the initial variable and outcome) has been disputed in the literature. In the strictest sense, it does not make sense to conduct mediational analyses in the absence of a significant treatment effect because there is no need to account for a nonexistent treatment effect (Baron & Kenny, 1986; McFall, 1978). Mediation is simply a special case of an indirect treatment effect on outcome, however. Indirect effects that pass through an intervening variable, such as a mediator, may be of substantive interest even in the absence of a direct effect of treatment on outcome (Holmbeck, 1997). Examination of the relations between a mediator and treatment and a mediator and outcome may suggest ways to improve treatments (e.g., by revising treatments to capitalize on the indirect effect) even in the absence of a significant direct effect. Some scholars have argued that establishing a significant treatment effect on outcome should not be a prerequisite for mediational analyses (Collins et al., 1998; MacKinnon, 2000; McFall, 1978) and have pointed out that this assumption is not appropriate in the case of small effect sizes or suppression (Shrout & Bolger, 2002). Similarly, if a treatment has a large effect on outcome that does not reach statistical significance due to low power, it may still be worthwhile to explore mediation. If a treatment is expected to have a distal effect on outcome that is mediated through complex causal chains or may be susceptible to other intervening influences (as is certainly the case when long-term abstinence rates are the target outcome), the investigator may wish to relax this criterion (Shrout & Bolger, 2002). If the treatment is expected to have a more proximal effect, then one may wish to retain this criterion (Shrout & Bolger, 2002). In addition, this criterion may not be appropriate in comparative treatment trials (Morgenstern & Longabaugh, 2000). In this study design, a treatment is compared to an alternative active treatment of known efficacy. Although the target
Chapter 06.qxd
146
3/28/2007
6:25 PM
Page 146
MCCARTHY, BOLT, BAKER
treatment may not perform significantly better than the alternative treatment, it may still influence outcome in a way that bears explaining by means of mediational analyses. In other words, a treatment should not go unexplored because it was compared against another efficacious treatment rather than against a straw-man or placebo control. If a new treatment fares as well as an accepted treatment, then much may be gained by studying and comparing the mechanisms of action of both. Kenny and colleagues suggest that mediational analyses in treatment failures may also yield interesting and informative results (Kenny et al., 1998). Some authors have suggested an additional criterion for mediation and a means to test the criterion: the significance of the indirect or mediated effect. MacKinnon and colleagues have adapted and evaluated numerous standard error estimates and significance testing approaches to identify tests of mediated effects that have the greatest power and best Type-I error rates (MacKinnon, 1994; MacKinnon & Lockwood, 2003; MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002). The magnitude of a mediated effect is equal to the product of two path coefficients (a and b from Fig. 6.1). In a simple, three-variable model, this is also equivalent to the difference between the direct effect estimated without the mediator in the model vs. with the mediator in the model (c–c′ from Fig. 6.1). This difference yields an estimate of the overall indirect effect in multiple mediator models. In an appropriately sized and powered model, it is possible to test the significance of the mediated effect in addition to testing the significance of Paths a, b, c, and c′ as suggested by Kenny and colleagues (Baron & Kenny, 1986; Judd & Kenny, 1981; Kenny et al., 1998). The mediated effect (ab) can be tested for significance using the formula z′ = ab/√a2 σb2 + b2 σ2α against an empirical z distribution (available on the web at http://www.public.asu.edu/~davidpm/ripl /freqdist.pdf) that takes into account deviations from normality that occur when one tests the product of coefficients (MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002). Based on a comparison of several pointestimation and significance testing procedures, MacKinnon and colleagues (MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002) recommended that investigators test the joint significance of Paths a and b and test the significance of the estimated mediated effect using the z′ formula above if both Paths a and b are significant.
Stage-Sequential Approach. Although Kenny and colleagues’ guidelines for mediation (Baron & Kenny, 1986; Judd & Kenny, 1981) are the most often cited and followed, alternative frameworks exist.
Chapter 06.qxd
3/28/2007
6:25 PM
Page 147
6. THE IMPORTANCE OF HOW
147
Collins and colleagues (Collins et al., 1998) have argued that mediation can be conceived as a cascade of events (coded categorically) initiated by the independent variable. Collins and colleagues (1998) define the following three independent criteria for establishing mediation: (a) the probability of an individual undergoing change in the mediator followed by change in the outcome is greater in the treatment group versus control group, (b) being in the treatment group increases the probability of change in the mediator occurring (when not already in the mediator stage), and (c) change in the mediator increases the probability of the outcome at every level of the initial, or treatment, variable (when not already in the outcome stage). This last criterion emphasizes that the mediator should be linked to the outcome regardless of treatment condition (i.e., this part of the chain reaction should exist regardless of whether the first domino is knocked over). In simpler terms, mediation is suggested when more people who experience the treatment (e.g., combination treatment vs. single modality treatments) and the specified (higher or lower) level of the mediator (e.g., urges to smoke) end up with the target outcome (e.g., 6-week abstinence). Collins and colleagues’ categorical framework has great illustrative value, but low statistical sophistication. Simple chi-square or logistic regression tests are conducted to examine whether treatment or mediator status influences subsequent outcomes in the hypothesized sequence of events. No estimate or test of a mediated effect is provided and all variables are treated as categorical. The Collins approach offers a way to characterize the potential clinical significance of a mediational effect, however, by depicting the proportions of individuals who are likely to experience a target sequence of treatment, mediator level, and outcome. We use this model to illustrate, rather than to test, mediational effects in the analyses described next. SAMPLE MEDIATIONAL ANALYSES The preceding review reflects significant advances in methods to test mediational hypotheses. To date, few of these advances have been applied to tobacco dependence treatment research. For example, little is known about how bupropion, the only nonnicotine agent currently approved as a firstline pharmacotherapy for tobacco cessation (Fiore et al., 2000), improves abstinence rates. Although some mediational research has been conducted regarding bupropion (see Lerman et al., 2002), past studies have not used state-of-the-art mediational analytical strategies, as described earlier.
Chapter 06.qxd
148
3/28/2007
6:25 PM
Page 148
MCCARTHY, BOLT, BAKER
In this chapter, we explore one possible mechanism (craving reduction) by which combined bupropion pharmacotherapy and individual smoking cessation counseling may increase short-term abstinence rates. We focus on this treatment combination for three primary reasons. First, bupropion efficacy has been established almost exclusively (but see Hall et al., 2002) in the context of co-occurring counseling (Richmond & Zwar, 2003). Second, problem-solving skills training and social support interventions are recommended in all smoking cessation interventions (Fiore et al., 2000), including those involving bupropion (Richmond & Zwar, 2003). Third, both bupropion and counseling are thought to work, in part, by reducing cravings for cigarettes, although through different pathways (e.g., bupropion may influence mesolimbic dopamine activity directly, whereas counseling may lead to avoidance of triggers and active coping with cravings). For these reasons, we present results of analyses testing the hypothesis that craving reduction mediates the effect of combined bupropion treatment and counseling on short-term abstinence rates. Although there are many other interventions and candidate mediators that likely influence successful cessation, we have selected this combined treatment condition and craving reduction primarily to illustrate different approaches to mediational analysis. We do not claim that craving reduction is the sole or primary mechanism of treatment action, although we have substantive reasons to investigate the potential mediating role of craving, in particular, as outlined next. Recent research employing real-time data collection methods has demonstrated that cravings or urges to smoke are tightly linked to subsequent smoking among abstainers and ad libitum smokers (Shapiro, Jamner, Davydov, & James, 2002; Shiffman et al., 1997, 2002, 2004). Some evidence suggests that craving is only weakly related to drug selfadministration and relapse, however (Tiffany, 1990). A more recent study using ecological momentary assessment reported that increases in craving on the quit day were predictive of point-prevalence smoking status at 3 months post-quit (McCarthy, Piasecki, Fiore, & Baker, 2006). In addition, a new cognitive neuroscience model of drug motivation affords a central role to craving as an index of conflict between competing response options (e.g., smoking and sitting in a movie theater; Curtin, McCarthy, Piper, & Baker, 2006). Thus, in light of recent research and theory, craving appears to be an important target for additional research. We next report the results of different tests of the hypothesis that the beneficial effects of combined bupropion and counseling treatment on
Chapter 06.qxd
3/28/2007
6:25 PM
Page 149
6. THE IMPORTANCE OF HOW
149
abstinence are mediated, in part, through reductions in craving. We use the traditional regression approach, SEMs with the mediators treated as latent variables, and then present results from SEMs using the stage-sequential approach to depict relations among treatment, craving, and abstinence using data from a recently completed randomized, placebo-controlled clinical trial of bupropion SR (sustained release) and counseling.
Current Study. Adult smokers who reported being motivated to quit were randomly assigned to receive either active or placebo bupropion SR in conjunction with eight sessions of brief (10-minute) individual cessation counseling or a no counseling, assessment control condition (McCarthy et al., in preparation). The study used a 2 (active drug vs. placebo) x 2 (counseling vs. no counseling) factorial design. Bupropion SR and placebo medication treatment began 1 week before quitting. Participants began taking one 150-mg pill in the morning 1 week before the quit day and then increased to two 150-mg pills per day at 4 days prior to quitting. Participants were instructed to continue taking 300 mg per day for 8 weeks post-quit. Counseling consisted of two prequit sessions, a session on the quit day, and five post-quit sessions over the first month of the quit attempt. Counseling focused on coping, problem solving, and intratreatment social support, in accordance with recommendations in the Treating of Tobacco Use and Dependence Clinical Practice Guideline (Fiore et al., 2000). Participants attended an information session and five office visits (including a baseline assessment session) in the 3 weeks prior to their quit date. Participants attended another eight office visits over 8 weeks following the quit date and then completed monthly follow-up phone calls, with office visits for biochemical verification of abstinence claims at 6 and 12 months post-quit. In addition to attending visits, participants carried electronic diaries (EDs) for 2 weeks preceding and 4 weeks following the target quit date. Participants were instructed to complete brief (2- to 3-minute) reports in response to prompts at wake-up, three to five randomly selected times throughout the day, and bedtime. Participants were regular smokers recruited via mass media who reported smoking at least 10 cigarettes per day and whose expired carbon monoxide (CO) level exceeded 9 parts per million at baseline. Potential participants were screened for serious psychopathology (bipolar disorder or psychosis), contraindications to use of bupropion SR (e.g., uncontrolled hypertension, history of seizure disorder, history of eating disorder,
Chapter 06.qxd
150
3/28/2007
6:25 PM
Page 150
MCCARTHY, BOLT, BAKER
current heavy drinking), and current depression. Four hundred and sixty-three participants passed all screening, enrolled in the study, and attended the first study visit. Participants provided ratings of affect, withdrawal symptoms, and smoking behavior each evening before bedtime. Smoking status was assessed at in-person visits via self-report and confirmed by CO testing at each visit. Mediational analyses were conducted using data from the 297 (64.1%) participants who did not relapse in the first week of the quit attempt. Relapse was defined as reporting smoking on three consecutive evening ED reports in the week beginning with the quit day. Participants who relapsed during the first week were excluded from mediational analyses to permit estimation of the mediator untainted by heavy and consistent smoking in the post-quit period. Candidate mediators were assessed each evening for 4 weeks following the quit day, via the ED. Specific withdrawal symptoms and affect ratings were collected via the ED nightly. For purposes of illustration, we focus on one candidate mediator: craving level at the outset of the quit attempt. Craving scores represent the average of two items derived from the Wisconsin Smoking Withdrawal Scale (Welsch et al., 1999) rated “on average since the last evening report” on an 11-point scale ranging from “No!!” to “Yes!!.” The items were: “Bothered by the desire to smoke” and “Urge(s) to smoke.” We examined whether the estimated level of craving on the quit day (i.e., the estimated quit day intercept) or the change in craving over the first week of the quit attempt mediated the effects of combined treatment versus all other treatments on point-prevalence abstinence 6 weeks post-quit. We tested this hypothesis using two approaches to mediational analyses. First, we present results from a simple regression approach. Second, we present contrasting structural equation models that were used to estimate the critical paths in the mediational model and test the significance of the mediated effect as recommended by MacKinnon and colleagues (MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002). We then present data from the SEM approach in the stage-sequential format to facilitate interpretation of the clinical significance of statistically significant effects. The mediator was assessed during the first week of the quit attempt, when none of the 297 participants eligible for these analyses had relapsed. Because we excluded all participants who relapsed in the first week of the quit attempt, we can be confident that estimates of craving severity closest to the quit day were not influenced by relapse. The
Chapter 06.qxd
3/28/2007
6:25 PM
Page 151
6. THE IMPORTANCE OF HOW
151
outcome in these analyses was biochemically verified 7-day point-prevalence abstinence 6 weeks post-quit (i.e., after the period of mediator assessment concluded). As such, we can have confidence that the mediator and outcome did not overlap in time. Treatment was ongoing during the mediator assessment period, however. Six-week abstinence rates among the 297 people who did not relapse in the first week were as follows: 30% in the placebo condition, 31% in the counseling-only condition, 44% in the medication-only condition, and 51% in the combination treatment condition. Receiving both bupropion SR and counseling was associated with a significant increase in the likelihood of abstinence χ2(1, N = 297) = 6.38, p < .02), relative to all other treatment conditions. Thus, the data meet one criterion for mediation: the treatment is related to the proximal (6-week) outcome, as one would expect. In the analyses reported next, we focus on the contrast between the combination treatment and the other three study conditions, captured using a single dummy-coded variable (0 = single or placebo treatment, 1 = combination treatment). We contrasted the combination treatment with the single-treatment conditions and the control condition because inspection of raw data suggested that the combination condition had unique relations with candidate mediators, whereas the single-treatment conditions were similar to the placebo condition. Simple Regression Analyses In the first set of mediational analyses, separate regression models were constructed to test Paths a, b, c, and c′, as depicted in Figure 6.1. The candidate mediator tested using regression was a difference score between the craving summary score on the seventh day of the quit attempt minus the craving summary score from the quit day, divided by 7 days. Only 243 participants who maintained abstinence for at least 1 week provided ratings on both Days 1 and 7 and were included in these analyses. Other researchers have used simple difference scores as candidate mediators in formal mediational analyses (Lerman et al., 2002). Path a linking treatment and the mediator was tested using linear regression. The combined treatment condition was not associated with craving difference scores. Thus, Path a did not reach significance in the regression modeling approach. Figure 6.1 depicts the significant standardized regression coefficient linking the combination condition versus all other treatments with the craving difference score.
Chapter 06.qxd
3/28/2007
6:25 PM
Page 152
152
MCCARTHY, BOLT, BAKER
Path b linking the difference score and 6-week relapse was tested using logistic regression, given the dichotomous outcome (0 = abstinent, 1 = smoking). Treatment was included as a control variable in this regression model. Logistic regression coefficients were standardized to permit estimation of the magnitude of the reduction in the direct effect of treatment on smoking outcome that occurred when controlling for the mediator in the regression analyses. The standardized logistic regression coefficient shown in Figure 6.1 was derived using the formula: betayx = (byxSx × R)/ slogity (Menard, 1995). Path b was significant (Unstandardized B = 1.02, SE = .37, Wald = 7.73, p < .005), suggesting that increased craving from the quit Day to 1 week post-quit was associated with higher risk of relapse between Weeks 2 and 6 post-quit. Path c is significant, as reported earlier (see Fig. 6.1 for the standardized coefficient for the combination treatment effect on relapse). Path c′ appears to be slightly reduced in magnitude (from -.17 to -.15, a 12% reduction) when the mediator is included in the logistic regression model. Given the lack of relationship between treatment and craving difference scores in these analyses, however, this reduction in the direct effect cannot be interpreted as evidence of mediation. In summary, regression analyses failed to establish one of the core criteria for mediation: evidence of a relationship between treatment and the mediator. Failure to find a treatment–mediator association in this analysis may reflect the crude nature of the mediator used here (a difference score) and the influence of error in the measurement of the mediator. ^
Structural Equation Modeling (SEM) In SEM, mediation is studied through effects represented statistically as latent variables. In the current analysis, latent variables represented features of change in craving scores for each evening of the first week following the quit date. The relationship between the predictor (combined treatment vs. all other treatment conditions) and criterion (relapse) was first assessed through a probit regression model. The resulting regression estimate of -0.39 (SE = 0.162, t = -2.43) was statistically significant (although modest in magnitude, standardized beta = -.18), suggesting a lower likelihood of relapse in the combined treatment condition. The mediational model tested attempted to explain the effects of the combined treatment on relapse through changes in craving following the quit attempt. The model thus considered up to seven post-quit craving scores, in addition to the combined treatment and relapse variables.
Chapter 06.qxd
3/28/2007
6:25 PM
Page 153
6. THE IMPORTANCE OF HOW
153
FIGURE 6.2 This figure depicts the structural equation model fit to the data. Loadings of craving scores collected nightly on Days 1 through 7 of the quit attempt were fixed at 1.0 for the quit-day intercept latent variable (crav_int). Loadings of craving Scores 1 and 7 were fixed at 0 and 6, respectively, for the latent slope variable (crav_slp). Intervening craving scores’ loadings were not fixed to allow for nonlinear change in craving over the first week post-quit. Craving scores were allowed to have correlated residuals to account for autocorrelation across repeated measures. Treatment represents the contrast between combined counseling and bupropion SR treatment versus all other treatment conditions. The outcome variable is relapse, a dichotomous variable indicating that a person did not achieve CO-confirmed 7-day point-prevalence abstinence 6 weeks post-quit. Residual correlations connect all observations, although only lag-1 autocorrelations were estimated.
Preliminary models applied only to the craving scores suggested highly nonlinear changes in craving scores during the first week post-quit. Moreover, even after accounting for individual variations in craving change, the residuals for craving scores collected closer in time tended to correlate positively, consistent with the presence of an autocorrelation structure. Thus, the SEM model used to represent the mediating effects of craving allowed for both nonlinear change (through use of estimated time scores among measures collected from Day 1 to Day 6 post-quit; Muthen & Muthen, 1998–2004, p. 83) and a first-order correlation structure among the craving score residuals (Muthen & Muthen, 1998–2004). The latent variables, denoted in Figure 6.2 as craving intercept and craving slope, represent the quit-day level of craving and the average daily change in craving during the first week after quitting, respectively.
Chapter 06.qxd
154
3/28/2007
6:25 PM
Page 154
MCCARTHY, BOLT, BAKER
Both latent variables were considered potential mediators of the effects of the combined treatment on smoking relapse. The model in Figure 6.2 was fitted using Mplus (Muthen & Muthen, 1998–2004). Due to the presence of missing data (some respondents had missing reports) and the use of categorical outcome variables, the weighted least squares with mean- and variance-adjustment (WMLSV) estimation method (the default in Mplus) was used. Based on standard goodness-of-fit criteria, the model provided a close approximation to the data (χ2 = 18.44, df = 12, p = .08; RMSEA = .043; CFI = .94; TLI = .98; WRMR = .450). Table 6.1 displays the model estimates associated with measurement of the latent mediators. The mean craving intercept and slope estimates represent the average growth trajectory across all smokers, and imply an average craving score of 7.66 on the quit day, and an average daily decline in craving of 0.21. The estimated time scores associated with the daily craving measures (displayed as the craving slope estimates) indicated the pattern of craving changed over the first week. For example, assuming 6 days of change from Day 1 to Day 7 (fixed loadings of 0 and 6 anchor the first and last measurements), it appears that the most substantial change on average occurred from Day 1 to Day 2 (2.57/6.00 = 43%), with slower rates of change in subsequent days. To evaluate the mediational effects of the craving intercept and slope, we examined the estimates reported in Table 6.2. For each parameter, the raw estimate, standard error, and standard estimates are reported, along with the ratio of the raw estimate to standard error, which can be interpreted as approximate z statistics. For the craving intercept, neither the path from combined treatment (aCINT) nor the path to relapse (bCINT) was statistically significant. By contrast, the corresponding paths for the craving slope were more substantial, with statistical significance attained for the path from craving slope to relapse (bCSLP) and a marginally significant estimate for the path from treatment to craving slope (aCSLP). This suggests a greater mediating role for craving slope. Specifically, the combined treatment leads to a greater decline in craving scores over the first week post-quitt, and this greater decline, in turn, results in a lower likelihood of relapse. The direct effect of the combined treatment on relapse also declined in magnitude to -.26 in the mediational model, an effect that was no longer statistically significant. To test the significance of the indirect effects associated with the two hypothesized mediating variables, we tested the two product coefficients,
Chapter 06.qxd
3/28/2007
6:25 PM
Page 155
155
6. THE IMPORTANCE OF HOW
TABLE 6.1 Mediational Model Estimates, Craving Measurement Parameters
Estimate Craving Intercept: Days 1-7 Craving Slope: Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Day 7 Residual Variance: Days 1-7 Adjacent Residual Covariance: Days 1-7 Mean Craving Intercept Mean Craving Slope Residual Variance, Craving Intercept Residual Variance, Craving Slope Residual Covariance, Craving Intercept and Slope
Std. Error Est./Std Error
1.00
…
…
0.00 2.57 3.44 4.02 5.18 5.90 6.00
… .40 .46 .51 .54 .49 …
… 6.43 7.46 7.91 9.68 12.14 …
2.05
.45
4.61
.64 7.66 −.21 3.10 .07 .00
.38 .18 .03 .69 .02 .06
1.71 43.48 −6.40 4.52 4.03 .01
aCINT x bCINT and aCSLP x bCSLP. For the craving intercept, the indirect effect estimate of -.001 had an estimated standard error of .017 using the delta method, whereas the craving slope indirect effect estimate of -.135 had an estimated standard error of .087. To evaluate the significance of the indirect effects, we compared the ratio of the indirect effect estimate to its standard error against z′ tables developed by MacKinnon (MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002). The ratio for craving slope = -.135/.087 = -1.56 exceeded in magnitude the critical value associated with α = .05 (approximately -1.00), suggesting a significant indirect effect through craving slope. By contrast, the ratio for the craving intercept = -.001/.017 = -.053 was not significant. In summary, SEM fit to the data supported the hypothesis that craving patterns over time mediate treatment effects on relapse in a tobacco quit attempt. Specifically, the combination of bupropion SR treatment and counseling was associated with steeper declines in craving ratings over the first week of the quit attempt, which were associated with reduced risk of relapse. When this indirect effect was taken into account, the
Chapter 06.qxd
3/28/2007
6:25 PM
Page 156
156
MCCARTHY, BOLT, BAKER TABLE 6.2 Mediational Model Estimates, Structural Parameters
Estimate
Std. Error
Est./Std. Error
Std. Estimate
Craving Intercept on Treatment (aCINT)
–.02
.31
–.05
–.00
Craving Slope on Treatment (aCSLP)
–.11
.06
–1.93
–.18
Relapse on Craving Intercept (bCINT)
.05
.06
.96
.09
Relapse on Craving Slope (bCSLP)
1.26
.44
2.89
.33
Relapse on Treatment (c)
–.26
.17
–1.53
–.12
direct effect of combination treatment on relapse was nonsignificant and nonessential to model fit. The estimated mediated effect was statistically significant for craving slope as well. As such, these data suggest that combination treatments increase early success rates by helping reduce cravings from peak levels quickly during the first week of a quit attempt. Stage Sequential Approach To illustrate the potential clinical significance of the significant mediated effect detected in SEM analyses, we performed a median split on factor scores derived from the SEM model and inspected the proportions of participants who experienced each possible stage sequence in the sample. These data are presented in Figure 6.3. The mediator is the factor score for the slope in craving over 1 week post-quit for each of the 293 individuals with enough data to be included in these analyses. A median split was performed on these factor scores. Scores below the median of -.23 indicated a rapid decline in craving over the first week post-quit. Scores above the median slope indicated persistent, worsening, or slow-to-resolve cravings. Presenting the data in the stage sequential framework suggests that the effect of combined treatment on standing on the mediator (estimated slope in craving over 1 week post-quit above vs. below the median) was substantial (a 20% difference across conditions) and of potentially great clinical significance. The 28% to 30% increase in the probability of abstinence associated with steeper declines in craving versus stable or increasing craving trends within each treatment group was also substantial and of likely clinical significance. Overall, these data suggest that the combined treatment was
Chapter 06.qxd
3/28/2007
6:25 PM
Page 157
6. THE IMPORTANCE OF HOW
157
FIGURE 6.3 Results of SEM analyses as stage-sequential processes. Path coefficients represent the probability that a person will enter each category. Those in the combined treatment condition were more likely to have steeper reductions in craving over the first week post-quitt (i.e., their factor scores fell below the median for changes in craving) than were those in the single treatment or placebo conditions. In both treatment groups, a more rapid decline in craving was associated with an increased probability of abstinence at 6 weeks post-quit, relative to those whose craving slope factor scores were above the median.
associated with an increase in the probability of the desired sequence of rapidly declining craving and abstinence occurring from 23% (in the single or no treatment groups) to 40%. This near doubling in the rate of rapidly reduced craving followed by short-term success in quitting suggests that the combined treatment may offer substantial clinical benefit, in addition to being statistically significant in SEM analyses.
Chapter 06.qxd
3/28/2007
6:25 PM
Page 158
158
MCCARTHY, BOLT, BAKER
CONCLUSIONS In the first part of this chapter, we presented diverse rationales for conducting mediational analyses and proposed diverse means of doing so. In the second part of the chapter, we provided examples of the divergent results that may emerge from application of different analytic methods to the same mediational hypothesis. Although we cannot be sure which pattern of results best approximates the true state of nature, we advocate for the latent variable approach that yielded positive results in this sample. The use of difference scores for the regression analyses was admittedly crude, particularly in comparison to the nonlinear growth in craving modeled in the SEM analyses. Even if we had aggregated several reports of craving, rather than using a difference score, measurement error would still have influenced results of our mediational tests. Only a latent variable approach can perform the critical step of removing error from the measurement of the mediator in social science research. In addition, SEM offers flexibility in the construction of the measurement model, permitting nonlinearity and addressing autocorrelation among repeated measures, as illustrated earlier. Performing a median split on factor scores on the latent craving slope construct allowed us to capture the potential clinical significance of the effects detected in the SEM analyses. Such presentation aids in the interpretation of mediational effects and their likely magnitude in populations of interest. In the illustrative analyses presented here, we did not adhere to all of the best practices guidelines for mediation. For example, we did not include in the SEM analyses control variables that likely influence standing on the mediator, outcome, or both. Removing variance that is unrelated to predictors of interest will increase the sensitivity of the target tests. Tobacco dependence level and cohabitation with a smoker are but two possible influences on craving and relapse. We also neglected to include potential moderators of the hypothesized mediational pathway, such as gender. A more thorough test of mediational hypotheses would include control variables and examine moderating effects. In addition, we permitted temporal overlap between treatment and the mediator in the analyses presented earlier, without controlling for pretreatment mediator levels in an effort to avoid excessive complexity in these illustrative analyses. Ideally, one could temporally divorce treatment and the mediator. In ongoing treatments such as bupropion SR therapy and multisession individual counseling, such temporal ordering can be difficult. In such cases, statistical control of pretreatment mediator levels is highly recommended.
Chapter 06.qxd
3/28/2007
6:25 PM
Page 159
6. THE IMPORTANCE OF HOW
159
In earnest mediational analyses, we might also elect to examine a continuous rather than dichotomous outcome to increase the sensitivity of tests of treatment and mediator effects. Rather than focusing on abstinence versus relapse, we might have focused on smoking heaviness, smoking trajectory, or latency to lapse or relapse. Although these continuous outcomes are of less vital public health importance, they may be related to treatment or mediators in such a way that suggests new action or conceptual models, and thus, treatment innovation. Each of the methods used in our sample analyses has associated costs and benefits. We have noted some of the potential limitations of the choices we made in our illustrative analyses, as well. We do not wish to suggest that the method we used is necessarily the best method. Instead, we hope that the preceding discussion and examples have highlighted some of the potential benefits of tackling mediation in research and the hidden costs of common analytic choices. Two other observations are warranted in closing. First, the results presented here are of substantive importance. Our analyses suggest that cigarette craving early in a quit attempt is an important influence on quitting success. Additionally, SEM analyses suggest that the evaluated treatments work, in part, by suppressing craving. This finding is important, if replicated, because it underscores the importance of craving, and it suggests that treatments may be improved by enhancing their ability to reduce craving. Second, the results are important because they constitute suggestive evidence that the field of tobacco research may now be in the position to heed some of McFall’s earlier counsel and explore the mechanisms by which treatments exert their effects. The availability of ecological momentary assessment data and powerful statistical techniques may now permit us to fulfill McFall’s vision for the field. REFERENCES Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Chen, H. T. (1990). Theory-driven evaluations. Newbury Park, CA: Sage. Cinciripini, P. M., Cinciripini, L. G., Wallfisch, A., Haque, W., & Van Vunakis, H. (1996). Behavior therapy and the transdermal nicotine patch effects on cessation outcome, affect and coping. Journal of Consulting and Clinical Psychology, 64, 314–323. Cinciripini, P. R., Wetter, D. W., Fouladi, R. T., Blalock, J. A., Carter, B. L., Cinciripini, L. G., & Baile, W. F. (2003). The effects of depressed mood on smoking
Chapter 06.qxd
160
3/28/2007
6:25 PM
Page 160
MCCARTHY, BOLT, BAKER
cessation: Mediation by postcessation self-efficacy. Journal of Consulting and Clinical Psychology, 71, 292–301. Cole, D. A., & Maxwell, S. E. (2003). Testing mediational models with longitudinal data: Questions and tips in the use of structural equation modeling. Journal of Abnormal Psychology, 112, 558–577. Collins, L. M., Graham, J. W., & Flaherty, B. P. (1998). An alternative framework for defining mediation. Multivariate Behavioral Research, 33, 295–312. Curtin, J. J., McCarthy, D. E., Piper, M. E., & Baker, T. B. (2006). Implicit and explicit drug motivational processes: A model of boundary conditions. In R. Weirs & A. W. Stacy (Eds.), Handbook of Implicit Cognition and Addiction. Thousand Oaks, CA: Sage. DeRubeis, R. J., & Feeley, M. (1990). Determinants of change in cognitive therapy for depression. Cognitive Therapy and Research, 14, 469–482. Eddy, J. M., Dishion, T. J., & Stoolmiller, M. (1998). The analysis of intervention change in children and families: Methodological and conceptual issues embedded in intervention studies. Journal of Abnormal Child Psychology, 26, 53–69. Feeley, M., DeRubeis, R. J., & Gelfand, L. A. (1999). The temporal relation of adherence and alliance to symptom change in cognitive therapy for depression. Journal of Consulting and Clinical Psychology, 67, 578–582. Fiore, M. C., Bailey, W. C., Cohen, S. J., et al. (2000). Treating tobacco dependence and use. Clinical practice guideline. Rockville, MD: U.S. Department of Health and Human Services, Public Health Service. Goldfried, M.R., Greenberg, L. S., & Marmar, C. (1990). Individual psychotherapy: Process and outcome. Annual Review of Psychology, 41, 659–688. Hall, S. M., Humfleet, G .L., Reus, V. I., Munoz, R. F., Hartz, D. T., & Maude-Griffin, R. (2002). Psychological intervention and antidepressant treatment in smoking cessation. Archives of General Psychiatry, 59, 929–937. Haynes, S. N. (1992). Models of causality in psychopathology: Toward dynamic, synthetic, and nonlinear models of behavior disorders. Needham Heights, MA: Allyn & Bacon. Herting, J. R. (2002). Evaluating and rejecting true mediation models: A cautionary note. Prevention Science, 3, 285–289. Holland, P. W. (1988). Causal inference, path analysis, and recursive structural equation models. Sociological Methodology, 18, 449–484. Holmbeck, G. (1997). Toward terminological, conceptual, and statistical clarity in the study of mediators and moderators: Examples from child-clinical and pediatric psychology literatures. Journal of Consulting and Clinical Psychology, 65, 599–610. Howard, K. I., Lueger, R. J., Maling, M. S., & Martinovich, Z. (1993). A phase model of psychotherapy outcome: Causal mediation of change. Journal of Consulting and Clinical Psychology, 61, 678–685. Hoyle, R. G., & Smith, G. T. (1994). Formulating clinical research hypotheses as structural equation models: A conceptual overview. Journal of Consulting and Clinical Psychology, 62, 429–440. Judd, C. M., & Kenny, D. A. (1981). Process analysis: Estimating mediation in treatment evaluations. Evaluation Review, 5, 602–619.
Chapter 06.qxd
3/28/2007
6:25 PM
Page 161
6. THE IMPORTANCE OF HOW
161
Kazdin, A. E., & Kendall, K. C. (1998). Current progress and future plans for developing effective treatments: Comments and perspectives. Journal of Clinical Child Psychology, 27, 217–226. Kazdin, A. E. (1999). Current (lack of ) status of theory in child and adolescent therapy research. Journal of Clinical Child Psychology, 28, 533–543. Kazdin, A. E. (2000). Understanding change: From description to explanation in child and adolescent psychotherapy research. Journal of School Psychology, 38, 337–347. Kazdin, A. E. (2001). Bridging the enormous gaps of theory with therapy research and practice. Journal of Clinical Child Psychology, 30, 59–66. Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in social psychology. In D. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), Handbook of social psychology (Vol. 1, 4th ed., pp. 233–265). New York: McGraw-Hill. Lerman, C., Roth, D., Kaufmann, V., Audrain, J., Hawk, L., Liu, W., Niaura, R., & Epstein, L. (2002). Mediating mechanisms for the impact of bupropion in smoking cessation treatment. Drug and Alcohol Dependence, 67, 219–223. Lipsey, M. W. (1993). Theory as method: Small theories of treatments. In L. B. Sechrest & A. G. Scot (Eds.), New directions for program evaluation (NIH Research Monograph No. 142, pp. 140–183). Rockville, MD: National Institute on Drug Abuse. MacCorquodale, K., & Meehl, P. E. (1948). On a distinction between hypothetical constructs and intervening variables. Psychological Review, 55, 95–107. MacKinnon, D. P. (2000). Contrasts in multiple mediator models. In J. S. Rose, L. Chassin, C. C. Presson, & S. J. Sherman (Eds.), Multivariate applications in substance use research: New methods for new questions (pp. 141–160). Mahwah, NJ: Lawrence Erlbaum Associates. MacKinnon, D. P. (1994). Analysis of mediating variables in prevention and intervention studies. In A. Cázares & L. Beatty (Eds.), Scientific methods for prevention intervention research (NIDA Monograph No. 139, pp. 127–153). Rockville, MD: National Institute on Drug Abuse. MacKinnon, D. P., & Lockwood, C. M. (2003). Advances in statistical methods for substance abuse prevention research. Prevention Science, 4, 155–171. MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002). A comparison of methods to test mediation and other intervening variable effects. Psychological Methods, 7, 83–104. MacKinnon, D. P., Taborga, M. P., & Morgan-Lopez, A. A. (2002). Mediation designs for tobacco prevention research. Drug and Alcohol Dependence, 68, S69–S83. Marston, A. R., & McFall, R. M. (1971). Comparison of behavior modification approaches to smoking reduction. Journal of Consulting and Clinical Psychology, 36, 153–162. McCarthy, D. E., Bolt, D. M., Lawrence, D. L., Piasecki, T. M., Fiore, M. C., & Baker, T. B. (in preparation). A randomized controlled clinical trial of bupropion SR and individual smoking cessation counseling. Manuscript in preparation. McCarthy, D. E., Piasecki, T. M., Fiore, M. C., & Baker, T. B. (2006). Life before and after quitting smoking: An electronic diary study. Journal of Abnormal Psychology, 115, 454–466. McFall, R. M. (1970). Effects of self-monitoring on normal smoking behavior. Journal of Consulting and Clinical Psychology, 35, 135–142.
Chapter 06.qxd
162
3/28/2007
6:25 PM
Page 162
MCCARTHY, BOLT, BAKER
McFall, R. M. (1978). Smoking cessation research. Journal of Consulting and Clinical Psychology, 46, 705–712. McFall, R. M., & Hammen, C. L. (1971). Motivation, structure, and self-monitoring: Role of nonspecific factors in smoking reduction. Journal of Consulting and Clinical Psychology, 37, 80–86. McFall, R. M., & Lillesand, D. B. (1971). Behavior rehearsal with modeling and coaching in assertion training. Journal of Abnormal Psychology, 77, 313–323. Meehl, P. E. (1978). Theoretical risks and tabular asterisks: Sir Karl, Sir Ronald, and the slow progress of soft psychology. Journal of Consulting and Clinical Psychology, 46, 806–834. Menard, S. (1995). Applied logistic regression analysis (Sage University Paper series on Quantitative Application in the Social Sciences, 07-106). Thousand Oaks, CA: Sage. Morgenstern, J., & Longabaugh, R. (2000). Cognitive-behavioral treatment for alcohol dependence: A review of evidence for its hypothesized mechanisms of action. Addiction, 95, 1475–1490. Muthen, L. K., & Muthen, B. O. (1998–2004). Mplus user’s guide (3rd ed.). Los Angeles, CA: Muthen & Muthen. Piasecki, T. M., Fiore, M. C., McCarthy, D. E., & Baker, T. B. (2002). Have we lost our way? The need for dynamic formulations of smoking relapse proneness. Addiction, 97, 1093–1108. Richmond, R., & Zwar, N. (2003). Review of bupropion for smoking cessation. Drug and Alcohol Review, 22, 203–220. Rozeboom, W. W. (1956). Mediation variables in scientific theory. Psychological Review, 63, 249–264. Rubin, D. B. (1974). Estimating causal effects of treatments in randomized and nonrandomized studies. Journal of Educational Psychology, 66, 688–701. Shadish, W. R., & Sweeney, R. B. (1991). Mediators and moderators in metaanalysis: There’s a reason we don’t let dodo birds tell us which psychotherapies should have prizes. Journal of Consulting and Clinical Psychology, 59, 883–893. Shapiro, D., Jamner, L. D., Davydov, D. M., & James, P. (2002). Situations and moods associated with smoking in everyday life. Psychology of Addictive Behaviors, 16, 342–345. Shiffman, S., Engberg, J. B., Paty, J. A., Perz, W. G., Gnys, M., Kassel, J. D., & Hickcox, M. (1997). A day at a time: Predicting smoking lapse from daily urge. Journal of Abnormal Psychology, 106, 104–116. Shiffman, S., Gwaltney, C. J., Balabanis, M. H., Liu, K. S., Paty, J. A., Kassel, J. D., Hickcox, M., & Gnys, M. (2002). Immediate antecedents of cigarettes smoking: An analysis from ecological momentary assessment. Journal of Abnormal Psychology, 111, 531–545. Shiffman, S., Paty, J. A., Gwaltney, C. J., & Dang, Q. (2004). Immediate antecedents of cigarette smoking: An analysis of unrestricted smoking patterns. Journal of Abnormal Psychology, 113, 166–171. Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: New procedures and recommendations. Psychological Methods, 7, 422–445. Smith, S. S., Jorenby, D. E., Leischow, S. J., Nides, M. A., Rennard, S. I., Johnston, J. A., Jamerson, B., Fiore, M. C., & Baker, T. B. (2003). Targeting smokers at
Chapter 06.qxd
3/28/2007
6:25 PM
Page 163
6. THE IMPORTANCE OF HOW
163
increased risk for relapse: Treating women and those with a history of depression. Nicotine and Tobacco Research, 5, 99–109. Tang, T. Z., & DeRubeis, R. J. (1999). Sudden gains and critical sessions in cognitivebehavioral therapy for depression. Journal of Consulting and Clinical Psychology, 67, 894–904. Tiffany, S. (1990). A cognitive model of drug urges and drug-use behavior: Role of automatic and nonautomatic processes. Psychological Review, 97, 147–168. Weersing, V. R., & Weisz, J. R. (2002). Mechanisms of action in youth psychotherapy. Journal of Child Psychology and Psychiatry, 43, 3–29. Welsch, S. K., Smith, S. S., Wetter, D. W., Jorenby, D. E., Fiore, M. C., & Baker, T. B (1999). Development and validation of the Wisconsin Smoking Withdrawal Scale. Experimental and Clinical Psychopharmacology, 4, 354–361. West, S. G., & Aiken, L. S. (1997). Toward understanding individual effects in multicomponent prevention programs: Design and analysis strategies. In K. J. Bryant, M. Windle, & S. G. West (Eds.), The science of prevention: Methodological advances from alcohol and substance abuse research (pp.167–210). Washington, DC: American Psychological Association Wiggins, J. S. (1973). Personality and prediction: Principles of personality assessment. Reading, MA: Addison-Wesley.
Chapter 06.qxd
3/28/2007
6:25 PM
Page 164
Chapter 07.qxd
3/28/2007
12:59 PM
Page 165
III MODEL CLINICAL SCIENCE RESEARCH PROGRAMS
Chapter 07.qxd
3/28/2007
12:59 PM
Page 166
Chapter 07.qxd
3/28/2007
12:59 PM
Page 167
7 Fear, Anxiety, Depression, and the Anxiety Disorder Spectrum: A Psychophysiological Analysis Peter J. Lang Lisa M. McTeague University of Florida
Bruce N. Cuthbert University of Minnesota
Dick McFall and I (PJL) were colleagues together for many years at the University of Wisconsin. We were both engaged in training students in clinical science. Our shared aim was to produce a cadre: to train research-oriented clinical psychologists who would advance understanding of psychopathology, who would develop research based treatments and assessment methods, and who would then in turn become teachers of the next generation of clinical investigators. In this endeavor, we were fortunate to be at a university with a superb faculty in basic experimental psychology—learning, psychophysiology, perception, statistical analysis. Thus, we were confident that when our students began experimental work, they had access to and consultant support in the substantive science that necessarily underlies any clinical science effort. In addition, however, we gave students specific instructions: When you address a clinical research problem, consider first its basic science and second, be sure that your experiment builds on this base.
167
Chapter 07.qxd
168
3/28/2007
12:59 PM
Page 168
LANG, MCTEAGUE, CUTHBERT
Increasingly, in his later years at Indiana University, Dick has been a beacon illuminating this fundamental concept. He has rightly railed against the clinical originals who invent parochial concepts and methods for clinical investigation, ignoring the well-established findings and procedures of the relevant basic science (see e.g., McFall, Treat, & Viken, 1997). More important, in his own work, Dick has built his clinical studies on the foundation of current cognitive science, often collaborating with colleagues in from this basic discipline, producing ideal examples of what is now called translational research (see chap. 12). In what follows, we describe research that might also be labeled translational— honoring Dick in presenting work guided by the principles for which he so nobly battles. Thus, we present here a view of diagnostic differences among the anxiety disorders from the perspectives of human psychophysiology and neuroscience research on the brain’s reflex circuits. The clinical material is first considered in terms of contemporary clinical interview and questionnaire methodologies. Subsequently, we reevaluate these patients, based on an analysis of their reflex physiology during threat, analyzing the covariation between clinical evaluation and a measure of fear based on research in basic neuroscience.
This chapter considers the concept of negative affect as it can be measured physiologically, and can serve as a unifying dimension across the anxiety disorder spectrum. It begins with an analysis of self-report data from nearly 300 anxiety patients and controls, considering three pertinent trait dimensions of psychopathology—fearfulness, anxiety, and depression—as they vary with principal diagnosis. Using a subsample of these participants, the relationship is assessed between status on these dimensions and a psychophysiological measure, potentiation of the probe startle reflex in the context of fearful imagery. The interview and questionnaire results suggest a hierarchy of disorders defined by increasing comorbid pathology and selfreport of negative affect. The psychophysiological findings show that, paradoxically, this ascending hierarchy of reported distress is directly associated with a reciprocal diminution in defensive reactivity to fear challenge. DIMENSIONS OF PATHOLOGICAL ANXIETY Studies of comorbidity have consistently found increasing psychopathology along a spectrum defined by principal diagnoses, ranging from least to most as follows: specific phobia → social phobia → panic disorder with agoraphobia (PDA) → generalized anxiety disorder (GAD). For example,
Chapter 07.qxd
3/28/2007
12:59 PM
Page 169
7. PSYCHOPHYSIOLOGY OF THE ANXIETY SPECTRUM
169
in a sample of 968 patients grouped according to principal anxiety disorder, Brown, Campbell, Lehmman, Grisham, and Mancil (2001) found that PDA (47%) and GAD (52%) showed the highest incidence of additional anxiety disorders; in contrast, specific phobia (27%) and social phobia (26%) showed significantly reduced risk. Rates for comorbid depression differed similarly across the anxiety spectrum: A higher incidence of mood disorder was found for PDA (33%) and GAD (36%; also, for a small sample with posttraumatic stress disorder, PTSD) than for specific phobia (10%)—with social phobia falling in between.
Self-Report of Symptoms. Questionnaire reports of symptoms show a similar spectrum pattern. Cuthbert, Lang, Strauss, Drobes, Patrick, and Bradley (2003) recently studied a sample of anxiety patients (N = 95: specific phobia, n = 28; social phobia, n = 30; PTSD, n = 22; PDA, n = 26) at the University of Florida Fear and Anxiety Disorders Clinic, assessing measures of anxiety, depression and fear: the Manifest Anxiety Questionnaire (Fenz & Epstein, 1965), Beck Depression Inventory (BDI; Beck, Ward, Medelsohn, Mock, & Erbaugh, 1961), a self-rating scale of phobic symptoms (Marks & Matthews, 1979), Fear Survey Schedule (FSS; Wolpe & Lang, 1964), and a temperament scale (EASI; Buss & Plomin, 1975). Multivariate analysis showed that controls and specific phobics did not differ; however, this phobic group endorsed significantly fewer symptoms than patients with PDA and PTSD. Social phobics again fell in the middle, showing more symptomatology than controls and specific phobics, but less than patients with PDA and PTSD. This same pattern emerged when the questionnaires were analyzed individually. Negative Affect. Some theorists view the dimensional pattern of self-report and interview-based comorbidity as representing a single diathesis factor that underlies both anxiety and depression (e.g., Krueger, 1999; Krueger & Finger, 2001; Mineka, Watson, & Clark, 1998). However, there is no consensus concerning its relationship to specific DSM–IV diagnostic categories (e.g., Brown, Chorpita, & Barlow, 1998; Clark & Watson, 1991; Krueger & Finger, 2001; Mineka et al., 1998). Clark and Watson (1991) proposed a tripartite model specifying overlapping and nonoverlapping symptoms of mood and anxiety that could be used to explore spectrum differences. The model combines nonspecific symptoms of distress common to both anxiety (e.g., nervousness) and
Chapter 07.qxd
170
3/28/2007
12:59 PM
Page 170
LANG, MCTEAGUE, CUTHBERT
depression (e.g., helplessness) in a general negative affect factor, including also cognitive (e.g., concentration difficulties) and somatic (e.g., insomnia) concerns (Watson, Weber, Assenheimer, Clark, Strauss, & McCormick, 1995). The authors proposed, furthermore, that depression and anxiety could be differentiated by two separate factors, involving unique features of the respective DSM diagnoses. Anhedonia or low positive affect (i.e., paucity of pleasurable emotional experiences) would index depression, whereas reports of chronic physiological arousal were more specific to anxiety. Watson and Clark (1991) developed a multiscale measure based on their model, the Mood and Anxiety Symptom Questionnaire (MASQ). The MASQ has five subscales designed to represent different symptom structures and hence, a range of discriminant validity. These scales were recently evaluated in a new sample of 273 participants seen at the University of Florida Fear and Anxiety Disorders Clinic: controls, n = 75, principal anxiety disorder, N = 198 (specific phobics, n = 37; social phobics, n = 59; PDA, n = 50; GAD, n = 52).1 The MASQ subscales, whether emphasizing anxiety or depression, proved to be highly correlated (i.e., r = .63–.85). Furthermore, similar to the depression and anxiety questionnaires used with the previous, smaller sample, all five symptom measures increased across disorders, from specific phobia, with the lowest scores, to the highest scores for PDA and GAD at the end of this anxiety spectrum [Diagnostic Group: Wilks λ = .29, F(20,790) = 17.82, p < .0001, η2 = .27, all univariate linear trends, p < .01]. PDA and GAD patients differed on only one scale, MASQ-Anxious Arousal, with PDA showing the higher mean score (see Fig. 7.1, right panel). Figure 7.1 in the left panel shows that this pattern of increasing scores from phobic to more generally anxious disorders reappears when other frequently used questionnaires are analyzed. The traits measured included anxiety (STAI; Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983), cognitive and somatic depression (BDI; Beck, Steer, & Brown, 1996), and anxiety sensitivity (ASI; Reiss & McNally, 1985) [Diagnostic Group; Wilks λ = .30, F(16,691) = 20.92, p < .0001, η2 = .27, all univariate linear trends, p < .01]. The findings just described are mirrored in the work of other researchers. For example, Krueger and Finger (2001) used the methods of 1 A PTSD and OCD sample is also being collected, but at the time of writing they are not yet of sufficient size for inclusion here.
Chapter 07.qxd
3/28/2007
12:59 PM
Page 171
7. PSYCHOPHYSIOLOGY OF THE ANXIETY SPECTRUM
171
FIGURE 7.1 The left panel shows the consistent pattern of increasing distress from phobic to more generally anxious disorders. The questionnaires measured include trait anxiety (STAI; Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983), cognitive and somatic depression (BDI; Beck. Steer, & Brown, 1996), and anxiety sensitivity (ASI; Reiss & McNally, 1985). The right panel shows that the same pattern emerged with the responses on the five scales of the Mood and Anxiety Symptom Questionnaire (MASQ): General Distress Anxiety, General Distress Depression, General Distress Mixed Symptoms of Anxiety and Depression, Anhedonic Depression, Anxious Arousal (Watson & Clark, 1991).
item response theory to factor-map the anxiety disorders. They found that PDA and GAD were highest on an internalizing factor (including both anxiety and depressive features). This factor was also associated with greater functional impairment and more inpatient admissions. In short, evidence exists, across laboratories and investigations, in support of an anxiety disorder severity continuum that is best characterized by a progressive increment in negative affect. Is Negative Affect a Physiological Dimension? The concept of an affective dimension that underlies anxiety disorder challenges traditional distinctions between subsyndromes, and even questions the diagnostic independence of anxiety and depression. A serious weakness of this theorizing, however, lies in the narrowness of its database. The relevant findings are derived entirely from self-report and symptom assessment at interview. It does not include behavioral task measures or, perhaps more important, objective measures of expressed emotion and
Chapter 07.qxd
172
3/28/2007
12:59 PM
Page 172
LANG, MCTEAGUE, CUTHBERT
physiological patterns that might be associated with a dimensional factor. Considering the low correlations that have been repeatedly shown between reports of physiological symptoms and actual symptom expression (e.g., Buss, 1962; Lang, 1978; Mandler, Mandler, Kremen, & Sholiton, 1961), this absence is a major concern. Put simply, do items on the Anxious Arousal scale of the MASQ—such as “Muscles twitched or trembled”; “Heart racing or pounding”; “Was trembling and shaking”; “Hands cold and sweaty”; “Was short of breath”; “Muscles were tense”— have anything to do with actual physiology? More important, is negative affect just language behavior, or is there a physiological core that also varies along a severity dimension? FEAR, STARTLE, AND THE MAMMALIAN DEFENSE SYSTEM The assumptions that anxiety disordered patients are fearful and that all display a physiology of arousal define the syndrome. The veracity of this view is generally accepted without reservation. Theorists may try to discriminate fear from anxiety—for example, fear is a response to imminent threat and anxiety is apprehension about future threat; or fear is specific and anxiety is vague; or fear is a cue response and anxiety is a response to context; or anxiety is simply fear of fear—but few argue that fear is not the core of the matter. The next section of this chapter describes recent research assessing the physiology of fear reactions in anxiety patients, considering how the physiological expression of fear may vary across diagnoses, self-ratings, and questionnaires. Prior to addressing the research, however, it will be important to define terms, physiological measures, and the theoretical context of the experiment.
The Mammalian Defense System. In the present view, fear is fundamentally an action disposition—as are other reported emotions. Fear is primitively associated with escape from danger or pain, secondarily with avoidance, and presumably also with counterattack and other mechanisms of defense. Following Konorski (1967) and many other theorists, it is presumed that a defensive motivational system evolved in the mammalian brain to ensure the survival of individuals and species, and further, that activation of this system accompanies the most reliable reports of fear experience. Our understanding of the brain’s defense circuitry comes primarily from neuroscience research with animals—mainly rodents—using relatively
Chapter 07.qxd
3/28/2007
12:59 PM
Page 173
7. PSYCHOPHYSIOLOGY OF THE ANXIETY SPECTRUM
173
FIGURE 7.2 Schematic diagram showing direct connection between the central nucleus of the amygdala and a variety of hypothalamic and brain stem target areas that may be involved in different animal tests of fear and anxiety. Adapted from Davis (1992) with permission.
simple, classical conditioning procedures. In this work, a nociceptive event (e.g., electric shock) is paired with a previously innocuous light or tone over repeated trials, until a connection is formed, such that the animal displays defensive reactions whenever the light or tone appears alone. Employing various neurosurgical, pharmacological, and electrophysiological tools, the links in the neural circuit are traced in the brain, starting from the sensory system, proceeding through the necessary connecting structures, and ending with the autonomic and motoric effector outputs. This research has repeatedly highlighted a small, almond-shaped structure located deep within the temporal lobe —the amygdala—as the center of a defense system mediating the acquisition and orchestrating the expression of conditioned fear (Davis, 1992; Gloor, 1960; Gray, 1989; Kapp & Pascoe, 1986; Kapp, Pascoe, & Bixler, 1984; LeDoux, 1987; Sarter & Markowitsch, 1985). Activation of the defense circuit begins when the lateral and basolateral nuclei of the amygdala receive threat-relevant information from any sensory system. These nuclei then innervate the amygdala’s central nucleus, which in turn projects to a variety of hypothalamic sites, the central gray, facial motor nucleus, and brainstem target areas that initiate a range of defensive behaviors and autonomic reactions that evolved to counter threats to survival (cf. Davis, 1992; see Fig. 7.2). These autonomic and somatic patterns have great variety. They can be functionally organized, however, into two broad output classes: defensive immobility (i.e., freezing, fear bradycardia, and hyperattentiveness) and
Chapter 07.qxd
174
3/28/2007
12:59 PM
Page 174
LANG, MCTEAGUE, CUTHBERT
defensive action (i.e., variations in patterns of fight/flight), stages in the normal, mammalian defense response (e.g., see Blanchard & Blanchard, 1989). It is probable, furthermore, that as representations of simple lights and tones can through association with an aversive event come to activate neural defense circuits in animals, more complex networks of information that characterize human cognition also engage the same defense system. Thus, at the level of human recall and recognition, it is proposed that emotion memory networks are defined by their connections to this primitive motivational circuitry—appetitive or defensive (Lang, 1994). Viewed from this perspective, the fear state is defined by defense system activation and its reflexive autonomic and somatic output (whether driven by external threat or internal association).
The Startle Reflex and Defense. The obligatory startle reflex may be elicited by any abrupt sensory stimulus. In many species, it appears to be a primitive escape response—such as the dramatic dispersal of drowsing fish when a rock is tossed into the pond. In mammals, startling stimuli prompt a similar whole-body flexor reaction, although its protective function is less clear. Research with rodents has shown, however, that the startle reflex is connected to the defense system and that the reflex is markedly enhanced when the animal is under threat, for example, when confronted with a formerly neutral cue that through conditioning has come to signal an imminent, aversive electric shock. In brief, startle is enhanced when the organism appears to be in a state of fear. Led by Michael Davis (2000; Davis & Lang, 2003), there has been extensive study of the fear conditioning paradigm and its effect on the startle reflex. In this procedure, a startle probe (a brief, abrupt acoustic stimulus) is presented during or shortly after the fear-conditioned cue, and the amplitude of the whole body reflex is measured through a stability meter under the floor of the cage. As indicated in Figure 7.2, the increment in startle observed under these shock-threat conditions depends on the activation of the amygdala and its direct projections to the pontine center of the normal startle reflex circuit. In this sense, the startle reflex provides a metric for the assessment of defense system activation and, indirectly, the fear state. Affective Modulation of the Startle Reflex in Humans. Founded on the same neurophysiological analysis, the startle reflex has been used extensively to probe the emotional state of human beings. It
Chapter 07.qxd
3/28/2007
12:59 PM
Page 175
7. PSYCHOPHYSIOLOGY OF THE ANXIETY SPECTRUM
175
has been demonstrated, first, that reactions to the conditioned shockfear paradigm are comparable to those found with animals. That is, during extinction, startle probe responses are greater in magnitude (potentiated) when presented during fear-conditioned stimuli than during control stimuli or during the intertrial interval (e.g., Hamm, Greenwald, Bradley, & Lang, 1993). It has also been shown that probe stimuli yield potentiated reflexes when participants are only threatened with shock—with no shocks actually delivered (e.g., Bradley, Moulder, & Lang, 2005; Grillon, Ameli, Woods, Merikangas, & Davis, 1991). Furthermor, participants reliably show augmented probe reflex responses when they look at pictures of unpleasant objects or events (Bradley, Codispoti, Cuthbert, & Lang, 2001; Bradley, Codispoti, Sabatinelli, & Lang, 2001) and when they imagine fearful scenarios (Vrana & Lang, 1990; Cuthbert et al., 2003). Finally, it is clear that participants with phobia show significantly greater startle probe potentiation than nonfearful participants when looking at pictures of phobic objects (Hamm, Cuthbert, Globisch, & Vaitl, 1997; Sabatinelli, Bradley, & Lang, 2001). As would be expected, the fear stimuli that prompt potentiated probe reflexes also directly instigate changes in heart rate, increases in sweat gland activity, negative facial muscle reactions, and other measures mediated by the defense circuit (e.g., Bradley et al., 2001). Furthermore, the same mediating neural structures, directly manipulated in the animal research, are also implicated in brain imaging studies of fear stimulus processing. A recent functional magnetic resonance imaging study from this laboratory (Sabatinelli, Bradley, Fitzsimmons, & Lang, 2005) provides an instructive example. Participants viewed pictures with emotionally arousing and neutral contents, while brain activation was assessed in inferotemporal visual areas and in the amygdala. Significantly larger activations occurred at both sites for pictures judged to be high in emotional arousal. This result is consistent with primate studies reported by Amaral, Price, Pitkanen, and Carmichael (1992) showing reentrant projections from the amygdala to the visual system, presumably enhancing attention and perceptual processing of motivationally relevant events. Of particular significance for the present discourse, snake phobics showed significantly greater amygdala activation than nonfearful participants (Fig. 7.3). In effect, fearful pictures enhanced sensory processing and engaged the same amygdaloid structure that mediates defense responses—including the probe startle reflex.
Chapter 07.qxd
176
3/28/2007
12:59 PM
Page 176
LANG, MCTEAGUE, CUTHBERT
FIGURE 7.3 Participants viewed pictures of snakes, along with a variety of other image contents, during functional magnetic resonance imaging (fMRI) of the brain. Both amygdala and temporal visual cortex were significantly activated during sensory processing of emotionally arousing pictures (Left panel: Talairach z = –7, N = 18). Brain activation in these regions to snake pictures (relative to pictures of nonthreatening animals) is graphed in the right panel (means and standard error bars): Snake fearful participants (N = 9) showed significantly greater amygdala activation than nonfearful participants (N = 9; Sabatinelli et al., 2005).
FEARFUL THOUGHTS AND IMAGES: ACTIVATING THE DEFENSE SYSTEM When prompted by a verbal cue, nearly all human beings directly process the meaning of the stimulus. Furthermore, if the cue refers to an object or event, they retrieve, with similar apparent automaticity, a memorial image. Thus, in the famous example attributed to Tolstoy, he tortured his younger brother by telling him he must not think about white bears. The brother could not, of course, not have ursine thoughts. Rather, once given the verbal cue, a parade of snowy mental bears marched through his mind, to the delight of his torturer (see Wegner, Schneider, Carter, & White, 1987). Studies of physiological reactions during imagery suggest that this phenomenon is not wholly cognitive. That is, when participants are told to imagine participating in some action, there is a subovert activation of the same muscles that would be the motors of the actual behavior. Similarly, when participants hear text describing fearful situations, they react with a reflex physiology of defense—increases in heart rate, skin conductance, facial muscle action, and in magnitude of the startle response
Chapter 07.qxd
3/28/2007
12:59 PM
Page 177
7. PSYCHOPHYSIOLOGY OF THE ANXIETY SPECTRUM
177
to an acoustic probe—similar in pattern to that evoked by actual fearful events. Thus, an apparatus that permits the amplification of small actions in the muscles and glands provides a needed window through which one can gauge the impact of cognitive phenomena.
Measuring Emotional Imagery. Psychophysiological research on imagery has a long history (Jacobson, 1931) involving many investigators. Perhaps the most sustained research program, however, has been carried forward by Lang and colleagues, based on a cognitive model that considers sensory, efferent, and semantic components in the image (e.g., Cuthbert et al., 2003; Lang, Levin, Miller, & Kozak, 1983; Miller et al., 1987; Vrana & Lang, 1990; Weerts & Lang, 1978). In this view (Lang, 1977, 1979), an episode is encoded in memory as an associative network of information units that includes stimulus representations (perceptual properties), response representations (behavior, physiology, and expressive language that occur in the stimulus context), and meaning representations (associated semantic information about the context). It is presumed that the network has a neural substrate, and that for emotional imagery, the response component would include activation of motivational systems (appetitive or aversive) and their reflex physiology (Lang, 1994). Sensory input—pictures, text, or other media—can prompt retrieval of an emotional episode in most individuals. For example, a sentence, “The large snake darts forward, fangs protruding, striking my leg in a flash of pain,” readily prompts a brief unpleasant image. That is, the verbal cues activate a network of representations that can prompt a transient psychophysiological change. However, if the same verbal cues are presented to a seriously phobic individual, whose snake network has high associative strength, we would anticipate a more forceful response—somatically, autonomically—in an elaborated reflex physiology of defense. Imagery and Anxiety Pathology. Almost by definition, all anxious patients are plagued by unpleasant, fearful thoughts and images and concerns about physiological symptoms. They invariably report persistent fear reactions that include pounding heart, palpitations, intense startle, breathing irregularities, sweaty palms, and tense muscles. Nevertheless, it is also true that these complaints are rarely accompanied by measurement of the described physiological reactions. Curiously, when psychophysiological studies have been conducted, many anxiety patients were found to be hyporeactive, not superreactors, to fearful cues (e.g., Hoehn-Saric, McLeod, Funderbunk, & Kowalski, 2004).
Chapter 07.qxd
178
3/28/2007
12:59 PM
Page 178
LANG, MCTEAGUE, CUTHBERT
Summarizing an initial series of imagery studies, Lang (1985) noted that although nearly all anxiety patients report comparable fearfulness and symptoms of anxious arousal, not all anxiety diagnoses showed an accompanying emotional physiology. Although physiological reactions were strong in specific phobic patients when imaging encounters with the phobic object, this was not always true for other diagnoses, most notably agoraphobia. Basing his interpretation on network theory, Lang proposed that fearfulness—defined as exaggerated defense reactions to specific cues in the environment—varies across a diagnostic spectrum, as noted at the outset of this chapter. In this view, specific phobics are held to be the most reactive. This responsiveness is ascribed to the high associative strength (coherence) of the mediating memory networks in the brain, particularly the strong association between neural representations of the cue stimulus (with associated semantic elaborations) and the neural representations that mediate physiological arousal and action. For panic disorder and generalized anxiety disorder, in contrast, representations of fear cues are embedded in networks of low associative strength and the activation of defensive reflexes is less reliably related to specific external stimuli or their internal representations. In effect, patients categorized by these diagnoses report persistent apprehension and distress, but paradoxically, from the perspective of physiological measurement, they appear to be less fearful. A series of imagery studies with anxiety patients and fearful volunteers has lent support to Lang’s (1985) conception (e.g., Cook, Melamed, Cuthbert, McNeil, & Lang, 1988; Lang et al., 1983; McNeil, Vrana, Melamed, Cuthbert, & Lang, 1993). More recently, Cuthbert et al. (2003) studied emotional imagery in more than 100 anxiety patients distributed over four principal diagnoses: specific phobia, social phobia, panic disorder with agoraphobia (PDA), and posttraumatic stress disorder (PTSD). Extending previous findings based on heart rate and skin conductance, the research assessed blink magnitude to a startle probe administered during imagery. Healthy controls, specific phobics, and social phobics all showed significant reflex potentiation to startle probes elicited during imagery of fearful situations (prompted by text crafted for each patient based on presenting complaint or, for controls, text describing the most fearful experiences they could recall); in contrast, both PTSD and PDA patients failed to show reliable fear potentiation to such scenes. As previously observed in a smaller sample (Cook et al., 1988), the diagnostic groups did not differ in the extent of self-reported vividness of their
Chapter 07.qxd
3/28/2007
12:59 PM
Page 179
7. PSYCHOPHYSIOLOGY OF THE ANXIETY SPECTRUM
179
imagery, the judged hedonic valence, or emotional arousal experienced during fearful images. The authors noted an inverse relationship over diagnoses between physiological reactivity and self-reported symptomatic distress, negative affect, and the frequency of mood-disorder comorbidity. Furthermore, internal analyses of the social phobia group showed that even within this diagnosis, as negative affect increased, physiological reactivity decreased. It was suggested that social phobia might be a transition diagnosis (perhaps because it includes both patients with specific performance fears and those with broader social anxiety) on the anxiety spectrum, positioned between the punctuated and reliable reactivity of focal phobics and the diffuse and unreliable reactivity of the more chronically anxious (i.e., PDA/PTSD). Given limitations of sample size and the specificity of the imagery task, many questions about the mechanism and significance of this apparent imagery deficit remain unanswered. For example, although there was some evidence that patients with PDA and PTSD might be generally hyporeactive to threat (not just to clinically relevant fear), data were limited. Furthermore, the previous experiment did not include a GAD group, which presumably would be the least physiologically responsive. To address these issues, to reexamine the role of depression as a moderator of physiological reactivity, and to increase sample size and the power of the analyses, a second imagery experiment was undertaken. IMAGERY OF SOCIAL AND SURVIVAL THREAT: PHYSIOLOGICAL DIFFERENCES AMONG THE ANXIETY DISORDERS The following preliminary research report describes a study of startle probe modulation in anxiety patients during imagery of threatening, fearful imagery. It is part of an ongoing multimodal assessment of anxiety disorder patients at the University of Florida Fear and Anxiety Disorders Clinic. (The questionnaire data from the full sample of 273 were described earlier in this chapter; the data reported here are from a subset of participants for which imagery data are currently available.) Participants were seen at the University of Florida Fear and Anxiety Disorders Clinic subsequent to clinical referral or newspaper or radio advertisement. Diagnostic interviews were conducted using the Anxiety Disorders Interview Schedule (ADIS–IV; Brown, DiNardo, & Barlow, 1994). The sample comprised 159 participants divided into four principal
Chapter 07.qxd
180
3/28/2007
12:59 PM
Page 180
LANG, MCTEAGUE, CUTHBERT
diagnostic categories (specific phobia, n = 30; social phobia, n = 36; panic disorder with agoraphobia [PDA], n = 27; generalized anxiety disorder [GAD], n = 26; see footnote 1, earlier in chapter); additionally, there was a demographically matched, nontreatment-seeking control group (n = 40), showing no clinically significant psychological disorder on the ADIS and recruited through newspaper advertisements. The majority of the sample was White (78%), approximately 67% of participants were female, and the mean age was 34.4 years (SD = 12.4). Comorbid depression as an additional diagnosis was present in 52% of the anxiety patients.
Experimental Procedure. After providing informed consent and filling out a questionnaire battery, participants were then interviewed with the ADIS–IV. On completion of the interview and a short break, the imagery protocol was introduced and sensors were placed on the participant. Participants were told they would hear a series of tones every several seconds, and that upon hearing a tone, they were to relax, to breathe slowly, and to silently repeat the word one. This was intended to function like secular meditation, reducing and stabilizing physiological activity between imagery trials (Benson, 1975). Participants were told that from time to time in the tone series they would hear a series of imagery scripts. They were to listen carefully to the scripts when presented. At stimulus offset the participant was to vividly imagine an active personal involvement in the situation described by the sentence—experiencing the scene as an active participant as opposed to an observer. She or he was told to maintain this active imagining until the tone series started again, and then to return to silently repeating the word one. Participants were instructed to maintain closed eyes throughout the entire session. Twenty-four sentences were constructed to correspond to 12 content categories and three superordinate valence categories. The sentences were 18 to 20 words long and reflected action and participation in the ongoing scene, as opposed to bystander observation. Sentences were written to communicate affective meaning within the first three to four words. The sentences were digitized into 6-second audio files and presented over headphones, as imagery prompts. Because the total data set is too extensive for presentation in this chapter, only the six threat sentences representing survival and social challenges are considered in the following analyses (see Fig. 7.4). A 12-second imagery period occurred immediately after script presentation, followed by a return to relaxation. Acoustic startle probes (95 dB[A]
Chapter 07.qxd
3/28/2007
12:59 PM
Page 181
7. PSYCHOPHYSIOLOGY OF THE ANXIETY SPECTRUM
181
FIGURE 7.4 Trial structure and sample scenes. Top: A single trial consisted of initial tone-cued, relaxation periods, then listening to a script followed by imagery, and finally a tone-cued return to relaxation. Startle probes were presented 3 to 10.5 seconds into the imagery period and during intertrial intervals. Bottom: Exemplar scripts depicting scenes of threat.
burst of white noise, 50-ms duration) were presented at varying times during imagery and inter-trial relaxation, and during an initial 5-minute period of baseline recording. The eyeblink component of the startle reflex was scored from orbicularis muscle action, as the maximum excursion from the level immediately preceding response onset. Trials with clear artifacts were rejected, whereas trials with no responses were scored as zero-magnitude blinks. Reactivity in corrugator muscle, skin conductance level, and heart rate were also recorded during the experiment. However, to better focus the issues in this brief presentation, analyses are restricted here to the startle probe reflex, which shows the clearest connection to defense system activation based on neuroscience research (e.g., Davis & Lang, 2003).
Diagnostic Differences in Startle Reactivity. Figure 7.5 illustrates the diagnostic group differences in blink magnitude to a startle probe administered during imagery of threatening scripts. As expected, diagnostic differences emerged, F(4, 158) = 2.63, p < .05, with larger
Chapter 07.qxd
182
3/28/2007
12:59 PM
Page 182
LANG, MCTEAGUE, CUTHBERT
FIGURE 7.5 Mean startle blink magnitude when imagining threatening scenes by principal anxiety diagnosis.
blinks elicited by specific phobics compared to controls (p < .05), PDA (p < .05), and GAD (p < .05). Social phobics also clearly showed larger blinks than PDA (p < .05) and, despite higher within-group variance, a strong trend for GAD (p = .08). Overall, the data support the distinction made by Cuthbert et al. (2003) between fearful and anxious disorders: As before, the specific and social phobic diagnoses represent the fearful, and PDA and GAD (new to this analysis) combine to form the anxious group. Again, it is the anxious who show significantly attenuated defensive reactivity during threatening or fearful imagery compared to the fearful specific and social phobics, F(1, 117) = 8.07, p < .01. In contrast to the group differences found during imagery of threat scenes, no differences emerged for startle reflexes elicited during a fiveminute baseline prior to the imagery protocol. Furthermore, no group differences emerged in the magnitude of startle responding to the probes delivered between trials during the brief relaxation period (i.e., 18- to 30second intertrial interval). Rather, the group differences appear to be a function of differences among the anxiety disorders in affective recruitment during imagery.
Imagery Self-Report. Is the diminished physiological reactivity of the anxious disorders due in part to reduced imaginal involvement? As measured by self-report ratings, there were no group differences in the rated unpleasantness (emotional valence) of the threat imagery. There was, however, an interesting group difference in rated emotional intensity (arousal) during threatening imagery, F(4, 156) = 5.25, p < .01: Social phobics, PDA, and GAD actually rated threatening scenes as significantly
Chapter 07.qxd
3/28/2007
12:59 PM
Page 183
7. PSYCHOPHYSIOLOGY OF THE ANXIETY SPECTRUM
183
FIGURE 7.6 The left panel depicts the trait anxiety (STAI; Spielberger et al., 1983) symptom-level differences and the right panel depicts depression (BDI; Beck et al., 1996) symptom-level differences in blink magnitude to a startle probe administered during imagery of threatening scripts. Lower levels of trait anxiety and depression predict larger and more robust defensive reactivity.
more arousing than did controls and specific phobics—the opposite pattern than would be expected from the physiological data. For ratings of dominance (the sense of being in control of one’s reactions) during fear imagery, a group difference also emerged, F(4, 156) = 3.94, p < .01: Social phobics, PDA, and GAD patients had a lesser sense of agency than controls or specific phobics.
Defense Reactivity and Negative Affect. The relationship between defense reactivity and questionnaire measures of anxiety and depression was assessed across the entire patient sample. The sample was first divided into three groups on trait anxiety scores (STAI; Spielberger et al., 1983), that is, the lower quartile, the middle two quartiles combined, and the upper quartile, yielding low, moderate, and high symptom groups. Significant differences were observed, F(2, 117) = 6.87, p < .01, with larger blinks elicited by low symptom endorsers compared to both moderate (p < .05) and high symptom groups (p < .05). Figure 7.6 (left panel) depicts the symptom level differences in blink magnitude to a startle probe administered during imagery of threatening scripts and shows an effect which might seem contradictory. That is, lower levels of trait anxiety predict larger and more robust defensive reactivity in contrast to increasingly reduced reactivity with increasing trait anxiety. Next, the anxiety patients were similarly divided on levels of cognitive and somatic symptoms of depression (BDI; Beck et al., 1996) into low, moderate, and high symptom groups. A trend for an overall group difference emerged, F(2, 117) = 2.86, p = .09. Planned comparisons revealed that the low-symptom group elicited larger blinks than the
Chapter 07.qxd
184
3/28/2007
12:59 PM
Page 184
LANG, MCTEAGUE, CUTHBERT
high-symptom group (p < .05). Figure 7.6 (right panel) illustrates differences in blink magnitude elicited during threatening imagery by level of depression symptoms. As shown in the previous analyses, lower levels of trait anxiety predict larger and more robust defensive reactivity and the same effect is evident here for depression. Fear-potentiated startle appears reduced with increasing levels of depression. The latter pattern is also observed if interview-diagnosed comorbid depression is used to predict group differences; that is, nondepressed anxiety patients show greater potentiated startle to threat than do those with comorbid depression, F(1, 117) = 7.85, p < .01.
Three Factors? Effects of Fear, Anxiety, and Depression. At this point, the results suggest that among anxiety patients both trait or generalized anxiety and depression attenuate defensive responding as assessed with the acoustically elicited startle reflex. Do anxiety and depression have separate effects on reactivity if both are simultaneously controlled? Is generalized anxiety or depression more influential in determining hyporeactivity? The following analyses investigate the selfreported and physiological reactivity of the putative fearful (i.e., specific and social phobics) and anxious (i.e., PDA and GAD) groups while also accounting for comorbid depression. Four patient groups are considered: fearful-nondepressed (n = 31), fearful-depressed (n = 32), anxiousnondepressed (n = 25), and anxious-depressed (n = 25) individuals. In a 2 × 2 analysis of variance (ANOVA) with fearfulness versus anxiety and presence/absence of depression as two independent variables predicting trait anxiety scores, a main effect for the fearful/anxious distinction emerged, F(1, 107) = 7.76, p < .01, owing to the elevated trait anxiety in the anxious compared with the fearful group. Similarly elevated trait anxiety among depressed compared with nondepressed patients resulted in a second main effect, F(1, 107) = 25.7, p < .01. Relative to the four patient groups, the fearful-nondepressed group showed the least trait anxiety (M = 37.0), followed by anxious-nondepressed (M = 44.2), then fearfuldepressed (M = 47.5), and finally anxious-depressed (M = 50.4) reporting the highest trait anxiety. Overall, anxious disorders are associated with more trait anxiety. However, comorbid depression is associated with higher trait anxiety for both fearful and anxious disorders. Similarly and not surprisingly, comorbid depression is also associated with significantly higher BDI scores for both fearful and anxious disorders. Employing the same 2 × 2 ANOVA with fearfulness versus anxiety and presence/absence of depression as two independent variables but this time
Chapter 07.qxd
3/28/2007
12:59 PM
Page 185
7. PSYCHOPHYSIOLOGY OF THE ANXIETY SPECTRUM
185
FIGURE 7.7 Illustrates differences in blink magnitude during threat imagery for fearful (specific and social phobics) versus anxious (PDA and GAD) disorders, with and without depression. The principal presenting problem (i.e., phobia or anxiety) interacts with comorbid depression to predict physiological reactivity.
predicting BDI scores, a trend for a main effect of fearfulness versus anxiety emerged, F(1, 111) = 3.51, p = .06, owing to the heightened dysphoria in the anxious compared with the fearful group. Not surprisingly, BDI scores were also higher in the depressed compared to nondepressed patients, resulting in an additional main effect, F(1, 111) = 56.57, p < .001. Comparing the four patient groups on BDI scores revealed the same pattern as trait anxiety: The fearful-nondepressed group rated the least symptomatology (M = 9.09), followed by anxious-nondepressed (M = 13.8), then fearful-depressed (M = 22), and finally anxious-depressed (M = 22.7) reporting the highest depression score. In short, both trait anxiety and depressive symptomatology increase in the presence of comorbid depression, regardless of whether the principal presenting problem is primarily one of fearfulness or anxiety. With these highly reliable self-report differences, how is blink magnitude affected? To investigate this question the same 2 × 2 analytical scheme was repeated, this time predicting blink magnitude. Figure 7.7 illustrates differences in blink magnitude during threat imagery for the same four patient groups: fearful versus anxious disorders, with and without depression. Looking first at the fearful and anxious disorders irrespective of comorbid depression, it is evident that those with circumscribed fear as opposed to generalized anxiety demonstrate augmented blink response, main effect, F(1, 112) = 7.37, p < .01. Looking next at the nondepressed and depressed groups revealed reduced startle reactivity in those with comorbid depression, main effect, F(1, 112) = 6.98, p < .01. Follow-up
Chapter 07.qxd
186
3/28/2007
12:59 PM
Page 186
LANG, MCTEAGUE, CUTHBERT
comparisons showed that the main effect of startle attenuation with comorbid depression is reliable for both fearful (p < .01) and anxious disorders (p < .01). Interestingly, in comparing both depressed groups, the fearful depressed group still exhibited greater startle potentiation than the anxious-depressed group (p < .05) suggesting that the principal presenting problem, phobia or anxiety, and presence or absence of comorbid depression additively modulate physiological reactivity.2 DISCUSSION AND CONCLUSIONS Similar to previous imagery studies (e.g., Cook et al., 1988; Cuthbert et al., 2003), the results support the spectrum model proposed by Lang (1985), in which phobics (specific and social) have the strongest defensive reaction to threatening images, and PDA and now also GAD show clear attenuation. More specifically, patients whose primary clinical complaint is a focused, cue-specific fear respond to threat imagery with robust startle potentiation; patients with a primary complaint of generalized distress (anxiety and depression) are unreactive. The present findings make it clear, furthermore, that this reflex pattern is not specific to fear cues that are related to the patient’s clinical problem (as highlighted in previous research: Cook et al., 1988; Cuthbert et al., 2003). The threat scenes analyzed here, describing events that most people find fearful, were administered to all participants. In fact, these same threat contents reliably prompt potentiated blink reflexes in healthy control participants (McTeague, Bradley, & Lang, 2002). For specific phobia, reflex potentiation is even more pronounced. Surprisingly, this defense reaction is markedly attenuated in PDA and effectively absent in GAD patients. These results highlight the marked discordance in some anxiety patients between psychophysiology and verbal report: PDA patients rated their threat images more emotionally arousing and unpleasant than did specific phobics (they also had the highest scores on the Anxious Arousal Scale: MASQ; the Anxiety Sensitivity Index, Fig. 7.1); nevertheless, they were physiologically, significantly less reactive. 2 It appears highly unlikely that the differences described here are attributable to variations in medication use. Although urine or blood assays were not performed on this sample, information on use and type of drug was obtained from patients at interview. When the overall anxiety patient sample was divided into current users and nonusers of psychotropic medication, no difference was found in blink reflex magnitude, F(1,171) = 1.37, ns. Furthermore, no differences were found in separate analyses of each diagnostic group (specific phobia, social phobia, PDA, GAD).
Chapter 07.qxd
3/28/2007
12:59 PM
Page 187
7. PSYCHOPHYSIOLOGY OF THE ANXIETY SPECTRUM
187
Fear, Anxiety, and Depression. A close relationship is demonstrated between diminished defense system reactivity and interview/questionnaire indexes of comorbid anxiety and depression, as in Figure 7.7. Blink magnitude decreases in a monotonic, stepwise function from the fearful patients (specific and social phobia) who are not depressed, to the fearful-depressed, to the anxious (PDA and GAD) but nondepressed, to the most severe, anxious-depressed patients, linear trend, p < .01. These results suggest an additive effect of increasing anxiety and depression. Given that these are the defining variables of negative affect (Clark & Watson, 1991), fear potentiated startle covaries systematically, but inversely. This is observed even within the phobic disorders; that is, although specific phobia shows greater threat reactivity than controls, defensive reflexes are diminished with increasing generalized anxiety and depression. In the other disorders (PDA and GAD), negative affect is the predominant symptomatology and true fearfulness is almost absent. The Structure of Psychopathology. The spectrum pattern defined by the defensive startle reflex resonates with the analyzed structure of psychopathology discerned in large, epidemiological samples. In a confirmatory factor analytic study of results from the National Comorbidity Survey (N = 8,098), Krueger (1999) analyzed comorbidity among a large number of disorders and defined a best-fit structure with two overarching factors, Externalizing and Internalizing. Of importance for this discussion, there were two Internalizing subfactors, Fear and AnxiousMisery. Furthermore, a virtually identical factor structure, with similar loadings, was reported in a separate sample from the NEMESIS study in the Netherlands (N = 7,076; Vollebergh, Iedema, Bijl, de Graaf, Smit, & Ormel, 2001). The new findings reported here add substantively to this picture: They emphasize a continuum from predominantly fearful to predominantly anxious diagnoses—a phenomenon not specifically suggested by the epidemiological studies. Furthermore, both Krueger (1999) and Vollebergh et al. (2001) placed agoraphobia within the Fear subfactor. The present findings—emphasizing physiological reactivity—suggest that panic/agoraphobia has a profile more similar to GAD, and thus more consistent with Anxious-Misery. Considered from this general perspective, a patient’s general symptom pattern might be determined by a single vector weighted with the separate strengths of the Fear and Anxious-misery factors. The Fear component would be more prominent in specific and social phobia, and the Anxious-Misery component more determinant in GAD and depression.
Chapter 07.qxd
188
3/28/2007
12:59 PM
Page 188
LANG, MCTEAGUE, CUTHBERT
Psychophysiological measures would contribute importantly to operational measurement of these constructs, and facilitate the eventual development of a practical, multifactor diagnostic system.
Is Physiological Hyporeactivity in Anxious and Depressed Patients Unique to the Imagery Paradigm? Although reduced physiological reactivity during fear imagery has been repeatedly demonstrated, it is still not clear whether it is unique to imagery (memory or language processing) or whether it reflects a broader deficit in response to any emotional challenge. Dichter, Tomarken, Shelton, and Sutton (2004) reported a similar startle hyporeactivity in depressed participants looking at emotionally evocative pictures. Although these patients evidenced normal attention to pictures, “the late-probe blink magnitudes of depressed patients were unrelated to picture valence” (p. 433). This absence of potentiation to unpleasant pictures is suggestive of the startle attenuation found here in depressed anxiety patients. Hyporeactivity to challenge has also been observed in a variety of different paradigms, using other, mainly autonomic measures. In this research, a similar discordance between verbal report and manifest physiology is also found (e.g., Friedman & Thayer, 1998, panic; Hoehn-Saric et al., 2004; Hoehn-Saric, McLeod, & Zimmerli, 1989, 1991, GAD and panic; Kirsch & Geer, 1988, premenstrual syndrome; Roemer & Borkovec, 1993, GAD). However, the mechanism of the effect remains unclear. Mechanism: Cognitive Avoidance. The startle response is an obligatory reflex; although it can be modulated by many factors, its intentional suppression has not been demonstrated. However, the possibility exists that participants can reduce probe reactions by somehow disattending or actively suppressing perception of unpleasant foreground stimuli. The notion of strategic cognitive avoidance is a common clinical hypothesis (e.g., Hayes, Strosahl, & Wilson, 1999), but the empirical literature is replete with contradictory findings. Thus, increased (not decreased) intrusive thoughts have been reported in a wide array of thought suppression studies (e.g., Wenzlaff & Wegner, 2000), and increased (not decreased) autonomic arousal has been found when participants, watching films, were told to suppress emotional expression (e.g., Gross & Levenson, 1993, 1997). Similarly, instructions to suppress or inhibit unpleasant emotions prompted by a biological challenge did not reduce
Chapter 07.qxd
3/28/2007
12:59 PM
Page 189
7. PSYCHOPHYSIOLOGY OF THE ANXIETY SPECTRUM
189
autonomic responding (Feldner, Zvolensky, Eifert, & Spira, 2003; Spira, Zvolensky, Eifert, & Feldner, 2004). Indeed, most studies of anxiety patients suggest that, rather than showing disattention to unpleasant cues, the attention of anxious participants is drawn to, even automatically captured by negative stimulation (Matthews & MacLeod, 1994). Nevertheless, it could be that image texts elicit a different response in generally anxious patients (i.e., worry) and not memory imagery (e.g.. Roemer & Borkovec, 1993). However, Cuthbert et al. (2003) found no relationship between reported imagery vividness and startle attenuation, and in the more recent study, the anxious diagnoses reported more (not less) intense emotion during threat imagery.
Mechanism: Physiology and Behavior. It is possible that the observed startle potentiation deficit is secondary to stress-related neurochemical changes that either interfere with image processing or, alternatively, directly alter the startle circuit and/or its connection to the defense system. For example, it is well known that the neurotransmitters norepinepherine and serotonin, when infused locally in the brain or introduced into the circulatory system, modulate startle amplitude (e.g., Davis, Astrachan, & Kass, 1980). Furthermore, neurochemical depletion has been associated with helplessness in animals and with motor retardation and cognitive symptoms in mood disorder. It is, of course, also possible that the normal defense system is simply not engaged in high anxiety and depression. Gray (1987) proposes that different motive circuits may be active in anxiety. He describes a behavioral activation circuit (BAS) that responds to imminent danger or pain, involving the amygdala. However, there is also an inhibitory septalhippocampal circuit that mediates passive avoidance. McNaughton and Gray (2000) suggest that when the amygdala is inactive, the inhibitory circuit mediates rumination and worry. There is not, however, evidence that this system attenuates potentiated startle. SUMMARY AND CONCLUSIONS 1.
There is strong evidence supporting the concept of an anxiety disorder spectrum characterized by increasing severity of psychopathology, and by a reciprocal diminution in fear reactivity. 2. Fear, defined here as a readiness to respond defensively to imagined threat, is a major spectrum dimension, permitting a meaningful
Chapter 07.qxd
190
3.
4.
5.
6.
3/28/2007
12:59 PM
Page 190
LANG, MCTEAGUE, CUTHBERT
distinction between fearful diagnoses (specific and social phobia) and anxious diagnoses (panic with agoraphobia and generalized anxiety). Generalized anxiety and comorbid depression additively attenuate startle potentiation (to imagined threat) in anxiety disordered patients. Discordance between physiological reactivity and verbal report of anxious arousal is marked in anxiety disorder, particularly in panic with agoraphobia, suggesting that reliance on interview and questionnaire assessment may obscure the diathesis of a disorder and its place in the factorial structure of negative affect and anxiety pathology. The physiological assessment of defensive reactivity contributes importantly to the diagnostic differentiation of anxiety disorders. The methodology could usefully supplement interview and questionnaires in routine psychological assessment. Determining the generality of blunted defense reactivity, across different tasks and reflex measures, merits continued research effort. Furthermore, the population sample must be extended to other diagnoses, such as obsessive-compulsive disorder and PTSD. In conclusion, it is critical to obtain an understanding of mechanisms, psychological and neurophysiological, that prompt anxious patients to show a paradoxical lack of vigor in fear reactivity. ACKNOWLEDGMENTS
This work was supported in part by National Institute of Mental Health grants MH37757 and P50-MH52384; an NIMH Behavioral Science grant to the Center for the Study of Emotion and Attention (CSEA), University of Florida, Gainesville; and an NRSA Research Fellowship F31-MH069048 to the second author. Many thanks to various members of the NIMH Center for the Study of Emotion and Attention for their assistance, especially Cyd Strauss, Marie-Claude Laplante, and Margaret Bradley. REFERENCES Amaral, D. G., Price, J. L., Pitkanen, A., & Carmichael, S. T. (1992). Anatomical organization of the primate amygdaloid complex. In J. P. Aggleton (Ed.), The amygdala: Neurobiological aspects of emotion, memory, and mental dysfunction (pp. 1–66). New York: Wiley.
Chapter 07.qxd
3/28/2007
12:59 PM
Page 191
7. PSYCHOPHYSIOLOGY OF THE ANXIETY SPECTRUM
191
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck Depression Inventory–second edition. San Antonio, TX: The Psychological Corporation. Beck, A. T., Ward, C. H., Mendelsohn, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 4, 561–571. Benson, H. (1975). The relaxation response. New York: Morrow. Blanchard, R. J., & Blanchard, D. C. (1989). Attack and defense in rodents as ethoexperimental models for the study of emotion. Progress in Neuro-Psychopharmacological & Biological Psychiatry, 13, 3–14. Bradley, M. M., Codispoti, M., Cuthbert, B. N., & Lang, P. J. (2001). Emotion and motivation: I. Defensive and appetitive reactions in picture processing. Emotion, 1, 276–298. Bradley, M. M., Codispoti, M., Sabatinelli, D., & Lang, P. J. (2001). Emotion and motivation: II. Sex differences in picture processing. Emotion, 1, 300–319. Bradley, M. M., Moulder, B., & Lang, P. J. (2005). When good things go bad: The reflex physiology of defense. Psychological Science, 16, 468–473. Brown, T. A., Campbell, L. A., Lehman, C. L., Grisham, J. R., & Mancil, R. B. (2001). Current and lifetime comorbidity of the DSM–IV anxiety and mood disorders in a large clinical sample. Journal of Abnormal Psychology, 110, 585–599. Brown, T. A., Chorpita, B. F., & Barlow, D. H. (1998). Structural relationships among dimensions of the DSM–IV anxiety and mood disorders and dimensions of negative affect, positive affect, and autonomic arousal. Journal of Abnormal Psychology, 107, 179–192. Brown, T. A., DiNardo, P. A., & Barlow, D. H. (1994). The Anxiety Disorder Interview Schedule for DSM–IV. Albany, NY: State University of New York, Albany, Center for Stress and Anxiety Disorders. Buss, A. H. (1962). Critique and notes: Two anxiety factors in psychiatric patients. Journal of Personality and Social Psychology, 65, 426–427. Buss, A. H., & Plomin, R. (1975). A temperament theory of personality development. New York: Wiley-Interscience. Clark, L. A., & Watson, D. (1991). Tripartite model of anxiety and depression: Psychometric evidence and taxonomic implications. Journal of Abnormal Psychology, 100, 316–336. Cook, E. W. III, Melamed, B. G., Cuthbert, B. N., McNeil, D. W., & Lang, P. J. (1988). Emotional imagery and the differential diagnosis of anxiety. Journal of Consulting and Clinical Psychology, 56, 734–740. Cuthbert, B. N., Lang, P. J., Strauss, C., Drobes, D., Patrick, C. J., & Bradley, M. M. (2003). The psychophysiology of anxiety disorder: Fear memory imagery. Psychophysiology, 40, 407–422. Davis, M. (1992). The role of the amygdala in conditioned fear. In J. Aggleton (Ed.), The amygdala: Neurobiological aspects of emotion, memory, and mental dysfunction (pp. 255–305). New York: Wiley. Davis, M. (2000). The role of the amygdala in conditioned and unconditioned fear and anxiety. In J. P. Aggleton (Ed.), The amygdala (Vol. 2, pp. 213–287). Oxford, England: Oxford University Press.
Chapter 07.qxd
192
3/28/2007
12:59 PM
Page 192
LANG, MCTEAGUE, CUTHBERT
Davis, M., Astrachan, D. I., & Kass, E. (1980). Excitatory and inhibitory effects of serotonin on sensorimotor reactivity measured with acoustic startle. Science, 209, 521–523. Davis, M., & Lang, P. J. (2003). Emotion. In M. Gallagher & R.J. Nelson (Eds.), Handbook of psychology: Volume 3. Biological psychology (pp. 405–439). New York: Wiley. Dichter, G. S., Tomarken, A. J., Shelton, R. C., & Sutton, S. K. (2004). Early- and late-onset startle modulation in unipolar depression. Psychophysiology, 41, 433–440. Feldner, M. T., Zvolensky, M. J., Eifert, G. H., & Spira, A. P. (2003). Emotional avoidance: An experimental test of individual differences and response suppression using biological challenge. Behavior Research & Therapy, 41, 403–411. Fenz, W. D., & Epstein, S. (1965). Manifest anxiety: Unifactorial or multifactorial composition? Perceptual and Motor Skills, 20, 773–780. Friedman, B. H., & Thayer, J. F. (1998). Autonomic balance revisited: Panic anxiety and heart rate variability. Journal of Psychosomatic Research, 44, 133–151. Gloor, P. (1960). Amygdala. In J. Field (Ed.), Handbook of physiology: Section 1. Neurophysiology (pp. 1395–1420). Washington, DC: American Physiological Society. Gray, J. A. (1987). The psychology of fear and stress (2nd ed.). Cambridge, England: Cambridge University Press. Gray, T. S. (1989). Autonomic neuropeptide connections of the amygdala. In Y. Tache, J. E. Morley, & M. R. Brown (Eds.), Neuropeptides and stress (pp. 92–106). New York: Springer. Grillon, C., Ameli, R., Woods, S. W., Merikangas, K., & Davis, M. (1991). Fearpotentiated startle in humans: Effects of anticipatory anxiety on the acoustic blink reflex. Psychophysiology, 28, 588–595. Gross, J. J., & Levenson, R. W. (1993). Emotional suppression: Physiology, selfreport, and expressive behavior. Journal of Personality and Social Psychology, 64, 970–986. Gross, J. J., & Levenson, R. W. (1997). Hiding feelings: The acute effects of inhibiting negative and positive emotion. Journal of Abnormal Psychology, 106, 95–103. Hamm, A. O., Cuthbert, B. N., Globisch, J., & Vaitl, D. (1997). Fear and startle reflex: Blink modulation and autonomic response patterns in animal and mutilation fearful subjects. Psychophysiology, 34, 97–107. Hamm, A. O., Greenwald, M. K., Bradley, M. M., & Lang, P. J. (1993). Emotional learning, hedonic change, and the startle probe. Journal of Abnormal Psychology, 102(3), 453–465. Hayes, S. C., Strosahl, K. D., & Wilson, K. G. (1999). Acceptance and commitment therapy. New York: Guilford. Hoehn-Saric, R., McLeod, D. R., Funderbunk, F., & Kowalski, P. (2004). Somatic symptoms and physiologic responses in generalized anxiety disorder and panic disorder: An ambulatory monitor study. Archives of General Psychiatry, 61, 913–921. Hoehn-Saric, R., McLeod, D. R., & Zimmerli, W. D. (1989). Somatic manifestations in women with generalized anxiety disorder. Archives of General Psychiatry, 46, 1113–1119.
Chapter 07.qxd
3/28/2007
12:59 PM
Page 193
7. PSYCHOPHYSIOLOGY OF THE ANXIETY SPECTRUM
193
Hoehn-Saric, R., McLeod, D. R., & Zimmerli, W. D. (1991). Psychophysiological response patterns in panic disorder. Acta Psychiatrica Scandinavica, 83, 4–11. Jacobson, E. (1931). Electrical measurements of neuromuscular states during mental activities: V. Variation of specific muscles contracting during imagination. American Journal of Physiology, 96, 115–121. Kapp, B. S., & Pascoe, J. P. (1986). Correlations aspects of learning and memory: Vertebrate model systems. In J. L. Martinez & R. P. Kesner (Eds.), Learning and memory: A biological view (pp. 399–440). New York: Academic Press. Kapp, B. S., Pascoe, J. P., & Bixler, M. A. (1984). The amygdala: A neuroanatomical systems approach to its contribution to aversive conditioning. In N. Butters & L.S. Squire (Eds.), The neuropsychology of memory (pp. 473–488). New York: Guilford. Kirsch, J. R., & Geer, J. H. (1988). Skin conductance and heart rate in women with premenstrual syndrome. Psychosomatic Medicine, 50, 175–182. Konorski, J. (1967). Integrative activity of the brain: An interdisciplinary approach. Chicago: University of Chicago Press. Krueger, R. F. (1999). The structure of common mental disorders. Archives of General Psychiatry, 56, 921–926. Krueger, R. F., & Finger, M. S. (2001). Using item response theory to understand comorbidity among anxiety and unipolar mood disorders. Psychological Assessment, 13, 140–151. Lang, P. J. (1977). Imagery in therapy: An information processing analysis of fear. Behavior Therapy, 8, 862–886. Lang, P. J. (1978). Language, imagery, and emotion. In P. Pliner, K. R. Blankstein, & I. M. Spigel (Eds.), Advances in study of emotion and affect: Vol. 5. Perceptions of emotion in self and others (pp. 107–117). New York: Plenum. Lang, P. J. (1979). A bio-informational theory of emotional imagery. Psychophysiology, 16, 495–512. Lang, P. J. (1985). The cognitive psychophysiology of emotion: Fear and anxiety. In A. H. Tuma & J. D. Maser (Eds.), Anxiety and the anxiety disorders (pp. 131–170). Hillsdale, NJ: Lawrence Erlbaum Associates. [Reprinted in Psychotherapeutisch Pasport, 3(2), 3–62 (1985)] Lang, P. J. (1994). The motivational organization of emotion: Affect-reflex connections. In S. VanGoozen, N. E. Van de Poll, & J. A. Sergeant (Eds.), Emotions: Essays on emotion theory (pp. 61–93). Hillsdale, NJ: Lawrence Erlbaum Associates. Lang, P. J., Levin, D. N., Miller, G. A., & Kozak, M. J. (1983). Fear imagery and the psychophysiology of emotion: The problem of affective response integration. Journal of Abnormal Psychology, 92, 276–306. LeDoux, J. E. (1987). Emotion. In V. B. Mountcastle, F. Plum, & St. R. Geiger (Eds.), Handbook of physiology: Section 1. The nervous system (Vol. 5, pp. 419–459). Bethesda, MD: American Physiological Association. Mandler, G., Mandler, J. M., Kremen, I., & Sholiton, R. (1961). The response to threat: Relations among verbal and physiological indices. Psychological Monographs, 75(Whole No. 513). Marks, I. M., & Mathews, A. M. (1979). Brief standard self-rating for phobic patients. Behaviour Research and Therapy, 17, 263–267.
Chapter 07.qxd
194
3/28/2007
12:59 PM
Page 194
LANG, MCTEAGUE, CUTHBERT
Matthews, A., & MacLeod, C. (1994). Cognitive approaches to emotions and emotional disorders. Annual Review of Psychology, 45, 25–50. McFall, R. M., Treat, T. A., & Viken, R. J. (1997). Contributions of cognitive theory to new behavioral treatments. Psychological Science, 8, 174–176. McNaughton, N., & Gray, J. A. (2000). Anxiolytic action on the behavioral inhibition system implies multiple types of arousal contribute to anxiety. Journal of Affective Disorders, 61, 161–176. McNeil, D. W., Vrana, S. R., Melamed, B. G., Cuthbert, B. N., & Lang, P. J. (1993). Emotional imagery in simple and social phobia: Fear versus anxiety. Journal of Abnormal Psychology, 102, 212–225. McTeague, L. M., Bradley, M. M., & Lang, P. J. (2002). Creating a mental image: Is a picture worth a thousand words? Psychophysiology, 39(Suppl. 1), S57. Miller, G. A., Levin, D. N., Kozak, M. J., Cook, E. W. III, McLean, A., Jr., & Lang, P. J. (1987). Individual differences in imagery and the psychophysiology of emotion. Cognition and Emotion, 1, 367–390. Mineka, S., Watson, D., & Clark, L. A. (1998). Comorbidity of anxiety and unipolar mood disorders. Annual Review of Psychology, 49, 377–412. Reiss, S., & McNally, R. J. (1985). Expectancy model of fear. In S. Reiss & R. R. Bootzin (Eds.), Theoretical issues in behavior therapy (pp. 107–121). San Diego, CA: Academic Press. Roemer, L., & Borkovec, T. D. (1993). Worry: Unwanted cognitive activity that controls unwanted somatic experience. In D. M. Wegner & J. W. Pennebaker (Eds.), Handbook of mental control. Century psychology series (pp. 220–238). Upper Saddle River, NJ: Prentice-Hall. Sabatinelli, D., Bradley, M. M., Fitzsimmons, J. R., & Lang, P. J. (2005). Parallel amygdala and inferotemporal activation reflect emotional intensity and fear relevance. NeuroImage, 24, 1265–1270. Sabatinelli, D., Bradley, M. M., & Lang, P. J. (2001). Affective startle modulation in anticipation and perception. Psychophysiology, 38, 719–722. Sarter, M., & Markowitsch, H. J. (1985). Involvement of the amygdala in learning and memory: A critical review, with emphasis on anatomical relations. Behavioral Neuroscience, 99, 342–380. Spielberger, C. D., Gorsuch, R. L., Lushene, P. R., Vagg, P. R., & Jacobs, G. A. (1983). Manual for the State-Trait Anxiety Inventory (STAI). Palo Alto, CA: Consulting Psychologists Press. Spira, A. P., Zvolensky, M. J., Eifert, G. H., & Feldner, M. T. (2004). Avoidanceoriented coping as a predictor of panic-related distress: A test using biological challenge. Journal of Anxiety Disorders, 18, 309–323. Vollebergh, W. A., Iedema, J., Bijl, R. V., de Graaf, R., Smit, F., & Ormel, J. (2001). The structure and stability of common mental disorders: The NEMESIS study. Archives of General Psychiatry, 58, 597–603. Vrana, S. R., & Lang, P. J. (1990). Fear imagery and the startle probe reflex. Journal of Abnormal Psychology, 99, 181–189. Watson, D., & Clark, L. A. (1991). The Mocd and Anviety Symptom Questionnaire. Unpublished manuscript, University of Iowa, Department of Psychology, Iowa City. Watson, D., Weber, K., Assenheimer, J. S., Clark, L. A., Strauss, M. E., & McCormick, R. A. (1995). Testing a tripartite model: I. Evaluating the convergent
Chapter 07.qxd
3/28/2007
12:59 PM
Page 195
7. PSYCHOPHYSIOLOGY OF THE ANXIETY SPECTRUM
195
and discriminant validity of anxiety and depression subscales. Journal of Abnormal Psychology, 104, 3–14. Weerts, T. C., & Lang, P. J. (1978). Psychophysiology of fear imagery: Differences between focal phobia and social performance anxiety. Journal of Consulting and Clinical Psychology, 46, 1157–1159. Wegner, D. M., Schneider, D. J., Carter, S. R., & White, T. L. (1987). Paradoxical effects of thought suppression. Journal of Personality & Social Psychology, 53, 5–13. Wenzlaff, R. M., & Wegner, D. M. (2000). Thought suppression. Annual Review of Psychology, 51, 59–91. Wolpe, J., & Lang, P. J. (1964). A fear survey schedule for use in behavior therapy. Behavior Research Therapy, 2, 27–30.
Chapter 07.qxd
3/28/2007
12:59 PM
Page 196
Chapter 08.qxd
3/28/2007
12:59 PM
Page 197
8 Behavior Therapy for Specific Fears and Phobias: Context Specificity of Fear Extinction Jayson L. Mystkowski University of California, Los Angeles
Susan Mineka Northwestern University
Behavior therapy had its origins in learning theory in the 1950s and 1960s. Principles of conditioning and learning initially studied in animals were applied to understanding the possible origins of human clinical problems such as anxiety and mood disorders. Although the search to understand the behavioral origins of such disorders continues today (e.g., Mineka & Zinbarg, 1996, 2006), early on investigators were also intent on the more immediate clinical problem of how best to treat people with these various disorders. For example, in his now classic book, Wolpe (1958) summarized animal research supporting the view that phobias and other anxiety disorders originate from classical conditioning of fear or anxiety (see also Watson & Rayner, 1920). He then extended his findings from animal studies using counterconditioning to extinguish these fear responses to the development of systematic
197
Chapter 08.qxd
198
3/28/2007
12:59 PM
Page 198
MYSTKOWSKI AND MINEKA
desensitization treatment, which then became widely used in the 1960s and 1970s for the treatment of phobias and other anxiety disorders.1 Later, certain researchers developed an avoidance learning model of anxiety disorders (rather than a simple classical conditioning one). This led to the idea that flooding, or prolonged exposure without escape or avoidance, would be a more straightforward and powerful treatment just as it is in the extinction of avoidance responses in animals2 (e.g., Mineka, 1979, 1985). Indeed, early findings in the 1970s reported remarkable successes using exposure therapy to treat a wider variety of more severe anxiety disorders like agoraphobia and obsessive-compulsive disorder, with which systematic desensitization was not particularly effective (e.g., Mathews, Gelder, & Johnston, 1981; Rachman & Hodgson, 1980). With both systematic desensitization and exposure therapy, it is easy to see how one’s model and theory of etiology can influence the development of successful treatments. This way of approaching clinical science is, of course, similar to the one long and forcefully advocated by Richard McFall (a former colleague of S. M.) because researchers had taken theoretical and empirical advances in the basic science of learning (or cognitive psychology in other cases) to enhance understanding of the etiology, assessment, or treatment of a given clinical disorder or problem. Unfortunately, the field of behavior therapy soon lost track of its roots in learning theory and became more of a truly applied enterprise. That is, advances continued to be made in developing more effective techniques to treat a wider range of disorders, but these advances were rarely founded in the basic learning science any more. As this happened, the close connections between the two fields waned, and they both advanced on more or less distinct and independent tracks (e.g., Ross, 1985). In the spirit of Richard McFall’s long-term advocacy for a continued dialogue between basic science and clinical science, we review in this 1 Wolpe’s choice of counterconditioning (i.e., using a hierarchy of stimuli from least to most anxiety-provoking and pairing these stimuli in succession with deep muscle relaxation) was dictated by his adherence to Hullian learning theory, which predicted that straightforward extinction would not work for complicated theoretical reasons not relevant here (Mineka, 1985). Unfortunately he did not seem to be aware that this theoretical approach to extinction had been clearly demonstrated to be wrong (e.g., Gleitman, Nachmias, & Neisser, 1954). 2 In flooding or exposure therapy, several key elements of systematic desensitization are not present. For example, it is not necessary to train in deep muscle relaxation, there is no attempt to countercondition any allegedly incompatible response, and the use of a hierarchy is also not necessary (although using a hierarchy is usually advisable because it makes the treatment less frightening for the individual undergoing it).
Chapter 08.qxd
3/28/2007
12:59 PM
Page 199
8. PHOBIA TREATMENT AND CONTEXT
199
chapter some of our attempts to illustrate how advances can still be made in behavioral treatments for anxiety disorders by integrating advances in theory and empirical research from the field of animal learning. In doing so, we review five studies from our laboratories that have been conducted with the ultimate goal of improving the long-term effectiveness of exposure therapies for anxiety disorders. These studies were inspired by, and built on, major theoretical and empirical developments that have been made in the study of extinction of classically conditioned responses in animals in the past two decades (e.g., Bouton, 1994; Bouton, Woods, Moody, Sunsay, & Garcia-Gutierrez, 2006). For our clinical analogue studies, we chose to focus on treatment of one specific fear/phobia—namely, spider fear. Spider fear/phobia was chosen because it is relatively common and because a highly effective one-session treatment already existed. The ultimate goal of these kinds of experiments is, of course, to extend what is found to the treatment of other more complex anxiety disorders in future clinical studies. HOW DO EFFECTIVE TREATMENTS FOR SPECIFIC FEARS AND PHOBIAS OPERATE? The treatment of choice for specific fears and phobias consists of repeated, systematic, and graded exposure to fear-provoking stimuli. In general, individuals are asked to confront their feared object or situation, preferably in real life or in vivo (although imaginal exposure also works), so they may learn that their fear will gradually extinguish or habituate with the passage of time (e.g., Foa & Kozak, 1986; Mineka & Thomas, 1999; Öst, 1997). According to Foa and Kozak’s (1986) emotional processing model, the habituation that naturally occurs during prolonged exposure to feared stimuli leads to a change in the person’s internal mental representations of both cognitive and affective information about the feared stimulus. Because their arousal typically subsides during prolonged exposure trials through a basic biologically based habituation process, clients typically realize that their fear levels do not get worse, but actually diminish. Moreover, as habituation of their emotional arousal occurs, there are opportunities for irrational cognitions about a feared stimulus to change (e.g., “The spider will jump on me and bite my arm” can change to “The spider does not jump or bite me when I touch it and pick it up”). In a somewhat different vein, Bandura (1977, 1986) and Williams (1996) emphasize that an individual’s beliefs about his or her ability to behave and interact
Chapter 08.qxd
200
3/28/2007
12:59 PM
Page 200
MYSTKOWSKI AND MINEKA
effectively with the environment also change during treatment; for these investigators this change in self-efficacy is the key mechanism through which these treatments work. Mineka and Thomas (1999) further proposed that another important element of what happens during exposure treatment is that the individual comes to develop a sense of perceived control over his or her fear or anxiety per se (i.e., not just their behavior). THE RETURN OF FEAR FOLLOWING EXPOSURE THERAPY Although there are perhaps hundreds of studies empirically supporting exposure treatments for various specific fears and phobias (see Craske, 1999, for a review), the effects of graded, exposure-based therapy for specific fears and phobias may be somewhat transient in that fear may partially return with the passage of time. That this occurs is not at all surprising, given the phenomenon of spontaneous recovery following extinction of classically conditioned responses first reported by Pavlov (1927). Spontaneous recovery usually occurs when a conditioned stimulus (CS) is presented following some delay after extinction of a conditioned response (CR), and is manifested by partial recovery of the fear CR to the previously extinguished CS. According to Rachman (1989, 1990), return of fear (the more common clinical term than spontaneous recovery) is the reappearance of some fear that has undergone full or partial extinction. Unfortunately, factors underlying the return of fear have proved difficult to determine. Researchers have explored a variety of variables that may predispose certain individuals to greater return of fear than other individuals. Most of the variables studied involved parametric variations on various aspects of the treatment rather than deriving from learning theory. These include pretreatment variables (such as baseline fear levels), treatment variables (such as exposure duration or intensity), and posttreatment variables that occur following treatment but before retesting at some follow-up interval (such as posttreatment cognitions or intervening stressors). Although some studies have seemed to provide promising clues, careful consideration of the results of these myriad studies reveals that none of these variables have had consistent effects on the return of fear (see Craske, 1999; Rodriguez, Craske, Mineka, & Hladek, 1999, for reviews). Yet until 1999, none of these published studies had tested one promising variable that might affect return of fear in humans that had been extensively studied in rats—namely, the effects of changing contexts between extinction and follow-up.
Chapter 08.qxd
3/28/2007
12:59 PM
Page 201
8. PHOBIA TREATMENT AND CONTEXT
201
CONTEXT SPECIFICITY OF EXTINCTION: ANIMAL RESEARCH For over two decades, animal conditioning research by Bouton and colleagues has investigated the effects of context on return of fear (e.g., Bouton, 1991, 1994; Bouton & Nelson, 1998). In general, their numerous studies on the contextual control of learning and memory have provided strong and consistent evidence that substantial fear recovery occurs in rats when a CS extinguished in one context (other than the one where acquisition occurred) is later presented in a different recovery context. In the basic paradigm used in such studies, rats typically undergo fear conditioning to a novel CS in one context (A). Then extinction of fear of that CS occurs in a different context (B), and fear retention is tested some time later (days or weeks) in context B, or in a third, novel context (C), or in the original conditioning context (A). Contexts are always counterbalanced so that none of the effects observed derives from the effects of a specific context. If extinction and fear retention testing occur in the same context (e.g., extinction in context B and retesting in context B), the magnitude of the fear recovery effect (typically called fear renewal in animal studies) is reduced relative to what is seen when retesting occurs in a different context such as context A (familiar) or C (unfamiliar; Bouton & King, 1983). Moreover, it is also important to note that the fear renewal effect occurs even when the animals have received 100 or more extinction trials (Bouton et al., 2006); thus, fear renewal is not simply occurring because extinction was incomplete. In general, Bouton’s work on extinction indicates that the learning that takes place during extinction does not cause unlearning of the first-learned information, but rather reflects new learning. In other words the CS “acquires a second ‘meaning’ that is available along with the first. In this sense, the current meaning of the signal…is ambiguous” (Bouton, 2002, p. 976). In other words, memories for both the original fear acquisition and fear extinction experiences are retained in the organism’s memory network. Depending on which memory is retrieved most strongly by the current context, fear will or will not be exhibited. When confronted with a context that is novel or different than the one encountered during fear extinction (e.g., context A or C at retention testing, when extinction occurred in context B), the contextual mismatch causes the organism to rely more on associations of fear acquisition (which were learned first) than when there is a contextual match and
Chapter 08.qxd
202
3/28/2007
12:59 PM
Page 202
MYSTKOWSKI AND MINEKA
fear extinction memories are more prominently retrieved. When this mismatch occurs then there is greater fear renewal (Bouton, 2002). Overall Bouton’s theoretical and empirical work with rats suggests the importance of studying contextual factors in human fear extinction because this work has potentially important clinical implications for the long-term effects of exposure therapy. If similar effects of context specificity occur in humans, it might suggest the importance of conducting exposure therapy in multiple contexts because of the variety of conditions in which a treated individual is likely to encounter phobic stimuli in the real world. For example, being treated only in a therapist’s office might not be helpful when later confronting a spider in an attic or garden. Alternatively or additionally, it might also suggest that teaching individuals techniques to increase retrieval of extinction memories acquired in a specific context might prevent or minimize return of fear when confronting previously feared objects or situations in different contexts. Finally, such findings would also have implications for understanding the mechanisms that underlie the effectiveness of such treatments. STUDYING THE EFFECTS OF EXTERNAL CONTEXT ON RETURN OF FEAR IN HUMANS Studying the effects of external context on return of fear is a relatively new area of study with regard to human fears and phobias. In this chapter we review the results of five of our own studies that have explored these effects in highly spider-fearful individuals. In each of our studies we used theory and principles derived from the basic learning science to enhance understanding of return of fear following graded exposure-based therapy techniques—the most widely used and validated treatment for phobias for the past several decades (e.g., Craske & Rowe, 1997; Öst, 1997). Dependent measures were primarily gathered during pretreatment, posttreatment, and follow-up behavioral approach tasks (BATs). During each behavioral approach, task participants were asked to approach, as closely as possible, a tarantula in an open container placed about 10 feet from the participant. The primary dependent measure was self-reported fear on a 0 to 100 subjective units of distress scale (SUDS; Wolpe, 1973). Additional dependent measures were heart rate (in three of the five studies) and behavioral avoidance as indexed by how close the participants could actually get to the phobic stimulus. Participants were originally invited to participate in these experiments based on high levels of self-reported fear assessed several weeks earlier using the Spider Phobia Questionnaire
Chapter 08.qxd
3/28/2007
12:59 PM
Page 203
8. PHOBIA TREATMENT AND CONTEXT
203
(Klorman, Hastings, Weerts, Melamed, & Lang, 1974). To be included in the study, participants also needed to report a SUDS rating of 70 or more (i.e., moderately severe spider fear) while attempting to get as close as possible to the tarantula during an initial BAT. To maximize the efficacy of these exposure techniques, we used a variant on these procedures developed by Bandura and colleagues (e.g., Bandura, Blanchard, & Ritter, 1969) called participant modeling. Specifically, during treatment, each individual was exposed to a live nonpoisonous Chilean rose-haired tarantula (Phrixotrichus spatulata; leg span approximately 6 in. [15.2 cm]) in a 14-step treatment hierarchy, with each successive step of increasing difficulty. During each step, the fearful participants were asked to perform a specific task that was first modeled by the experimental therapist (hence the term participant modeling). For example, the therapist first modeled standing 5 feet from the tarantula in its closed container and encouraged the participants to stand as close as possible also (Step 1); later they were taught how to move the tarantula gently by touching it five times with a small paintbrush while it moved around in a plastic basin (Step 7). The final task (Step 14) involved the participant letting the tarantula walk over his or her bare hand (first modeled by the experimental therapist). Treatment was only considered to be complete when participants successfully completed Step 14. The time allotted for treatment was 90 minutes, and nearly all participants completed treatment in less than that amount of time. Indeed, there were never more than a few participants in each study that failed to complete treatment in the allotted time. In addition, in excess of 95% of all participants were able to at least touch the spider with their bare fingertip in the first posttreatment BAT conducted shortly after the end of treatment.
Study 1. In the first study, Rodriguez, Craske, Mineka, and Hladek (1999) investigated an undergraduate sample of 65 students with moderate to severe levels of spider fear. Graded, exposure-based therapy with participant modeling as described earlier was conducted in either context A or B, followed by a test for fear renewal or recovery two weeks later in either context A or B (see Table 8.1). Context was operationally defined by three environmental cues: (a) the particular experimental therapist used for treatment, (b) location of the room for treatment, and (c) salient visual cues associated with both the therapist and the room (e.g., color of therapist lab coat, room furnishings, wall hangings). Rodriguez et al. (1999) demonstrated some preliminary support for the hypothesis that contextual shifts can affect return of fear in that the
Chapter 08.qxd
204
3/28/2007
12:59 PM
Page 204
MYSTKOWSKI AND MINEKA
heart-rate results indicated greater return of fear when follow-up testing occurred in a different context other than where treatment had occurred (relative to when testing occurred in the treatment context). However, this effect (which only occurred with heart rate) was only significant under one of two conditions. In this study, the size and speed of movement of the tarantula used was also varied from treatment to follow-up context for half the participants. For reasons that are unclear, greater return of fear only occurred in a different context when a small, fast-moving tarantula was confronted during the follow-up BAT (but not when a large, slow-moving tarantula was used at follow-up). Unfortunately, selfreport and behavioral dependent measures did not reveal significant effects of context shifts on return of fear. There were, however, two potentially important methodological limitations of the Rodriguez et al. study that may have decreased the likelihood of finding stronger effects of contextual shifts on return of fear. First, all participants had exposure to both experimental contexts (i.e., pretreatment BATs were conducted in contexts A and B for all participants). Consequently, exposure to the tarantula in both contexts during the pretreatment BATs may have served as initial extinction trials in each context; this in turn may have minimized differences between later context shifts from treatment to follow-up because some extinction had already occurred at the outset in the context not used in treatment. Second, the two contexts may not have been sufficiently different. The physical dimensions and appearances of the treatment rooms did not differ dramatically, and both rooms were part of the same laboratory. For example, although the experiment attempted to manipulate context using female experimental therapists with different color lab coats, it did not use gender of the experimental therapist as an additional potentially stronger manipulation of context. In summary, the Rodriguez et al. (1999) investigation may have been limited by a relatively weak context manipulation that decreased the chances of observing a context-based return of fear.
Study 2. Given the limitations observed in the Rodriguez et al. (1999) investigation, Mineka, Mystkowski, Hladek, and Rodriguez (1999) sought to replicate the Rodriguez et al. study using a similar undergraduate spider-fearful sample, but with an improved paradigm. Thirty-six spider-fearful participants all received graded, exposure-based therapy with participant modeling in one of two treatment contexts (A or B), and were followed up 1 week later in either the same treatment context or a new and different context. Again, SUDS and behavioral
A and B
A or B A or B
A(caf) or B (plac) A or B
Rodriguez et al. (1999)
Mineka et al. (1999)
Mystkowski et al. (2002)
Mystkowski et al. (2003)2
Mystkowski et al. (in press)3 A or B
A or B
A or B
A and B
A or B
A or B
Follow-Up Context
SUDS− Behav− HR+ SUDS+ Behav− SUDS+ Behav− HR− SUDS+ Behav− SUDS+ Behav− HR−
RoF Results
12:59 PM
A or B
A or B
A or B
A or B
Treatment Context
3/28/2007
Note. “SUDS+” RoF (i.e., return of fear) corresponds to self-reported fear during the follow-up BAT. “Behav+” and “HR+” RoF correspond to increased behavioral avoidance and heartrate, respectively, when confronting a spider in a different context at the follow-up Behavioral Approach Task (BAT), compared to individuals reassessed in the original treatment context. A minus sign indicates a lack of RoF on a given measure. 1 The pretreatment context refers to where the baseline BAT took place, and is the same context as the treatment context, except for the Rodriguez et al. study, in which both contexts were used for the baseline BAT. 2 The Mystkowski et al. (2003) study used drug state as the treatment context manipulation (i.e., caffeine = caf; placebo = plac), where as the earlier studies manipulated physical context. 3 As described later in the chapter, the Mystkowski et al. (2006) study had half of the participants mentally reinstate the treatment context before engaging in the follow-up BAT.
Pretreatment Context1
Study
TABLE 8.1 Spider Phobia Studies: Designs and Dependent Variables
Chapter 08.qxd Page 205
205
Chapter 08.qxd
206
3/28/2007
12:59 PM
Page 206
MYSTKOWSKI AND MINEKA
avoidance ratings during the behavioral approach tasks (BATs) were the primary dependent measures (heart rate was not obtained in this study). However, the pretreatment behavioral approach task (BAT) only took place in what would become the treatment context for a given participant. Thus, there was no exposure to the tarantula in the second completely novel context before the contextual shift at follow-up to avoid potential contamination of the context manipulation as in Rodriguez et al. (1999). In addition, the Mineka et al. (1999) study increased the distinctiveness of the experimental contexts, by using contexts that differed in more substantial ways than than the ones used in the Rodriguez et al. (1999) study, with the rationale being that a more powerful contextual manipulation might enhance the likelihood of obtaining significant contextual effects on return of fear. Specifically, the 18 participants who underwent a context shift received a different experimental therapist (one male, one female), one of two rooms (small or large) on the same floor of the same building (but only one was in a lab and the other was a small seminar room), and same or different salient visual cues in the therapy context at treatment and follow-up (e.g., different color paintbrushes to move the spider, spider cages, gloves to handle the spider, and varying pictures on the context walls). In addition, the use of a completely novel context as the second context in this study should have enhanced external validity because individuals who encounter a phobic stimulus after therapy are probably at least as likely to do so in novel contexts as in contexts where they have previously encountered their phobic stimulus. For the second session approximately 1 week later, participants were tested for return of fear in either the previously experienced treatment context or in the novel context (see Table 8.1). The results using selfreported fear data as the dependent measure (i.e., SUDS ratings) obtained during the follow-up BAT confirm that treatment participants tested in a novel context at follow-up showed a greater return of fear than participants tested in the same context. Results were not significant with the behavioral approach measure, but this may well have been because of a ceiling effect (i.e., as mentioned earlier, in excess of 95% of the participants in all four groups showed no behavioral avoidance at posttreatment or follow-up BATs, in that they were able to touch the spider with their bare fingertip during these BATs). See Figure 8.1 for the primary results. Although Mineka et al. observed significant effects of contextual shifts on return of fear, it should be noted that the observed effects (etasquared = 0.16) were not as large as those seen in the animal conditioning
Chapter 08.qxd
3/28/2007
12:59 PM
Page 207
8. PHOBIA TREATMENT AND CONTEXT
207
FIGURE 8.1 Mean self-reported fear (SUDs) for each group (measured during the BATs) at pretreatment (Pre-Tx), posttreatment (Post-Tx), and follow-up. From Mineka, Mystkowski, Hladek, and Rodriguez (1999), with permission.
literature. Certain considerations from that basic science literature, as well as others, led us to hypothesize that there may be three potential explanations as to why the effects of context in the Rodriguez et al. (1999) and Mineka et al. (1999) studies were relatively small compared to those seen in animals. First, post hoc power analyses of the Mineka et al. (1999) findings indicated only moderate statistical power (0.67) for finding context effects on return of fear with the design used. Upon reflection we realized that a paradigm incorporating a within-subjects factor would yield greater power to detect contextual shifts in return of fear. Such a design, in which all participants would be tested at follow-up in both the same and different contexts, would also be more similar to the experimental designs sometimes used in animal studies by Bouton and colleagues that have revealed especially large effects (M. E. Bouton, personal communication, 1998; Bouton & Brooks, 1993). A second potential explanation for why the first two studies may have produced relatively small effects may stem from some inherent differences in the human and animal paradigms. In the animal literature, fear is first conditioned in one context (e.g., A), and then extinguished in either context A or B. Renewal effects are later tested in A, B, or C (a novel context). When extinction occurs in context B, renewal effects are generally stronger when the animal is tested for return of fear in the
Chapter 08.qxd
208
3/28/2007
12:59 PM
Page 208
MYSTKOWSKI AND MINEKA
context in which the fear was originally conditioned (A) than when tested in a new context (C) (Bouton & Brooks, 1993; Harris, Jones, Bailey, & Westbrook, 2000). This is not surprising given that there should be more direct competition between memories for fear acquisition and fear extinction in the original context than in a novel context. However, the equivalent of this paradigm is not possible in human treatment research, such as in the studies presented here, because the original context(s) in which fear learning occurred (A) is generally not known and/or not available during treatment. Moreover, testing for fear renewal in novel contexts (or at least ones not associated with the original learning) has greater external validity in terms of the situations individuals undergoing exposure treatments are likely to encounter in the real world following treatment. Third and finally, although Mineka et al. (1999) attempted to make the two contexts quite distinctive, we realized that the two contexts may not have been functionally different from the standpoint of the participants. Here it is important to consider that what researchers designate to be physically distinctive contexts may not be what the participants themselves consider as the most salient contextual cues. For example, with both physically distinctive contexts, the participants had to participate in the same psychology experiment (which may be functionally the most salient context to the participant) on the same floor of the same building on campus, and with another undergraduate experimental therapist (albeit of a different gender and color of clothing). In addition, participants also came to the same instruction room at both treatment and follow-up, where similar questionnaire measures were administered. It is possible, therefore, that making the two contexts more functionally different for the participants might increase the magnitude of the context effects obtained. Two issues seem particularly relevant in the choice of contexts used in the studies discussed thus far that may help to explain their somewhat limited findings. One possible issue is that some contexts may have a larger degree of relatedness or fit with fear of certain stimuli than do other contexts, somewhat similar to the notion of conditioned stimulus (CS)–unconditioned stimulus (US) belongingness or preparedness in the acquisition of fears or phobias (e.g., Hamm, Vaitl, & Lang, 1989; Öhman & Mineka, 2001). As such, one might expect that fear would be higher in some settings than others, or that inherent differences in the nature of the contexts might affect treatment outcome or return of fear (M. E. Bouton, personal communication, 2003). For example, outdoor settings may be
Chapter 08.qxd
3/28/2007
12:59 PM
Page 209
8. PHOBIA TREATMENT AND CONTEXT
209
more fear-provoking than indoor settings because animals such as spiders belong more outdoors than indoors. This might lead to greater return of fear in outdoor settings. Alternatively, the use of indoor and outdoor setting might not differentially affect fear levels per se, but rather simply provide greater distinctiveness of contexts. The second issue to consider regarding the choice of contexts used in the aforementioned studies is whether some contexts outshine or overshadow others. Specifically, according to Smith’s (1988) outshining hypothesis, the impact of memory-enhancing cues is weakened by the presence of other competing cues during testing. Conceivably, starting the follow-up session in the same instruction room with similar questionnaires before testing in either context A or B, as was done in the Mineka et al. (1999) study, could dominate the perception of context and outshine or overshadow the other novel cues that subsequently followed for the groups tested in the novel context. Thus, it remains quite possible that if participants were not given most of the questionnaires and study instructions/details in one common instruction room (which is different than either context), the effects of context on return of fear might be greater than those observed in the first two studies.
Study 3. With several of these possibilities in mind, Mystkowski, Craske, and Echiverri (2002) further examined whether participants have more return of fear at follow-up when tested in a novel versus familiar treatment context. To address the possible limitations of the studies described already, the differences between the two experimental contexts were increased by using an outside context in addition to the standard indoor laboratory context. This also allowed examination of whether a context that may belong more with the feared spider stimulus would affect fear extinction and/or fear renewal effects. Additionally, as in the Mineka et al. (1999) study, each context was differentiated by the gender of the experimenter and salient visual cues (i.e., color of treatment materials and therapist lab coats). Moderately to severely spider-fearful undergraduates participated in a two-session study. For the first session, 46 participants received graded, exposure-based therapy in either an inside (A) or outside (B) context (see Table 8.1). The outside context was an outdoor patio setting, in which trees surrounded a secluded, concrete path that had a spider cage on a table at the end of a 10-foot path. One week later, participants were tested for return of fear in both the original treatment context and the different context, using a mixed design.
Chapter 08.qxd
210
3/28/2007
12:59 PM
Page 210
MYSTKOWSKI AND MINEKA
Participants treated in either A or B were then first followed up 1 week later in either A or B, and then were followed up a second time immediately afterward in B or A, in counterbalanced order. Thus, there were four groups: Group A–A/B; Group B–B/A; Group B–B/A, and Group B–A/B (the first letter indicates which context the treatment occurred in, and the second and third letters indicate where and in what order the two follow-up tests occurred; see Table 8.1). As in Rodriguez et al. (1999), Mystkowski et al. (2002) measured fear during the BATs with selfreported SUDS ratings, behavioral avoidance, and heart rate. Fear measurements were taken in the treatment context at pretreatment and posttreatment, and in both the treatment and different contexts at follow-up, in counterbalanced order. To address the possibility that the instruction context might have overshadowed a novel context at follow-up, all self-report fear measures were administered in the original treatment context throughout the study. In agreement with Mineka et al. (1999), self-report SUDS results reveal that a return of fear was more likely to occur when an individual was confronted with a previously feared stimulus in a different context than in the treatment/extinction context. Interestingly, Mystkowski et al. (2002) found no support for the idea that that the outdoor context belonged together better with the fear stimulus. That is, neither treatment nor followup results were affected by the use of an outdoor versus an indoor context. Finally, with respect to the limitations in statistical power found in the prior studies, the results obtained by Mystkowski et al. (2002) demonstrated a somewhat larger effect size for return of fear on self-reported SUDS (eta-squared = 0.29), compared to the Mineka et al. (1999) study using the same measure. Specifically, fear levels that had been greatly reduced during treatment increased by about 30% when follow-up testing occurred in a different context, but by only about 7.5% when testing occurred in the same context. These results are illustrated in Figure 8.2. DRUG STATE AS INTERNAL CONTEXT Contexts can vary in ways other than just an individual’s physical surroundings. Indeed, internal states can also serve as contexts, as has long been known in the demonstration of drug-state-dependent learning (e.g., Cunningham, 1979; Overton, 1978; Shulz, Sosnik, Ego, Hairdarliu, Ahissar, 2000; Slot & Colpaert, 1999). In the general case of drug-statedependent learning, material learned while under one drug state (e.g., active
Chapter 08.qxd
3/28/2007
12:59 PM
Page 211
8. PHOBIA TREATMENT AND CONTEXT
211
FIGURE 8.2 Mean self-reported fear (SUDs) for each group from the pretreatment to posttreatment, and the two follow-up BATs (same and different). From Mystkowski, Craske, and Echiverri (2002), with permission.
drug vs. placebo) is better retained when subsequent retention tests are administered in the same drug state (drug–drug or placebo–placebo) than when retention tests are administered in the different drug state (drug–placebo or placebo–drug). With regard to internal context specificity of extinction, several studies with rats indicate that renewal of fear following extinction is influenced by manipulation of internal states induced by benzodiazepines such as diazepam or alprazolam, versus placebo. Specifically, fear renewal is greater when fear retesting at follow-up occurs under the influence of the other drug state, rather than the same drug state in which extinction occurred (e.g., Bouton, Kenney, & Rosengard, 1990). The Bouton et al. results using diazepam and placebo in rats thus suggested that a mismatch of internal states experienced during treatment and follow-up can lead to significantly greater return of fear just as mismatches between physical environmental cues do (Bouton & Swartzentruber, 1991). Generalizing to humans with clinically diagnosed anxiety disorders, the widespread use of benzodiazepine medications during exposure therapy for anxiety disorders should be expected to lead to greater return of fear (or relapse) when drugs are discontinued than when drugs were never administered during treatment. Results showing this have indeed been
Chapter 08.qxd
212
3/28/2007
12:59 PM
Page 212
MYSTKOWSKI AND MINEKA
gathered by Marks et al. (1972, 1993) and Otto, Pollack, and Sabatino (1996). In these cases, a mismatch occurs from the drug state present during anxiety treatment and the drug state present following active treatment. Although these results are highly suggestive of drug-state-dependent learning effects, clinical studies on this issue generally do not test all four of the conditions necessary to unequivocally demonstrate such effects. That is, clinical participants typically receive exposure therapy with or without benzodiazepine treatment, and all are followed up off medication as occurs in clinical practice. In other words, there are no groups treated on placebo and followed up under the influence of benzodiazepines. Such groups would be required to know that the interference demonstrated in the benzodiazepine/placebo condition is not simply due to benzodiazepines having interfered with the learning that occurs during exposure treatment (rather than true state-dependent learning). An initial analogue clinical study by Zoellner and Craske (1998) did test all four conditions using drug state as a context variable for treatment of spider fear. Drug state was manipulated by conducting exposure therapy, as well as a later follow-up session, with or without alprazolam (i.e., alprazolam or placebo during exposure, and alprazolam or placebo during followup testing). However, the authors found no effect of shifting from a drug to a no-drug state, or vice versa. One likely possible account of these null results is that the dosages of alprazolam (0.25 mg) used in the Zoellner and Craske (1998) study may not have generated a sufficiently salient context.
Study 4. Nevertheless, evidence for some internal context specificity of extinction was demonstrated in another analogue clinical study that manipulated drug state through caffeine versus placebo ingestion for individuals who were highly fearful of spiders (Mystkowski, Mineka, Vernon, & Zinbarg, 2003). Forty-three participants were first assessed using a behavioral approach task following which they ingested one of two solutions in a double-blind manner, using an orange-flavored drink (Tang) to disguise the flavor of the two solutions: caffeine citrate and flat quinine water (placebo). (The caffeine dose was substantial—the equivalent of about three cups of coffee consumed quickly in one glass of the orange-flavored drink.) Thirty minutes later (allowing time for the caffeine to take effect), a second BAT was administered to assess possible effects of caffeine on fear levels. Following the second BAT, graded exposure-based therapy as described in prior studies was conducted, followed by a third posttreatment BAT. At follow-up, 1 week later, participants first ingested a drink mixture (caffeine or placebo) that was either the same as or different from the drink ingested
Chapter 08.qxd
3/28/2007
12:59 PM
Page 213
8. PHOBIA TREATMENT AND CONTEXT
213
FIGURE 8.3 Mean self-reported fear (SUDs) for each group from the two pretreatment BATs (before and after caffeine/placebo ingestion) to posttreatment and follow-up BATs. From Mystkowski, Mineka, Vernon, and Zinbarg (2003), with permission.
during the treatment session (see Table 8.1). (Thus, drink status was completely crossed during the extinction and testing sessions.) Approximately 30 minutes later, they received a fourth follow-up BAT. As predicted, participants experiencing incongruent drug states exhibited a small but significantly greater self-reported return of fear (i.e., SUDS ratings), measured during a behavioral approach task, from posttreatment to follow-up, than those participants experiencing congruent drug states during treatment and follow-up (eta-squared = 0.11). Thus, the basic prediction that changes in internal state (at least with caffeine) can affect return of fear was confirmed. Interestingly the effects were comparable when drink status changed from placebo to caffeine as when it changed from caffeine to placebo. The primary results are presented in Figure 8.3. RETURN OF FEAR USING EXTINCTION CUES AT RETENTION TESTING The studies discussed thus far have demonstrated that changing external, as well as internal, contexts increases return of fear when participants are confronted with an unfamiliar physical surrounding or drug state at
Chapter 08.qxd
214
3/28/2007
12:59 PM
Page 214
MYSTKOWSKI AND MINEKA
follow-up. However, none of these controlled studies has examined possible ways of counteracting a contextually based return of fear. Preventing or minimizing fear renewal in new or unfamiliar situations could have farreaching implications for enhancing the treatment of anxiety disorders, which can sometimes pose a chronic and costly struggle for individuals who have them. In the animal conditioning literature, there is evidence that supports the use of physical retrieval cues for attenuating renewal effects (e.g., Brooks & Bouton, 1994). Using an appetitive conditioning paradigm, Brooks and Bouton (1994) first conditioned rats to eat food pellets (US) when preceded by a tone (CS) in context A. They then extinguished that CS in a new context (B), with or without the presence of what would become a retrieval cue (L or N: light or noise) in the followup retention phase. The results reveal that retrieval cues presented during retention testing of the tone CS in the original acquisition context (A), or in a new context (C), decreased the magnitude of the spontaneous recovery of food eating CRs relative to the magnitude seen when no retrieval cues were present during retention testing (Brooks & Bouton, 1994). A second reason that further study of retrieval cues is worthwhile is that Collins and Brandon (2002) also showed the expected beneficial effects of retrieval cues in a group of binge drinkers who underwent an extinction procedure for conditioned craving. In this study, visual and olfactory retrieval cues present during extinction of craving (and salivary secretion) to alcohol-related cues (such as the sight of a can of beer) later reliably reduced the effects of context shifts on retention of extinction of these cravings during a follow-up test. Thus, perhaps the use of retrieval cues in studies of fear renewal could be effective with a more robust conditioned fear response that takes longer to extinguish, and/or with the use of more naturalistic retrieval cues (e.g., olfactory, as in Collins & Brandon, 2002). Nevertheless, the use of visual or even olfactory retrieval cues in facilitating the retention of fear extinction memories across contexts is unlikely to be very practical in humans who have undergone exposure therapy for their anxiety disorders. This is because such visual or olfactory cues are not likely to be portable, and therefore are likely to be unavailable at the times they are needed when previously feared objects or situations are encountered in new contexts. An alternative and perhaps more promising idea would be to use mental reinstatement of extinction cues because such techniques would be highly portable.
Chapter 08.qxd
3/28/2007
12:59 PM
Page 215
8. PHOBIA TREATMENT AND CONTEXT
215
EFFECTS OF MENTAL REINSTATEMENT OF CONTEXTUAL CUES FROM EXTINCTION ON RETURN OF FEAR Just as there is individual variability in the magnitude of fear extinction effects, so too is there individual variability in the magnitude of fear renewal or recovery effects. One potential reason some people may show minimal return of fear even following a contextual shift is because they mentally reinstate the treatment context. For example, some people undergoing cognitive-behavioral therapy for panic disorder have stated that they coped with unexpected panic symptoms or highly frightening situations outside of the treatment context by imagining what their therapist would have said, or by imagining the therapist with them in difficult situations (Craske, 1999). Therefore, perhaps some people mentally reinstate extinction contexts where their exposure therapy had occurred through the use of imagery (e.g., the therapist, treatment information, and/or the physical surroundings where treatment took place). This use of imagery could potentially override veridical context shifts, helping the person deal with new or unfamiliar settings in which they confront anxiety-provoking stimuli (Craske, 1999). Until recently, however, the possibility that mental reinstatement of context might minimize a contextually based return of fear has not been studied systematically. Nevertheless, there are a substantial number of potentially relevant studies on how mental reinstatement of context enhances retrieval of environmental context (EC)-dependent verbal memory. Research on EC-dependent memory has indicated that verbal learning that takes place in one context is better recalled when tested in the original encoding context than when tested in a new or unfamiliar environmental context (i.e., physical reinstatement; Smith, 1979, 1988). Smith (1979) concluded that EC-dependent memory occurs as a result of contextual associations to the original learning context that are not available when tested in a different context. However, Smith also found that familiar visual retrieval cues from the original learning context may not be sufficient to counteract EC memory effects when simply presented passively during learning and retention (Smith, 1979). Instead, Smith hypothesized that what may be required is a cognitive strategy through which the individuals actively reinstate the original learning context (Smith, 1979). Thus, if participants in an unfamiliar environment are instructed to recall the original learning environment just prior to free recall of a list of words in an unfamiliar
Chapter 08.qxd
216
3/28/2007
12:59 PM
Page 216
MYSTKOWSKI AND MINEKA
environment, a release from contextual dependence is observed, and performance is identical to that of participants tested in the original learning environment (Smith, 1979). Since Smith’s (1979) original investigation, other studies have verified that EC-dependent verbal memory, tested by cued and free recall, is enhanced when participants are asked to recall the original learning context when confronted with a novel test context (e.g., Fisher, Geiselman, Holland, & MacKinnon, 1984; Smith, 1984).
Study 5. Based on such findings from the animal conditioning and cognitive psychology literatures, Mystkowski, Craske, Echiverri, and Labus (2006) sought to investigate whether a contextually based return of fear could also be counteracted via mental reinstatement when participants were confronted with a novel context. As in Mystkowski et al. (2002), two groups of spider-fearful participants were treated and followed up in the same context, and two groups were treated and followed up in different contexts (one context was inside and the other was outside; the total sample was 48). Half of the participants in each of the groups were instructed to mentally reinstate the treatment context and the material learned in that context before they entered the test context (same or different) at follow-up; the other half of the participants in each of the four groups were asked to recall a neutral (treatment-unrelated) scenario before entering the test context at follow-up (see Table 8.1). Fear measurements included subjective, behavioral, and physiological measures taken during BATs before and after treatment and at follow-up. As in prior studies, the participants treated and followed up in the same context were expected to display less return of fear than the participants treated and followed up in different contexts. Second, it was hypothesized that participants given treatment-related mental reinstatement instructions (denoted by a “+” sign after the group name; please refer to Fig. 8.4) prior to follow-up testing in a different context (DIFF+) would show less return of fear than those given treatment-unrelated instructions (denoted by a “−” sign following the group name) prior to follow-up testing in a different context (DIFF–). Thus, the DIFF+ group was expected to show less return of fear than the DIFF– group. Moreover, if treatment-related mental reinstatement instructions were sufficiently powerful, it was expected that participants in the DIFF+ group with treatment-related mental reinstatement would show no greater return of fear than those followed up in the SAME context without treatment-related instructions (i.e., group SAME–).
Chapter 08.qxd
3/28/2007
12:59 PM
Page 217
8. PHOBIA TREATMENT AND CONTEXT
217
FIGURE 8.4 Mean self-reported fear (SUDs) for each group from the pretreatment to posttreatment and follow-up BAT. A plus or minus after the group name corresponds to the type of mental reinstatement instruction: (+) = treatmentrelated mental reinstatement and (-) = treatment-unrelated mental reinstatement. From Mystkowski, Craske, Echiverri, and Labus (2006), with permission.
The results of this study did replicate and extend recent studies examining the effects of changing contexts on enhancing return of fear (eta-squared = 0.28) in humans discussed earlier (e.g., Mineka et al., 1999; Mystkowski et al., 2002; Mystkowski et al., 2003), as well as in animals (e.g., Bouton, 1993; Bouton & Nelson, 1998; Bouton et al., in press). More important, however, the results of this study also indicate, as predicted, that mental reinstatement of the treatment context attenuated selfreported return of fear in participants who were treated and retested in different contexts (i.e., the DIFF+ group showed less return of fear than the DIFF– group; eta-squared = 0.19). In addition, mental reinstatement of context reduced fear among participants confronted with a different context at follow-up to a level comparable to participants who were retested in the same treatment context, without treatment-related mental reinstatement instructions. That is, participants with matched contexts across sessions without mental reinstatement of the treatment context (i.e., group SAME–) were equivalent to participants with mismatched contexts across sessions with mental reinstatement of the treatment context (DIFF+). Interestingly enough, although not predicted, results also demonstrate that mental reinstatement of context enhanced the effect of treatment (i.e., promoted further fear reduction) even for participants
Chapter 08.qxd
218
3/28/2007
12:59 PM
Page 218
MYSTKOWSKI AND MINEKA
with a context match from posttreatment to follow-up (i.e., the SAME + group showed less return of fear than the SAME– group). These exciting results are the first of their kind in the fear extinction literature. Thus, they extend the work in cognitive psychology on ECdependent memory and mental reinstatement of context to emotional learning in humans, namely, the treatment of human fears and anxiety (e.g., Bjork & Richardson-Klavehn, 1989; Smith, 1979, 1984, 1988). Although these results were robust, they are obviously in need of replication. Nevertheless, they do provide promising support for the clinical observations that some clients who have undergone exposure therapy report that they seem to spontaneously use mental reinstatement of the context and/or the therapists’ instructions to help them in situations where they begin to get anxious. CONCLUSIONS In this chapter, we reviewed the results of five studies on the effects of contextual changes on increasing the return of fear in individuals who have undergone exposure therapy for moderately severe spider fear. Each of these studies was inspired by theory and findings from the basic conditioning literature (as well as from the verbal memory literature for Study 5). Many of the implications of this literature have not yet been tested in the treatment of human fear and anxiety. Unfortunately, however, some of the clinical intuitions that might seem to derive from such results become somewhat less promising when one closely examines the most recent animal literature on this topic, suggesting the need for continued interplay between these basic and clinical literatures. For example, at first glance, the implications of these studies seem to suggest that one way to decrease the effects of contextual change on enhancing return of fear would be to conduct exposure therapy in multiple contexts so that extinction memories can be cued by multiple contexts. However, examination of the animal literature on this topic reveals that the results of the studies testing such ideas in animals have been inconsistent (see Bouton et al., 2006, for a review). Some studies have found that conducting extinction in multiple contexts has reduced fear renewal in a final novel context relative to conducting extinction in only one context, but other studies have revealed no differences. In other words, “extinction in multiple contexts is not a ‘magic bullet’ that prevents the renewal effect” (Bouton et al., 2006, p. 189). Another possible
Chapter 08.qxd
3/28/2007
12:59 PM
Page 219
8. PHOBIA TREATMENT AND CONTEXT
219
example of clinical intuitions that one could derive from our results might be that counterconditioning treatments might be more effective than exposure-based treatments in mitigating against the effects of contextual shifts on return of fear. This is because in counterconditioning (somewhat similar to Wolpe’s systematic desensitization) a new response may be conditioned. However, examination of the animal literature shows that the effects of context shifts following counterconditioning on fear renewal are still robust (see Bouton et al., 2006, for a review). Although different findings might be obtained in human studies than in these two examples, the animal work nevertheless suggests caution in drawing any quick conclusions. Animal research devoted to this topic has found that nearly all attempts to make extinction less context specific have not had substantial or consistent effects (Bouton et al., 2006). Although the current studies suggest that the effects of context shifts on human fear extinction may not be as large as in animals, they have been demonstrated consistently in four studies. Perhaps human researchers should build on Bouton’s suggestions that more promising lines of work may involve developing bridges between treatment and follow-up contexts. In this regard, the results of Study 5 (Mystkowski et al., 2006) on the effects of mental reinstatement may have some of the strongest implications for the best ways to proceed with humans undergoing exposure therapy. We encourage others interested in this topic to pursue these and other promising leads in the spirit of Richard McFall’s call for continuing to pursue the integration of basic and clinical science. REFERENCES Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191–215. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall. Bandura, A., Blanchard, E., & Ritter, B. (1969). Desensitization and modeling to induce behavioral, affective, and attitudinal change. Journal of Personality and Social Psychology, 13, 173–179. Bjork, R. A., & Richardson-Klavehn, A. (1989). On the puzzling relationship between environmental context and human memory. In C. Izawa (Ed.), Current issues in cognitive processes: The Tulane Flowerree Symposium of Cognition (pp. 313–344). Hillsdale, NJ: Lawrence Erlbaum Associates. Bouton, M. E. (1991). A contextual analysis of fear extinction. In P. R. Martin (Ed.), Handbook of behavior therapy and psychological science: An integrative approach (pp. 435–453). Elmsford, NY: Pergamon.
Chapter 08.qxd
220
3/28/2007
12:59 PM
Page 220
MYSTKOWSKI AND MINEKA
Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian learning. Psychological Bulletin, 114, 80–99. Bouton, M. E. (1994). Conditioning, remembering, and forgetting. Journal of Experimental Psychology: Animal Behavior Processes, 20, 219–231. Bouton, M. E. (2002). Context, ambiguity, and unlearning: Sources of relapse after behavioral extinction. Biological Psychiatry, 52, 976–986. Bouton, M. E., & Brooks, D. C. (1993). Time and context effects on performance in a Pavlovian discrimination reversal. Journal of Experimental Psychology: Animal Behavior Processes, 19, 165–179. Bouton, M. E., Kenney, F. A., & Rosengard, C. (1990). State-dependent fear extinction with two benzodiazepine tranquilizers. Behavioral Neuroscience, 104, 44–45. Bouton, M. E., & King, D. A. (1983). Contextual control of the extinction of conditioned fear: Tests for the associative value of the context. Journal of Experimental Psychology: Animal Behavior Processes, 9, 248–265. Bouton, M. E., & Nelson, J. B. (1998). The role of context in classical conditioning: Some implications for cognitive behavior therapy. In W. T. O’Donohue (Ed.), Learning theory and behavior therapy (pp. 59–83). Boston: Allyn & Bacon. Bouton, M. E., & Swartzentruber, D. (1991). Sources of relapse after extinction in Pavlovian and instrumental learning. Clinical Psychology Review, 11, 123–140. Bouton, M. E., Woods, A. M., Moody, E. W, Sunsay, C., & Garcia-Gutierrez, A. (2006). Counteracting the context-dependence of extinction: Relapse and some tests of possible methods of relapse prevention. In M. G. Craske, D. Hermans, & D. Vansteenwegen (Eds.), Fear and learning: Basic science to clinical application (pp. 175–196). Washington, DC: APA Books. Brooks, D. C., & Bouton, M. E. (1994). A retrieval cue for extinction attenuates response recovery (renewal) caused by a return to the conditioning context. Journal of Experimental Psychology: Animal Behavior Processes, 20, 366–379. Collins, B. N., & Brandon, T. H. (2002). Effects of extinction context and retrieval cues on alcohol cue reactivity among nonalcoholic drinkers. Journal of Consulting and Clinical Psychology, 70, 390–397. Craske, M. G. (1999). Anxiety disorders: Psychological approaches to theory and treatment. Boulder, CO: Westview. Craske, M. G., & Rowe, M. (1997). A comparison of behavioral and cognitive treatments of phobias. In G. Davey (Ed.), Phobias: A handbook of theory, research, and treatment (pp. 247–280). Chichester, England: Wiley. Cunningham, C. L. (1979). Alcohol as a cue for extinction: State dependency produced by conditioned inhibition. Animal Learning & Behavior, 7, 45–52. Fisher, R. P., Geiselman, R. E., Holland, H. L., & MacKinnon, D. P. (1984). Hypnotic and cognitive interviews to enhance memory of eyewitnesses to crime. International Journal of Investigative and Forensic Hypnosis, 7, 28–31. Foa, E., & Kozak, M. (1986). Emotional processing of fear: Exposure to corrective information. Psychological Bulletin, 99, 20–35. Gleitman, H., Nachmias, J., & Neisser, U. (1954). The S-R reinforcement theory of extinction. Psychological Review, 61, 23–33.
Chapter 08.qxd
3/28/2007
12:59 PM
Page 221
8. PHOBIA TREATMENT AND CONTEXT
221
Hamm, A. O., Vaitl, D., & Lang, P. J. (1989). Fear conditioning, meaning, and belongingness: A selective association analysis. Journal of Abnormal Psychology, 98, 395–406. Harris, J. A., Jones, M. L., Bailey, G. K., & Westbrook, F. R. (2000). Contextual control over conditioned responding in an extinction paradigm. Journal of Experimental Psychology: Animal Behavior Processes, 26, 174–185. Klorman, R., Hastings, J. E., Weerts, T. C., Melamed, B. G., & Lang, P. J. (1974). Psychometric description of some specific-fear questionnaires. Behavior Therapy, 5, 401–409. Marks, I. M., Swinson, R. P., Basoglu, M., Kuch, K., Noshirvani, H., & O’Sullivan, G. (1993). Alprazolam and exposure alone and combined in panic disorder with agoraphobia: A controlled study in London and Toronto. British Journal of Psychiatry, 162, 776–787. Marks, I. M., Viswanathan, R., Lipsedge, M. S., & Gardner, R. (1972). Enhanced relief of phobias following flooding during waning diazepam effect. British Journal of Psychiatry, 121, 493–505. Mathews, A. M., Gelder, M. G., & Johnston, D. W. (1981). Agoraphobia: Nature and treatment. New York: Guilford. Mineka, S. (1979). The role of fear in theories of avoidance learning, flooding and extinction. Psychological Bulletin, 86, 986–1010. Mineka, S. (1985). Animal models of anxiety-based disorders: Their usefulness and limitations. In A. H. Tuma & J. Maser (Eds.), Anxiety and the anxiety disorders (pp. 199–244). Hillsdale, NJ: Lawrence Erlbaum Associates. Mineka, S., Mystkowski, J., Hladek, D., & Rodriguez, B. (1999). The effects of changing contexts on return of fear following exposure treatment for spider fear. Journal of Consulting and Clinical Psychology, 67, 599–604. Mineka, S., & Thomas, C. (1999). Mechanisms of change during exposure treatments for anxiety disorders. In T. Dalgleish & M. Power (Ed.), Handbook of cognition and emotion (pp. 747–764). Chichester, England: Wiley. Mineka, S., & Zinbarg, R. (1996). Conditioning and ethological models of anxiety disorders: Stress-in-Dynamic-Context Anxiety models. In D. Hope (Ed.), Perspectives on anxiety, panic, and fear. 43rd Annual Nebraska Symposium on Motivation (pp. 135–211). Lincoln: University of Nebraska Press. Mineka, S., & Zinbarg, R. E. (2006). A contemporary learning theory perspective on the etiology of anxiety disorders: It’s not what you thought it was. American Psychologist, 61, 10–26. Mystkowski, J. L., Craske, M. G., & Echiverri, A. M. (2002). Treatment context and return of fear in spider phobia, Behavior Therapy, 33, 399–416. Mystkowski, J. L., Craske, M. G., Echiverri, A. M., & Labus, J. S. (2006). Mental reinstatement of context and return of fear in spider fearful participants. Behavior Therapy, 37, 49–60. Mystkowski, J. L., Mineka, S., Vernon, L. L., & Zinbarg, R. E. (2003). Changes in caffeine state enhance return of fear in spider phobia. Journal of Consulting and Clinical Psychology, 71, 243–250. Öhman, A., & Mineka, S. (2001). Fears, phobias, and preparedness: Toward an evolved module of fear and fear learning. Psychological Review, 108, 483–522.
Chapter 08.qxd
222
3/28/2007
12:59 PM
Page 222
MYSTKOWSKI AND MINEKA
Öst, L. G. (1997). Rapid treatment of specific phobias. In G. Davey (Ed.), Phobias: A handbook of theory, research, and treatment (pp. 227–246). Chichester, England: Wiley. Otto, M. W., Pollack, M. H., & Sabatino, S. A. (1996). Maintenance of remission following cognitive behavioral therapy for panic disorder: Possible deleterious effects of concurrent medication treatment. Behavior Therapy, 27, 473–482. Overton, D. A. (1978). Major theories of state dependent learning. In B. T. Ho, D. W. Richards III, & D. L. Chute (Eds.), Drug discrimination and state dependent learning (pp. 283–318). New York: Academic Press. Pavlov, I. P. (1927). Conditioned reflexes. London: Oxford University Press. Rachman, S. (1989). The return of fear: Review and prospect. Clinical Psychology Review, 9, 147–168. Rachman, S. (1990). Fear and courage (2nd ed.). New York: W. H. Freeman. Rachman, S., & Hodgson, R. (1980). Obsessions and compulsions. Englewood Cliffs, NJ: Prentice-Hall. Rodriguez, B. I., Craske, M. G., Mineka, S., & Hladek, D. (1999). Context-specificity of relapse: Effects of therapist and environmental context on return of fear. Behaviour Research and Therapy, 37, 845–862. Ross, A. O. (1985). To form a more perfect union: It is time to stop standing still. Behavior Therapy, 16, 195–204. Shulz, D. E., Sosnik, R., Ego, V., Hairdarliu, S., & Ahissar, E. (2000). A neuronal analogue of state-dependent learning. Nature, 403, 549–552. Slot, L. A., & Colpaert, F. C. (1999). Recall rendered dependent on opiate state. Behavioral Neuroscience, 113, 337–344. Smith, S. M. (1979). Remembering in and out of context. Journal of Experimental Psychology: Human Learning and Memory, 5, 460–471. Smith, S. M. (1984). A comparison of two techniques for reducing context-dependent forgetting. Memory & Cognition, 12, 477–482. Smith, S. M. (1988). Environmental context-dependent memory. In G. M. Davies & D. M. Thomson (Eds.), Memory in context: Context in memory (pp. 13–34). New York: Wiley. Watson, J. B., & Rayner, R. (1920). Conditioned emotional reactions. Journal of Experimental Psychology, 3, 1–14. Williams, S. L. (1996). Therapeutic changes in phobic behavior are mediated by changes in perceived self-efficacy. In R. Rapee (Ed.), Current controversies in the anxiety disorders (pp. 344–368). New York: Guilford. Wolpe, J. (1958). Psychotherapy by reciprocal inhibition. Stanford, CA: Stanford University Press. Wolpe, J. (1973). The practice of behavior therapy. New York: Pergamon. Zoellner, L. A., & Craske, M. G. (1998, November). Contextual effects of alprazolam on exposure therapy. Poster presented at the annual conference of the Association for Advancement of Behavior Therapy.
Chapter 09.qxd
3/28/2007
12:59 PM
Page 223
9 Assessment of Mental Architecture in Clinical/Cognitive Research James T. Townsend and Mario Fific′ Indiana University
Richard W. J. Neufeld University of Western Ontario
The nature of mental structures and perceptual and cognitive processing modes has long been of concern in clinical psychology. However, in recent years, there has been a movement toward more rigorous descriptions and even predictions within a cognitive setting (e.g., Granholm, Asarnow, & Marder, 1996a, 1996b). In this chapter, we consider the strategic issue of mental architecture. Mental architecture refers to the organization of a set of mental processes. Two special cases of great importance are parallel processing, which means the simultaneous processing of items, and serial processing, which means the sequential and nonoverlapping processing of items. For instance, whether certain syndromes cause or are associated with a change from parallel to serial processing has often been a question of interest to clinical scientists (e.g., Knight, Manoach, Elliott, & Hershenson, 2000; Magaro, 1983). Mental architecture is one of a set of critical issues contained in our general theoretical approach (e.g., Townsend, 1974; Townsend & Ashby, 1983; Townsend & Wenger, 2004a). Processing refers to some perceptual or cognitive operation such as 223
Chapter 09.qxd
224
3/28/2007
12:59 PM
Page 224
TOWNSEND, FIFIC′ , NEUFELD
search, comparison, evaluation, or the like. More specific examples are given later. As an apt example that relates to mental architecture, the dichotomy of automatic versus controlled processing is virtually omnipresent throughout cognitive science (e.g., see review by Shiffrin, 1988) and has been widely employed in clinical science (e.g., Carter, Robertson, Chaderjian, Celaya, & Nordahl, 1992; Hartlage, Alloy, Vázquez, & Dykman, 1993). Roughly, automatic processing is assumed to be effortless, requiring little or no attention, whereas controlled processing is assumed to be effortful with high attentional demands. In addition, all would agree that automatic processing is parallel, but whether controlled processing must be serial, rather than an inefficient form of parallel processing, would in recent times be more controversial. Now as Shiffrin (1988) intimates, it would be impossible to render the notion rigorous at the level of generality that is commensurate with its ubiquity. Nonetheless, in delimited settings it can often be interpreted in a rigorous and even mathematical fashion. In fact, we show in a later section that automaticity should be characterized, in each experimental milieu, in terms of the other critical issues as well as the parallel versus serial distinction. Although we treat the theme of mental architecture in a relatively general way, we subsequently briefly indicate the relationship of our developments to the notion of automaticity. We stress that the major dependent variable with which we labor in this exposition is that of response times. The implementation of quantitative signatures of mental architecture in clinical studies is much in the spirit of integrative psychological science, a movement whose most forceful and articulate proponent has been Richard McFall. His own work, in collaboration with students and coworkers, has exemplified the merits of such synthesis (e.g., Treat, McFall, Viken, & Kruschke, 2001; Treat et al., 2002). Dick has indicted an excessive reliance in clinical cognitive science on assemblies of offthe-shelf measures, or tasks contrived essentially according to clinical hunch, in lieu of choice cognitive-science developments, especially formal versions. He has forcefully taken the discipline to task for what has often amounted to reinvention of the clinical-science wheel, at best, or discharging its scientific mandate with compromised measurement methods, at worst. Throughout, Dick has put the welfare of the ultimate consumer of clinical science’s offerings, specifically clients with problems in living,
Chapter 09.qxd
3/28/2007
12:59 PM
Page 225
9. ASSESSMENT OF MENTAL ARCHITECTURE
225
first and foremost. Practice based on the best psychological science has to offer, and impelled to prove itself in the court of outcome research (witness the prominence of evidence-based practice), has found its most formidable advocate in Dick McFall. IMPORTANCE OF DISCERNING STATUS OF MENTAL ARCHITECTURE IN RELATION TO PSYCHOPATHOLOGY Evaluative reviews of the literature on applications of cognitive psychology’s information-processing models, most prominent in schizophrenia research, have concluded that these models provide a valid framework for interpreting performance deviations (e.g., Neufeld & Broga, 1981). It follows that the structure of processing systems deemed to bear on cognitive tasks is tantamount to a faculty that is spared with the advent of disorder. Such conclusions, however, have been based on verbal conjecture, rather than mathematically derived diagnostics of processing-system design. The importance of architectural aspects of processing in clinical cognitive science beckons the use of contemporary quantitative signatures, whose paradigms in principle can be appropriated in clinical studies (e.g., Neufeld & McCarty, 1994; Vollick, 1994). Integrity of mental architecture is of obvious interest in its own right. It is important to know if psychopathology impinges on the usual operation of processing structure, including its apparent adaptation to selected variations in task composition (Townsend & Fific, 2004). As intimated already (and further elaborated later in this chapter), cognitive architecture is but one component of the automatic-controlled processing construct, ubiquitously invoked in clinical studies. Any one or some combination of this construct’s components may effect changes in observed performance. It therefore becomes important to ascertain cognitive architecture’s contribution to performance deviations through methods isolating the design of the processing system. If evidently unaltered, for example, other sources can be scrutinized with greater confidence in assumed architectural intactness. Alternatives include overall processing capacity (Neufeld, Townsend, & Jetté, in press) and its parametric constituents (Neufeld, Carter, Vollick, Boksman, Levy, & Jetté, in press). Evidence bearing on cognitive architecture furthermore is important to complementing analyses of cognitive performance, when a certain architecture (e.g., parallel, serial, or hybrid) is purported to prevail among
Chapter 09.qxd
226
3/28/2007
12:59 PM
Page 226
TOWNSEND, FIFIC′ , NEUFELD
normal and symptomatic participants alike (e.g., Carter & Neufeld, 1999). Moreover, Bayesian-based approaches to mediating group-level findings to individual participants have deemed selected structures of processing systems as common to the studied groups (Neufeld, Vollick, Carter, Boksman, & Jetté, 2002; Neufeld, 2005). Finally, in research on functional neurocircuitry (e.g., via functional magnetic resonance imaging [fMRI]), it seems imperative to anchor imputed cognitive functions, notably the temporal arrangement of constituent operations, in mathematically illumined behavioral terms. Apart from being armed with freestanding cognitive-behavioral signatures of mental architecture, we become vulnerable to the circularity inherent in inferring the functions that are at work from the investigated neurocircuitry. When it comes to treatment interventions aimed at improving information processing, rigorous profiles of clients’ strengths and weaknesses of cognitive faculties seem indispensable. The efficiency of biological and psychological interventions in principle can be improved by targeting and monitoring such profiles’ disorder-affected elements (e.g., Broga & Neufeld, 1981). Psychological interventions ideally can exploit spared elements, such as the parallel, serial, or other structural aspects of mentation (Townsend & Wenger, 2004a, 2004b). Moreover, proposed cognitive-science entrenched computational methods of assessing individuals’ functioning over the course of treatment, and plotting treatment groups’ trajectories of response to pharmacological agents, invoke specific parametric-model architectures that fall into the classes articulated here (Neufeld, in press-a). The preceding are but a sampling of reasons that should motivate delving into quantitative developments for ascertaining cognitive architecture. The exposition that follows is devoted to the most prominent division, parallel versus serial transaction of task elements. BASIC PROCESSING CHARACTERISTICS AND AVOIDING PITFALLS Certain fundamental characteristics of human information processing have been known for some time (Townsend, 1974, 1990a; Townsend & Wenger, 2004a). These characteristics, although logically distinct, can interact in ways that can dupe or confound unwary researchers. We briefly outline the major concepts here and delve into more detail subsequently.
Chapter 09.qxd
3/28/2007
12:59 PM
Page 227
9. ASSESSMENT OF MENTAL ARCHITECTURE
227
In addition to (a) serial versus parallel processing, it is necessary to consider (b) the decision or stopping rule, (c) the question of independence versus dependence of item or channel processing times, and (d) capacity, or how efficiently processes function, especially as the workload is increased. The major reason for the potential of going astray mentioned in the first paragraph, is that different combinations of values of the listed characteristics can mimic one another. For instance, one of the first outcomes of mathematical research on parallel and serial processing was that limited-capacity parallel processing (i.e., each parallel channel is slowed as more items are being processed in other channels) could so perfectly mimic standard serial processing in the popular experimental designs that the two forms were mathematically identical and thus could not be distinguished in those designs (Townsend, 1969, 1971). Even today, one finds confounding between capacity (efficiency of processing; more on this soon) and architecture (e.g., parallel vs. serial processing; again, more detail later). It is of the utmost importance to observe that the mathematical identity of certain parallel and serial models does not imply that the underlying physical mechanisms, whether neural, electronic, or mechanical, are equivalent! Rather, they simply look and act like one another (in fact, like identical twins raised in the same environment!) in certain experimental settings and under certain assumptions. To avoid the sloughs of methodological despond threatening the psychological scientist, it is necessary to consider all of the characteristics together and in a rigorous framework. As clinical science moves inexorably into the realms of hard science, it would seem desirable that it not repeat the same mistakes that cognitive psychology has already encountered and begun to surmount. We must also pay heed to the fact that because mental functions are probabilistic, not deterministic, even at the neural level, it is required that theory and theory-driven methodology be couched in stochastic language. Deterministic models can sometimes yield helpful intuition but they must be engaged with great caution, because sometimes their predictions are at odds with the true stochastic interpretations. The next section builds up, or reminds the reader of, some needed quantitative tools. Certain of the material may seem overly simple to some readers, but we prefer to be as inclusive as possible. Any of it may be skipped at the reader’s discretion.
Chapter 09.qxd
3/28/2007
12:59 PM
Page 228
228
TOWNSEND, FIFIC′ , NEUFELD
Some Required Quantitative Tools The first notion required is that of what the well-known probabilist Emanuel Parzen refers to as a probability law (Parzen, 1960). The probability law is a general term used to designate any of a number of formulas that define how the underlying random aspects will appear. Many investigators employ the alternate term distribution to refer to the probability law. We use distribution and probability law interchangeably. The term distribution can also be employed in a more detailed form, as shown later. Perhaps the most common formula or designation of a probability law is the frequency function, that is, the idea from elementary statistics that counts up the times or relatively frequency that an event occurs (e.g., proportion of students testing at such-and-such an IQ, etc.). In the ideal or theoretical case, these frequency functions may be continuous, such as the normal curve or the exponential distribution. Probabilists and modelers call the ideal frequency functions probability densities, although the concept has nothing to do with the usual physical concept of density. We write a probability density or frequency function (hereafter density) f(t) where t is, of course, time because we are focusing on response times, and naturally t is greater than (or equal to) 0 and less than infinity. Another useful designator of a probability law is the cumulative distribution function (note the special use of distribution here), which is the sum (in a discrete probability law) or integral (in a continuous probability law) from the lower limit (usually 0 in reaction time models) to an arbitrary value of the independent variable (usually time = t here). Thus, if we wish to know the probability that the response time was less than or equal to some t (rather than being exactly t), we calculate F(t) = ∫ f(t′) dt′ integrating (summing in a continuous way) from 0 to the value of interest, t. In response-time research, the so-called survivor function (from actuarial theory) is of value, indicating the likelihood that the response time or processing time is not yet finished. It is S(t) = 1 − F(t)= ∫ f(t′) dt′, this time integrating from t to infinity. There are an infinite number of probability laws and therefore densities and quite a few useful ones, such as the normal and exponential. Everyone is familiar with the normal. A figure of the exponential distribution is shown in Figure 9.1. Let exp represent the exponential number 2.7182… (like π, exp goes on forever, without repeating). Then the formula for the exponential density is f(t) = a exp(–at), where, as usual, two symbols being
Chapter 09.qxd
3/28/2007
12:59 PM
Page 229
9. ASSESSMENT OF MENTAL ARCHITECTURE
229
FIGURE 9.1 The exponential probability density function f(t) (left), its corresponding survivor function S(t) (middle), and its hazard function h(t) (right). Note that the hazard function of the exponential density has a constant value.
placed next to one another indicates multiplication. The variable a is the rate of processing in a model of response times. In the case of the exponential distribution, S(t) = exp(–at), and F(t) = 1 – exp(–at). The mean of a distribution, also called the expectation, is just E(T) = ∫ f(t′)t′dt′ this time integrating over all possible values from t = 0 to infinity. (Note that it is a convention to use capital T instead of lower case here, to indicate that it stands for any possible value, being what is known in the trade as a random variable.) Finally, we need a finer grain statistic known as the hazard function. The concept, like that of survivor function, comes from actuarial statistics, where it gives the probability that, say, a person will die in the next short time, given that the person has survived until the present moment. It is written as the ratio of the density over the survivor function, which does, indeed, condition on the event not yet having occurred. In response-time theory, of course, it refers to, say, an item finishing in the next instant, given it is not yet completed. Its formula is h(t) = f(t)/S(t). The hazard function for the exponential distribution is the elementally and unique h(t) = a, a constant. The fact that it is a constant indicates that the instantaneous conditional rate of completion neither increases nor decreases. Figure 9.1 shows the various formulae associated with the exponential probability law. The mean of the exponential distribution is simply 1/a. We next outline the basic processing characteristics involved in psychological systems. Architecture: The Serial Versus Parallel Issue Serial processing means processing things one at a time or sequentially, with no overlap among the successive processing times. Processing might mean search for a target among a set of distractors in memory or in a display,
Chapter 09.qxd
230
3/28/2007
12:59 PM
Page 230
TOWNSEND, FIFIC′ , NEUFELD
FIGURE 9.2 (a) A serial system, and (b) a parallel system. The input is a source of information for the system, for example a face or a nonface stimulus. “A” and “B” denote two channels of processing, two processes, or two units. For example, “A” and “B” could be face-feature detectors (responding to the presence of eyes and lips). In a serial system both channels process the input information in a nonoverlapping manner, whereas in a parallel system the channels operate simultaneously. After all channels finish processing (for example, the recognition of a face feature) the decision is generated. In other words, upon the positive recognition of all face features the response “I see a face” is generated. Otherwise the response “This is not a face” is generated.
solving facets of a problem, deciding among a set of objects, and so on. Parallel processing means processing all things simultaneously, although it is allowed that they may finish at different times. Although the term architecture might seem to imply rigid structure, we may employ it to refer to such, or to more flexible arrangements. Thus, it might be asserted that certain neural systems are, at least by adulthood, fairly wired in and that they act in parallel (or in some cases, in serial). However, a person might scan the newspaper for, say, two terms, one at a time, that, is serially or, by dint of will, might try to scan for them in parallel. Although parallel versus serial processing is in some sense the most elemental pair of architectures, much more complexity can be imagined and, indeed, investigated theoretically and empirically (e.g., Schweickert, 1978; Schweickert & Townsend, 1989). Figure 9.2 illustrates the flow diagrams associated with serial and parallel processing. If we are dealing with only one or two channels or items, we shall often just refer to these as a or b, but if we must consider the general case of arbitrary n items or channels, we list them as 1, 2,…, n – 1, n. So if n = 2, and a and b are stochastically independent (see later material for more on this issue), then the density of the sum of the two serial times is the
Chapter 09.qxd
3/28/2007
12:59 PM
Page 231
9. ASSESSMENT OF MENTAL ARCHITECTURE
231
so-called convolution of the separate densities. We can’t go into the details here (but see e.g., Townsend & Ashby, 1983), but simply note that this new density is designated as fa(t) * fb(t), where the asterisk denotes convolution and a and b are processed serially. The mean or expectation of the sum E[Ta + Tb] = E[Ta] + E[Tb], that is, the old result we were taught in statistics that the mean of the sum is the sum of the means: The overall completion time for serial processes is the sum of all the individual means. The standard serial model requires that fa(t) = fb(t), which in turn implies that E[Ta] = E[Tb] = E[T], and E[Ta + Tb] = 2 E[T] In more general settings, one might need to allow for a or b to take different amounts of time, depending on which is processed first. For simplicity, we do not consider that situation here, but even so, it may matter, depending on the stopping rule (see just below), which order is taken. Hence, we can assume that with probability p, a is done first and with probability 1 – p, b is done first. Figure 9.2 shows the simplest case where, say, item a is always processed first. In parallel processing, assuming again stochastic independence across the items or channels, the overall completion time for both items has to be the last, or maximum finishing time for either item. Thus, the density that measures the last finishing time is fmax(t) = fa(t)Fb(t) + fb(t)Fa(t). This formula has an easy interpretation that either a finishes last at time t and b is already done by then, or b finishes last at time t and a is already done by then. In this case, the mean is not so easy to write from first principles. Nonetheless, we can use a trick to do it. It is a very nice fact that the mean of a positive variable T is the integral of the survivor function: E[T] = ∫ S(t′) dt′, integrating from 0 to infinity. The survivor function in the present situation is S(t) = 1 – Fa(t)Fb(t) and the mean can be calculated from there using the already given integral. Stopping or Decision Rule: When Does Processing Cease? No predictions can be made about processing times until the model designer has a rule for when processing stops. In some high-accuracy situations, such as search tasks, it is usually possible to define a set of events, any one of which will allow the processor to stop without error. In search for a set of targets then, the detection of any one of them can serve as a signal to cease processing. A special case ensues when exactly one soughtfor target is present. In any task where a subset of the display or memory items is sufficient to stop without error, and the system processor is capable
Chapter 09.qxd
232
3/28/2007
12:59 PM
Page 232
TOWNSEND, FIFIC′ , NEUFELD
of stopping (not all may be), the processor is said to be capable of selftermination (like many terms of specialized argot, this one could perhaps be more descriptive). Because many earlier (e.g., Sternberg, 1966) investigations studied exhaustive versus single-target search, self-termination was often employed to refer to the latter, although it can also have generic meaning and convey, say, first-termination when the completion of any of the present items suffices to stop processing. The latter case is often called an OR design because completion of any of a set of presented items is sufficient to stop processing and ensure a correct response (e.g., Egeth, 1966; Townsend & Nozawa, 1995). If all items or channels must be processed to ensure a correct response, then exhaustive processing is entailed. For instance, on no-target (i.e., nothing present but distractors or noise) trials, every item must be examined to guarantee no targets are present. In an experiment where, say, all n items in the search set must be a certain kind of target, called an AND design, exhaustive processing is forced on the observer (e.g., Sternberg, 1966; Townsend & Nozawa, 1995). Nevertheless, as intimated earlier, some systems may by their very design have to process everything in the search set, so the question is of interest even when, in principle, selftermination is a possibility. Hence, in summary, there are three cases of especial interest: (a) minimum time, OR, or first-terminating processing, where the first item to complete stops processing; (b) single-target self-termination, where there is one target among n – 1 other items and processing can cease when it is found; and (c) exhaustive or AND processing, where all items or channels are processed. Figure 9.3 depicts AND (exhaustive) and OR (first-terminating) processing in a serial system, whereas Figure 9.4 does the same for a parallel system. Suppose again there are just two items or channels to process, a and b, and serial processing is being deployed. Assume that a is processed first. Then the minimum time processing density is simply fmin(t) = fa(t), that is, naturally just the density of a itself. Assume now there is a single target present in channel a and one distractor is in channel b, and self-terminating (ST) serial processing is in force. Then the predicted density is fst(t) = pfa(t) + (1 – p)fb(t) * fa(t). That is, if a happens to be checked first, which occurs with probability p, then the processing stops. On the other hand, if b is processed first and a distractor is found (as it must be), then a has to be processed also so the second term is the convolution of the a and b densities. In the event that both items must be processed (or an
Chapter 09.qxd
3/28/2007
12:59 PM
Page 233
9. ASSESSMENT OF MENTAL ARCHITECTURE
233
FIGURE 9.3 Schematics of stopping rules in a serial system. (a) A diagram of the standard serial system in the case of “AND” (exhaustive) processing. (b) The stopping rule in the serial system is depicted as an additional arrow which goes from the output of “A” directly to the decision box, allowing for the possibility of bypassing process “B.” When the evidence accumulated by process “A” is enough to make a decision then the processing can terminate, and additional processing of “B” is unnecessary.
inflexible serial processor cannot do otherwise), then the prediction is just that given earlier: fmax(t) = fa(t) * fb(t). When processing is independent parallel, the minimum time rule delivers a horse race to the finish, with the winning channel determining the processing time (Fig. 9.4a). The density is just fmin(t) = fa(t)Sb(t) + fb(t)Sa(t). This formula possesses the nice interpretation that a can finish at time t, but b is not yet done (indicated by b’s survivor function), or the reverse can happen. If processing is single-target self-terminating with the target in channel a, parallel independence predicts that the density is the simple fst(t) = fa(t)! Finally, if processing is exhaustive (maximum time) and independent, then processing is the same as shown in the introduction to parallel processing, fmax(t) = fa(t)Fb(t) + fb(t)Fa(t) (Fig. 9.4b). Independence Versus Dependence Of Channel or Item Processing The next important issue to discuss is that of independence versus dependence of channels, stages, or subsystems (these terms can be used interchangeably although stages is sometimes restricted to serial systems and channels to parallel systems).
Chapter 09.qxd
234
3/28/2007
12:59 PM
Page 234
TOWNSEND, FIFIC′ , NEUFELD
FIGURE 9.4 Schematics of stopping rules in a parallel system. (a) A diagram of the standard parallel system in the case of “AND” (exhaustive) processing. (b) In the “OR” case, the evidence accumulated by process “B” is enough to make a decision and processing can terminate, even though “A” has not yet finished processing (this is indicated by the short arrow).
We have been explicitly assuming independence of the processing times, whether they are serial or parallel. For the present tutorial purposes, somewhat limited space, and without assuming significantly more mathematical background of the reader, we prefer to circumambulate this issue as far as writing out the technical equations goes. Nonetheless, it may be pertinent to give some indication of where it matters. In serial processing, if the successive items are dependent then what happens on a, say, can affect the processing time for b. Although it is still true that the overall mean exhaustive time will be the sum of the two means, the second, say b, will depend on a’s processing time. Hence, if, say, a is speeded up, then ordinarily that will affect even the mean time of b. Figure 9.5 indicates independence versus dependence in serial systems. In parallel processing too, the processing times could be dependent. For instance, because they are being processed simultaneously, ongoing inhibition or facilitation (or both!) can take place during a single trial and while processing is ongoing. Townsend and Wenger (2004b) discuss this topic in detail. Figure 9.6 illustrates the concepts of independent versus dependent processing in parallel systems. It is interesting to note that the above prediction of independent parallel processing in ST situations will no longer strictly hold. However, it will still be true even if processing is dependent that the predicted ST density will be the average or expected value (i.e., known in probability jargon as the marginal) of the density in the channel where the sought-for target is
Chapter 09.qxd
3/28/2007
12:59 PM
Page 235
9. ASSESSMENT OF MENTAL ARCHITECTURE
235
FIGURE 9.5 Dependency between “A” and “B” in a serial system: (a) the standard serial system, and (b) a positively dependent serial system: Duration of “B” depends positively on duration of “A,” that is faster processing in “A” will produce facilitation or faster processing in “B” and vice versa. For example, in a face-recognition task faster recognition of the first face feature could give some “confidence” to a second process to speed up processing of a second feature. (c) A negatively dependent serial system: The processing time of “B” is inversely related to a processing time of “A.” Faster processing of “A” produces slow processing of “B”; that is, “A” inhibits “B,” and vice versa. Overall, a positively dependent system with facilitation exhibits the fastest reaction time (500 msec), whereas a negative dependent system with inhibition exhibits the slowest reaction time (1,000 msec).
located, E[Ta]. Only in the nonindependent situation, this expectation has to be taken over all the potential influences from the surrounding channels. The speed-ups or slow-downs shown in Figures 9.5 and 9.6 can be interpreted in terms of the notion of capacity, which we discuss next. Capacity: Various Speeds on a Speed Continuum Capacity refers generally to the speed of processing in response-time tasks. We first provide an informal sketch of the major concepts and then turn to a more rigorous exposition. For greater mathematical detail and indepth discussion see Townsend and Ashby (1978), Townsend and Nozawa (1995), and Townsend and Wenger (2004b). Wenger and Townsend (2000) offer an explicit tutorial and instructions on how to carry out a capacity analysis.
Chapter 09.qxd
236
3/28/2007
12:59 PM
Page 236
TOWNSEND, FIFIC′ , NEUFELD
FIGURE 9.6 Dependency between “A” and “B” in a parallel system. (a) The standard parallel system, and (b) a positively dependent parallel system: The positive sign arrow from “A” to “B” indicates positive facilitation. That is, faster processing of one channel speeds up processing in the other channel (as depicted in the figure), and vice versa. (c) A negatively dependent parallel system: The processing time of “A” is inversely related to the processing time of “B.” Faster processing of “A” will produce longer processing of “B”; that is, “A” inhibits “B” (as depicted in the figure), and vice versa. Overall, a positively dependent system with the facilitation exhibits the fastest reaction time (500 msec), while a negatively dependent system with the inhibition exhibits the slowest reaction time (1,000 msec).
Informally, the notion of unlimited capacity refers to the situation when the finishing time of a subsystem (item, channel, etc.) is identical to that of a standard parallel system (described in more detail later); that is, the finishing times of the distinct subsystems are parallel, probabilistically independent, and the finishing times of each do not depend on how many others are engaged (e.g., in a search task the finishing time of one item is invariant over the total number of items being searched). Limited capacity refers to the situation when item or channel finishing times are less than what would be expected in a standard parallel system. Supercapacity indicates that individual channels are processing at a rate even faster than standard parallel processing. Figure 9.7 illustrates the general intuitions accorded these concepts, again in an informal manner. We pause to observe that, although the stopping rule obviously affects overall processing times, as indicated in Figure 9.8 for both serial and
Chapter 09.qxd
3/28/2007
12:59 PM
Page 237
9. ASSESSMENT OF MENTAL ARCHITECTURE
237
FIGURE 9.7 Graphical intuition of a system’s behavior under different capacity bounds: limited capacity, unlimited capacity, and supercapacity. The total system’s capacity resource remains the same across all conditions. (a) In the limited-capacity case, the total capacity is split between two channels. (b) In the case of the unlimited capacity, each channel receives the total capacity. (c) In the supercapacity case, the capacity devoted to each channel exceeds the total system capacity. Note that an increase in channel capacity produces faster processing for that channel.
parallel systems, we assess capacity (i.e., efficiency of processing speed) in comparison with standard parallel processing with specification of a particular stopping rule. Thus, although the minimum time (firstterminating or OR processing) decreases as a function of the number of items undergoing processing (because all items are targets), the system is merely unlimited, not super, because the actual predictions are from a standard parallel model (i.e., unlimited capacity with independent channels). But observe that each of the serial predictions would be measured as limited capacity because for each stopping rule, they are slower than the predictions from standard parallel processing. Figure 9.8 indicates mean response times as a function of workload. Workload refers to the quantity of labor required in a task. Most often, workload is given by the number of items that must be operated on in some fashion. For instance, workload could refer to the number of items in a visual display that must be compared with a target or memory item. Although Figure 9.8 indicates speed of processing through the mean response times, there are various ways of measuring this speed. The mean = E[T] is a rather coarse level of capacity measurement. A stronger gauge is found in the cumulative distribution function F(t), and the hazard
Chapter 09.qxd
238
3/28/2007
12:59 PM
Page 238
TOWNSEND, FIFIC′ , NEUFELD
FIGURE 9.8. Expected processing time as a function of load-set size for different stopping rules (exhaustive, self-terminating, and minimum) for (a) the standard serial model, and (b) the parallel unlimited capacity processing model. The loadset size is defined as number of processing units of information: usually number of memorized items or items in the visual filed that have to be searched.
function (to be discussed momentarily) is an even more powerful and fine-grained measure. This kind of ordering is a special case of a hierarchy on the strengths of a vital set of statistics (Townsend, 1990b; Townsend & Ashby, 1978). The ordering establishes a hierarchy of power because, say, if Fa(t) > Fb(t) then the mean of a is less than the mean of b. However, the reverse implication does not hold (the means being ordered does not imply an order of their cumulative distribution functions). Similarly if ha(t) > hb(t) then Fa(t) > Fb(t), but not vice versa, and so on. Obviously, if the cumulative distribution functions are ordered then so are the survivor functions. That is, Fa(t) > Fb(t) implies Sa(t) < Sb(t). There is a useful measure that is at the same strength level as F or S. This measure is defined as –loge[S(t)], that is, minus one times the natural logarithm of the survivor function. It turns out that this is actually the integral of the hazard function h(t′) from 0 to t (e.g., Wenger & Townsend, 2000; illustrative uses of this and related measures, below, in clinical science are presented in Neufeld, Townsend, & Jetté, in press). We thus write the integrated hazard function as H(t) = –loge[S(t)]. Although it is of the same level of strength as S(t), it has some very helpful properties not directly shared by S(t). Now we are in a position to compare two or more experimental situations by comparing their statistics. For example we might compare Condition 1
Chapter 09.qxd
3/28/2007
12:59 PM
Page 239
9. ASSESSMENT OF MENTAL ARCHITECTURE
239
to Condition 2 as in S1(t) versus S2(t), or say, H1(t) versus H2(t)—or,more easily, consider H1(t)/H2(t). If this ratio is greater than 1 for all reasonable values of t, then we know that the capacity (efficiency, speed) of Condition 1 is greater than that in Condition 2 and in a quite strong sense. As a special case of great import, assume now that we change the number of channels or items that must be processed, in the present context, say, n goes from 1 to 2. Suppose we wish to measure the effect of this increase in workload in a situation where an efficient system can stop in the first-terminating (minimum) time. We require a measuring instrument, in a sense, because there is no elementary ruler we can use for arbitrary capacity measurement. Our measuring instrument is that of the set of predictions by unlimited-capacity independent parallel processing. Unlimited capacity means here that each parallel channel processes its input (item, etc.) just as fast when there are other surrounding channels working (i.e., with greater n) as when it is the only channel being forced to process information. Now it has been demonstrated that when processing is of this form, then the sum of the integrated hazard functions for each item presented alone is precisely the value, for all times t, of the integrated hazard function when both items are presented together (Townsend & Nozawa, 1995). That is, Ha(t) + Hb(t) = Hab(t). This intriguing fact suggests the formulation of a new capacity measure, which the latter authors called the capacity coefficient C(t) and set it equal to C(t) = Hab(t)/[Ha(t) + Hb(t)], that is, the ratio of the double item condition over the sum of the single item conditions. If this ratio is identical to 1 for all t, then the processing is identical to that of an unlimited capacity independent parallel model. If C(t) is less than 1 for some value of t, then we call processing limited. For instance, either serial processing of the ordinary kind or a fixed-capacity parallel model that spreads the capacity equally across a and b predicts C(t) = ½ for all times t > 0. If C(t) > 1 at a time (or any, or maybe all times t) t, then we call the system supercapacity for those times. A tutorial on capacity and how to assess it in experimental data is offered in Wenger and Townsend (2000), with clinical-science applications being illustrated in Neufeld, Townsend, and Jetté (in press). In a recent extension of these notions, we have shown that if configural parallel processing is interpreted as positively interactive parallel channels (thus being dependent or positively correlated rather than independent), then configural processing can produce striking supercapacity (Townsend & Wenger, 2004b).
Chapter 09.qxd
240
3/28/2007
12:59 PM
Page 240
TOWNSEND, FIFIC′ , NEUFELD
Up to this point, in addition to providing motivation and relevance to clinical science, we have reviewed some required probability tools and, more important, introduced a set of critical dimensions from our information processing theory. These included the focus of this chapter: parallel versus serial processing. It is possible to construct a huge variety of processing systems simply by combining different values from each dimension (e.g., moderately limited capacity parallel processing with negatively dependent channels and with an exhaustive processing rule imposed). Nonetheless, certain particular systems have gained almost archetypal status in cognitive science. In fact, this is so much so that many times investigators seem to operate as if they are the only available or possible systems in nature. Of course, this is far from true, but these prototypical models bear special consideration on our part. We next take up these prominent models. After that,we return to a more in-depth discussion of automatic versus controlled processing. This section is followed by presentation of a relatively recent and powerful experimental approach that permits assessment of the basic dimensions of information processing, using response times. PROMINENT ARCHITECTURES The first of the major architectures comprises the standard serial class of model. This class has been quantitatively well understood at least since Sternberg’s (1966) initial papers. The standard parallel class of models seems to be less well comprehended at large, although a number of psychological notions, for instance, automaticity, can be captured by this type of processing. The third architecture, coactive parallel processing, is a relatively recent contender—as we show later, these models permit performance that is superior even to standard parallel processing. These models make distinct predictions even at the level of mean response times, as a function of workload, as indicated in this section. However, they are still open to problems with model mimicking at this level. A later section, Experimental Testing of Parallel and Serial Architectures, shows how to effectively circumvent the model-mimicking dilemma. It is worth pointing out that none of these models’ characteristic predictions rely on any special kind of probability density—that is, the predictions are a quantitative form of qualitative, and are thus so-called distribution-free and, of course, are parameter-free as well. The latter stipulation simply means that the predictions do not depend on any particular
Chapter 09.qxd
3/28/2007
12:59 PM
Page 241
9. ASSESSMENT OF MENTAL ARCHITECTURE
241
mean or variance, say of the underlying distributions, although naturally the actual means or variances would depend on the parameters. Standard Serial Models This type of model is what most people mean when they only say serial unadorned. Thus, it is the model advocated by S. Sternberg in many of his papers (e.g., Sternberg, 1966, 1969, 1975). To reach it in the case that n = 2, simply let fa(t) = fb(t) = f(t)—that is, the probability densities are the same across items or positions and even n. The latter means that f(t) defines the length of time taken on an item or channel no matter how big or little the entire set of operating items or channels is. Furthermore, it is assumed in the standard serial model that each successive processing time is independent of all others. So, if a is second, say, its time does not depend on how long the preceding item (e.g., b) took to complete its processing. Note, however, that we still allow in general that different paths through the items might be followed. We also do not confine the stopping rule to a single variety. Now, S. Sternberg’s preferred model did assume that exhaustive processing was used even in target-present trials. But because this seems like a secondary issue we allow the standard model to follow other, sometimes more optimal, rules of cessation. Because all the n densities are now the same we can simply write the nth order convolution for exhaustive processing in symbolic form as fmax(t) = f*(n)(t). The exhaustive mean processing time is then Emax[T1 + T2 + … + Tn] = n E[T]. Next consider the situation where exactly one target is present among n – 1 distractors and the system is self-terminating. Again, it is assumed that the target is placed with probability 1/n in any of the n locations. Then it follows that fst(t) = (1/n) ∑ f*(i) where the summation goes from i = 1 to i = n. The mean processing time in this case is the well-known Est[T] = (n + 1)E[T]/2. This formula can be interpreted that on average, it takes the searcher approximately one-half of the set of items to find the target and cease processing. Finally, when processing stops as soon as the first item is finished, then we have the result fmin(t) = f(t) and the elemental Emin[T] = E[T]. Standard Parallel Models The standard parallel model assumes independence again among the processing items, but this time in a simultaneous sense. At this point, we
Chapter 09.qxd
3/28/2007
12:59 PM
Page 242
242
TOWNSEND, FIFIC′ , NEUFELD
have to make the decision of whether to force all the channels to process items at the same (stochastic) speed. Why not? After all, we did just that with the standard serial model. However, the standard serial model could still produce so-called position effects. Position effects are produced when distinct locations of a target produce different mean processing times and, hence, are associated with distinct densities (e.g., Sternberg, 1966; Van Zandt & Townsend, 1993). Standard serial models can do this by letting different processing paths through the items occur with different probabilities. For instance, if n = 3, the path
might be taken with probability 1/3, the path with probability 1/6, the path with probability ½, and all the other paths never occur. This distribution on orders would make shorter times on average of a over b and c and the c times would tend to be shorter than those of b. The only way an independent parallel model can generate position effects is if the distinct channels or items have different densities, as in fa(t) ≠ fb(t). Hence, this provision is usually allowed in standard parallel models. In any event, because we always assumed independence in the above treatment of parallel models, the formulas for n = 2 stay the same. However, for simplicity, take the special case where the densities are all the same. Then E[MAX(T1 , T2,…,Tn)] = ∫ [1 – Fn(t’)] dt’ with the integral being taken from 0 to infinity. It is straightforward to show that in this case, the curve of mean processing times (and therefore response times) is always increasing but with a concave-down shape. In this especially simple case, the single self-terminating target case, among n – 1 distractors, is just E[T], the time required for any single item to complete. Finally, the time required for the minimum or first-terminating time, that is, the time of the winning horse, is E[MIN(T1 , T2,…,Tn) = ∫ [1 – F(t′)]n dt’ and again the integral is from 0 to infinity. Here, it can be demonstrated that the mean times in this kind of model (and this is true even if not all the distributions are identical), the curve of mean times is always decreasing, concave up. Egeth (1966) employed this characteristic frequently to argue for parallel processing because it is an unnatural prediction for serial models. The proofs of the theorems on concavity of reaction time curves as a function of load are provided by Townsend and Ashby (1983). Coactive Parallel Models Starting in the late 1980s, J. Miller (1982, 1986) began to produce data that seemed to indicate that processing could be even better, more capacious somehow, than even ordinary (or we would now say, standard parallel
Chapter 09.qxd
3/28/2007
12:59 PM
Page 243
9. ASSESSMENT OF MENTAL ARCHITECTURE
243
FIGURE 9.9 Schematics of (a) a parallel independent system, and (b) a coactive multiple channels processing systems. A coactive model assumes that an input from separate parallel channels is consolidated into a resultant common processor before a decision is made.
action). His main line of argument used a clever probability inequality that ordinary parallel processing would have to satisfy but that extraordinary or, as we might say today, supercapacity operations could violate. Before long, a number of investigators, including Miller, commenced to develop actual nonstandard parallel models that indeed would violate the inequality (e.g., Diederich, 1992, 1995; Diederich & Colonius, 1991; Schwarz, 1994; Townsend & Nozawa, 1995; Townsend & Wenger, 2004b). All these models possess the property that activity in the separate channels was summated or pooled into a final common channel before a detection decision was made. In standard and in fact, any parallel model where separate detection decisions are made in their individual channels, this pooling does not occur. Figure 9.9 exhibits a comparison between ordinary parallel processing where separate decisions (detections, etc.) are made on the distinct channels as opposed to coactive processing where the activations on the several channels are combined, for instance, summed arithmetically. Subsequently, a general theory of capacity was formulated that permitted the measurement of processing efficiency for all times during a trial (Townsend & Nozawa, 1995). Employing standard parallel processing as a
Chapter 09.qxd
244
3/28/2007
12:59 PM
Page 244
TOWNSEND, FIFIC′ , NEUFELD
cornerstone, the theory defined unlimited capacity as efficiency identical to that of standard parallel processing in which case the measure is C(t) = 1. It defined limited capacity as efficiency slower than standard parallel processing. For instance, standard serial processing produces a measure of capacity of C(t) = .5. And finally, the theory defined supercapacity as processing with greater efficiency than standard parallel models could produce, that is, C(t) > 1. It was then proven that a very broad class of coactive parallel models, which included all the special ones constructed before that, would imply C(t) > 1 and would inevitably violate Miller’s inequality. Because our focus here must be on architecture, we have not been able to expend much space on capacity, but the reader is referred to Wenger and Townsend (2000) for a tutorial on capacity measurement, Townsend and Ashby (1978) for the early work on this concept, and Townsend and Wenger (2004b) for the latest theoretical results on capacity in interactive (i.e., not independent) parallel systems. Note that the capacity construct has been prominent in clinical cognitive science. Initial extensions of mathematical treatments of the construct to this arena of study are reported in Neufeld, Vollick, and Highgate (1993). Subsequent developments are enumerated in Neufeld (in press-a) and Neufeld, Carter, Vollick, Boksman, Levy, and Jetté (in press), and inaugural implementations in this field of the integrated hazard function H(t) and the capacity coefficient appear in Neufeld, Townsend, and Jetté (in press). We have provided a succinct overview of three highly important kinds of processing models: standard serial, standard parallel, and coactive parallel. The standard serial model is least efficient because it entails sequential processing of each item with average processing times on each always being the same. Standard parallel processing implies quite efficient processing because each item can be processed at the same rate, regardless of how many others are being operated on, and all are processed simultaneously. Coactive parallel processing can be exceedingly fast by virtue of all of the item channels pouring their activation into a single final conduit. The next section revisits the very popular topic of automatic processing. AUTOMATICITY: INTERPRETATION VIA ARCHITECTURE AND CAPACITY The notion of automaticity has rarely, if ever, been given a rigorous mathematical definition. One primary correlate has been superior efficiency of processing as expressed in response times.
Chapter 09.qxd
3/28/2007
12:59 PM
Page 245
9. ASSESSMENT OF MENTAL ARCHITECTURE
245
In fact, often, the criterion investigators use to ascribe processing as automatic, is simply that a mean response-time function of workload be flat. Because, as we have seen, response speed depends on all the major characteristics of a system, the investigator must be wary when drawing conclusions about automaticity even from this simple standard. Thus, even a standard serial model predicts a flat response-time function for a first-terminating stopping rule. Furthermore, a standard parallel model with its independent, unlimited capacity, will predict a flat response-time curve when the system is searching for a single target among n – 1 distractors. We further witnessed that with first-termination, such a model will actually predict a decrease in mean response times. It is not clear to the authors whether the class of standard parallel models are sufficiently free of capacity limitations to qualify as automatic processing in the minds of most investigators of this topic. It would seem that a coactive model, with channels left undegraded as workload n increases, should merit that assessment. This question is primarily one of convention, but it would be propitious if agreement on a rigorous set of criteria could be reached. We also have to take account of the differing experimental paradigms where the phenomenon is concluded to exist. The modern instantiation of this concept stems from the Schneider and Shiffrin (1977) studies on visual and memory search mentioned in the introduction. We haven’t the space to consider the wide array of experimental conditions they ran and in particular must ignore their results on accuracy in favor of the responsetime dependent variable under very high accuracy. Their basic findings, replicated scores of times, were that both the single target present (i.e., potentially single-target self-terminating) and target absent (forced exhaustive processing) response times were flat or almost flat. Standard parallel processing models easily predict the first result but not the second. If mean response times for exhaustive processing are flat, then processing is highly supercapacity. For instance, Townsend and Ashby (1983) show that a model whose channels are unlimited capacity at the start of a trial, but that is capable of reallocating capacity from completed channels to uncompleted ones, predicts flat exhaustive mean response-time curves. However, either processing is also exhaustive on the target present trials, or is less super in capacity on the latter—an unlimited capacity parallel model with reallocation would predict decreasing mean response-time functions. A similar but not identical topic arises when one speaks of gestalt figures or holistic processing. A unified Gestalt or holistic percept might
Chapter 09.qxd
246
3/28/2007
12:59 PM
Page 246
TOWNSEND, FIFIC′ , NEUFELD
be more or less wired in, or at least installed early in life, as in face perception, or learned for some purpose later on. One expression for the welding of several initially separate parts into a whole is unitization, a term employed by Czerwinski, Lightfoot, and Shiffrin (1992) in their study of this kind of learning. Another example stems from Goldstone’s (2000) experiments, where observers learned to weld together a set of originally meaningless squiggles into a perceptually holistic object. All of the squiggle features in the designated object had to be perceived in order to be correct, forcing exhaustive processing. Employing concepts from cognitive stochastic process theory (e.g., Townsend & Ashby, 1983), he showed that the various parts of the holistic object were perceived in a supercapacity (i.e., better than standard parallel processing) fashion. Subsequent replicative experiments employing new measures of capacity by Blaha and Townsend (2004a) have confirmed and strengthened Goldstone’s conclusions. Blaha and Townsend (2004b) have further developed a neuralistic model based on a dynamic system instantiation of Hebbian concepts, which produces the massive supercapacity found in these investigations. Another experimental example of a kind of automatization may be that of the well-known and documented pop-out effect. This effect occurs whenever a target that is sufficiently distinct from all the distractors is used, as, for instance, when a colored object is placed among a set of gray distractors or a green object is placed among a set of red distractors (e.g., Treisman & Gormican, 1988; see also Van Zandt & Townsend, 1993). Here the emphasis is obviously on the target present case. Response times on these trials are flat across set size n. Hence, standard parallel processing can account for these results without having to posit super capacity. In summary, as noted before, the literature and phenomena of automatization are vast. Nonetheless, it appears that at least in certain publicized cases, and in particular where a small set of targets is identified with the same speed across increasing workload n (i.e., increasing the number of present distractors), standard parallel processing is a sufficient explanation for automatization. However, when processing has to be exhaustive and yet the response-time curves are flat across n, the system has to be exceedingly supercapacity. Such cases appear in the Schneider and Shiffrin (1977) target absent data as well as the Goldstone (2000) and later the Townsend and Blaha (2004a, 2004b) investigation. Positively interactive parallel processing or coactive parallel processing can readily produce such findings (Townsend & Wenger, 2004b).
Chapter 09.qxd
3/28/2007
12:59 PM
Page 247
9. ASSESSMENT OF MENTAL ARCHITECTURE
247
It is apparent that a simple and universal isomorphism between the concept of automaticity and rigorous information-processing dimensions is a chimera. It is especially vital to understand that there may be more than one way to skin the cat, so to speak—that is, to avoid confusing sufficiency of a model with its necessity. For instance, Kahneman (1973) promoted the idea that systems might in some situations be capable of drawing on extra resources when workload is increased. Such extra resources could transform standard parallel processing into supercapacity (and apparently automatic) parallel processing. In any event, in each individual research area, our taxonomy can be applied to endow this exceedingly efficient type of processing with precise meaning and experimental implications. EXPERIMENTAL TESTING OF PARALLEL VERSUS SERIAL ARCHITECTURES There now exist several experimental assays of mental architecture. Most of these circumvent the major impediment of the ability of limited capacity parallel models to mimic the behavior of standard serial models. However, before embarking on the primary target of this section, we observe that serial models that can mimic unlimited or supercapacity parallel models are typically quite unintuitive, so that evidence of such behavior can be taken as falsifying serial architectures (e.g., Townsend, 1971, 1974). For more complete reviews of, and references to, the available panoply of serial–parallel assessment techniques, the reader is pointed to Townsend and Wenger (2004a), Townsend (1990a), and Townsend and Ashby (1983). One major set of strategies has sprung from S. Sternberg’s (1969) additive factors method. This method and its descendents, rests on the assumption of selective influence. Selective influence assumes that specified experimental variables separably affect distinct processing systems. In the additive factors method, the subsystems are often called stages. The stages consume some random amount of time and it is postulated that these processing times do not overlap, although it could be that switching times between stages could add some additional time. Suppose we are concerned with just two subsystems (stages for serial processing). Let us refer to these as a and b as usual and name the experimental factors that selectively affect them A and B. Then the next step is to perform a factorial experiment with factors Ai × Bj , where i and j
Chapter 09.qxd
248
3/28/2007
12:59 PM
Page 248
TOWNSEND, FIFIC′ , NEUFELD
FIGURE 9.10 The time course of processing of two items (left column), the corresponding deterministic (middle column), and stochastic (right column) mean interaction contrast (MIC), across different architectures and stopping rules (rows). The time course of processing depicts the change in total processing time for different factorial conditions (HH, HL, LH, LL) for different architectures. Each upright bold arrow in the graph corresponds to total processing of one unit (in the left column), which could be at the H (high) or L (low) level. A dotted upright arrow indicates a process that did not complete because the processing terminated on a completion of the previous process. The deterministic MIC, in the middle column, represents the duration or the sum of process times (as indicated on the y-axis in the first column). Note that we are not able to directly observe the deterministic MIC in experiments because in a real system processing components will add some variability or noise. The stochastic MIC is an observable measure and is obtained when some variability or noise is added to the overall processing. Error bars around each mean condition represent standard error statistic (added here arbitrarily for the sake of presentation). Also observe also that each architecture combined with a different stopping rule exhibits a different MIC value.
Chapter 09.qxd
3/28/2007
12:59 PM
Page 249
9. ASSESSMENT OF MENTAL ARCHITECTURE
249
indexes run over the various levels of each factor (e.g., i or j = 1, 2). Because of selective influence and the successive nonoverlapping times from the stages, it is then predicted that the two (or more) experimental factors will exhibit additive effects in the mean response times garnered in the factorial experiment. Let the random times for the two stages be named Ta(Ai) and Tb(Bj), respectively, and set the overall mean (=expected) exhaustive processing time under condition Ai × Bj as E[Tij]. Then this overall mean exhaustive processing time for condition Ai × Bj in this serial model is E[Tij] = E[Ta(Ai) + Tb(Bj)] and by virtue of the aforementioned elementary statistical fact that the mean of the sum equals the sum of the means we arrive at E[Ta(Ai) + Tb(Bj)] = E[Ta(Ai)] + E[Tb(Bj)]. By adopting the convention that E[Ta(Ai)] = ta(Ai), it is straightforward to compute the mean interaction contrast (MIC) as MIC = E[T11] – E[T12] – {E[T21] – E[T22]} = E[T11] – E[T12] – E[T21] + E[T22] = ta(A1) + tb(B1) – [ta(A2) + tb(B1)] − [ta(A1) + tb(B2)] + [ta(A2) + tb(B2)] = 0. This little operation demonstrates the additivity, and therefore the zero MIC, of the serial model under selective influence. Figure 9.10 shows mean reaction time predictions for each mental architecture combined with different stopping rules. For each model time, course of activation is depicted in the first column. Second and third columns of Figure 9.10 show both additive and stochastic MIC predictions for each model, while the stochastic MIC is empirically observed only. For the aforementioned serial exhaustive processing, it is evident that indeed E[Tll] – E[Tlh] = E[Thl] – E[Thh], eventuating in the MIC of zero. Interestingly, first- and single-target terminating stopping rules also result in additive (implying MIC = 0) response-time factors. However, we show later that a more penetrating statistic is able to distinguish the stopping rules. To a number of theorists, the appearance of the original method in 1969 raised a vital question: What kinds of predictions would nonserial architectures make? Schweickert (1978, 1983), in his latent mental network theory, contributed the first major extension of the additive factors method, involving more complex architectures under the assumption of selective influence. This theory was very general, including serial and parallel systems as special cases. Taking a different approach, Townsend and Ashby (1983) found that the mean interaction contrast distinguished parallel and serial stochastic models when selective influence was assumed. Stopping rule matters here,
Chapter 09.qxd
250
3/28/2007
12:59 PM
Page 250
TOWNSEND, FIFIC′ , NEUFELD
too. Interestingly, the sign of the mean interaction contrast depends not only on the architecture, but also on the stopping rule. Thus, parallel exhaustive processing exhibits a negative contrast, whereas a parallel race model (i.e., first-terminating parallel processing) will evidence a positive contrast (see Townsend & Nozawa, 1995). The case of single-target termination has not been much discussed or applied, but it is easy to demonstrate that, intriguingly enough, this case implies additive factors. Thus, parallel predictions are shown in Figure 9.10. Note that if only a single target condition were employed, the serial and parallel predictions are the same. The general set of experimental strategies that include nonserial architecture and also statistics other than the mean (discussed next) has been called systems factorial technology. Systems Technology Response-Time Factorial Distributions As suggested by Townsend (1990a, 1990b), certain aspects of probability distributions are more powerful than others. That is, knowledge of some aspects always implies knowledge of others, but not vice versa. In particular, the entire cumulative probability distribution (i.e., the integral of the density from 0 to t) function on response times is more powerful than the means alone. We observe that statistics at the distributional level deliver much deeper and more conclusive information about processing architecture and stopping rules than was possible with mean response times. For instance, a test case presents itself in the question as to whether coactive parallel processing can be distinguished, using factorial methods, from ordinary parallel processing. It turns out that at the level of mean response times (RTs) and within an OR design, coactive parallel processing cannot be distinguished from ordinary parallel processing with an OR stopping rule (i.e., with a race between the two channels determining when the process is completed). Specifically, the MIC is positive just as in a parallel horse race. However, if the factorial interaction concept is extended to the entire RT distribution (as in Townsend & Nozawa, 1995), it turns out that it is possible to distinguish a coactive model from a standard parallel model with an OR gate. In principle, either the cumulative probability distribution [P(T < t) = F(t)] or the survivor function [S(t) = 1 – F(t)] can be used (e.g., Schweickert, Giorgini, & Dzhafarov, 2000). Because the original derivations were in terms of the survivor
Chapter 09.qxd
3/28/2007
12:59 PM
Page 251
9. ASSESSMENT OF MENTAL ARCHITECTURE
251
FIGURE 9.11 An ordering of joint survivor functions for different factorial conditions (HH, HL, LH, LL) (left column) and the survivor interaction contrast (SIC) (right column) across different architectures and stopping rules (rows). Note that each SIC function is calculated using SIC(t) = Sll(t) – Slh(t) – Shl(t) + Shh(t). Each joint survivor function on the right-hand side is estimated from data (displayed in the left column). Note that each combination of architecture and stopping rule exhibits a unique SIC function. The shapes of these different SIC functions are independent of the form of the probability density function.
Chapter 09.qxd
3/28/2007
12:59 PM
Page 252
TOWNSEND, FIFIC′ , NEUFELD
252
function (Townsend & Nozawa, 1995), we employ that one here, with the interaction contrast for the survivor function (SIC): SIC(t) = Sll(t) – Slh(t) – [Shl(t) – Shh(t)] = Sll(t) – Slh(t) – Shl(t) + Shh(t) Figure 9.11 displays the predictions for the various models. Observe that ordinary parallel-processing SICs reveal total positivity in the case of OR conditions, but total negativity in the case of AND conditions. Furthermore, OR parallel and coactive parallel processing now are distinguished by their respective SICs: The contrast for OR parallel processing is consistently positive, whereas the contrast for the coactive model possesses a small negative blip at the earliest times before going positive. Because MIC must be positive in coactivation, the positive portion of the SIC always has to exceed the negative portion. Calculation of the SIC function in reaction time experiments could be a laborious job when using some standard statistical packages. In the appendix, we provide guidelines for calculating SIC function. Two scripts that calculate and display the survivor interaction contrast function are available for download on the Psychology Press Web site (http://www.psypress.com/brainscans-etc), written for the Mathematica and Matlab environments. Details of implementations are displayed as the comment sections within each script. The advantages associated with the use of both the SIC and the MIC go beyond the ability to distinguish coactive from parallel processing. It is also intriguing that the OR and the AND serial stopping rules are now experimentally distinguishable, because in the OR case SIC = 0 always, but in the AND case there is a large negative portion of the SIC, followed by an equally large positive portion. Thus, both the architecture and the stopping rule are experimentally determinable by the factorial tests carried out at the distributional level. The general applicability of the distributional approach has benefited from theoretical extensions by Schweickert and colleagues (2000) to general feedforward architectures, which contain parallel and serial subsystems, and from advances in methods of estimating entire RT distributions (see in particular, Van Zandt, 2000, 2002). CONCLUSION In this brief space, we have come all the way from underscoring the importance of discerning mental architecture in clinical cognitive science, through enumeration of the chief issues encountered in engaging this challenge, to mathematical-theory spawned technology for resolution.
Chapter 09.qxd
3/28/2007
12:59 PM
Page 253
9. ASSESSMENT OF MENTAL ARCHITECTURE
253
Developments in contemporary mathematical cognitive science represent compelling supplements, and even alternatives, to currently proffered batteries of measures aimed at mapping cognitive functioning among clinical samples (e.g., Heinrichs, 2005), or monitoring response to treatment (Nuechterlein, Barch, Gold, Goldberg, Green & Heaton, 2004; cf. McFall & Townsend, 1998). Application in clinical science is not without its problems (Neufeld, 2001). In-depth analyses of clinical-setting exigencies, and tacks to overcoming them, whose exposition is beyond the present scope and space, are available in auxiliary sources (Carter, Neufeld, & Benn 1998; Neufeld, in press-b). We anticipate that advances in quantitative cognitive science will continue to make inroads on the clinical scene, eventuating in significant improvements in cognitive assessment and intervention, and that tutorials of this nature will serve to accelerate the process. APPENDIX: A GUIDE TO CONSTRUCTING AND USING THE SURVIVOR INTERACTION CONTRAST FUNCTION 1. For each condition, remove the RTs that correspond to errors, equipment failures, anticipatory reactions, and lapses of observer attention. 2. Determine the reaction time bin size. Usually we use 10 msec bin size. The range and the size of the bins vary according to the nature of the task. 3. Count the number of observations in each bin. This step generates a frequency distribution function. 4. Divide each counted frequency (bin) by the total number of observations. This produces relative frequency function (i.e., an empirical probability density function). 5. Calculate the empirical cumulative distribution function (CDF) F(t) for each stimulus condition (HH, HL, LH, LL) by accumulating the empirical probabilities from the lowest to the highest valued bin. This will produce four vectors of data, each describing empirical cumulative distribution function for particular condition (Fll(t), Flh(t), Fhl(t), Fhh(t)). 6. Calculate an empirical survivor function for each condition by subtracting the value of F(t) for each condition from 1; that is, calculate 1 – F(t). Do this for four data sets that correspond to each factorial condition. This step will produce four vectors of data,
Chapter 09.qxd
254
3/28/2007
12:59 PM
Page 254
TOWNSEND, FIFIC′ , NEUFELD
each describing empirical survivor distribution function for each condition (Sll(t), Slh(t), Shl(t), Shh(t)). 7. Calculate the survivor interaction function (SIC) by taking the double difference between survivor data vectors, SIC(t) = Sll(t) – Slh(t) – Shl(t) + Shh(t), and form the final SIC data vector.
ACKNOWLEDGMENTS Sources of support for this research included a National Institute of Mental Health grant awarded to Dr. James Townsend and Dr. Mario Fific (NIMH-R01 MH57717-04A1), a Social Sciences and Humanities Research Council of Canada operating grant awarded to Richard W. J. Neufeld, Workplace Safety and Insurance Board, and Canadian Institutes of Health Research operating grants (Richard W. J. Neufeld, co-investigator), and a Canadian Institutes of Health Research New Emerging Teams grant (Richard W. J. Neufeld, co-investigator). REFERENCES Blaha, L. M., & Townsend, J. T. (2004a, July). From nonsense to gestalt: The influence of configural learning on processing capacity. Paper presented at Society for Mathematical Psychology annual meeting, University of Michigan, Ann Arbor. Blaha, L. M., & Townsend, J. T. (2004b, November). A dynamic hebbian model of perceptual learning. Poster presented at the Psychonomic Society annual meeting, Minneapolis, MN. Broga, M. I., & Neufeld, R. W. J. (1981). Multivariate cognitive performance levels and response styles among paranoid and nonparanoid schizophrenics. Journal of Abnormal Psychology, 90, 495–509. Carter, C. S., Robertson, L. C., Chaderjian, M. R., Celaya, L. J., & Nordahl, T. E. (1992). Attentional asymmetry in schizophrenia: Controlled and automatic processes. Biological Psychiatry, 31, 909–918. Carter, J. R., & Neufeld, R. W. J. (1999). Cognitive processing of multidimensional stimuli in schizophrenia: Formal modeling of judgment speed and content. Journal of Abnormal Psychology, 108, 633–654. Carter, J. R., Neufeld, R. W. J., & Benn, K. D. (1998). Application of process models in assessment psychology: Potential assets and challenges. Psychological Assessment, 10, 379–395. Czerwinski, M., Lightfoot, N., & Shiffrin, R. M. (1992). Automatization and training in visual search. American Journal of Psychology. Special Views and varieties of automaticity, 105(2), 271–315. Diederich, A. (1992). Probability inequalities for testing separate activation models of divided attention. Perception & Psychophysics, 52(6), 714–716.
Chapter 09.qxd
3/28/2007
12:59 PM
Page 255
9. ASSESSMENT OF MENTAL ARCHITECTURE
255
Diederich, A. (1995). Intersensory facilitation of reaction time: Evaluation of counter and diffusion coactivation models. Journal of Mathematical Psychology, 39(2), 197–215. Diederich, A., & Colonius, H. (1991). A further test of the superposition model for the redundant-signals effect in bimodal detection. Perception & Psychophysics, 50(1), 83–86. Egeth, H. E. (1966). Parallel versus serial processes in multidimensional stimulus discrimination. Perception & Psychophysics, 1, 245–252. Goldstone, R. L. (2000). Unitization during category learning. Journal of Experimental Psychology: Human Perception and Performance, 26, 86–112. Granholm, E., Asarnow, R. F., & Marter, S. R. (1996a). Dual-task performance operating characteristics, resource limitations, and automatic processing in schizophrenia. Neuropsychology, 10, 11–21. Granholm, E., Asarnow, R. F., & Marder, S. R. (1996b). Display visual angle and attentional scanpaths on the span of apprehension task in schizophrenia. Journal of Abnormal Psychology, 105, 17–24. Hartlage, S., Alloy, L. B., Vázquez, C., & Dykman, B. (1993). Automatic and effortful processing in depression. Psychological Bulletin, 113, 247–278. Heinrichs, R. W. (2005). The primacy of cognition in schizophrenia. American Psychologist, 60, 229–242. Kahneman, D. (1973). Attention and effort. Englewood Cliffs, NJ: Prentice-Hall. Knight, R. A., Manoach, D. S., Elliott, D. S., & Hershenson, M. (2000). Perceptual organization in schizophrenia: The processing of symmetrical configurations. Journal of Abnormal Psychology, 109, 575–587. Magaro, P. A. (1983). Psychosis and schizophrenia. In W. D. Spaulding & J. K. Cole (Eds.), Theories of schizophrenia and psychosis: Nebraska Symposium on Motivation (Vol. 31, pp. 157–229). Lincoln: University of Nebraska Press. McFall, R. M., & Townsend, J. T. (1998). Foundations of psychological assessment: Implications for cognitive assessment in clinical science. Psychological Assessment, 10(4), 316–330. Miller, J. (1982). Divided attention: Evidence for coactivation with redundant signals. Cognitive Psychology, 14, 247–279. Miller, J. (1986). Timecourse of coactivation in bimodal divided attention. Perception & Psychophysics, 40(5), 331–343. Neufeld, R. W. J. (2001, April). Formal models in explanation and measurement of cognitive psychopathology. Paper presented at Mount Sinai Workshop on Cognition in Schizophrenia: International Congress on Schizophrenia Research, Whistler, British Columbia, Canada. Neufeld, R. W. J. (in press-a). Composition and uses of formal clinical cognitive science. In W. Spaulding & J. Poland (Eds.), Modeling complex systems: Motivation, cognition and social processes: Nebraska Symposium on Motivation (pp. ). Lincoln: University of Nebraska Press. Neufeld, R.W.J. (in press-b). Introduction. In R. W. J. Neufeld (Ed.), Advances in clinical cognitive sciences: Formal modeling and assessment of processes and symptoms. Washington, DC: American Psychological Association Publications. Neufeld, R. W. J., & Broga, M. I. (1981). Evaluation of information sequential aspects of schizophrenic performance: II. Research strategies and methodological issues. Journal of Nervous and Mental Disease, 169, 569–579.
Chapter 09.qxd
256
3/28/2007
12:59 PM
Page 256
TOWNSEND, FIFIC′ , NEUFELD
Neufeld, R. W. J., Carter, J. R., Vollick, D., Boksman, K., Levy, L., & Jetté, J. (in press). A mathematical process-representation of stimulus encoding subserving memory search and other collateral processes, among paranoid and nonparanoid schizophrenic patients. In R. W. J. Neufeld (Ed.), Advances in clinical cognitive science: Formal modeling and assessment of processes and symptoms. Washington, DC: American Psychological Association. Neufeld, R. W. J., & McCarty, T. S. (1994). A formal analysis of stressor and stressproneness effects on simple information processing. British Journal of Mathematical & Statistical Psychology, 47(2), 193–226. Neufeld, R. W. J., Townsend, J. T., & Jetté, J. (in press). Quantitative response time technology for measuring cognitive-processing capacity in clinical studies. In R. W. J. Neufeld (Ed.), Advances in clinical cognitive science: Formal modeling and assessment of processes and symptoms. Washington, DC: American Psychological Association. Neufeld, R. W. J., Vollick, D., Carter, J. R., Boksman, K., & Jetté, J. (2002). Application of stochastic modeling to the assessment of group and individual differences in cognitive functioning. Psychological Assessment, 14(3), 279–298. Neufeld, R. W. J., Vollick, D., & Highgate, S. (1993). Stochastic modelling of stimulus encoding and memory search in paranoid schizophrenia: Clinical and theoretical implications. In R. L. Cromwell & R. C. Snyder (Eds.), Schizophrenia: Origins, processes, treatment, and outcome: The Second Kansas Series in Clinical Psychology (pp. 176–196). Oxford, England: Oxford University Press. Nuechterlein, K. H., Barch, D. M., Gold, J. M., Goldberg, T. E., Green, M. F., & Heaton, R. K. (2004). Identification of separable cognitive factors in schizophrenia. Schizophrenia Research, 72, 29–39. Parzen, E. (1960). Modern probability theory and its applications: Oxford, England: Wiley. Schneider, W., & Shiffrin, R. M. (1977). Controlled and automatic human information processing: I. Detection, search, and attention. Psychological Review, 84(1), 1–66. Schwarz, W. (1994). Diffusion, superposition, and the redundant-targets effect. Journal of Mathematical Psychology, 38(4), 504–520. Schweickert, R. (1978). A critical path generalization of the additive factor method: Analysis of a Stroop task. Journal of Mathematical Psychology, 18(2), 105–139. Schweickert, R. (1983). Latency network theory: Scheduling of processes in sentence verification and the Stroop effect. Journal of Experimental Psychology: Learning, Memory, & Cognition, 9(3), 353–383. Schweickert, R., Giorgini, M., & Dzhafarov, E. (2000). Selective influence and response time cumulative distribution functions in serial-parallel task networks. Journal of Mathematical Psychology, 44(4), 504–535. Schweickert, R., & Townsend, J. T. (1989). A trichotomy: Interactions of factors prolonging sequential and concurrent mental processes in stochastic discrete mental (PERT) networks. Journal of Mathematical Psychology, 33(3), 328–347. Shiffrin, R. M. (1988). Attention. In R. C. Atkinson & R. J. Herrnstein (Eds.), Stevens’ handbook of experimental psychology: Vol. 1. Perception and motivation: Vol. 2. Learning and cognition (2nd ed., pp. 739–811). Oxford, England: Wiley. Sternberg, S. (1966). High-speed scanning in human memory. Science, 153, 652–654.
Chapter 09.qxd
3/28/2007
12:59 PM
Page 257
9. ASSESSMENT OF MENTAL ARCHITECTURE
257
Sternberg, S. (1969). The discovery of processing stages: Extensions of Donders’ method. In W.G. Koster (Ed.), Attention and performance (Vol. 2, pp. 276–315). Amsterdam: North Holland. Sternberg, S. (1975). Memory scanning: New findings and current controversies. Quarterly Journal of Experimental Psychology, 27(1), 1–32. Townsend, J. T. (1969). Mock parallel and serial models and experimental detection of these. Paper presented at the Purdue Centennial Symposium on Information Processing, Purdue University. Townsend, J. T. (1971). A note on the identifiability of parallel and serial processes. Perception & Psychophysics, 10, 161–163. Townsend, J. T. (1974). Issues and models concerning the processing of a finite number of inputs. In B. H. Kantowitz (Ed.), Human information processing: Tutorials in performance and cognition (pp. 133–168). Hillsdale, NJ: Lawrence Erlbaum Associates. Townsend, J. T. (1990a). Serial versus parallel processing: Sometimes they look like Tweedledum and Tweedledee but they can (and should) be distinguished. Psychological Science, 1(1), 46–54. Townsend, J. T. (1990b). Truth and consequences of ordinal differences in statistical distributions: Toward a theory of hierarchical inference. Psychological Bulletin, 108(3), 551–567. Townsend, J. T., & Ashby, F. G. (1978). Methods of modeling capacity in simple processing systems. In J. Castellan & F. Restle (Eds.), Cognitive theory (Vol. III, pp. 200–239): Hillsdale, NJ: Lawrence Erlbaum Associates. Townsend, J. T., & Ashby, F. G. (1983). The stochastic modeling of elementary psychological processes. Cambridge, England: Cambridge University Press. Townsend, J. T., & Fific, M. (2004). Parallel versus serial processing and individual differences in high-speed search in human memory. Perception & Psychophysics, 66(6), 953–962. Townsend, J. T., & Nozawa, G. (1995). Spatio-temporal properties of elementary perception: An investigation of parallel, serial, and coactive theories. Journal of Mathematical Psychology, 39(4), 321–359. Townsend, J. T., & Wenger, M. J. (2004a). The serial-parallel dilemma: A case study in a linkage of theory and method. Psychonomic Bulletin & Review, 11(3), 391–418. Townsend, J. T., & Wenger, M. J. (2004b). A theory of interactive parallel processing: New capacity measures and predictions for a response time inequality series. Psychological Review, 111(4), 1003–1035. Treat, T. A., McFall, R. M., Viken, R. J., & Kruschke, J. K. (2001). Using cognitive science methods to assess the role of social information processing in sexually coercive behavior. Psychological Assessment, 13(4), 549–565. Treat, T. A., McFall, R. M., Viken, R. J., Nosofsky, R. M., MacKay, D. B., & Kruschke, J. K. (2002). Assessing clinically relevant perceptual organization with multidimensional scaling techniques. Psychological Assessment, 14(3), 239–252. Treisman, A., & Gormican, S. (1988). Feature analysis in early vision: Evidence from search asymmetries. Psychological Review, 95(1), 15–48. Van Zandt, T. (2000). How to fit a response time distribution. Psychonomic Bulletin & Review, 7(3), 424–465.
Chapter 09.qxd
258
3/28/2007
12:59 PM
Page 258
TOWNSEND, FIFIC′ , NEUFELD
Van Zandt, T. (2002). Analysis of response time distributions. In J. W. H. Pashler (Ed.), Stevens’ handbook of experimental psychology: Vol. 4. Methodology in experimental psychology (3rd ed., pp. 461–516): New York: Wiley. Van Zandt, T., & Townsend, J. T. (1993). Self-terminating versus exhaustive processes in rapid visual and memory search: An evaluative review. Perception & Psychophysics, 53(5), 563–580. Vollick, D. N. (1994). Stochastic models of encoding-latency means and variances in paranoid schizophrenia. Unpublished doctoral dissertation, University of Western Ontario, London, Ontario, Canada. Wenger, M. J., & Townsend, J. T. (2000). Basic response time tools for studying general processing capacity in attention, perception, and cognition. Journal of General Psychology, 127(1), 67–99.
Chapter 10.qxd
3/28/2007
1:00 PM
Page 259
10 Using a Simple Associative Learning Procedure to Study Clinical Disorders and Related Brain Function Joseph E. Steinmetz Indiana University
Richard McFall has spent the lion’s share of his research and academic career actively promoting what we now call translational research. As longtime director of clinical training at Indiana University, Dick McFall instilled into both undergraduate and graduate students the value of multi- and interdisciplinary research and clinical training (which he often refers to as hybrid training), especially relating clinical science to the fields of cognitive science and neuroscience. Typical of Dick, he was far ahead of the field. Translational research, defined as the marriage of basic and clinical research and involving a variety of research perspectives and approaches, is deservingly now at the forefront of clinical science. It is clear that those students and seasoned investigators who were trained to actively collaborate with others in their research endeavors, as Dick envisioned years ago, are well positioned to lead the field in exciting new directions in the years to come. Dick McFall has had a significant impact on the development of my research interests and directions, as recently my laboratory has become involved in a translational research program involving colleagues at 259
Chapter 10.qxd
3/28/2007
1:00 PM
Page 260
260
STEINMETZ
Indiana University and other institutions around the country. Over the last 10 years or so, we have used simple associative learning procedures to study clinical disorders and brain processes related to normal and pathological behavioral function. More specifically, we have used eyeblink classical conditioning, a basic form of associative motor learning, to explore sensory, motor, learning, and emotional processing in a variety of populations of subjects. Arguably, more is known about eyeblink conditioning at the behavioral and neural levels than any other learning procedure available. Due to the elegant work of Gormezano and his colleagues, among others, the effects of systematically manipulating parametric features of eyeblink conditioning on learning and performance of the response are well understood (see Gormezano, Kehoe, & Marshall, 1983, for review). Equally important is the fact that we currently know a lot about how the brain encodes this form of learning, thanks largely to the efforts of Thompson and his colleagues (see Steinmetz, 2000, for review). Given the advanced state of knowledge of behavioral and neural aspects of eyeblink conditioning, relatively specific predictions about brain–behavior correlates are now possible, including predictions about how behavioral and neural correlates may underlie or be related to clinical disorders. In this chapter, I present this translational approach by describing the eyeblink classical conditioning procedure, describing what we know about how the brain encodes this type of learning, and then describing how we have applied this knowledge and approach to study four seemingly unrelated disorders—fetal alcohol syndrome, autism, obsessive-compulsive disorder (OCD), and schizophrenia. EYEBLINK CLASSICAL CONDITIONING Eyeblink classical conditioning involves presentations of two stimuli: a conditioned stimulus (CS) that is followed in time by an unconditioned stimulus (US). Typically tones, lights, or tactile stimuli are used as CSs, and these stimuli do not produce overt movements by the subjects when presented early in training. Air puffs directed toward the cornea of the eye or mild electric shocks are most often used as USs, and these stimuli produce reliable eyeblinks when presented. The reflexive eyeblink produced by the presentation of the US is the unconditioned response (UR). Training involves paired presentations of the CS and US with onsets of the two stimuli separated by 150 to 2,500 msec. During early phases of conditioning, the US reliably elicits the reflexive UR. With 50 to 100 CS–US pairings,
Chapter 10.qxd
3/28/2007
1:00 PM
Page 261
10. CONDITIONING AND CLINICAL DISORDERS
261
an anticipatory eyeblink to the CS appears, which is called a conditioned response (CR). With enough CS–US pairings, a CR is established that rivals the UR in amplitude and duration (and, in fact, is often larger than the UR). The learned CR has another interesting property—it is exquisitely well timed. That is, in a well-trained subject, the peak amplitude of the CR occurs at the precise time when the US is presented. Indeed, if the CS–US interval is changed (i.e., either shortened or lengthened), the CR moves accordingly in just a few training trials so that the eyeblink amplitude is maximal at the precise time of US delivery. A number of variations of the basic conditioning procedure have been used to explore behavioral and neural function. The simplest procedure is delay conditioning and involves presentation of the CS with a totally or partially overlapping US. In trace conditioning, the CS is turned on and then off, a period of time is allowed to elapse, and then the US is presented. This slightly more complicated version of the paradigm introduces an additional requirement; because the CS and US do not overlap, the subject must remember the CS or hold a trace of the CS in memory. Other commonly used manipulations of the conditioning procedure include simple discrimination procedures where two different CSs are presented: one followed by a US and the other not followed by the US. The goal of this procedure is to establish CRs to one CS but not the other. Another discrimination procedure that has been used recently is interstimulus interval (ISI) discrimination. In this procedure, two CSs are presented: one that is followed by the US at one ISI and the other followed by the US at a second ISI. Using this procedure, it is possible to train subjects to perform two CRs appropriately timed for two different CSs, a useful procedure for studying behavioral and neural timing mechanisms. Other procedures that have been extensively studied include extinction, latent inhibition, conditioned inhibition, blocking, overshadowing, sensory preconditioning, conditional discrimination, second-order conditioning, and many other variations of the basic conditioning procedure (see Lavond & Steinmetz, 2002, for review). In summary, these studies have provided a rich database concerning this basic form of associative learning. THE NEURAL SUBSTRATES OF EYEBLINK CONDITIONING How the brain encodes learning and memory phenomena has been of interest to psychologists and brain scientists for more than a century. Our
Chapter 10.qxd
262
3/28/2007
1:00 PM
Page 262
STEINMETZ
understanding of the brain correlates of learning and memory has only recently taken giant steps forward, in part due to the development of technologies and experimental approaches that have advanced our ability to study brain function and in part due to the growing body of behavioral data, such as that described in the preceding section, from human and nonhuman experiments. Eyeblink classical conditioning has proved useful for studying the neural correlates of simple learning and memory. Indeed, the great deal of experimental control, the discrete-trial nature of the procedure, the relatively rapid rate of learning (one to two sessions), and the conservation across species of behavioral and neural aspects of the learning have made this procedure extremely amenable to studies of brain substrates of learning (Lavond & Steinmetz, 2002). In the early 1980s, Richard Thompson and his colleagues made a series of seminal discoveries that identified brain regions that are now thought to be the necessary and sufficient neural circuitry for this elementary form of learning. Using lesion, recording, anatomy, stimulation, and pharmacology methods, Thompson and associates showed that regions of the cerebellum contained neurons that are involved in encoding the acquisition and performance of eyeblink conditioned responses (e.g., Mamounas, Thompson, & Madden, 1987; McCormick & Thompson, 1984a, 1984b; Steinmetz, Lavond, & Thompson, 1989). Over the last 20 years or so, efforts in a number of laboratories have further delineated the neural circuitry and have begun studying network properties and molecular and cellular processes involved in establishing and maintaining plasticity in neurons and synapses involved in the conditioning process (Bao, Chen, & Thompson, 1998; Cooke, Attwell, & Yeo, 2004; Gould & Steinmetz, 1996; Mauk, Steele, & Medina, 1997; Moore, Desmond, & Berthier, 1989; Schreurs, Sanchez-Andres, & Alkon, 1991, 1992; Tracy, Britton, & Steinmetz, 2001). Figure 10.1 shows what is thought to be the primary essential neural circuitry for eyeblink conditioning involving a tone CS and an air puff US. Information concerning tone CS presentation is thought to be relayed from primary auditory nuclei to neurons in the basilar pontine nuclei. These pontine neurons send mossy fiber axons to discrete regions of cerebellar cortex and also collateral axons to the interpositus nucleus. Information concerning air puff US presentation is thought to be relayed from the cornea of the eye to neurons in the trigeminal nucleus, where two routes for US input emerge. One route is to cranial nerve nuclei that participate in eyeblinking via brainstem reticular neurons. This route is
Chapter 10.qxd
3/28/2007
1:00 PM
Page 263
10. CONDITIONING AND CLINICAL DISORDERS
263
FIGURE 10.1 Essential cerebellar and brainstem circuitry thought to be involved in eyeblink classical conditioning. Basic plasticity that underlies the acquisition and performance of the eyeblink CR is thought to occur in regions of the cerebellar cortex and the interpositus nucleus where CS inputs and US inputs converge. From Steinmetz (2000), with permission.
thought to mediate UR execution. The second route is to neurons in the inferior olivary complex that in turn project climbing fiber axons to cerebellar cortex and the interpositus nucleus. Although there are some differences of opinion as to how cortical and nuclear areas interact to produce conditioned responding, some neural network models of conditioning posit that plasticity is established in both regions of the cerebellum where CS and US inputs (i.e., pontine mossy fibers and inferior olive climbing fibers) converge (Katz & Steinmetz, 2002; Mauk & Donegan, 1997; Steinmetz, 2000). Interestingly, cerebellar cortical regions that receive convergent CS–US input (which include Larsell’s lobule HVI and discrete areas of the anterior lobe) project inhibitory input to regions of the interpositus nucleus that receive convergent
Chapter 10.qxd
264
3/28/2007
1:00 PM
Page 264
STEINMETZ
CS–US input. One hypothesis being actively explored is that conditioning establishes plasticity that involves nuclear neurons that act like a gate to drive motor neurons in brainstem nuclei that produce eyeblinks (Steinmetz, 2000). Cortical plasticity in this model is thought to modulate activity in nuclear neurons, which affects important gains and timing mechanisms associated with the CR. Feedback from the interpositus nucleus onto CS and US inputs to cortex (and the nucleus, for that matter) may be important for dynamically controlling timing and other corticalmediated features of the CRs. Much of our understanding of this basic neural circuitry for eyeblink conditioning comes from lesion and recording experiments involving mainly rabbits and rats. Initially, electrolytic lesions of the interpositus nucleus were shown to completely abolish previously learned eyeblink CRs and also to prevent acquisition of CRs when the lesions were administered before training (Lincoln, McCormick, & Thompson, 1982; McCormick, Lavond, Clark, Kettner, Rising, & Thompson, 1981). Later studies showed that chemical lesions (e.g., with kainic acid) and temporary inactivation (e.g., with muscimol or brain cooling) produced the same effect as electrolytic lesions, demonstrating that interpositus cells were critically involved in the conditioning (Clark, Zhang, & Lavond, 1992; Krupa, Thompson, & Thompson, 1993; Lavond, Hembree, & Thompson, 1985). The effect is permanent—rabbits trained daily for more than 12 months after the lesion show no signs of recovery of CRs (Steinmetz, Logue, & Steinmetz, 1992). Lesions of cerebellar cortex have produced more variable results. For example, lesions of lobule HVI have produced CR abolition in some laboratories (e.g., Yeo & Hardiman, 1992; Yeo, Hardiman, & Glickstein, 1985) and more temporary and less severe effects in other laboratories (Lavond & Steinmetz, 1989; Lavond, Steinmetz, Yokaitis, & Thompson, 1987). Lesions of the anterior lobe have consistently produced mistimed CRs, suggesting that this cortical region may be crucial for normal CR timing (Perrett, Ruiz, & Mauk, 1993). Single-unit brain-recording experiments have also provided data suggesting the involvement of the cerebellum in eyeblink conditioning. Recordings made in the interpositus nucleus have revealed populations of neurons that respond to the presentation of the CS and the US (e.g., Berthier & Moore, 1990; Katz & Steinmetz, 1997). More important, neurons in the interpositus nucleus show firing patterns that are strongly related to performance of the CR. These neurons form tight amplitude-time
Chapter 10.qxd
3/28/2007
1:00 PM
Page 265
10. CONDITIONING AND CLINICAL DISORDERS
265
course models of the CR and are activated 30 to 50 msec prior to onset of the behavioral response, precisely what would be expected given synaptic delays for a population of cells responsible for driving brain stem motor neurons that are responsible for generating the eyeblink. Learningrelated unit responses can also be seen in records taken from cerebellar cortex (Berthier & Moore, 1986; Gould & Steinmetz, 1996). Cortical neurons show CS- and US-related responses as well as spiking patterns that appear to be related to execution of the CR. Interestingly, both excitatory and inhibitory patterns of unit discharge can be seen—an observation that is compatible with a modulatory role for the cerebellar cortex in eyeblink conditioning. Together with the lesion data, the singleunit recording data have helped build a compelling case for the essential involvement of the cerebellum and brainstem in eyeblink conditioning. Indeed, it appears that plasticity in populations of neurons in the cerebellum and related circuitry may be the essential neural substrate of eyeblink conditioning. One of the major reasons that lower brain areas were explored for their involvement in eyeblink conditioning was the fact that many previous studies had shown that higher brain regions were not necessary for acquisition and performance of delay eyeblink classical conditioning. Aspiration lesions of all or part of cerebral cortex (including the hippocampus and other limbic areas) had little effect on eyeblink conditioned responding, and conditioning was found to be possible after decerebration (Mauk & Thompson, 1987; Oakley & Russell, 1972, 1976). Do these findings demonstrate that higher brain regions are not involved in eyeblink conditioning? The answer to this question appears to be no. Several studies indicate that a number of higher brain regions are engaged and play important roles in eyeblink conditioning. For example, single-unit recordings taken from the hippocampus have shown that the firing patterns of many neurons there are changed by paired CS-US presentations (e.g., Berger & Thompson, 1978). Moreover, lesions of the hippocampus have been shown to severely impair variations of eyeblink conditioning, including trace conditioning and reversal of a simple discrimination (Moyer, Deyo, & Disterhoft, 1990; Orr & Berger, 1985). In addition, more recent studies involving other areas such as the amygdala, striatum, thalamus, and frontal cortex have consistently demonstrated a role for these (and other) brain regions in conditioning (Blankenship, Huckfeldt, Steinmetz, & Steinmetz, 2005; McLaughlin, Powell, & White, 2002; White, Miller, White, Dike, Rebec, & Steinmetz, 1994). The consensus in the field seems to be that,
Chapter 10.qxd
266
3/28/2007
1:00 PM
Page 266
STEINMETZ
although plasticity in the cerebellum and associated brainstem pathways is necessary for basic associative response formation (and perhaps optimal response timing), a variety of higher brain areas may be engaged when the conditioning procedure becomes more complex, such as in trace and discrimination conditioning. Much of the work that is currently underway in laboratories around the country is aimed at delineating the specific conditions under which different brain regions are engaged and critically involved in learning the eyeblink CR. Indeed, these studies may go a long way in advancing our understanding of the role of these higher brain areas in other behavioral and cognitive tasks. PARALLELS OF HUMAN AND NONHUMAN EYEBLINK CONDITIONING The great majority of research on eyeblink conditioning that has been conducted over the last 40 to 50 years has involved nonhuman animals. Much of this research has used the rabbit as an experimental subject, although more recently the rat and mouse have gained popularity as subjects. Interestingly, however, human subjects were first used in eyeblink conditioning nearly a century ago, long before nonhuman animals were used. Various phenomena were studied, including basic motor processing, emotional responding and pathology (including anxiety disorders), involuntary learning, the influence of instructions on learning, and aging effects (see Woodruff-Pak & Steinmetz, 2000a, for review). Rabbits replaced humans as subjects in the early 1960s for a variety of reasons (Woodruff-Pak & Steinmetz, 2000b). For example, humans demonstrate high spontaneous blink rates and are greatly affected by instructions that are given before conditioning. Additionally, humans can use voluntary (declarative) learning and memory mechanisms to learn and remember the eyeblink conditioned response, and this kind of learning may be somewhat different from and engage different neural substrates than more procedurally based, involuntary, eyeblink conditioning. In short, the rabbit proved to be a better model for studying behavioral correlates and laws associated with learning, and, as pointed out earlier, this model has proved extremely useful for advancing our understanding of how the brain encodes this type of associative learning. These latter studies could not be conducted with human subjects. Over the last 10 to 15 years, however, there has been a renewed interest in using eyeblink conditioning to study behavioral, cognitive,
Chapter 10.qxd
3/28/2007
1:00 PM
Page 267
10. CONDITIONING AND CLINICAL DISORDERS
267
and neural functions in humans (Steinmetz, 1999; Woodruff-Pak & Steinmetz, 2000b). The number of human eyeblink conditioning studies has grown exponentially over this time period and includes investigations of brain damage, degenerative brain disorders, normal aging effects, information-processing capabilities, and sensory processing, to list a few examples (see Woodruff-Pak & Steinmetz, 2000a). Why has this resurgence of interest in human eyeblink conditioning occurred? The answer may be twofold. First, there appears to be a close parallel between humans and nonhumans in behavioral aspects of eyeblink conditioning, especially when relatively short ISIs are used (as longer ISIs allow humans to use voluntary responding strategies). At some combinations of conditioning parameters, humans learn the conditioned response at the same rate as other mammals, including rabbits, rats and mice, and also show similar effects when the basic delay procedure is manipulated (e.g., when trace conditioning, simple discrimination learning, or inhibitory conditioning procedures are introduced). In essence, behavioral responding in eyeblink conditioning appears to be conserved across mammalian species. A second reason for the renewed interest in human eyeblink conditioning may be the parallels that have been discovered between humans and nonhumans in the neurobiology of eyeblink conditioning. Data collected to date suggest that the same basic neural systems are critical for eyeblink conditioning in all mammals, including humans. For example, humans with cerebellar damage show severe impairments in all forms of eyeblink conditioning (Daum & Schugens, 1996; Daum et al., 1993), and humans with limbic-system damage that includes the hippocampus and related cortical structures appear to show some impairments in relatively complex variations of eyeblink conditioning, such as trace conditioning (McGlinchey-Berroth, Carrillo, Gabrieli, Brawn, & Disterhoft, 1997; but see Woodruff-Pak, 1993). Further, recent functional magnetic resonance imaging (fMRI) studies have shown that cerebellar and limbic-system areas are activated during conditioning as predicted from the animal studies (e.g., Lemieux & Woodruff-Pak, 2000). These data suggest that, similar to behavioral responding, the neural substrates of eyeblink conditioning are conserved across mammalian species. Therefore, a strong case can be made for the translation of basic animal eyeblink conditioning data for use in exploring behavioral and neural correlates of clinical disorders involving humans.
Chapter 10.qxd
3/28/2007
1:00 PM
Page 268
268
STEINMETZ
USING EYEBLINK CONDITIONING TO STUDY CLINICAL DISORDERS At first glance, it may seem odd that one would attempt to use eyeblink classical conditioning, a simple associative learning task, to study rather complex clinical disorders, except for, of course, its obvious use to study disorders of procedural learning and memory. I would argue, however, that given the relatively advanced state of knowledge that exists concerning the behavioral and neural correlates of eyeblink conditioning, rather specific predictions can be made about behavioral and neural pathologies that may be related to a variety of clinical disorders. In the realm of learning and memory disorders, for example, Squire and his colleagues have conducted an elegant series of eyeblink conditioning studies that have examined the role of awareness in learning and memory in amnesic and normal subjects (see Clark & Squire, 2000, for review). These studies were based largely on the findings that a dissociation between brain areas involved in declarative versus nondeclarative learning could be explored using eyeblink conditioning, as previous studies showed that delay conditioning requires the cerebellum whereas trace conditioning requires the cerebellum and hippocampus. They predicted (and subsequently observed) that learning trace conditioning required awareness of the CS–US contingency, as reported by the subjects, whereas delay conditioning did not. In this example, knowledge of the brain areas involved in variations of eyeblink conditioning was used in combination with parametric manipulations of the paradigm to explore features of a clinical disorder (awareness and conditioning in amnesics as well as normal adults). Using a similar strategy, we have used eyeblink classical conditioning to explore brain and behavioral correlates of a number of clinical disorders. Presented here is a brief summary of our research involving four rather different disorders—fetal alcohol syndrome, autism, schizophrenia, and OCD. Before summarizing data we have collected concerning these four disorders, a brief description of each disorder is presented along with our reasoning as to why we used eyeblink conditioning to explore biobehavioral aspects of each disorder. AN ANIMAL MODEL OF FETAL ALCOHOL SYNDROME Exposure to alcohol during critical periods of prenatal development can subsequently result in a variety of behavioral and physiological effects
Chapter 10.qxd
3/28/2007
1:00 PM
Page 269
10. CONDITIONING AND CLINICAL DISORDERS
269
that are collectively known as fetal alcohol syndrome (FAS). Physiological effects of prenatal alcohol exposure include central nervous system (CNS) damage, facial abnormalities, and growth deficits (e.g., Roebuck, Mattson, & Riley, 1998), whereas behavioral effects include disruptions in motor development, learning and memory impairments, decreased IQ, and hyperactivity (e.g., Mattson & Riley, 1998; Streissguth & O’Malley, 2000). The timing of alcohol exposure appears to determine, to a large extent, the physiological and behavioral abnormalities that are seen. For example, exposure during the third trimester is associated with abnormal cerebellar cellular development, as well as deficits in hippocampal pyramidal-cell development (Guerri, 1998; Hannigan, 1996). A number of animal models of FAS have been developed and used successfully to study behavioral and physiological correlates of early alcohol exposure. One of these preparations, developed by West, Goodlett, and their associates, has proved useful (e.g., Marcussan, Goodlett, Mahoney, & West, 1994). In this model, rats are exposed to alcohol during a portion of the first 10 postnatal days. This postnatal period in the rat is roughly equivalent to the third trimester in humans, as much neural development in the rat takes place during the first 2 weeks after birth. Studies using this rat model have found that both the hippocampus and cerebellum (along with some other brain areas) are profoundly impacted by alcohol exposure given on postnatal days 4 to 9. For example, cell-counting experiments have shown a reduction in the number of cerebellar Purkinje cells when blood alcohol levels greater than 200 mg/dl were produced in the rats pups during exposure, most likely due to the interruption of Purkinje cell differentiation, which is ongoing during this postnatal period (Bonthius & West, 1990, 1991; Hamre & West, 1993). The period of greatest vulnerability appears to be around postnatal days 4 to 5 (Goodlett & Eilers, 1997). In addition to cerebellar cell loss, reductions in inferior olivary cell numbers have also been reported (Napper & West, 1995). The neonatal alcohol exposure has also been found to impair behavioral performance. For example, performance in the rotating rod task has been studied in rats neonatally exposed to alcohol (Goodlett, Thomas, & West, 1991). In this task, a rat is placed on a stationary rod that is set in motion at increasing speeds, and the time the rat remains on the rod is noted. Neonatal alcohol-exposed rats were severely impaired in this task, a result that may not be surprising—the cerebellum is critically involved in coordinating the movements necessary for this task, and this brain region is significantly affected by postnatal alcohol exposure.
Chapter 10.qxd
270
3/28/2007
1:00 PM
Page 270
STEINMETZ
FIGURE 10.2 Acquisition and extinction of eyeblink conditioning in rats give ethanol (circles), sham intubations (triangles), or no intubations (squares) on postnatal days 4 to 9. The rats were conditioned at age 5 to 6 months. Bars = SEM. From Green, Rogers, Goodlett, and Steinmetz (2000), with permission.
Given the large cell loss noted in the cerebellum after alcohol exposure on postnatal days 4 to 9 and given the critical dependence of eyeblink conditioning on cerebellar function, we hypothesized that lasting deficits in eyeblink conditioning would be seen in rats exposed to alcohol neonatally. This was indeed confirmed in an initial experiment we conducted (Green et al., 2000). In this study, three groups of rats were used. One group received daily intragastric infusions of alcohol in milk formula on postnatal days 4 to 9. A second group received the milk infusions with no alcohol on postnatal days 4 to 9. A third group was untreated. The rats were then set aside until they reached 5 to 6 months of age and then were given 10 days of eyeblink conditioning training with a tone CS and periorbital shock US followed by 4 days of CS-alone extinction. Our results were quite clear. The rats that were neonatally exposed to ethanol 6 months prior to training showed a severe deficit in learning when compared with the sham-intubated and nonintubated control rats (see Fig. 10.2). That is, the early alcohol exposure had a profound and lasting effect on behavior that critically involved the cerebellum.
Chapter 10.qxd
3/28/2007
1:00 PM
Page 271
10. CONDITIONING AND CLINICAL DISORDERS
271
To directly test the hypothesis that deficits in eyeblink conditioning observed in our initial study were caused by deficits in cerebellar processing, we conducted two follow-up studies. In the first of these studies, we exposed rats to alcohol on postnatal days 4 to 9, allowed them to mature to adulthood, trained them in eyeblink conditioning using a tone CS and periorbital shock US, then removed the rats’ brains and counted the number of neurons present in the deep cerebellar nuclei (Green, Tran, Steinmetz, & Goodlett, 2002). Our behavioral data were similar to our initial study: A severe deficit in conditioning was seen in the alcoholexposed rats. Further, we found that the alcohol-exposed rats had a 50% reduction in deep nuclear cells compared with the two control groups, thus suggesting that the conditioning deficit may be caused by loss of cerebellar cells. In a second study, we exposed rats to alcohol neonatally; when they were adults, we implanted recording electrodes into the interpositus nucleus and then trained them in the eyeblink conditioning task while concomitantly recording from the nucleus (Green, Johnson, Goodlett, & Steinmetz, 2002). Again, we replicated the basic behavioral deficit in the alcohol-exposed rats. Our interpositus nucleus recording revealed normal learning-related responses in cells that were isolated. However, not as many cells were found as in the control groups. We interpreted this to indicate that, although the cells found in the alcoholexposed rats functioned normally, there were not enough of the neurons present to drive a conditioned response via activation of brainstem motor neurons. This use of classical eyeblink conditioning to study behavioral and neural function associated with this rat model of FAS was initiated by the observation that this third-trimester exposure produced severe deficits in cerebellar anatomy and physiology and by our knowledge that the cerebellum was essential for eyeblink conditioning. In this application, a known brain pathology associated with a clinical disorder (i.e., cerebellar cell loss due to early alcohol exposure) was thought to involve neurons that were critically involved in the acquisition and performance of a learned response (i.e., eyeblink classical conditioning). Eyeblink conditioning affords a great deal of experimental control and opportunities for varying the basic procedure. With this in mind, we are currently following up these studies with other experiments designed to assess potential cerebellar-mediated timing abnormalities in conditioning using ISI shift procedures, to explore dose-response relationships at exposure time, and also to study hippocampal-mediated conditioning deficits that may be produced by using slightly later exposure times to alcohol and by using
Chapter 10.qxd
3/28/2007
1:00 PM
Page 272
272
STEINMETZ
trace versus delay conditioning tasks. In the future, we also hope to explore eyeblink conditioning in children with FAS, as we expect they may have difficulties with this task along with other cerebellar-mediated tasks. AUTISM Our earliest attempt to use eyeblink conditioning to study a clinical disorder was to study the acquisition and performance of eyeblink CRs in persons with autism. Autism is characterized by severe impairments in communication and social relationships and is often accompanied by ritualistic and repetitive patterns of behavior (American Psychiatric Association, 1994). Although the disorder is known to be associated with a variety of neurological pathologies, a growing body of literature began to appear in the 1980s that suggested that the cerebellum may be pathological in autism. Histoanatomical and brain-imaging studies have shown hypoplasia of the posterior cerebellar vermis and hemispheres and an age-related variation of cell numbers present in the deep cerebellar nuclei (e.g., Bauman, 1991; Bauman & Kemper, 1985; Courchesne, Yeung-Courchesne, Press, Hesselink, & Jernigan, 1988). Some other recent data have shown some critical period-dependent pathology in the medullas of persons with autism that may impact the development and later function of the cerebellum (e.g., Rodier, Bryson, & Welch, 1997). Based on these data suggesting cerebellar pathologies in autistics, we made a simple prediction and designed an experiment to test our hypothesis: Because eyeblink conditioning is heavily dependent on cerebellar function and autistics appear to have cerebellar pathologies, we predicted that we would see abnormal eyeblink conditioning when persons with autism were trained in a simple delay procedure (Sears, Finn, & Steinmetz, 1994; Sears & Steinmetz, 2000). We recruited 11 persons with autism for this study who ranged in age from 7 to 22 years. Eleven age-, gender-, and IQ-matched control subjects were also used. All subjects were given two sessions of eyeblink conditioning training using a tone CS and air puff US. CS-alone extinction training was given during the second half of the second session. Before collecting the data, we expected to see a decreased level of conditioning in the autistic subjects given the cerebellar pathology purported to exist in these individuals. We were quite surprised by the results: The autistic group showed a facilitation of conditioning relative to the control group (see Fig. 10.3). Initial extinction measured over the first block of training
Chapter 10.qxd
3/28/2007
1:00 PM
Page 273
10. CONDITIONING AND CLINICAL DISORDERS
273
FIGURE 10.3 Percent CRs recorded in autistic (filled circles) and control (open circles) subjects during 2 days of eyeblink classical conditioning. A1, A2, and A3 were blocks of training involving presentations of the US alone. The autistic subjects showed facilitated acquisition relative to the control subjects. Bars = SEM. From Sears, Finn, and Steinmetz (1994), with permission.
was more rapid in the autistic group, but overall both groups declined to similar levels by the end of extinction training. We examined UR amplitudes and found no differences, suggesting that processing of the aversive air puff US was similar in the two groups. Analysis of CR timing characteristics, however, revealed an interesting and consistent difference between the two groups (see Fig. 10.4). The onset and peak latencies of the CRs were shifted forward in time in the autistic subjects, such that these subjects began and ended their learned blinks earlier than controls. In essence, the autistic subjects mistimed their CRs—they closed their eyes and began opening them again before the air puff was presented (i.e., 350 msec after tone onset). Further, there was a clear age-related trend in the autistic data, as the facilitated conditioning and mistimed responses were more prevalent in younger subjects than in older subjects. We have speculated that the facilitated and mistimed CRs may be the result of a development-related imbalance involving the cerebellar cortex and the deep cerebellar nuclei. Normally, cerebellar Purkinje cells exert tonic inhibitory control on neurons in the deep cerebellar nuclei that contributes to the overall excitability of the nuclear output. With fewer
Chapter 10.qxd
274
3/28/2007
1:00 PM
Page 274
STEINMETZ
FIGURE 10.4 Onset latencies (top graph) and peak latencies (bottom graph) of CRs performed by autistic (filled circles) and control (subjects) during 2 days of eyeblink classical conditioning. The three blocks labeled “A” were blocks of training involving presentations of the US alone. The autistic subjects showed CRs with earlier onset and peak latencies. Bars = SEM. From Sears, Finn, and Steinmetz (1994), with permission.
Purkinje cells in autistics, it is possible that there is a reduced tonic inhibitory influence of the cortex on the nuclei, thus resulting in a hyperexcitable nucleus. Pairing-specific plasticity in the nucleus may not be checked by cortical inhibition, resulting in facilitated acquisition and mistimed response. Alternatively, it is possible that Purkinje cell loss is selective to cortical areas involved in modulating CR timing (e.g., anterior lobe neurons), and the mistimed CRs are due to the lack of this important modulatory input to the deep nuclei. To date, we have not followed up these studies with others designed to explore this effect more closely. It seems important to conduct studies in
Chapter 10.qxd
3/28/2007
1:00 PM
Page 275
10. CONDITIONING AND CLINICAL DISORDERS
275
two areas. First, future studies should examine the effects of manipulating CR timing via ISI shift and ISI discrimination procedures and also manipulating CS and US intensity to explore the specificity of potential cerebellar deficits associated with response timing and stimulus processing. Second, assuming that the facilitation and mistiming effects are replicated, it is important to establish that these are actually related to cerebellar pathologies in autistics. This could be done using brain imaging techniques to examine potential structural deficits in the cerebellum that may be related to deficits in conditioning. The use of eyeblink conditioning to study behavioral and neurobiological correlates of autism was based on a growing literature suggesting that persons with autism have cerebellar abnormalities and a very large literature showing that the cerebellum is essential for eyeblink conditioning. Interestingly, although the wholesale deficits in conditioning that were expected were not seen, the paradigm was sensitive enough to pick up acquisition and timing differences that may help explain the specific deficits (behavioral and neural) present in autistics. We do not know what causes autism. However, it seems very likely that problems in neural development underlie the disorder. From a neurological perspective there appear to be similarities between FAS and autism, especially when cerebellar function is considered. Indeed, it is tempting to speculate that both disorders are the result of interruptions in normal neural development, albeit by different causal factors (e.g., alcohol for FAS; unknown factors for autism) and perhaps at different developmental times. Eyeblink conditioning may prove useful for exploring this possibility. SCHIZOPHRENIA In collaboration with my colleague William Hetrick and his associates, my laboratory has recently begun using eyeblink conditioning to explore behavioral and neural correlates of schizophrenia (Brown et al., 2005). This research line is based largely on the idea that the classic symptoms of schizophrenia (formal thought dissorder, disorganized and bizarre behavior, and other neurological signs) may be the result of a disturbance of the temporal coordination of information processing in the brain. Supporting this idea is the observation of timing deficits in schizophrenia. For example, deficits in time estimation tasks and in temporal production and reproduction tasks have been reported (Densen, 1977; Johnson & Petzel, 1971; Tysk, 1983; Volz et al., 2001). We have reasoned that, given the participation of the cerebellum in tasks that are highly time dependent,
Chapter 10.qxd
276
3/28/2007
1:00 PM
Page 276
STEINMETZ
it is possible the cerebellum and related structures may be pathological in schizophrenia, an idea that has been entertained by others. For example, Andreasen, Paradiso, and O’Leary’s (1998) cognitive dysmetria model posits that the symptoms of schizophrenia may be the result of a disturbance in the fluid temporal coordination of motor, perceptual, and cognitive sequences of behavior that is controlled by a neural circuit composed of a cortico-cerebello-thalamo-cortico loop through the brain. We have begun systematically exploring the cerebellar link in this processing loop with the idea that disorders in timing and temporal coordination in schizophrenia may originate in part from cerebellar dysfunction. The cerebellum has often been linked to response timing (e.g., Fiala, Grossberg, & Bullock, 1996; Ivry, Keele, & Diener, 1989; Steinmetz, 2000), and some have suggested it may contribute to cognitive functioning (e.g., Leiner, Leiner, & Dow, 1991). Some studies have reported cerebellar abnormalities in schizophrenia such as decreases in cerebellar volume (e.g., Weinberger, Torrey, & Wyatt, 1979). Therefore, it appears that the cerebellum may be worthy of further study for its potential role in mediating the symptoms of schizophrenia. We recently completed our initial study of eyeblink conditioning in schizophrenics, and these early results suggest that this approach may be promising for advancing our understanding of this disorder (Brown et al., 2005). For this study, we used 13 participants with schizophrenia and 13 age- and sex-matched nonpsychiatric comparison subjects. All schizophrenic participants were taking antipsychotic medication at the time of training. All subjects received 100 acquisition trials comprised of paired tone CS and air puff US presentations (ISI = 350 msec), followed by 25 CS alone and 25 US-alone extinction trials. Our results indicate an impairment of conditioning in the schizophrenic participants. During early stages of acquisition, performance was similar in the two groups. With additional training, however, control subjects showed significantly more CRs than the schizophrenics. We looked at the timing of the CRs in the two groups. No significant differences in average onset or peak latencies were seen when the groups were compared. However, when coefficients of variability for onset latencies were calculated across training, group differences were apparent—although the control subjects showed a large decrease in response timing variability across training, the schizophrenia subjects did not. No differences in extinction training were found when these data were analyzed. The high within-subject variability in response timing that we observed may be the cause of the relatively poor
Chapter 10.qxd
3/28/2007
1:00 PM
Page 277
10. CONDITIONING AND CLINICAL DISORDERS
277
learning that we observed in the schizophrenia. In addition, this high variability in response timing is consistent with models of schizophrenia in which timing deficits underlie information processing abnormalities and clinical features of the disorder. Others have attempted to use eyeblink conditioning to study schizophrenia, and the results have been mixed. For example, Sears, Andreasen, and O’Leary (2000) found accelerated acquisition of CRs in schizophrenics, whereas Hofer, Doby, Anderer, and Dantendorfer (2001) found impaired learning in a group of schizophrenics trained on an eyeblink conditional discrimination paradigm. One important difference between these studies is that the Sears et al. study involved nonmedicated subjects whereas the Hofer et al. study, similar to the present study, used medicated subjects. Importantly, neither study thoroughly examined potential variability in CR timing, which may be very important for understanding differences in CR acquisition and performance seen in schizophrenics. In ongoing work, we are systematically analyzing several factors that contribute to the conditioning process in schizophrenic and comparison subjects, such as response timing issues, stimulus intensity factors, task complexity effects (e.g., delay vs. trace conditioning effects), and the effects of medication on eyeblink CR learning and performance. In addition, we are conducting a parallel set of rat eyeblink conditioning studies looking at some of these same issues in hopes of exploring network and cellular mechanisms in an animal model that may contribute to the temporal processing deficits noted in schizophrenia. The use of eyeblink classical conditioning to study behavioral and neural correlates of schizophrenia was inspired by a series of interconnected findings. First, the cerebellum is centrally involved in eyeblink conditioning. Second, cerebellar abnormalities have been observed in schizophrenics. Third, cerebellar pathology or damage is known to cause deficits in temporal response processing. Fourth, temporal processing deficits have been observed in schizophrenics. Thus, similar to the animal model created for FAS and our work with autistics, our interest in using eyeblink classical conditioning to study schizophrenics is based largely on the brain–behavioral correlates we have established between the cerebellum and this form of associative learning. Further, in many ways, we believe that this line of research serves as a good example of translational research as it involves different levels of analysis, uses basic science as well as clinical science approaches, and it is likely that the basic side of the research will inform the clinical side of the research and vice versa.
Chapter 10.qxd
3/28/2007
1:00 PM
Page 278
278
STEINMETZ
OBSESSIVE-COMPULSIVE DISORDER Obsessive-compulsive disorder (OCD) is characterized by unwanted, intrusive, and uncontrollable thoughts, images, or urges that are often accompanied by repetitive behaviors or mental acts that the person feels compelled to perform. It is an anxiety disorder with a lifetime prevalence of 2% to 3% in the United States. Many have suggested that OCD symptoms are the result of associative learning and thus are not easily extinguished. In fact, one of the leading treatments for OCD is exposure with response prevention (ERP) that is based in part on Mowrer’s (1960) two-stage theory. This theory states that the acquisition and maintenance of fear and avoidance arises because neutral objects become associated with fear and anxiety through association with a negative event. Fear and anxiety are thought to be subsequently reduced by engaging in repetitive behavior that is negatively reinforcing. ERP may work by breaking the maladaptive associations such that the compulsive behavior is reduced or extinguished. What further can be learned by using eyeblink classical conditioning to study OCD? The answer lies in the fact that Mowrer’s theory (and, for that matter, the use of ERP as a treatment) involves predictions that come directly from learning theory. We reasoned that if persons with OCD had a propensity to form associations between neutral stimuli or contexts and negative events, it might be possible to see this in a controlled laboratory environment using eyeblink conditioning procedures, which require subjects to make associations between a neutral stimulus (a tone) and an aversive stimulus (an air puff). In short, we predicted that persons showing symptoms related to OCD would demonstrate facilitated rates of eyeblink conditioning compared to age-matched control subjects. We conducted two experiments, both involving undergraduate college student subjects selected from a pool of introductory psychology classes (Tracy, Ghose, Stecher, McFall, & Steinmetz, 1999). Individuals who scored in the top 4% on the Maudsley Obsessional-Compulsive Inventory (MOCI) were chosen as experimental subjects, whereas control subjects were gender matched from the lower 75% of the MOCI that was administered. In the first experiment, subjects were given 100 paired presentations of a tone CS and air puff US (delay paradigm; 400 msec ISI) followed by 30 CS-alone extinction trials. During training, subjects performed a visual search task designed to maintain attentional and arousal levels during the relatively boring eyeblink conditioning procedure. For this task, subjects
Chapter 10.qxd
3/28/2007
1:00 PM
Page 279
10. CONDITIONING AND CLINICAL DISORDERS
279
FIGURE 10.5 Averaged traces of eyelid movement recorded over 10 blocks of eyeblink classical conditioning for control subjects (left column) and persons with OCD-like symptoms (right column) during two experiments (see text for details). Block 1 responses are represented by the top trace in each panel. The first vertical line in each panel denotes CS onset whereas the second vertical line denotes US onset. Note the similarity in responding across groups during Experiment 1 and the relative facilitation of conditioning seen during early blocks in the OCD group in Experiment 2. From Tracy, Ghose, Stecher, McFall, and Steinmetz (1999), with permission.
were presented with a series of 5 × 5 matrixes on a computer screen composed of 25 blue Ds and green Os randomly presented in the matrix. One half of the matrices contained one green D. The subjects were given 10 sec to determine whether the green D was present in the matrix. There was no correlation between the presentation of the matrices and the presentation of eyeblink conditioning trials. We examined the usual conditioning variables in the study: CR frequency, CR amplitude, onset latency, and peak latency. The results of this experiment were unambiguous—no differences in rates or levels of eyeblink conditioning were found when the OCD-like and control subjects were compared (see Fig. 10.5). Further, there were no differences in performance of the visual search task when the two groups were compared. The results of the first experiments suggested that OCD-like subjects did not show the facilitated conditioning that we hypothesized, given the Mowrer model and success of the ERP treatment regimen. However, before abandoning this idea altogether, we did note an intriguing potential flaw in our design that was worthy of further pursuit. We came up with the visual search task to combat the relative boredom associated with the
Chapter 10.qxd
280
3/28/2007
1:00 PM
Page 280
STEINMETZ
eyeblink conditioning task (actually as a more engaging alternative to silent movies that are often shown during training). However, some data brought to our attention indicated that OCD subjects have significantly increased evoked response potentials, relative to control subjects, during visual search tasks similar to the one we used in the first experiment (J. Allen, personal communication, January 18, 1998). Given this result, we speculated that the background task we used might not be as neutral as we thought for the two groups of subjects. Thus, we decided to repeat the experiment using a more passive viewing task. Using the same selection criterion described for the first experiment, we formed two groups of subjects and gave them eyeblink classical conditioning using methods and procedures identical to the first experiment, except for the background task. For the second experiment, instead of actively searching for a specific stimulus, subjected viewed a series of neutral pictures and responded with a computer key press (yes or no) to whether they found each picture pleasing. Overall, no differences in percent CRs during acquisition or extinction were seen when OCD-like and control subjects were compared, although the OCD-like subjects showed lower UR amplitudes. When rates of conditioning were compared overall for the two experiments, conditioning was found to be superior in the first experiment when the active visual search task was used. We did note a very interesting effect when the first blocks of conditioning trials were examined—the OCD-like subjects showed a significant number of CRs in the first block, suggesting that they learned the task very quickly relative to controls (see Fig. 10.5). This effect was also noted when CR amplitudes were analyzed. We interpreted these data to indicate that OCD-like subjects can show facilitated eyeblink classical conditioning. However, there is an interaction between conditioning rate and the background environment and context in which the training trials are delivered. The use of eyeblink conditioning to compare OCD-like subjects to control subjects was not the result of known neural correlates of the disorder (as was the case for the FAS model, autism, and schizophrenia). Rather, the use of eyeblink conditioning to study OCD was based on a theoretical model (e.g., Mowrer’s model) that predicts a facilitation of aversive associative learning in individuals with OCD. In this case, eyeblink conditioning is simply a behavioral tool that may prove useful for examining how aversive conditioning occurs in persons with OCD. Indeed, unlike the other clinical disorders that we have studied, it is
Chapter 10.qxd
3/28/2007
1:00 PM
Page 281
10. CONDITIONING AND CLINICAL DISORDERS
281
unlikely that cerebellar pathologies underlie OCD. It is more likely that disorders in forebrain systems (including limbic areas like the amygdala and hippocampus) may be central in this disorder. The involvement of these areas could be systematically explored by varying task complexity (e.g., delay vs. trace conditioning) or varying context and fear aspects of the conditioning process, thus potentially linking neurology with the behavioral deficits that are observed. Our studies with OCD-like subjects have sensitized us to one very important aspect of the conditioning process: the importance of considering arousal or background activation during the task. To this end, we have begun conducting studies to systematically look at the influence of context and arousal on eyeblink conditioning in humans (e.g., Tracy, McFall, & Steinmetz, 2005). Our early results show that this oft-overlooked aspect of conditioning may be quite important in determining the basic conditioning rate we see and that this aspect of conditioning must be explored in depth for us to gain a fuller understanding of associative learning in clinical populations. SUMMARY AND CONCLUSIONS In this chapter, I attempted to show how eyeblink classical conditioning, a relatively simple associative motor learning procedure, has been used to study a variety of clinical disorders that at least on the surface do not seem to be related to each other. Similar to most animal models, the rat model of FAS was created as a means to systematically explore, at the neuronal and brain systems level, the debilitating effects of in utero alcohol exposure. The model appears to do a good job capturing the neurological pathologies that are seen in FAS children, which include rather extensive loss of neurons in regions of the cerebellum that are critical for eyeblink conditioning. Our work to date suggests that this is a very promising research approach and that use of eyeblink conditioning in conjunction with the array of neurobiological tools that are available to study brain function should help us understand more fully brain–behavior correlates associated with the disorder. Our use of eyeblink conditioning to study autism was based on a growing research literature suggesting that autistics have neurological pathologies that involve regions of the cerebellum. We actually predicted poor conditioning in these individuals, assuming that cerebellar neuronal losses would always lead to deficits in rates of acquisition. We were wrong on this prediction, as facilitated rates of conditioning were observed.
Chapter 10.qxd
282
3/28/2007
1:00 PM
Page 282
STEINMETZ
However, the mistimed, early CRs that we observed are similar to more selective cerebellar deficits that have been shown in animal models (e.g., cerebellar anterior lobe lesions and some inactivation studies involving cerebellar cortex). Specific testable hypotheses can be forwarded based on data from humans and nonhumans and these should prove useful for framing future studies. Our studies of eyeblink conditioning in schizophrenics have roots in hypothesized cognitive/behavioral function as well as known neurological pathologies purported to be hallmarks of the disorder. Behaviorally, problems in temporal processing have been hypothesized to perhaps be the cause of symptoms that are seen with the disorder. Further, some studies have suggested that disorders in the cerebro-cerebello-thalamo-cerebro loop, which has recently garnered some attention for its roles in movement as well as cognition, have been hypothesized to cause the temporal processing difficulties. Eyeblink conditioning should be useful for studying this potentially pathological system, as variations in the procedure that perturb sensory processing as well as response timing processes can be introduced. In addition, higher brain systems can be engaged by varying the procedure to tax more complex processing (e.g., complex discrimination and reversal conditioning, contextual shift paradigms, etc.). Unlike the first three examples, the use of eyeblink conditioning to study OCD was not based on known neurological deficits nor on behavioral deficits specific to acquisition and performance of the eyeblink CR. Rather, we chose to use eyeblink conditioning with the assumption that it is one of a number of conditioning procedures that might reveal general facilitation of learning that involved pairing a neutral stimulus with an aversive stimulus or event. The observed facilitation of conditioning is promising, especially in light of its dependency on contextual/cognitive processing that is ongoing during the training. When more is understood about the neurobiology of OCD and also more is understood about how the brain encodes contextual, arousal, attention and related processes, further eyeblink conditioning studies would seem useful for advancing our understanding of brain–behavioral correlates associated with this disorder. I began this chapter by pointing out that Richard McFall has been a pioneer in insisting that the cause and treatment of clinical disorders be studied from an interdisciplinary perspective that requires the combined talents of scientists who use different approaches and different levels of analysis. Dick McFall has also been a strong proponent of the hybrid training of the next generation of graduate students who can appreciate
Chapter 10.qxd
3/28/2007
1:00 PM
Page 283
10. CONDITIONING AND CLINICAL DISORDERS
283
the contributions that fields like cognitive science and neuroscience can make to the development of a comprehensive understanding of the cause and treatment of clinical disorders. In this chapter, I attempted to provide the reader with a flavor of how basic associative learning methods and standard neuroscience techniques can be used to explore pathologies of behavior and cognition. The potential power of translational approaches, such as the research described here, seems enormous. I thank the forwardthinking Richard McFall for nudging some of us in this important and exciting direction. REFERENCES American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Andreasen, N. C., Paradiso, S., & O’Leary, D. S. (1998). “Cognitive dysmetria” as an integrative theory of schizophrenia. Schizophrenia Bulletin, 24(2), 203–212. Bao, S., Chen, L., & Thompson, R. F. (1998). Classical eyeblink conditioning in two strains of mice: Conditioned responses, sensitization, and spontaneous eyeblinks. Behavioral Neuroscience, 112(3), 714–718. Bauman, M. (1991). Microscopic neuroanatomic abnormalities in autism. Pediatrics, Supplement, 1, 791–796. Bauman, M., & Kemper, T. (1985). Histoanatomic observations of the brain in early infantile autism. Neurology, 35, 866–874. Berger, T. W., & Thompson, R. F. (1978). Neuronal plasticity in the limbic system during classical conditioning of the rabbit nictitating membrane response: I. The hippocampus. Brain Research, 145, 323–346. Berthier, N. E., & Moore, J. W. (1986). Cerebellar Purkinje cell activity related to the classically conditioned nictitating membrane response. Experimental Brain Research, 63, 341–350. Berthier, N. E., & Moore, J. W. (1990). Activity of deep cerebellar nuclear cells during classical conditioning of nictitating membrane extension in rabbits. Experimental Brain Research, 83, 44–54. Blankenship, M. R., Huckfeldt, R., Steinmetz, J. J., & Steinmetz, J. E. (2005). The effects of amygdala lesions on hippocampal activity and classical eyeblink conditioning in rats. Brain Research, 1035(2), 120–130. Bonthius, D. J., & West, J. R. (1990). Alcohol-induced neuronal loss in developing rats: Increased brain damage with binge exposure. Alcoholism: Clinical and Experimental Research, 14, 107–118. Bonthius, D. J., & West, J. R. (1991). Permanent neuronal deficits in rats exposed to alcohol during the brain growth spurt. Teratology, 44, 147–163. Brown, S. M., Kieffaber, P. D., Vohs, J. L., Carroll, C. A., Tracy, J. A., Shekhar, A., O’Donnell, B. F., Steinmetz, J. E., & Hetrick, W. P. (2005). Eye-blink conditioning deficits indicate timing and cerebellar abnormalities in schizophrenia. Brain and Cognition, 58, 94–108.
Chapter 10.qxd
284
3/28/2007
1:00 PM
Page 284
STEINMETZ
Clark, R. E., & Squire, L. R. (2000). Awareness and the conditioned eyeblink response. In D. S. Woodruff-Pak & J. E. Steinmetz (Eds.), Eyeblink classical conditioning: Vol. 1. Human applications (pp. 229–251). Boston: Kluwer. Clark, R. E., Zhang, A. A., & Lavond, D. G. (1992). Reversible lesions of the cerebellar interpositus nucleus during acquisition and retention of a classically conditioned behavior. Behavioral Neuroscience, 106, 879–888. Cooke, S. F., Attwell, P. J. E., & Yeo, C. H. (2004). Temporal properties of cerebellardependent memory consolidation. Journal of Neuroscience, 24(12), 2934–2941. Courchesne, E., Yeung-Courchesne, R., Press, G. A., Hesselink, J. R., & Jernigan, T. L. (1988). Hypoplasia of cerebellar vermal lobules VI and VII in autism. New England Journal of Medicine, 318(21), 1349–1354. Daum, I., & Schugens, M. M. (1996). On the cerebellum and classical conditioning. Current Directions in Psychological Science, 5(2), 58–61. Daum, I., Schugens, M. M., Ackermann, H., Lutzenberger, W., Dichgans, J., & Birbaumer, N. (1993). Classical conditioning after cerebellar lesions in humans. Behavioral Neuroscience, 107(5), 748–756. Densen, M. E. (1977). Time perception and schizophrenia. Perceptual and Motor Skills, 44(2), 436–438. Fiala, J. C., Grossberg, S., & Bullock, D. (1996). Metabotropic glutamate receptor activation in cerebellar Purkinje cells as substrate for adaptive timing of the classically conditioning eye-blink response. Journal of Neuroscience, 16(11), 3760–3774. Goodlett, C. R., & Eilers, A. T. (1997). Alcohol-induced Purkinje cell loss with a single binge exposure in neonatal rats: A stereological study of temporal windows of vulnerability. Alcoholism: Clinical and Experimental Research, 21, 738–744. Goodlett, C. R., Thomas, J. D., & West, J. R. (1991). Long-term deficits in cerebellar growth and rotarod performance in rats following “binge-like” alcohol exposure during the neonatal brain growth spurt. Neurotoxicology and Teratology, 13, 69–74. Gormezano, I., Kehoe, E. J., & Marshall, B. S. (1983). Twenty years of classical conditioning with the rabbit. Progress in Psychobiology and Physiological Psychology, 10, 197–275. Gould, T. J., & Steinmetz, J. E. (1996). Changes in rabbit cerebellar cortical and interpositus nucleus activity during acquisition, extinction and backward classical conditioning. Neurobiology of Learning and Memory, 65, 17–34. Green, J. T., Johnson, T. B., Goodlett, C. R., & Steinmetz, J. E. (2002). Eyeblink classical conditioning and interpositus nucleus activity are disrupted in adult rats exposed to ethanol as neonates. Learning and Memory, 9, 304–320. Green, J. T., Rogers, R. F., Goodlett, C. R., & Steinmetz, J. E. (2000). Impairment in eyeblink classical conditioning in adult rats exposed to ethanol as neonates. Alcoholism: Clinical and Experimental Research, 24(4), 438–447. Green, J. T., Tran, T., Steinmetz, J. E., & Goodlett, C. R. (2002). Neonatal ethanol produces cerebellar deep nuclear cell loss and correlated disruption of eyeblink conditioning in adult rats. Brain Research, 956(2), 302–311. Guerri, C. (1998). Neuroanatomical and neurophysiological mechanisms involved in central nervous system dysfunction inducted by prenatal alcohol exposure. Alcoholism: Clinical and Experimental Research, 22, 304–312.
Chapter 10.qxd
3/28/2007
1:00 PM
Page 285
10. CONDITIONING AND CLINICAL DISORDERS
285
Hamre, K. M., & West, J. R. (1993). The effects of timing of ethanol exposure during the brain growth spurt on the number of cerebellar Purkinje cells and granule cell nuclear profiles. Alcoholism: Clinical and Experimental Research, 17, 610–622. Hannigan, J. H. (1996). What research with animals is telling us about alcoholrelated neurodevelopmental disorders. Pharmacology, Biochemistry and Behavior, 55, 489–499. Hofer, E., Doby, D., Anderer, P., & Dantendorfer, K. (2001). Impaired conditional discrimination learning in schizophrenia. Schizophrenia Research, 51, 127–136. Ivry, R. B., Keele, S. W., & Diener, H. C. (1988). Dissociation of the lateral and medial cerebellum in movement timing and movement execution. Experimental Brain Research, 73, 167–180. Johnson, J. E., & Petzel, T. P. (1971). Temporal orientation and time estimation in chronic schizophrenics. Journal of Clinical Psychology, 27(2), 194–196. Katz, D. B., & Steinmetz, J. E. (1997). Single-unit evidence for eyeblink conditioning in cerebellar cortex is altered, but not eliminated, by interpositus nucleus lesions. Learning & Memory, 4(1), 88–104. Katz, D. B., & Steinmetz, J. E. (2002). Psychological functions of the cerebellum. Behavioral and Cognitive Neuroscience Reviews, 1(3), 229–241. Krupa, D. J., Thompson, J. K., & Thompson, R. F. (1993). Localization of a memory trace in the mammalian brain. Science, 260, 989–991. Lavond, D. G., Hembree, T. L., & Thompson, R. F. (1985). Effect of kainic acid lesions of the cerebellar interpositus nucleus on eyelid conditioning in the rabbit. Brain Research, 326, 179–182. Lavond, D. G., & Steinmetz, J. E. (1989). Acquisition of classical conditioning without cerebellar cortex. Behavioural Brain Research, 33, 113–164. Lavond, D. G., & Steinmetz, J. E. (2002). Handbook of classical conditioning. Boston: Kluwer. Lavond, D. G., Steinmetz, J. E., Yokaitis, M. H., & Thompson, R. F. (1987). Reacquisition of classical conditioning after removal of cerebellar cortex. Experimental Brain Research, 67, 569–593. Leiner, H. C., Leiner, A. L., & Dow, R. S. (1991). The human cerebro-cerebellar system: Its computing, cognitive, and language skills. Behavioral Brain Research, 44, 113–128. Lemieux, S. K., & Woodruff-Pak, D. S. (2000). Functional MRI studies of eyeblink classical conditioning. In D. S. Woodruff-Pak & J. E. Steinmetz (Eds.), Eyeblink classical conditioning: Vol. 1. Human applications (pp. 71–93). Boston: Kluwer. Lincoln, J. S., McCormick, D. A., & Thompson, R. F. (1982). Ipsilateral cerebellar lesions prevent learning of the classically conditioned nictitating membrane/eyelid response. Brain Research, 242, 190–193. Mamounas, L. A., Thompson, R. F., & Madden, J. (1987). Cerebellar GABAergic processes: Evidence for critical involvement in a form of simple associative learning in the rabbit. Proceedings of the National Academy of Science (USA), 84, 2101–2105. Marcussan, B. L., Goodlett, C. R., Mahoney, J. C., & West, J. R. (1994). Developing rat Purkinje cells are more vulnerable to alcohol-induced depletion during differentiation than during neurogenesis. Alcohol, 11, 147–156.
Chapter 10.qxd
286
3/28/2007
1:00 PM
Page 286
STEINMETZ
Mattson, S. N., & Riley, E. P. (1998). A review of the neurobehavioral deficits in children with fetal alcohol syndrome or prenatal exposure to alcohol. Alcoholism: Clinical and Experimental Research, 22, 279–294. Mauk, M. D., & Donegan, N. H. (1997). A model of Pavlovian eyelid conditioning based on the synaptic organization of the cerebellum. Learning & Memory, 3, 130–158. Mauk, M. D., Steele, P. M., & Medina, J. F. (1997). Cerebellar involvement in motor learning. Neuroscientist, 3(5), 303–313. Mauk, M. D., & Thompson, R. F. (1987). Retention of classically conditioned eyelid responses following acute decerebration. Brain Research, 403, 89–95. McCormick, D. A., Lavond, D. G., Clark, G. A., Kettner, R. R., Rising, C. E., & Thompson, R. F. (1981). The engram found? Role of the cerebellum in classical conditioning of nictitating membrane and eyelid responses. Bulletin of the Psychonomic Society, 18, 103–105. McCormick, D. A., & Thompson, R. F. (1984a). Cerebellum: Essential involvement in the classically conditioned eyelid response. Science, 223, 296–299. McCormick, D. A., & Thompson, R. F. (1984b). Neuronal responses of the rabbit cerebellum during acquisition and performance of a classically conditioned nictitating membrane-eyelid responses. Journal of Neuroscience, 4, 2811–2822. McGlinchey-Berroth, R., Carrillo, M. C., Gabrieli, J. D. E., Brawn, C. M., & Disterhoft, J. F. (1997). Impaired trace eyelid conditioning in bilateral medial temporal lobe amnesia. Behavioral Neuroscience, 111, 873–882. McLaughlin, J., Powell, D. A., & White, J. D. (2002). Characterization of the neuronal changes in the medial prefrontal cortex during jaw movement and eyeblink Pavlovian conditioning in the rabbit. Behavioral Brain Research, 132, 117–133. Moore, J. W., Desmond, J. E., & Berthier, N. E. (1989). Adaptively timed conditioned responses and the cerebellum: A neural network approach. Biological Cybernetics, 62, 17–28. Mowrer, O. H. (1960). Learning theory and behavior. New York: Wiley. Moyer, J. R., Deyo, R. A., & Disterhoft, J. F. (1990). Hippocampectomy disrupts trace eye-blink conditioning in rabbits. Behavioral Neuroscience, 104, 243–252. Napper, R. M. A., & West, J. R. (1995). Permanent neuronal cell loss in the inferior olive of adult rats exposed to alcohol during the brain growth spurt: A stereological investigation. Alcoholism: Clinical and Experimental Research, 19, 1321–1326. Oakley, D. A., & Russell, I. S. (1972). Neocortical lesions and Pavlovian conditioning. Physiology & Behavior, 8, 915–926. Oakley, D. A., & Russell, I. S. (1976). Subcortical nature of Pavlovian differentiation in the rabbit. Physiology & Behavior, 17, 947–954. Orr, W. B., & Berger, T. W. (1985). Hippocampectomy disrupts the topography of conditioned nictitating membrane responses during reversal learning. Behavioral Neuroscience, 99, 35–45. Perrett, S. P., Ruiz, B. P., & Mauk, M. D. (1993). Cerebellar cortex lesions disrupt learning-dependent timing of conditioned eyelid responses. Journal of Neuroscience, 13, 1708–1718. Rodier, P. M., Bryson, S. E., & Welch, J. P. (1997). Minor malformations and physical measurements in autism: Data from Nova Scotia. Teratology, 55, 319–325.
Chapter 10.qxd
3/28/2007
1:00 PM
Page 287
10. CONDITIONING AND CLINICAL DISORDERS
287
Roebuck, T. M., Mattson, S. N., & Riley, E. P. (1998). A review of the neuroanatomical findings in children with fetal alcohol syndrome or prenatal exposure to alcohol. Alcoholism: Clinical and Experimental Research, 22, 339–344. Schreurs, B. G., Sanchez-Andres, J. V., & Alkon, D. L. (1991). Learning-specific differences in Purkinje-cell dendrites of lobule HVI (lobulus simplex): Intracellular recording in a rabbit cerebellar slice. Brain Research, 548, 18–22. Schreurs, B. G., Sanchez-Andres, J. V., & Alkon, D. L. (1992). GABA-induced responses in Purkinje cells dendrites of the rabbit cerebellar slice. Brain Research, 597, 99–107. Sears, L. L., Andreasen, N. C., & O’Leary, D. S. (2000). Cerebellar functional abnormalities in schizophrenia are suggested by classical eyeblink conditioning. Society of Biological Psychiatry, 48, 204–209. Sears, L. L., Finn, P. R., & Steinmetz, J. E. (1994). Abnormal classical eyeblink conditioning in autism. Journal of Autism and Developmental Disorders, 24(6), 737–751. Sears, L. L., & Steinmetz, J. E. (2000). Classical eyeblink conditioning in normal and autistic children. In D. S. Woodruff-Pak & J. E. Steinmetz (Eds.), Eyeblink classical conditioning: Vol. 1. Human applications (pp. 143–162). Boston: Kluwer. Steinmetz, J. E. (1999). A renewed interest in human classical eye-blink conditioning. Psychological Science, 10(1), 24–25. Steinmetz, J. E. (2000). Brain substrates of classical eyeblink conditioning: A highly localized but also distributed system. Behavioural Brain Research, 110, 13–24. Steinmetz, J. E., Lavond, D. G., & Thompson, R. F. (1989). Classical conditioning in rabbits using pontine nucleus stimulation as a conditioned stimulus and inferior olive stimulation as an unconditioned stimulus. Synapse, 3(3), 225–233. Steinmetz, J. E., Logue, S. F., & Steinmetz, S. S. (1992). Rabbit classically conditioned eyelid responses do not reappear after interpositus nucleus lesion and extensive post-lesion training. Behavioural Brain Research, 51, 103–114. Streissguth, A. P., & O’Malley, K. (2000). Neuropsychiatric implications and longterm consequences of fetal alcohol spectrum disorders. Seminars in Clinical Neuropsychiatry, 5, 177–190. Tracy, J. A., Britton, G. B., & Steinmetz, J. E. (2001). Comparisons of single unit responses to tone, light and compound conditioned stimuli during rabbit classical eyeblink conditioning. Neurobiology of Learning and Memory, 76, 253–267. Tracy, J. A., Ghose, S. S., Stecher, T., McFall, R. M., & Steinmetz, J. E. (1999). Classical conditioning in a nonclinical obsessive-compulsive population. Psychological Science, 10(1), 9–13. Tracy, J. A., McFall, R. M., & Steinmetz, J. E. (2005). Effects of emotional valence and arousal manipulation on eyeblink classical conditioning and autonomic measures. Integrative Physiological and Behavioral Sciences, 40(1), 45–54. Tysk, L. (1983). Time estimation by healthy subjects and schizophrenic patients: A methodological study. Perceptual and Motor Skills, 56, 983–988. Volz, H.-P., Nenadic, I., Gaser, C., Rammsayer, T., Hager, F., & Sauer, H. (2001). Time estimation in schizophrenia: An fMRI study at adjusted levels of difficulty. NeuroReport, 12(2), 313–316. Weinberger, D. R., Torrey, E. F., & Wyatt, J. R. (1979). Cerebellar atrophy in chronic schizophrenia. Lancet, 1, 718–719.
Chapter 10.qxd
288
3/28/2007
1:00 PM
Page 288
STEINMETZ
White, I. M., Miller, D. P., White, W., Dike, G. L., Rebec, G. V., & Steinmetz, J. E. (1994). Neuronal activity in rabbit neostriatum during classical eyelid conditioning. Experimental Brain Research, 99, 179–190. Woodruff-Pak, D. S. (1993). Eyeblink classical conditioning in H.M.: Delay and trace paradigms. Behavioral Neuroscience, 107(6), 911–925. Woodruff-Pak, D. S., & Steinmetz, J. E. (Eds.). (2000a). Eyeblink classical conditioning: Vol. 1. Human applications. Boston: Kluwer. Woodruff-Pak, D. S., & Steinmetz, J. E. (2000b). Past, present and future of human eyeblink classical conditioning. In D. S. Woodruff-Pak & J. E. Steinmetz (Eds.), Eyeblink classical conditioning: Vol. 1. Human applications (pp. 1–17). Boston: Kluwer. Yeo, C. H., & Hardiman, M. J. (1992). Cerebellar cortex and eyeblink conditioning: A reexamination. Experimental Brain Research, 88, 623–638. Yeo, C. H., Hardiman, M. J., & Glickstein, M. (1985). Classical conditioning of the nictitating membrane response of the rabbit: II. Lesions of the cerebellar cortex. Experimental Brain Research, 60, 99–113.
Chapter 11.qxd
4/6/2007
2:30 PM
Page 289
11 Integrating Clinical and Cognitive Science Teresa A. Treat Melanie A. Dirks Yale University
Throughout his career, Richard McFall has been an articulate and visionary spokesperson for what has become known as integrative psychological science (IPS), an approach to the conduct and application of psychological research that draws on the best available theoretical, measurement, and analytical models across areas of psychology and other relevant fields. McFall has argued that greater integration not only would benefit clinical scientists, who far too often have operated in isolation from their colleagues’ basic knowledge about normative cognitive, neural, social, and developmental processes, but also basic psychological scientists, who typically have not capitalized on clinical researchers’ expertise in the conceptualization and measurement of individual differences and in the operation of socially complex and clinically relevant processes (McFall, 2006; McFall, chap. 14, this volume; McFall & Townsend, 1998; McFall, Treat, & Viken, 1997, 1998). This perspective led McFall and his colleagues at Indiana University to develop the first National Institute of Mental Health (NIMH)-funded IPS training model, in which clinical and nonclinical students become legitimate scholars in, and significant contributors to, both clinical science and either neural, cognitive, social, or developmental science. This integrative model extends the National Institutes of Health (NIH) translational model (National Advisory Mental Health Council 289
Chapter 11.qxd
290
4/6/2007
2:30 PM
Page 290
TREAT AND DIRKS
Behavioral Science Workgroup, 2000), which brings together basic behavioral scientists and clinical scientists to enrich the scientific vision and scope of NIH research, by advocating for the integration of basic and applied expertise within a single individual. McFall and his colleagues also have advanced IPS by conducting pioneering research that brings the best of contemporary cognitive science and neuroscience to bear on clinical questions and considers the implications of clinically relevant individual differences for the evaluation and extension of normative models of cognitive and neural processing. This chapter provides an overview of the contributions of this work to the advancement of quantitative clinical-cognitive science. Readers are urged to consult Steinmetz (chap. 10, this volume) for an analogous overview of McFall and colleagues’ work within the realm of clinical neuroscience. FROM PERSONAL CONSTRUCT THEORY TO SOCIAL INFORMATION PROCESSING McFall’s interest in the role of cognitive processing in clinical phenomena can be traced to his graduate training at Ohio State University with George Kelly (1955), whose personal construct theory highlighted the importance of individual differences in construal to our understanding of the development, maintenance, and treatment of psychopathology. Many aspects of Kelly’s theoretical model anticipated notions central to contemporary cognitive science, but empirical evaluation of this early information-processing model was limited by a dearth of adequate measurement models for the assessment of individual differences in construal. As a result, McFall’s research career has focused, in no small part, on (a) fleshing out theoretical models that articulate the role of cognitive processing in psychopathology, and (b) searching for measurement models that afford more rigorous assessment of processing constructs. In 1982, McFall published one of the field’s first social informationprocessing (SIP) models, in which he argued that individual differences in the operation of cascading decoding, decision making, and enactment processes contribute to variation in persons’ perceived effectiveness at completing specific contextualized tasks that are relevant to clinical phenomena (see also McFall, 1990). For example, individual differences in how skillfully men decode women’s sexual-interest cues should be linked to the likelihood that men will exhibit acquaintance-initiated heterosexual
Chapter 11.qxd
4/6/2007
2:30 PM
Page 291
11. INTEGRATING CLINICAL AND COGNITIVE SCIENCE
291
aggression (see Lipton, McDonel, & McFall, 1987). Similarly, variation in how skillfully college women make decisions about how to respond in common, problematic, and difficult interpersonal situations should predict both their competence in managing these situations and the likelihood of various psychological symptoms (see Goddard & McFall, 1992). McFall’s decoding process loosely corresponded to Kelly’s construal construct, but McFall differentiated among reception, perception, and interpretative subprocesses of decoding. McFall’s model further extended Kelly’s theoretical model by including other central cognitive operations, such as decision processes (i.e., response search, test, and selection; repertoire search and utility evaluation) and enactment processes (i.e., execution and self-monitoring). Moreover, McFall (1982, 1990; McFall et al., 1997, 1998; McFall & Townsend, 1998) urged the development and use of performance-based assessments of cognitive processing, in which researchers draw inferences about the operation of participants’ cognitive processes by observing their performance on information-processing tasks, rather than relying on participants’ verbal reports of the operations or products of such processes. McFall pointed toward the potential utility of information-processing paradigms developed by cognitive psychologists for this purpose. It would be difficult to overstate the impact of McFall’s SIP model on clinical science because this model and extensions of it have been applied in almost every area of psychopathology (see e.g., Crick & Dodge, 1994; Holtzworth-Munroe, 1992; Milner, 1993; O’Donohue & Rudman, 1999; Sayette, Wilson, & Elias, 1993). Nonetheless, McFall recognized the potential utility of (a) incorporating other cognitive processes beyond decoding, decision making, and enactment into theoretical models of clinically relevant cognitive processing; and (b) continuing to search for performance-based measurement approaches that would afford more valid assessments of cognitive processing. To this end, he began attending Indiana University’s weekly cognitive lunch series in the mid-1980s. He began to consider both the potential transportability of the commonly discussed models and methods to clinical psychology and to raise questions about the extent to which these models and methods could be extended to account for individual differences in non-normative processing of complex stimuli (i.e., in real-world processing). In essence, McFall’s pioneering efforts to bridge clinical and quantitative cognitive psychology helped to create the now burgeoning integrative field of quantitative clinical-cognitive science.
Chapter 11.qxd
292
4/6/2007
2:30 PM
Page 292
TREAT AND DIRKS
INTEGRATING CLINICAL SCIENCE AND QUANTITATIVE COGNITIVE SCIENCE Quantitative cognitive science focuses on the development and evaluation of formal process models of normative cognitive processing, which specify mathematically the hypothesized influence of component cognitive processes (e.g., perceptual organization, attention, classification, identification, memory, learning, etc.) on participants’ observed behavior when completing performance-based tasks. Thus, the theoretical models of quantitative cognitive science specify relevant component cognitive processes and their mathematical interrelationships with one another, as well as the way in which they affect participant behavior; the measurement models comprise the relevant performance-based tasks and stimuli; and the analytical models evaluate quantitatively the extent to which the theoretical models account for participant behavior within a particular measurement context. Although quantitative cognitive science endeavors to characterize normative processing of simple, typically artificial stimuli, quantitative clinical-cognitive science aims to quantify formally the clinically relevant individual differences in contextualized cognitive processing of complex, socially relevant stimuli. Surprisingly, clinical scientists have translated few of the theoretical, measurement, and analytical models of quantitative cognitive science for the purpose of characterizing clinically relevant individual differences in cognitive processing, although altered cognitive processing has been implicated in the development and maintenance of numerous forms of psychopathology. Moreover, the concepts and procedures employed in cognitive therapy bear little resemblance to the constructs and methods of contemporary cognitive science, because the cognitive component of evidence-based treatments for many clinical disorders typically emphasizes the careful identification, considered evaluation, and deliberate modification of distorted thinking patterns. Consequently, the potential of an integrative discipline of clinical-cognitive science remains largely untapped. This is unfortunate because process modeling of cognitive influences on psychopathology may challenge and advance our theoretical models, as well as suggest novel, performancebased intervention strategies that aim to modify deficient cognitive processing directly. Quantitative cognitive science and clinical-cognitive science share a focus on component cognitive processing and an insistence on both performance-based assessment with standardized stimuli and quantitatively
Chapter 11.qxd
4/6/2007
2:30 PM
Page 293
11. INTEGRATING CLINICAL AND COGNITIVE SCIENCE
293
rigorous specification, evaluation, and comparison of theoretical models. Quantitative clinical-cognitive scientists incorporate both systematic individual differences in cognitive processing and elements of psychopathology in their theoretical models, however, which necessitates wrestling with difficulties surrounding the representation and estimation of individual differences in participant performance in cognitive scientists’ computational models (Neufeld, 1998, 2002, 2005). Should processing parameters be estimated separately for each (a) participant, (b) externally identified group of participants (e.g., high vs. low symptom), or (c) analytically identified group of participants (e.g., via Bayesian approaches that identify distinct classes of participants with similar parameter estimates and then provide class-specific parameter estimates)? Quantitative cognitive scientists historically have treated multiple participants as replications of one another, but recently have begun to incorporate individual differences into their analytical models at one of these three levels, given concerns about the effects of the averaging process on the validity of their inferences (e.g., Ashby, Maddox, & Lee, 1994; Lee, 2001; Lee & Pope, 2003; Lee & Webb, 2005; Nosofsky & Palmeri, 1998; Rouder, Lu, Speckman, Sun, & Jiang, 2005). Thus, examination of clinically relevant individual differences in processing provides quantitative cognitive scientists a ready opportunity to evaluate the validity and utility of their models (Neufeld, 1998, 2002, 2005). Apart from their explicit focus on individual differences, quantitative clinical-cognitive scientists also differ from their cognitive colleagues in their interest in participant processing of more complex, socially relevant, and ecologically valid stimuli. Cognitive scientists evaluate their theoretical models of normative and decontextualized information processing most readily by relying on simple, artificial stimulus sets that vary along a finite number of orthogonal, easily identified, and uniformly perceived dimensions (e.g., geometric figures that vary in size and shape, or color patches that vary in hue and saturation). In contrast, answering clinicalcognitive scientists’ research questions frequently necessitates the creation of a stimulus set that both captures real-world variation along the small number of dimensions of greatest theoretical interest and eliminates or constrains the real-world variation along the dimensions of lesser theoretical interest. To evaluate hypothesized links between college women’s problematic eating patterns and their processing of other women’s bodysize and facial-affect information, for example, Viken, Treat, Nosofsky, McFall, and Palmeri (2002) developed a stimulus set of women’s photos that varied markedly along the body-size and facial-affect dimensions of
Chapter 11.qxd
294
4/6/2007
2:30 PM
Page 294
TREAT AND DIRKS
primary theoretical interest. Variation along other dimensions that might influence participant processing, but that were not of theoretical interest, was minimized by photographing each woman in similar clothing, in the same position, and under standard lighting conditions in front of a uniform background. Many clinical scientists, in our experience, view this as an overcontrolled and highly artificial stimulus set with limited ecological validity, whereas quantitative cognitive scientists often are somewhat aghast at what they perceive to be an uncontrolled and impossibly complex stimulus set. The latter reaction reflects concerns that cognitive scientists’ process models may not account well for processing of socially relevant stimuli that vary along a potentially infinite number of interrelated dimensions that are identified much less easily and perceived much less uniformly. Most process models assume the absence of such complexities, which may attenuate the validity and strength of the inferences drawn about clinically relevant processing from such efforts. Thus, integrating clinical and quantitative cognitive science clearly presents both challenges and opportunities. Generalization of the theoretical, measurement, and analytical models of cognitive science to the messiness of clinical phenomenon is by no means a foregone conclusion, in part for the reasons already detailed. Examination of clinically relevant processing provides quantitative cognitive scientists a ready opportunity, however, to examine and enhance the extent to which their formal process models generalize across the numerous facets of variation between clinical and cognitive science. In fact, variation in process models’ flexibility in accounting meaningfully for individual differences in real-world processing may serve as a useful novel criterion for evaluation of and selection among competing process models (Neufeld, 1998, 2002, 2005). A dearth of integrative training models and opportunities, such as the ones that McFall and colleagues developed at Indiana University, also threatens the advancement of quantitative clinical-cognitive science. Translational models of training and research arguably discourage the development of truly integrative researchers because few incentives exist to motivate students and researchers to duplicate within themselves the expertise of a readily available collaborator. Thus, translational research teams readily can maximize their efficiency by passing the work at hand to the most expert member of the team in that particular area, such that members do not learn much about and cannot necessarily talk articulately about what is going on across levels of analysis, multiple measurement or analytical methods, or basic and applied aspects of the phenomenon
Chapter 11.qxd
4/6/2007
2:30 PM
Page 295
11. INTEGRATING CLINICAL AND COGNITIVE SCIENCE
295
of interest. Integrative training models, in contrast, facilitate the professional development of researchers who can converse with experts in and make substantive contributions to multiple areas of psychological science. Such hybrid scholars view psychopathology from a novel vantage point that may enhance our understanding, assessment, and treatment of psychological problems. The development and proliferation of additional training opportunities are particularly key within the realm of quantitative clinical-cognitive science, as the interdisciplinary domains of clinical neuroscience and clinical-cognitive neuroscience currently are populated much more richly. Overcoming these challenges to the integration of clinical and cognitive science provides compelling opportunities to advance our understanding of the role of cognitive processing in the development and maintenance of psychopathology. Moreover, we may be able to leverage cognitive scientists’ extensive theoretical, measurement, and analytical models of learning processes to develop novel cognitive therapeutic techniques that remediate cognitive processing deficits by modifying them directly, rather than indirectly via verbally mediated procedures. For example, dot-probe paradigms that modify attentional patterns implicitly already are showing promise for the treatment of anxiety (Campbell, Rutherford, & MacLeod, 2002; MacLeod, Campbell, Rutherford, & Wilson, E., 2004; MacLeod, Rutherford, Campbell, Ebsworthy, & Holker, 2002; Mathews & MacLeod, 2002). A wealth of other learning paradigms could be used for this or similar purposes. Finally, integrating clinical and cognitive science staves off the need to posit separate and qualitatively distinct models for normative and non-normative processing, instead facilitating the development of a unified theory of human cognition that accounts for both normal and abnormal processing and behavior. QUANTITATIVE CLINICAL-COGNITIVE SCIENCE: ILLUSTRATIONS McFall and colleagues’ efforts to transport formal process models of quantitative cognitive science to clinical science have relied primarily on a unified class of theoretical, measurement, and analytical models that treat participants’ perceptual organizations of stimuli as a representational base for the operation of other component cognitive processes, such as classification, memory, and learning (e.g., Kruschke, 1992; Kruschke & Johansen, 1999; Nosofsky, 1991, 1992a, 1992b). This integrated class of
Chapter 11.qxd
296
4/6/2007
2:30 PM
Page 296
TREAT AND DIRKS
process models estimates psychologically meaningful parameters that should be of interest to clinical researchers and accounts well for participants’ normative processing of simple, artificial stimuli. Nonetheless, the validity and utility of adopting this approach for the examination of clinically relevant processing initially were unknown, given the plethora of necessary modifications and extensions to the theoretical, measurement, and analytical models employed by cognitive scientists. To facilitate evaluation of the generalizability of this class of models to clinically relevant processing, we are applying them to two different clinical phenomena for which several well-established theoretical models implicate a central role for cognitive processing deficiencies: eating disorders and sexual aggression (e.g., McFall, 1990; Schewe & O’Donohue, 1993; Vitousek, 1996; Ward, Hudson, Johnston, & Marshall, 1997; Wilson, 1999). Both research lines necessitated the generation and norming of novel photo stimulus sets of young women. The former line of research examines the extent to which college women who struggle with eating disorders show increased attention to, memory for, and learning about other women’s shape- and weight-related information relative to their processing of other women’s facial-affect information. The latter program evaluates whether college men who perceive that continued heterosexual advances are justified in the face of increasingly negative feedback show greater attention to, better memory for, and faster learning about female acquaintances’ physicalappearance information than their facial-affect information. In both cases, the overarching hypothesis is that impoverished or inaccurate processing of others’ affective information, secondary to more elaborate or enhanced processing of more salient information, contributes to the maintenance of the clinical phenomena. This section illustrates our application of a coherent class of cognitive-processing models to these clinically relevant phenomena. Mapping Participants’ Perceptual Organizations The first step in our approach to quantitative clinical-cognitive science aims to characterize individual differences in participants’ perceptual organizations (POs) of clinically relevant stimuli that vary along the primary dimensions of theoretical interest and control variation along secondary dimensions (see Treat, McFall, Viken, Kruschke, Nosofsky, & Wang, in press, for further information on the stimulus-generation process). PO refers to the way in which participants cognitively represent the stimuli, as inferred from their reported perceptions of pairwise stimulus similarity.
Chapter 11.qxd
4/6/2007
2:30 PM
Page 297
11. INTEGRATING CLINICAL AND COGNITIVE SCIENCE
297
FIGURE 11.1 Multidimensional scaling representation of group psychological space and body-size-oriented participant’s psychological space. Adapted from Viken et al. (2002), with permission.
In a similarity-ratings task, participants judge the similarity of all possible pairs of stimuli on a 10-point scale (1 = very different and 10 = very similar). On a single similarity-ratings trial, for example, a participant might evaluate the similarity of a normatively heavy and happy woman to a normatively heavy and unhappy woman; a rating of 1 suggests that the participant is attending relatively more to affect than to body size, whereas a rating of 10 suggests the opposite. Note that the participant is not directed to attend to any particular stimulus characteristics when
Chapter 11.qxd
298
4/6/2007
2:30 PM
Page 298
TREAT AND DIRKS
making her similarity ratings, so the resulting assessment of PO is relatively implicit. Multidimensional scaling (MDS) analyses of the similarity-ratings data then provide a multidimensional spatial representation (or psychological space) of participants’ POs, in which stimuli perceived to be similar are scaled close together and stimuli perceived to be different are scaled far apart (see Treat, McFall, Viken, Nosofsky, MacKay, & Kruschke, 2002, for further information on the use of MDS to map clinically relevant POs). The top panel of Figure 11.1 illustrates a group psychological space similar to that which emerged for college women’s perceptions of the photo stimuli in Viken et al. (2002). As expected, two dimensions best described as affect and body size underlay the optimal group scaling solution, presumably secondary to our efforts to balance internal and external validity by accentuating variation along theoretically relevant dimensions and minimizing but not eliminating variation along numerous other dimensions. Weighted MDS (WMDS) analyses of similarity-ratings data can be used to characterize not only the group-level psychological space, but also individual differences in participants’ attention to the dimensions of the group-level PO. Dimension-specific attention weights, which stretch and shrink the dimensions of the group space, are estimated for each participant. The bottom panel of Figure 11.1 illustrates the individual-specific psychological space for a body-size-oriented participant for whom WMDS estimates a large attention weight for the body-size dimension and a small attention weight for the affect dimension. The stretching and shrinking of the body-size and affect dimensions, respectively, reflects the body-sizeoriented participant’s perception that normatively heavier and lighter women are quite dissimilar to one another, whereas normatively happier and sadder women are much more similar. As expected, Viken et al. (2002) observed marked individual differences in participants’ attention weights, and the presence of self-reported bulimic symptoms accounted well for the extent to which participants attended relatively more to other women’s body size than to affect information. Similarly, Treat, McFall, Viken, and Kruschke (2001) demonstrated that college men who perceived unwanted sexual advances to be justified, relative to men who perceived the opposite, showed relatively greater attention to women’s physical-appearance cues than to their facial-affect cues. Across numerous evaluations of hypotheses about clinically relevant POs, we have observed that the fit of MDS models to participants’
Chapter 11.qxd
4/6/2007
2:30 PM
Page 299
11. INTEGRATING CLINICAL AND COGNITIVE SCIENCE
299
similarity-ratings data is much poorer than typically observed in the cognitive literature (Lease, McFall, Treat, & Viken, 2003; Lease, McFall, & Viken, 2003; Treat et al., 2001, 2002; Viken et al., 2002; Wang, Treat, & Brownell, under review). This is not surprising given the complexity of the stimuli and participants’ perceptions of them, as well as the simplicity of the process models. For example, even the WMDS model, which explicitly represents individual differences in dimensional attention, assumes that all participants (a) attend only to the dimensions that underlie the group psychological space, (b) perceive the multiple dimensions to be orthogonal to one another, (c) organize the stimuli the same way within a dimension (e.g., perceive the ordering of the values of different stimuli along the body-size dimension to be the same), and (d) perceive the values of the same stimulus along each dimension identically across trials (i.e., deterministically, rather than probabilistically). Each of these assumptions almost certainly is false. More complex models allow researchers to relax these assumptions (see Treat et al., 2002, in press), but more data are needed to provide precise parameter estimates than a researcher typically is able to obtain when addressing clinically relevant questions. Thus, the consistency of the observed findings with researchers’ a priori theoretical expectations becomes a more significant criterion by which to evaluate the appropriateness of MDS analytical models than the absolute magnitude of traditional fit indices. To date, we have found that the MDS-based theoretical, measurement, and analytical models of PO generalize well to examination of research questions about clinically relevant individual differences in the PO of much more complex, socially relevant information than cognitive psychologists typically study (Lease et al., 2003; Treat et al., 2001, 2002; Treat, Kruschke, & McFall, in preparation; Treat, Kruschke, Viken, & McFall, in preparation; Viken et al., 2002; Wang et al., under review). This is encouraging news because theoretical models across many areas of psychopathology articulate a central role for the representation and organization of incoming information in the development and maintenance of clinical phenomena (e.g., Beck, 1976; Kelly, 1955; McFall, 1982; Williams, Watts, MacLeod, & Mathews, 1997), but measurement and analytical models that provide relatively implicit characterizations of PO have been lacking. Moreover, the MDS-based approach to mapping PO provides the representational input to other component cognitiveprocessing mechanisms, according to several formal process models of
Chapter 11.qxd
300
4/6/2007
2:30 PM
Page 300
TREAT AND DIRKS
identification, classification, memory, decision making, category-learning, and so on. We turn first to an examination of recognition-memory processes. Investigating Recognition-Memory Processes Having defined participants’ psychological representations of the stimuli, we are well situated to examine clinically relevant variation in the operation of other higher order cognitive processes. Recognition memory refers to the accuracy of participants’ explicit reports of whether they previously have been exposed to particular stimuli. Numerous theoretical frameworks suggest that distorted or impoverished memory for relevant information contributes to the maintenance of clinical phenomena (e.g., Beck, 1976; Williams et al., 1997). For example, the likelihood that a college-age man will exhibit sexually aggressive behavior against a female acquaintance may increase not only as a function of his reduced attention to women’s sexual-interest cues, but also as a function of his poor memory for this information because this may impede accurate decision making. Countless studies have investigated recognition-memory correlates of clinical symptoms across a wide range of clinical phenomena. These studies typically either assess memory processes in isolation (i.e., apart from other component cognitive processes) or examine attention and memory processes in the absence of an overarching process model that specifies the expected mathematical links between the processes in a rigorous fashion (i.e., the model indicates more than that performance on the two tasks should correlate). Recently, Gotlib et al. (2004) raised questions about the marked incoherence of attention and memory processes as indexed by extremely weak correlations between performance in commonly used attention paradigms, such as the emotional Stroop and dotprobe tasks, and frequently used explicit-memory paradigms, such as free-recall and recognition-memory tasks. Gotlib and colleagues (2004) also questioned the lack of coherence across multiple putative indicators of the same process (see also Dalgleish et al., 2003; Lim & Kim, 2005; Rinck & Becker, 2005), and they concluded their paper by calling for “a stronger and more explicit conceptual and methodological focus on the similarities and differences among various measures of cognitive biases” (pp. 396–397). These difficulties may be attributed, in part, to clinical researchers’ tendency to view cognitive processing in a relatively simplistic, atheoretical, and reified fashion, where all indicators of attention should hang together well and indicators of attention and memory also
Chapter 11.qxd
4/6/2007
2:30 PM
Page 301
11. INTEGRATING CLINICAL AND COGNITIVE SCIENCE
301
should converge. Cognitive scientists do not construe cognition—or attention or memory, for that matter—as a monolithic construct, however, anymore than clinical scientists construe psychopathology—or anxiety or depression—as an undifferentiated construct. Thus, our understanding of the role of attention and memory processes in clinical phenomena will advance more rapidly to the extent that we do more than pull cognitive paradigms off the cognitive shelf and apply them with limited awareness of the theoretically, methodologically, and analytically rich contexts in which they are embedded. The translation of classes of integrated process models for multiple processes may prove to be particularly fruitful in this regard because they specify mathematically both the way in which the processes operate and how they interrelate. A recently completed study (Treat, Kruschke, Viken, & McFall, in preparation) examined whether college women who reported clinically significant eating-disorder symptoms, relative to college women who reported either no or some symptoms, showed enhanced attention to and memory for other women’s shape- and weight-related information and impoverished attention to and memory for other women’s affective information. Participants first viewed 28 photos of unique women that varied along body-size and facial-affect dimensions for 3 seconds apiece. Participants then completed a prototype-classification task with these 28 stimuli, which allowed estimation of each participant’s relative attention to affect and body size (see Viken et al., 2002, for more information). In the recognition-memory task, participants viewed 56 photos one at a time and indicated whether they had seen the identical photo previously, as well as how confident they were in their judgment. Half of the photos (n = 28) were old; that is, they were identical to those that participants initially studied and then classified in the attention task. The remaining photos (n = 28) depicted the same women, but either their body size or facial affect differed (e.g., an old normatively heavy and happy woman might correspond to a new normatively heavy and sad woman). Warping techniques were used to make realistic modification to women’s body sizes. As in Viken et al. (2002), high-symptom participants attended more to body size and less to affect, compared with medium- and low-symptom participants. Additionally, despite the extremely sparse data, participantspecific signal-detection theory (SDT) analyses (Macmillan & Creelman, 2005) of the confidence-rating data revealed that high-symptom participants, relative to medium- and low-symptom participants, showed similar memory for other women’s shape- and weight-related information, but
Chapter 11.qxd
302
4/6/2007
2:30 PM
Page 302
TREAT AND DIRKS
much poorer memory for other women’s affective information (Treat, Kruschke, Viken, & McFall, in preparation). The absence of group differences in memory for body size suggests that the memory deficit for affect may be secondary to high-symptom participants’ decreased attention to affect. Consideration of the body-size findings suggests that future studies should examine whether symptomatic women store or retrieve body-size information less well. Efforts to ameliorate high-symptom women’s memory deficit for affective information also should be developed and evaluated, as retaining impoverished or inaccurate information about other women’s affect readily could lead to interpersonal difficulties and bouts of negative affect, which are well-recognized correlates of and triggers for eating-disordered behavior (e.g., McFall, Eason, Edmondson, & Treat, 1999). Although SDT quantifies only the decision processes underlying participant responses, Nosofsky’s (1991, 1992a, 1992b) process model instantiates the representational and memory processes involved in recognition memory. Nosofsky’s model formalizes the notion that stimuli are more likely to feel familiar and be classified as old (i.e., to clear an estimated familiarity threshold) when a participant perceives them as more similar overall to relevant stored stimuli—in other words, when they are closer to relevant old stimuli in the participant’s psychological space. The model also assumes that stimuli may be stored with different strengths in memory. Consider Figure 11.2, which presents a pair of old and new stimuli that vary along the affect dimension. Although the affect-oriented participant, whose perceptual representation of these stimuli is depicted in the top panel, perceives the old and new stimuli to be dissimilar (i.e., they are scaled further apart in her PO), the body-sizeoriented participant, whose PO is portrayed in the bottom panel, perceives the stimuli to be quite similar. As Nosofsky’s model specifies that the familiarity of new stimuli results in part from the perceived similarity of the relevant new and old stimuli, the model could predict that affectoriented participants would distinguish better than body-size-oriented participants between old and new stimuli that vary along the affect dimension. To the extent that high-symptom women continue to attend less to affect in the memory task than low-symptom women, therefore, the model could predict that affect-oriented participants will show better memory than body-size-oriented participants for other women’s affect. The model also could account for the similar memory for body size across groups by assuming that participants across groups weight body-size
Chapter 11.qxd
4/6/2007
2:30 PM
Page 303
11. INTEGRATING CLINICAL AND COGNITIVE SCIENCE
303
FIGURE 11.2 Placement of sample old and new-affect stimuli in affect-oriented and body-size-oriented participants’ psychological spaces.
information similarly when retrieving this information in the context of a memory task, even though the high-symptom women attend more to body-size information when initially encoding the stimuli. Whether the modeling bears out these assumptions remains to be seen, but this brief sketch of Nosofsky’s process model and potential findings highlights the utility of translating a theoretically coherent class of process models to examine research questions about clinically relevant processing.
Chapter 11.qxd
4/6/2007
2:30 PM
Page 304
304
TREAT AND DIRKS
These interrelated theoretical, measurement, and analytical models can be used to obtain and interpret mathematically precise characterizations of both normative and non-normative cognitive processing of both highly artificial and more complex, socially relevant stimuli. In addition, this class of models estimates psychologically meaningful parameters that map well onto constructs of interest to clinical scientists. Moreover, Nosofsky’s model and other process models in this class assume that higher order processes, such as recognition memory, operate on the same underlying spatial representation of the stimuli, which formalizes the intuitively appealing notion that PO constrains but does not wholly determine the operation of other cognitive processes. Thus, our examinations of clinically relevant PO and memory are not conducted in isolation; rather, our MDSbased description of the former informs our evaluation of the latter. Next, we turn to an overview of the conceptualization and measurement of clinically relevant category-learning processes. Examining Category-Learning Processes Category learning refers to the increasingly accurate placement of stimuli into categories, as participants either explicitly or implicitly learn the category structure driving the trial-by-trial feedback that they receive on the accuracy of their classifications. In one study, college women first completed a prototype-classification task with the photo stimulus set described previously, which allowed us to categorize their POs as affect-oriented, both-oriented, or body-size-oriented (Treat, Kruschke, & McFall, in preparation). Next, they completed a multiple-phase category-learning task, in which they learned either an affect category structure and then a body-size category structure (the Affect Initial condition) or the reverse sequence of structures (the Body-Size Initial condition). All participants were told that they would be classifying individual pictures of women, each of whom was a member of the arbitrarily labeled Category F or Category J. Participants were instructed to guess the category membership of the stimuli at first until the trial-by-trial feedback helped them figure out the category structure. Participants also were told that the basis for the feedback might change. For example, a participant in the Affect Initial condition would learn initially to place normatively sad women in Category F and normatively happy women in Category J; an unannounced shift to the body-size category structure then would necessitate that the participant learn to classify normatively heavy women in Category F and normatively light women in Category J.
Chapter 11.qxd
4/6/2007
2:30 PM
Page 305
11. INTEGRATING CLINICAL AND COGNITIVE SCIENCE
305
Clinical scientists have devoted surprisingly little attention to the study of clinically relevant variation in the operation of category-learning processes on complex, socially relevant stimuli. This is unfortunate, as category-learning tasks provide a window into the dynamics of cognitive processing. Clinical researchers have tended to focus on static characterizations of cognitive processing, rather than to examine the way in which processing varies, whether naturally over different time scales, as a function of feedback, or in response to a theoretically relevant manipulation that is associated with exacerbation of the clinical phenomenon (e.g., taboo food consumption in the case of eating disorders, alcohol ingestion in the case of sexual aggression). In the present case, variation in how rapidly and well participants learn a clinically relevant category structure may be more diagnostic than static indices of processing. For example, we already have demonstrated that college women who struggle with eatingdisorder symptoms attend more to other women’s body size and less to their affective information (Treat, Kruschke, Viken, & McFall, in preparation; Viken et al., 2002). The women of greatest interest clinically may be those who not only show this static PO, however, but also learn a bodysize category structure much more rapidly and an affect category much more slowly. In other words, the impairing aspect of high-symptom women’s cognitive processing patterns may be less their elevated attention to other women’s shape and weight than their difficulty shifting their attention to other women’s affect when required. Similarly, decreased attention to women’s sexual interest cues may be less consequential for college men who exhibit sexually aggressive behavior toward female acquaintances than slowness to shift attention to women’s affective cues when necessary. Cognitive scientists have developed numerous process models that account well for normative category learning with simple artificial stimuli and that estimate psychologically meaningful parameters that map well onto constructs of interest to clinical researchers (e.g., Ashby, AlfonsoReese, Turken, & Waldron, 1998; Erickson & Kruschke, 1998; Nosofsky & Palmeri, 1998). We have relied on a class of well-established process models developed by Kruschke and colleagues that assumes that categorylearning processes operate on an MDS-derived spatial representation of the stimuli (Erickson & Kruschke, 1998; Kruschke & Johansen, 1999; Kruschke, 1992, 2001; Nosofsky, Kruschke, & McKinley, 1992). These models predict—and studies have confirmed—that participants learn category structures based on more salient stimulus dimensions faster than category structures based on less salient stimulus dimensions, because the
Chapter 11.qxd
4/6/2007
306
2:30 PM
Page 306
TREAT AND DIRKS
FIGURE 11.3 Placement of category boundary for affect category structure in affect-oriented and body-size-oriented participants’ psychological spaces.
exemplars in the same categories are construed as more similar (i.e., they are closer in psychological space) than the exemplars in different categories. The study described earlier in this section was designed to evaluate the generalizability of this claim about the link between PO and category learning to the study of individual differences in the processing of complex, socially relevant stimuli (Treat, Kruschke, & McFall, in preparation). Earlier examinations of this claim relied on the experimental manipulation of stimulus dimensions that participants perceived uniformly. The current study, in contrast, investigated this claim quasi-experimentally, by
Chapter 11.qxd
4/6/2007
2:30 PM
Page 307
11. INTEGRATING CLINICAL AND COGNITIVE SCIENCE
307
capitalizing on the presence of naturally occurring individual differences in the perceived salience of the stimulus dimensions. We anticipated that participants’ static POs (i.e., affect-oriented, both-oriented, or body-sizeoriented) would account for substantial variation in their performance on the category-learning tasks, such that participants would learn either the earlier or later category structure much more quickly when the structure was congruent with their PO (e.g., affect-oriented participants, relative to the remaining participants, would learn the affect category structure more quickly, regardless of whether it was first or second). Figure 11.3 illustrates the hypothesized relationship between the perceived salience of the stimulus dimensions and category-learning performance. This figure superimposes the affect category structure on affect- and body-size-oriented participants’ POs; participants must learn to classify the normatively sad stimuli to the left of the category boundary as Fs and the normatively happy stimuli to the right of the boundary as Js. Affect-oriented participants, who perceive affect to be much more salient than body size, should learn the affect structure much more quickly, because they view withincategory exemplars as much more similar than between-category exemplars. These participants should learn a body-size category structure much more slowly, however, given the marked perceived similarity of the women placed in the same and different categories. In contrast, bodysize-oriented participants should struggle to learn the affect category structure that is incongruent with their PO but learn the body-size category structure quickly. As predicted by the process models, participants learned the initial highcongruent category structure much more quickly than the initial lowcongruent category structure, with an average proportion correct across blocks of .96 and .83, respectively. The congruence effect was even more pronounced when learning the second category structure, with average proportion correct scores of .88 and .68, respectively. In fact, although all participants learned the initial category structure, albeit at different rates, a number of them were unable to learn the second category structure when it was incongruent with their PO. This effect emerged regardless of whether the second category structure was body size or affect—in both cases, participants who perceived the dimension underlying the initial category structure to be extremely salient struggled to learn a category structure based on a dimension to which they initially attended little. Only one of the two subgroups is of potential clinical interest in the current context—that is, the body-size-oriented participants, who rapidly learned the body-size structure, but then struggled with the affect
Chapter 11.qxd
308
4/6/2007
2:30 PM
Page 308
TREAT AND DIRKS
structure—but the process hypothesized to underlie their difficulties does not differ from that proposed to account for affect-oriented participants’ difficulties when tackling the body-size structure after the affect structure. Future research needs to examine whether high-symptom women are overrepresented among the subgroup of participants who experienced particular difficulty learning the affect structure, particularly after learning the body-size structure. If this proves to be the case, then we potentially can account for both normative and non-normative processing by varying parameter estimates within a single model of category learning, rather than by proposing qualitatively distinct mechanisms to account for qualitatively distinct processing patterns. The predicted link between normative PO and category learning with simple stimuli generalized well to our evaluation of the link between individual differences in PO and category learning with much more complex stimuli. Analogously to Nosofsky’s process model for recognition memory, Kruschke and colleagues’ process models for category learning assume that the underlying spatial representation of the stimuli constrains, but does not wholly determine, the operation of higher order cognitive processes. In this case, PO prior to completion of the learning task is assumed to facilitate or inhibit performance, depending on the congruence of the participant’s PO with the category structure to be learned. Kruschke and Johansen’s (1999) model proposes two other mechanisms, in addition to the congruence of the initial PO with the category structure, that may account for individual variability in participant performance. The first indicates that participants learn a particular category structure by shifting their attention away from nondiagnostic dimensions and toward diagnostic dimensions. For example, the affect-oriented participant might learn the body-size category structure by decreasing her attention to affect and increasing her attention to body size, such that normatively heavier and lighter women are well-separated in her psychological space. The second suggests that participants simply learn to associate particular regions of the psychological space with the correct category labels. Thus, the affect-oriented participants might solve the body-size category structure by mapping normatively heavier women to the category label F and normatively lighter women to the label J. Distinguishing among the relative contributions of these three learning mechanisms to observed performance not only promises to enhance our understanding of why learning sometimes proceeds slowly, but also lays some of the basic-research foundation necessary for the development of more targeted and efficient treatment strategies.
Chapter 11.qxd
4/6/2007
2:30 PM
Page 309
11. INTEGRATING CLINICAL AND COGNITIVE SCIENCE
309
We evaluated the relative importance of these three mechanisms to college men’s performance when learning category structures based on women’s physical-appearance and facial-affect cues (Treat et al., 2001). Physical appearance refers to the level of physical exposure exhibited by the woman in the photograph; women who received high normative ratings along this dimension typically wore revealing or tight-fitting clothing. As in the eating-disorder-relevant study described in this section, participants viewed photos of 26 women one at a time, classified them as members of Category F or Category J, and received trial-by-trial feedback. Participants completed four blocks of 26 trials of either the physical-exposure or facial-affect category structure and then four blocks of 26 trials of the other structure; the shift between the structures was unannounced. Prior to completing the learning task, participants completed a similarityratings task, which allowed us to classify them as either exposure-oriented or affect-oriented. As mentioned previously, exposure-oriented college men were significantly more likely to judge unwanted sexual advances to be justified. We fitted Kruschke and Johansen’s (1999) RASHNL model to the exposure- and affect-oriented groups’ data and allowed parameter estimates of the three learning mechanisms either to vary or to be fixed across the two groups (see Treat et al., 2001, for details). The best-fitting model suggested that group-specific differences in initial PO accounted for substantial variability in performance, such that exposure-oriented participants learned the exposure category structure more quickly than affectoriented participants, who learned the affect category structure more quickly than their exposure-oriented counterparts. Moreover, although both groups learned the incongruent category structure, they did so not by shifting their attention away from the irrelevant dimension and toward the relevant dimension, but rather by learning to associate particular regions of their psychological space with the correct category labels. These findings highlight again the extent to which individual differences in PO constrain category learning: Participants learned congruent structures much more quickly than incongruent structures. Contrary to the findings in numerous category-learning studies in the cognitive literature, however, the results reported in Treat et al. (2001) indicated that participants’ psychological spaces were not amenable to change in response to feedback (i.e., participants did not learn by shifting attention to relevant stimulus dimensions). This discrepancy may result from differences in the way that the typical stimuli in cognitive and clinical science are perceived. Almost all category-learning studies in cognitive science rely
Chapter 11.qxd
310
4/6/2007
2:30 PM
Page 310
TREAT AND DIRKS
on simple, artificial stimuli composed of separable dimensions that readily are processed independently of one another, enhancing the ease with which participants can shift their attention toward relevant dimensions and away from irrelevant dimensions across multiple category structures. In contrast, participants process the photo stimuli that we have used in both lines of our research in a more integral fashion (i.e., more holistically). Nosofsky and Palmeri (1996) demonstrated that dimensional attention shifting is much less likely to occur when integral-dimension stimuli are used (e.g., color rectangles that vary in hue, brightness, and saturation) than when separable-dimension stimuli are used (e.g., objects that vary in shape, color, and size). Dimensional attention shifting, however, can facilitate much more rapid acquisition of incongruent category structures than learning to map regions of the psychological space to particular category labels, which produces more gradual increments in performance (Kruschke & Johansen, 1999; Nosofsky & Palmeri, 1996). Being able to shift attention rapidly to important stimulus information could be critical under some conditions, such as when a man who is interested sexually in a woman needs to attend more to her affect than to how revealing her clothing is, or when a woman in conversation with a female friend needs to attend more to her affect than to her shape and weight. Thus, both clinical and cognitive scientists will benefit from examining the conditions under which participants can and do shift attention to relevant stimulus dimensions when they process the stimuli in a more integral fashion. Using cognitive scientists’ process models to examine clinically relevant category-learning processes not only facilitates clinical scientists’ understanding of the role of more dynamic aspects of processing in clinical phenomena, but also may contribute to the development of novel prevention or intervention strategies that directly target distorted or deficient cognitive processing. Currently, the concepts and procedures employed in cognitive therapy bear little resemblance to the constructs and methods of contemporary cognitive science (McFall et al., 1997, 1998; Treat et al., in press), but a new form of cognitive therapy could draw on the plethora of learning paradigms developed by cognitive scientists to modify maladaptive perceptual organizations or attentional patterns, and to facilitate the acquisition of important category structures.1 1 In a conceptually related line of research, MacLeod and colleagues have used implicit-learning methods to retrain anxious persons’ attention toward or away from disorder-relevant word stimuli (Campbell, Rutherford, & MacLeod , 2002; MacLeod, Campbell, Rutherford, & Wilson, E., 2004;
Chapter 11.qxd
4/6/2007
2:30 PM
Page 311
11. INTEGRATING CLINICAL AND COGNITIVE SCIENCE
311
Goldstone and colleagues have examined the extent to which category learning influences participants’ normative perceptual representations of simple, artificial stimuli (Goldstone, 1994; Goldstone, Lippa, & Shiffrin, 2001; Goldstone & Steyvers, 2001). Researchers have yet to evaluate the generalizability of their basic findings to individual differences in learning about more complex, socially relevant information, however, or to consider the applicability of this basic research to amelioration of cognitiveprocessing deficits in psychopathology. We currently are completing a study that examines the impact of learning either a body-size or affect category structure on college women’s (a) perceptual organizations for stimuli evaluated in a later similarity-ratings task, and (b) recognition memory for body-size and affect information. We also will evaluate whether eating-related difficulties moderate any observed effects. Preliminary WMDS analyses indicate that both category-learning conditions enhanced participants’ later attention to the stimulus dimension on which the feedback was based, and that this effect generalized to novel stimuli. Thus, unlike the findings reported in Treat et al. (2001), category learning appeared to modify participants’ dimensional attention patterns, as assessed in a later similarity-ratings task, rather than in formal modeling of performance on the learning task. Numerous potential explanations for this disparity of inferences will be evaluated in future research, as we continue to explore the conditions under which category-learning paradigms might be used to modify problematic aspects of a person’s PO. Because PO is assumed to constrain the operation of other higher order processes within the current class of process models described in this chapter, we also investigate the extent to which observed changes in PO result in desirable changes to the operation of classification, memory, decision making, and other higher order processes.
MacLeod, Rutherford, Campbell, Ebsworthy, & Holker, 2002; Mathews & MacLeod, 2002). For example, Campbell et al. (2002) used a dot-probe procedure to train high trait-anxious individuals to avoid threat-relevant words by placing the target to be detected consistently near a neutral word (rather than a threatening word) across thousands of training trials. Individuals assigned randomly to a control condition in which no contingencies were introduced to the dot-probe procedure showed no decrease in their attention toward threat-relevant information or in their trait anxiety. In contrast, participants in the experimental condition showed clinically significant reductions in their attention toward threat-relevant information and in their trait anxiety. Recently, these findings have been extended to social phobia (Malcom, 2003, as reported in MacLeod et al., 2004). This line of research illustrates the potential therapeutic utility of experimental cognitive methods that act directly on deficient cognitive processing.
Chapter 11.qxd
4/6/2007
2:30 PM
312
Page 312
TREAT AND DIRKS
Future Directions The unified class of process models described here, which cognitive scientists developed to account for normative processing of simple artificial stimuli, appears to generalize quite well to examinations of clinically relevant processing of more complex stimuli. Investigating the role of cognitive processing in psychopathology with these integrated theoretical, measurement, and analytical models allows us to proceed in a more theoretically informed, performance-based, and analytically rigorous fashion than is typical in this area. In turn, this strategy affords stronger inferences about the role of cognitive operations in clinical problems. Moreover, it positions us well to develop more targeted prevention and intervention techniques, perhaps using some of the many well-established learning paradigms in cognitive science as a novel form of cognitive therapy that targets cognitive deficits directly rather than relies on verbally mediated approaches. Our overarching approach can be extended in a number of directions. First, numerous variants of the similarity-ratings, recognition-memory, and category-learning paradigms could be used to examine the operation of PO, memory, and learning processes under theoretically interesting altered conditions. For example, we might compare college women’s perceptual representations when they are directed to attend to body-size information while making similarity ratings, so that we could evaluate whether symptomatic women perceive body-size information in a more discretized and less continuous fashion than asymptomatic women. Alternatively, we might examine whether instructional manipulations that enhance college women’s attention to other women’s affect eliminate symptomatic women’s memory deficit for affect. We also could evaluate the effect of probabilistic, rather than deterministic, feedback on college men’s ability to learn a category structure based on women’s sexualinterest cues (e.g., each woman might be classified in the same category 80% of the time, rather than 100% of the time); college men who exhibit particular difficulties learning a sexual-interest structure when receiving this more ecologically valid form of feedback should be at greater risk of exhibiting sexually aggressive behavior toward female acquaintances. Second, additional process models within this class can be applied to examine clinically relevant variation in the operation of numerous other processes, such as identification, classification, and decision making. In the latter case, for example, we currently are using preference-scaling paradigms and analytical approaches to examine the link between college women’s eating-disorder symptoms and their utilization of body size, affect,
Chapter 11.qxd
4/6/2007
2:30 PM
Page 313
11. INTEGRATING CLINICAL AND COGNITIVE SCIENCE
313
and attractiveness information when defining their real and ideal selves (Treat, Viken, Wang, McFall, & Weierich, in preparation; see also Wang et al., under review, for an alternative example of this approach). Finally, we need to examine whether theoretically relevant manipulations that we know are associated with an altered likelihood of the clinical phenomena of interest to us also are associated with predictable changes in cognitive processing. Both eating-disorder symptoms and sexually aggressive behavior are highly contextualized. Thus, observed processing deficits should be exacerbated in the presence of these contextual manipulations, such as consumption of taboo food or interpersonal rejection in the case of eating disorders, and consumption of alcohol or sexual-arousal manipulations in the case of acquaintance-initiated sexual aggression. Future research also should examine whether effective treatment for the clinical problem remediates the processing deficits. CLOSING COMMENTS Psychological scientists have but scratched the surface of what quantitative clinical-cognitive science has to offer (a) clinical scientists’ efforts to characterize and ameliorate processing deficits associated with psychopathology, and (b) quantitative cognitive scientists’ efforts to evaluate the generalizability of their process models in more complex, socially relevant circumstances. A growing body of work overwhelmingly suggests that the formal models that cognitive scientists have developed to account for normative processing of simple artificial stimuli generalize quite well to examinations of clinically relevant processing (e.g., Busemeyer & Stout, 2002; Carter & Neufeld, 1999; Filoteo & Maddox, 1999; Neufeld, 2005, in press; Neufeld, Vollick, Carter, Boksman, & Jetté, 2002; Nosofsky & Zaki, 1998), and our own work suggests that this claim also extends to the processing of more complex stimuli (e.g., Farris, Viken, Treat, & McFall, 2006; Treat et al., 2001, 2002, in press; Viken et al., 2002). Moreover, the unified class of process models described in this chapter provides only a single example of the numerous process models developed by cognitive scientists that might be translated to examine clinically relevant processing.2 Thus, the time is ripe for the rapid 2 General recognition theory (e.g., Ashby & Townsend, 1986; Kadlec & Townsend, 1992), a multidimensional generalization of signal-detection theory, provides an alternative integrated class of theoretical, measurement, and analytical models that also could be used to examine clinically relevant individual differences in PO and other higher order cognitive processes.
Chapter 11.qxd
4/6/2007
2:30 PM
Page 314
314
TREAT AND DIRKS
exploration of this relatively unmapped region of what McFall commonly refers to as integrative psychological science. Clinical scientists, and psychological scientists more generally, owe McFall a profound debt of gratitude for his invaluable contributions to the advancement of integrative psychological science. From his prescient vision and articulation of the potential of integrative approaches, to his development of a premier interdisciplinary training program, to his pioneering research program at the conjunction of clinical, cognitive, and neural science, McFall always seems to be well out in front of the rest of us. Fortunately, as we struggle to catch on and catch up, McFall continues to challenge, encourage, and inspire us, for which we are indeed thankful. REFERENCES Ashby, F. G., Alfonso-Reese, L. A., Turken, A. U., & Waldron, E. M. (1998). A neuropsychological theory of multiple systems in category learning. Psychological Review, 105, 442–481. Ashby, F. G., Maddox, W. T., & Lee, W. W. (1994). On the dangers of averaging across subjects when using multidimensional scaling or the similarity-choice model. Psychological Science, 5, 144–151. Ashby, F. G., & Townsend, J.T. (1986). Varieties of perceptual independence. Psychological Review, 93, 154–179. Beck, A. T. (1976). Cognitive theory and the emotional disorders. New York: International Universities Press. Busemeyer, J. R., & Stout, J. D. (2002). A contribution of cognitive decision models to clinical assessment: Decomposing performance on the Bechara gambling task. Psychological Assessment, 14, 253–262. Campbell, L., Rutherford, E. M., & MacLeod, C. (2002, November). Practice makes perfect: The reduction of trait anxiety through the extended retraining of attentional response to threat. Paper presented at the 36th AABT annual convention, Reno, NV. Carter, J. R., & Neufeld, R. W. J. (1999). Cognitive processing of multidimensional stimuli in schizophrenia: Formal modeling of judgment speed and content. Journal of Abnormal Psychology, 108, 633–654. Crick, N. R., & Dodge, K. A. (1994). A review and reformulation of social-information processing mechanisms in children’s development. Psychological Bulletin, 115, 74–101. Dalgleish, T., Taghavi, R., Neshat-Doost, H., Moradi, A., Canterbury, R., & Yule, W. (2003). Patterns of processing bias for emotional information across clinical disorders: A comparison of attention, memory, and prospective cognition in children and adolescents with depression, generalized anxiety, and posttraumatic stress disorder. Journal of Clinical Child and Adolescent Psychology, 32, 10–21. Erickson, M. A., & Kruschke, J. K. (1998). Rules and exemplars in category learning. Journal of Experimental Psychology: General, 127, 107–140.
Chapter 11.qxd
4/6/2007
2:30 PM
Page 315
11. INTEGRATING CLINICAL AND COGNITIVE SCIENCE
315
Farris, C., Viken, R. J., Treat, T. A., & McFall, R. M. (2006). Heterosocial perceptual organization: A choice model application to sexual coercion. Psychological Science, 17, 869–875. Filoteo, J. V., & Maddox, W. T. (1999). Quantitative modeling of visual attention processes in patients with Parkinson’s disease: Effects of stimulus integrality on selective attention and dimensional integration. Neuropsychology, 13, 206–222. Goddard, P., & McFall, R. M. (1992). Decision-making skills and heterosocial competence in college women: An information-processing analysis. Journal of Social and Clinical Psychology, 11, 401–425. Goldstone, R. (1994). Influences of categorization on perceptual discrimination. Journal of Experimental Psychology, 123(2), 178–200. Goldstone, R. L., Lippa, Y., & Shiffrin, R. (2001). Altering object representations through category learning. Cognition, 78, 27–43. Goldstone, R. L., & Steyvers, M. (2001). The sensitization and differentiation of dimensions during category learning. Journal of Experimental Psychology, 130(1), 16–139. Gotlib, I. H., Kasch, K. L., Traill, S., Joormann, J., Arnow, B. A., & Johnson, S. L. (2004). Coherence and specificity of information-processing biases in depression and social phobia. Journal of Abnormal Psychology, 113, 386–398. Holtzworth-Munroe, A. (1992). Social skill deficits in maritally violent men: Interpreting the data using a social information processing model. Clinical Psychology Review, 12, 605–617. Kadlec, H., & Townsend, J. T. (1992). Signal detection analyses of dimensional interactions. In F. G. Ashby (Ed.), Multidimensional models of perception and cognition. Scientific psychology series (pp. 181–227). Hillsdale, NJ: Lawrence Erlbaum Associates. Kelly, G. A. (1955). The psychology of personal constructs. New York: Norton. Kruschke, J. K. (1992). ALCOVE: An exemplar-based connectionist model of category learning. Psychological Review, 99, 22–44. Kruschke, J. K. (2001). Toward a unified model of attention in associative learning. Journal of Mathematical Psychology, 45, 812–863. Kruschke, J. K., & Johansen, M. K. (1999). A model of probabilistic category learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 25, 1083–1119. Lease, A. M., McFall, R. M., Treat, T. A., & Viken, R. J. (2003). Assessing children’s representations of their peer group using a multidimensional scaling technique. Journal of Social and Personal Relationships, 20, 707–728. Lease, A. M., McFall, R. M., & Viken, R. J. (2003). Distance from peers in the group’s perceived organizational structure: Relation to individual characteristics. Journal of Early Adolescence, 23, 194–217. Lee, M. D. (2001). Determining the dimensionality of multidimensional scaling representations for cognitive modeling. Journal of Mathematical Psychology, 45, 149–166. Lee, M. D., & Pope, K. J. (2003). Avoiding the dangers of averaging across subjects when using multidimensional scaling. Journal of Mathematical Psychology, 47, 32–46. Lee, M. D., & Webb, M. R. (2005). Modeling individual differences in cognition. Psychonomic Bulletin & Review, 12, 605–621.
Chapter 11.qxd
316
4/6/2007
2:30 PM
Page 316
TREAT AND DIRKS
Lim, S., & Kim, J. (2005). Cognitive processing of emotional information in depression, panic, and somatoform disorder. Journal of Abnormal Psychology, 114, 50–61. Lipton, D. N., McDonel, E. C., & McFall, R. M. (1987). Heterosocial perception in rapists. Journal of Consulting and Clinical Psychology, 55, 17–21. MacLeod, C., Campbell, L., Rutherford, E., & Wilson, E. (2004). The causal status of anxiety-linked attentional and interpretive bias. In J. Yiend (Ed.), Cognition, emotion, and psychopathology: Theoretical, empirical and clinical directions (pp. 68–85). Cambridge, England: Cambridge University Press. MacLeod, C., Rutherford, E., Campbell, L., Ebsworthy, G., & Holker, L. (2002). Selective attention and emotional vulnerability: Assessing the causal basis of their association through the experimental manipulation of attentional bias. Journal of Abnormal Psychology, 111, 107–123. Macmillan, N. A., & Creelman, C. D. (2005). Detection theory: A user’s guide (2nd ed.). Mahwah, NJ: Lawrence Erlbaum Associates. Mathews, A., & MacLeod, C. (2002). Induced processing biases have causal effects on anxiety. Cognition and Emotion, 16, 331–354. McFall, R. M. (1982). A review and reformulation of the concept of social skills. Behavioral Assessment, 4, 1–33. McFall, R. M. (1990). The enhancement of social skills: An information-processing analysis. In W. L. Marshall & D. R. Laws (Eds.), Handbook of sexual assault: Issues, theories, and treatment of the offender (pp. 311–330). New York: Plenum. McFall, R. M. (2006). Doctoral training in clinical psychology. Annual Review of Clinical Psychology, 2, 21–49. McFall, R. M., Eason, B. J., Edmondson, C. B., & Treat, T. A. (1999). Social competence and eating disorders: Development and validation of the Anorexia and Bulimia Problem Inventory. Journal of Psychopathology and Behavioral Assessment, 21, 365–394. McFall, R. M., & Townsend, J. T. (1998). Foundations of psychological assessment: Implications for cognitive assessment in clinical science. Psychological Assessment, 10, 316–330. McFall, R. M., Treat, T. A., & Viken, R. J. (1997). Contributions of cognitive theory to new behavioral treatments. Psychological Science, 8, 174–176. McFall, R. M., Treat, T. A., & Viken, R. J. (1998). Contemporary cognitive approaches to studying clinical problems. In D. K. Routh & R. J. DeRubeis (Eds.), The science of clinical psychology: Accomplishments and future directions (pp. 163–197). Washington, DC: American Psychological Association. Milner, J. S. (1993). Social information processing and physical child abuse. Clinical Psychology Review, 13, 275–294. National Advisory Mental Health Council Behavioral Science Workgroup. (2000). Translating behavioral science into action: Report of the National Advisory Mental Health Council Behavioral Science Workgroup (NIH Publication No. 00-4699). Bethesda, MD: National Institutes of Health/National Institute of Mental Health. Neufeld, R. W. J. (1998). Intersections and disjunctions in process-model applications. Psychological Assessment, 10, 396–398. Neufeld, R. W. J. (2002). Introduction to the special section on cognitive science and psychological assessment. Psychological Assessment, 14, 235–238.
Chapter 11.qxd
4/6/2007
2:30 PM
Page 317
11. INTEGRATING CLINICAL AND COGNITIVE SCIENCE
317
Neufeld, R. W. J. (2005). Composition and uses of formal clinical cognitive science. In W. Spaulding & J. Poland (Eds.), Modeling complex systems: Motivation, cognition and social processes: Nebraska Symposium on Motivation. Lincoln, Nebraska: University of Nebraska Press. Neufeld, R. W. J. (Ed.). (in press). Advances in clinical-cognitive science: Formal modeling and assessment of processes and symptoms. Washington, DC: APA Books. Neufeld, R. W. J., Vollick, D., Carter, J. R., Boksman, K., & Jetté, J. (2002). Application of stochastic modeling to the assessment of group and individual differences in cognitive functioning. Psychological Assessment, 14, 279–298. Nosofsky, R. M. (1991). Tests of an exemplar model for relating perceptual classification and recognition memory. Journal of Experimental Psychology: Human Perception and Performance, 17, 3–27. Nosofsky, R. M. (1992a). Similarity scaling and cognitive process models. Annual Review of Psychology, 43, 25–53. Nosofsky, R. M. (1992b). Exemplar-based approach to relating categorization, identification, and recognition. In F. G. Ashby (Ed.), Multidimensional models of perception and cognition (pp. 363–393). Hillsdale, NJ: Lawrence Erlbaum Associates. Nosofsky, R. M., Kruschke, J. K., & McKinley, S. C. (1992). Combining exemplarbased category representations and connectionist learning rules. Journal of Experimental Psychology: Learning, Memory, & Cognition, 18, 211–233. Nosofsky, R. M., & Palmeri, T.J. (1996). Learning to classify integral-dimension stimuli. Psychonomic Bulletin & Review, 3, 222–226. Nosofsky, R. M., & Palmeri, T. J. (1998). A rule-plus-exception model for classifying objects in continuous-dimension spaces. Psychonomic Bulletin & Review, 5, 345–369. Nosofsky, R. M., & Zaki, S. R. (1998). Dissociations between categorization and recognition in amnesic and normal individuals: An exemplar-based interpretation. Psychological Science, 9, 247–255. O’Donohue, W., & Rudman, J. C. (1999). Social relations of sexually abused children: A social information processing analysis. Aggression and Violent Behavior, 4, 29–39. Rinck, M., & Becker, E. S. (2005). A comparison of attentional biases and memory biases in women with social phobia and major depression. Journal of Abnormal Psychology, 114, 62–74. Rouder, J. N., Lu, J., Speckman, P., Sun, D., & Jiang, Y. (2005). A hierarchical model for estimating response time distributions. Psychonomic Bulletin & Review, 12, 195–223. Sayette, M. A., Wilson, G. T., & Elias, M. J. (1993). Alcohol and aggression: A social information processing analysis. Journal of Studies on Alcohol, 54, 399–407. Schewe, P. A., & O’Donohue, W. T. (1993). Rape prevention: Methodological problems and new directions. Clinical Psychology Review, 13, 667–682. Treat, T. A., Kruschke, J. K., & McFall, R. M. (in preparation). Individual differences in blocking of socially relevant stimuli. Treat, T. A., Kruschke, J. K., Viken, R. J., & McFall, R. M. (in preparation). The role of attention, memory, and correlation-detection processes in eating disorders.
Chapter 11.qxd
318
4/6/2007
2:30 PM
Page 318
TREAT AND DIRKS
Treat, T. A., McFall, R. M., Viken, R. J., & Kruschke, J. K. (2001). Using cognitive science methods to assess the role of social information processing in sexually coercive behavior. Psychological Assessment, 13, 549–565. Treat, T. A., McFall, R. M., Viken, R. J., Kruschke, J. K., Nosofsky, R. M., & Wang, S. S. (in press). Clinical-cognitive science: Applying quantitative models of cognitive processing to examination of cognitive aspects of psychopathology. In R. W. J. Neufeld (Ed.), Advances in clinical-cognitive science: Formal modeling and assessment of processes and symptoms. Washington, DC: APA Books. Treat, T. A., McFall, R. M., Viken, R. J., Nosfosky, R. M., MacKay, D. B., & Kruschke, J. K. (2002). Assessing clinically relevant perceptual organization with multidimensional scaling techniques. Psychological Assessment, 14, 239–252. Treat, T. A., Viken, R. J., Wang, S. S., McFall, R. M., & Weierich, M. R. (in preparation). A preference-scaling analysis of the link between eating disorders and self perception. Viken, R. J., Treat, T. A., Nosfosky, R. M., McFall, R. M., & Palmeri, T. (2002). Modeling individual differences in perceptual and attentional processes related to bulimic symptoms. Journal of Abnormal Psychology, 111, 598–609. Vitousek, K. B. (1996). The current status of cognitive-behavioral models of anorexia nervosa and bulimia nervosa. In P. M. Salkovskis (Ed.), Frontiers of cognitive therapy (pp. 383–418). New York: Guilford. Wang, S. W., Treat, T. A., & Brownell, K. D. (under review). Cognitive processing in the classroom: Teachers’ attention to and utilization of girls’ body size, ethnicity, attractiveness, and facial affect in classroom contexts. Ward, T., Hudson, S. M., Johnston, L., & Marshall, W. L. (1997). Cognitive distortions in sex offenders: An integrative review. Clinical Psychology Review, 17, 479–507. Williams, J. M., Watts, F. N., MacLeod, C., & Mathews, A. (1997). Cognitive psychology and emotional disorders. Chichester, England: Wiley. Wilson, G. T. (1999). Cognitive behavior therapy for eating disorders: Progress and problems. Behaviour Research and Therapy, 37(Suppl. 1), S79–S95.
Chapter 12.qxd
3/28/2007
1:00 PM
Page 319
IV FUTURE DIRECTIONS FOR RESEARCH, APPLICATION, AND TRAINING
Chapter 12.qxd
3/28/2007
1:00 PM
Page 320
Chapter 12.qxd
3/28/2007
1:00 PM
Page 321
12 Translational Research and the Future of Psychological Clinical Science Bruce N. Cuthbert University of Minnesota
The Society for a Science of Clinical Psychology (SSCP; Section III, Division 12 of the American Psychological Association [APA]) was established in 1966. Considering that the Journal of Abnormal Psychology was founded in 1906, sixty years seems at first glance a rather long gestation. In retrospect, however, it is easy to see why such a society could not come into existence until then. Listening to the reminiscences of senior colleagues, it is apparent that a systematic science of clinical psychology was largely nonexistent even in the 1940s and 1950s. Notwithstanding the scientist-practitioner model for clinical training espoused by the Boulder Conference, there was little for clinical science to do in 1949. Like psychiatry, clinical psychology was dominated by psychodynamic theory, with its largely untestable claims and hypotheses. Professors of clinical psychology, for the most part, conducted research programs in basic areas such as perception or conditioning, returning to the clinic to train students in interviewing and therapy. The appeal of behavior therapy, starting in the late 1950s, was not simply that it offered an alternative to psychodynamic theory as such; it was that the emphasis on overt behavior as an appropriate subject matter permitted (and indeed, demanded) experiments that conformed to good scientific 321
Chapter 12.qxd
322
3/28/2007
1:00 PM
Page 322
CUTHBERT
practice—that is, with procedures that could be readily specified and replicated by other scientists, and measures that were readily quantifiable and amenable to standard tests for reliability and validity. Richard McFall, whose festschrift occasions this special volume, entered the field just as the modern science of clinical psychology was coming into being, receiving his PhD from Ohio State University 1 year before the establishment of SSCP. This was a fortuitous coincidence: Indisputably, the development of our field across the next 40 years has been in no small part a product of the scope and force of McFall’s vision. Dick McFall’s legacy will be one of insistence that clinical science in psychology be grounded in methodologies and theories every bit as rigorous as in other areas of experimental psychology. In fact, a cornerstone of clinical science is the tenet that abnormal psychology may largely be construed in terms of the same fundamental psychological principles— learning, cognition, conditioning, and so on—that characterize experimental psychology. Notwithstanding a long stream of papers and talks, the ultimate testament of McFall’s career will without doubt be the Manifesto for a Science of Clinical Psychology (the Manifesto), appearing as the reprint of a Section III presidential address some 15 years ago (McFall, 1991). This document staked out a unifying claim to what psychological science should be, and has provoked a lively and continuing controversy ever since. The Manifesto laid out some fundamental principles—so fundamental, in fact, that McFall said later that he had not anticipated the prolonged debate evoked by its publication (although one wonders if this is merely an indication of his somewhat puckish pleasure in provoking intellectual debate). The Manifesto contained a single theme and two corollaries. The cardinal principle stated that “Scientific clinical psychology is the only legitimate and acceptable form of clinical psychology” (McFall, 1991, p. 76) The first corollary concerned treatment: “Psychological services should not be administered to the public (except under strict experimental control) until they have satisfied four minimal criteria” (p. 79; viz., describe the exact nature of the services clearly, state benefits explicitly, validate the benefits scientifically, and empirically rule out overweening side effects). The second addressed training: “The primary and overriding objective of doctoral training programs in clinical psychology must be to produce the most competent clinical scientists possible” (p. 82).
Chapter 12.qxd
3/28/2007
1:00 PM
Page 323
12. TRANSLATIONAL RESEARCH AND THE FUTURE
323
It has now been approximately 15 years since the publication of the Manifesto. During this period, clinical science in psychology has made better progress than we might have expected, but not, perhaps, as much as we might have hoped. On the plus side, psychological clinical science has prospered. A large number of graduate programs are staunch adherents of clinical science. The Academy of Psychological Clinical Science (APCS), founded in 1994 to represent training programs that basically hew to the principles laid out in the Manifesto, is a well-established and stable organization that, in addition to over 40 graduate programs, counts an increasing number of clinical internship sites among its members. Faculty members in clinical science programs comprise an important and thriving component of the research portfolios at relevant institutes of the National Institutes of Health (NIH), and its members are heavily represented in grant review committees, NIH policy sessions, and advisory councils. On the other hand, the Manifesto has hardly been greeted with unanimous acclaim. In psychology and other parts of the mental health treatment community, many practitioners have decried the Manifesto as a threat to a sizable portion of the practice community. Empirically based treatments are regarded as cookbook approaches that are not suitable for the complex mix of problems observed in actual clinical practice. From the other end of the mental health spectrum, the research base in areas of brain science such as genetics, molecular biology, and cellular processes has grown at a dizzying pace in the past decade, with significant new discoveries emerging practically every month. This explosion of knowledge has led some researchers in the neuroscience community to conclude that genetic and molecular/cellular approaches to mental disorders are necessary and sufficient for explaining their etiology and for developing new (psychopharmacologic) treatments—relegating to psychological clinical science a status that ranges from second-rate to superfluous. Thus, the field finds itself receiving critical scrutiny from different quarters of the mental health community. On this occasion, some 15 years after publication of the Manifesto, I want to speculate about what the next 15 to 20 years may hold for the directions that clinical science in psychology might take. The preamble for these speculations is a brief review of the concerns about psychological clinical science both from practitioner and from strong biological reductionist perspectives. I argue that although the nature of the commentary from these two viewpoints is markedly disparate, as would be
Chapter 12.qxd
3/28/2007
1:00 PM
Page 324
324
CUTHBERT
expected, the future directions and goals that psychological clinical science needs to undertake in response are largely the same in either case. My own perspective comes from my position as a program official at the National Institute of Mental Health (NIMH) for the past 7 years, which has provided an opportunity to observe trends and opinions in a wide variety of fields integral to the mental health research enterprise. PSYCHOLOGICAL CLINICAL SCIENCE AND CLINICAL PRACTICE Soon after publication of the Manifesto, Division 12 of the APA, under David Barlow’s leadership, formed a task force to develop a list of empirically validated therapies. The first of several versions of this list was published in 1995 (Task Force on Promotion and Dissemination of Psychological Procedures, 1995). The task force continued to develop its work over the course of several years; the list was revised on a periodic basis to incorporate new evidence, and today exists as a standing link on the Division 12 Web site. The major criterion for including therapies on the list of empirically validated treatments included at least two “good between group design experiments” (Chambless et al., 1998, p. 4) that incorporated a manualized procedure, and that demonstrated either statistical superiority to another treatment (or a placebo) or equivalent efficacy to another already established treatment. Although the term was not used in the original reports, this criterion has generally been interpreted to refer to a randomized clinical trial (RCT), in which clients are assigned at random to treatment or control groups. An alternative criterion was a series of 10 or more single-case experimental designs that also compared the intervention to another treatment. Since the original publication of the list, however, RCTs have generally been considered as the gold standard for making decisions about empirical validation. Discussion and criticism regarding the empirically supported treatment movement have been widespread and included many different aspects. Commentary has come from many different perspectives, including the health policy field (e.g., Tanenbaum, 2003, 2005), psychotherapy and psychopathology researchers (e.g., Beutler, 2004; Westen, Novotny, & Thompson-Brenner, 2004), and practicing clinicians, including an incoming president of the APA (Levant, 2004). However, a preponderance of the concerns has centered on various aspects of the use of RCTs to define evidence about valid treatments.
Chapter 12.qxd
3/28/2007
1:00 PM
Page 325
12. TRANSLATIONAL RESEARCH AND THE FUTURE
325
A chief complaint is that patients seen in actual clinical practice do not conform to the clean patients typically seen in clinical trials, in which persons seeking treatment are assessed carefully in order to ensure that they exhibit only the disorder under study; clinicians argue that comorbidity is more typical than not in real-world practice. Further, an experienced clinician observes significant psychopathology in many patients whose problems fail to fit any one DSM category neatly, and must be assigned to the NOS (not otherwise specified) specifier in the mood, anxiety, personality, or other categories. The clinician, of course, lacks the luxury of declining treatment for these patients on the basis of their failure to fit neatly into a predetermined category. Another point of contention has been the requirement for manualized procedures in order for therapies to be designated as empirically validated. Clinicians and some clinical researchers claim that manuals cannot take into account the flexibility that is needed to treat the idiosyncratic problems encountered in real-world practice. Rather, an experienced clinician has the ability to evaluate critically the particular set of problems presented by any individual patient, and select an eclectic mix of treatment procedures that s/he knows from experience will be optimal for that patient. Further, manuals are only written for a single, DSM-specified diagnosis, which is insufficient for the majority of cases in which patients present with multiple problems. Finally, some commentators have taken the position that much of the effectiveness of any psychological treatment derives from nonspecific factors such as the relationship between the clinician and patient. These factors are difficult to measure, and often are amenable only to the kind of qualitative data analysis that does not fit into the statistical design of an RCT. Relatedly, therapies that depend heavily on the relationship between the therapist and patient, and involve procedures that explicitly focus on the therapist–patient interactional process, are seen as particularly disadvantaged because the relevant variables are so difficult to measure. These concerns have received extensive rebuttal and discussion from the proponents of empirically validated treatments (see, e.g., CritsChristoph, Wilson, & Hollon, 2005; Stirman, Crits-Cristoph, & DeRubeis, 2004; Weisz, Weersing, & Henggeler, 2005). Other clinical researchers have suggested various areas of middle ground (e.g., Beutler, Moleiro, & Talebi, 2002; Rosen & Davison, 2003). It is not the point of this essay,
Chapter 12.qxd
326
3/28/2007
1:00 PM
Page 326
CUTHBERT
however, to review in detail the lively and ongoing debate. Rather, a brief review of the issues is a necessary background for some comments about the ways in which future directions in psychological clinical science may help resolve some of these areas of contention. PSYCHOLOGICAL CLINICAL SCIENCE AND NEUROSCIENCE Another challenge to clinical science in psychology comes from a quite different source—the exponential explosion of new knowledge about the brain and its relationship to mental disorders. Mental disorders are increasingly viewed as brain disorders, in a more-or-less neurological sense. In this view, psychological accounts of disorders are largely superfluous. It follows that psychological science risks being left behind by the rapid advance of knowledge in the neurosciences, and finding itself in the position of a second-rate and marginalized, at best, discipline. A prime example of these developments is the area of neurogenesis. It has been definitively established over the past decade that new neurons are created in the adult brain, particularly in the hippocampus and related structures. This finding of adult neurogenesis has completely overturned the long-held conventional wisdom that all creation of new neurons takes place by or shortly after birth, and sparked an understandably intense burgeoning of interest for its implications for mental and neurological disorders (e.g., see the excellent review by Eisch & Nestler, 2002; but see also Rakic, 2002, for a cautionary evaluation). The initial reports were marked more by a general promise than any findings specifically germane to mental disorders. However, recently a report from René Hen’s group provided results highly relevant to depression (Santarelli et al., 2003), using a mouse model in which feeding is suppressed while the animals are situated in a novel environment. The investigators first showed that the effects of increased eating latencies and reduced consumption were largely blocked by antidepressant drugs at 28 days but not at 5 days, paralleling the time course of antidepressant action observed in clinical patients; further, measures of hippocampal neurogenesis showed a 60% increase after 28 days, but no change after 5 days. In a subsequent experiment, neurogenesis was blocked in one group of mice by irradiating the hippocampus; this group failed to show the ameliorative antidepressant effects in the novel feeding test, with appropriate controls showing that the results were not due to generalized disruptions in feeding behavior.
Chapter 12.qxd
3/28/2007
1:00 PM
Page 327
12. TRANSLATIONAL RESEARCH AND THE FUTURE
327
This report prompted great excitement that neurogenesis may be the major pathway through which antidepressant drugs, and particularly the class of selective serotonin reuptake inhibitors (SSRIs), act therapeutically. The results also provide an attractive model for the development of newer agents that may target neurogenesis even more specifically. The completion of the human genome and the almost weekly proliferation of new and exciting genetic findings have produced another area that has already resulted in rapidly shifting views about the brain basis of mental disorders. Large-scale efforts such as the International Hapmap Project, originated to specify the common patterns of DNA sequence variation across the entire human genome, offer a potentially powerful way to rapidly zero in on areas of the genome strongly related to particular diseases (International HapMap Consortium, 2003). This would greatly reduce the time needed to find particular candidate genes that are implicated in specific disorders. Perhaps the most salient example of genetic findings at this writing is a polymorphism in the promoter region of the serotonin transporter (5-HTT) gene, which regulates the reuptake of serotonin from synapses. A remarkably large literature has emerged over the last decade showing that individuals with two short alleles in this region demonstrate higher levels of anxiety and negative affect as measured in various ways, and are at higher risk for mood and anxiety disorders (e.g., Greenberg et al., 2000; Sen, Burmeister, & Ghosh, 2004). Further, animal models of this effect have begun to reveal the potential developmental pathways that are involved (Ansorge, Zhou, Lira, Hen, & Gingrich, 2004). A recent epidemiological report from the Dunedin epidemiological study particularly catalyzed interest in this effect (Caspi et al., 2003). The authors reported a strong gene by environment interaction: Subjects with two short copies of the serotonin transporter gene showed an elevated risk for major depression as a function of the number of stressful life events (e.g., loss of job, death of close family member, moving) that were experienced, whereas this relationship was not observed in subjects with two long copies of the gene. This relationship has received at least one independent replication (Kendler, Kuhn, Vittum, Prescott, & Riley, 2005). Neuroimaging research has quickly provided data regarding the brain circuitry that is involved with such effects. In one noteworthy example, Danny Weinberger’s group at NIMH recently reported exciting new findings about the relationship of the 5-HTT gene to coupling between
Chapter 12.qxd
3/28/2007
328
1:00 PM
Page 328
CUTHBERT
the amygdala and emotionally relevant areas of the anterior cingulate cortex (ACC; Pezawas et al., 2005). Structural neuroimaging data showed that the only two areas of the brain that demonstrated morphological changes in subjects with the short-short allele, as compared to the long-long allele, were in the amygdala and the perigenual ACC, both known to be important for the regulation of emotional behavior; followup analyses showed a significant correlation between structural measures in these two areas. Further, functional imaging data in a task requiring subjects to view emotional faces indicated a tight coupling of response between the amygdala and ACC for subjects with the long-long allele; however, a relative uncoupling was observed for subjects with the shortshort allele. These data provide strong evidence for a major role of the serotonin transporter gene in understanding phenomena implicated in both anxiety and depression, and provide support for a conclusion by many neuroscientists that these lines of research, when brought to full fruition, will be necessary and sufficient to produce new genetic and psychopharmacologic therapies for brain disorders. That is to say, psychological conceptions of these disorders may be superfluous when direct access to the brain circuitry involved is available. Summary Clinical science in psychology has thus encountered criticism both from practitioners in psychology and from scientists who primarily adopt a reductionistic and neuroscience-oriented approach to mental disorders. It is perhaps only a slight overgeneralization to say that the former criticizes clinical science for being too scientific, and the latter for not being scientific enough. The reader might object that I have created a pair of straw men with these characterizations; however, the foregoing sections reflect my experience since coming to NIMH that both sets of viewpoints are strongly held, and quite prevalent. What might the next decade or two of clinical science in psychology look like, as the field responds and reacts to these near diametrically opposite influences? Anticipating the future is always perilous, and depends to a large extent on one’s vantage point. The following suppositions about where we are likely to go—and where we need to go—in the future reflect an attempt to synthesize the viewpoints of both a laboratory scientist and an NIH program official tasked with developing new research portfolios.
Chapter 12.qxd
3/28/2007
1:00 PM
Page 329
12. TRANSLATIONAL RESEARCH AND THE FUTURE
329
TRANSLATIONAL RESEARCH Translational research has occupied a core position as one of the dominant research themes at the NIH over the past decade. The idea of speeding the pace at which discoveries in basic science are applied to clinical practice has obvious appeal for scientists, clinicians, the public, and politicians. Curiously, for a term that has gained such widespread use, translational research has never been exactly defined. One typically sees phrases such as “bench to bedside” or “science to practice.” In fact, the term translational research covers a wide variety of activities that differ considerably, with at least three distinct meanings. The first is used to refer to the ways in which research on basic mechanisms of behavior or physiology may inform the study of abnormal behavior or pathophysiology. This usage was best defined in the recent report of an NIMH Working Group for Basic Behavioral Research: “Translational research in the behavioral and social sciences addresses how basic behavioral processes inform the diagnosis, prevention, treatment, and delivery of services for mental illness, and, conversely, how knowledge of mental illness increases our understanding of basic behavioral processes” (NIMH, 2000, p. iii). This sense clearly reflects one of the fundamental tenets of most contemporary research in psychopathology: the idea that psychopathology can best be understood as perturbations of normal behavioral processes in areas such as cognition, emotion, mood, social interactions, language, and so forth. This postulate seems rather obvious, but it remains true that many popular therapies that enjoy considerable empirical support—for example, cognitive behavioral therapy—are not derived systematically from any body of basic behavioral research. The second, related sense of translational research refers to the development of new treatment or prevention approaches—whether behavioral or pharmacological—by an enhanced understanding of the mechanisms or pathophysiology of disorder. A recent example has been the use of propranolol, a beta-adrenergic antagonist, to prevent the onset of posttraumatic stress disorder (PTSD) following a trauma (e.g., Pitman et al., 2002). Finally, a third sense refers to the dissemination of newly found treatments throughout the primary care and public health sectors—a process typically referred to as dissemination research in some quarters of NIH. For the most part, because mental disorders are so poorly understood, work at NIMH has concentrated on the first sense of translational
Chapter 12.qxd
330
3/28/2007
1:00 PM
Page 330
CUTHBERT
research. This effort has been applied in both neuroscience and behavioral science arenas. In the neurosciences, the Silvio O. Conte Centers for the Neuroscience of Mental Disorders have provided a significant source of support for large-scale, multifaceted centers that apply molecular and cellular neurosocience, neuroimaging, and genetics to studying the pathophysiology of various mental disorders; this program has been seen at NIMH as highly successful in terms of leveraging basic neuroscience findings to studies of mental illness. In behavioral science, the application of basic research to clinical issues has tended to lag considerably as compared to the neurosciences. In response to this perceived need, NIMH developed in the early part of the decade a slate of programs designed to encourage and foster translational research in the behavioral sciences. The creation of these programs stemmed from the perception that basic behavioral scientists and clinical researchers frequently worked in different settings and so faced unique barriers to translation. Basic researchers lack the opportunity to become familiar with clinical settings and the kinds of issues most germane to clinical research, whereas it is often difficult for clinical scientists to glean from the literature sufficient information regarding the subtleties and critical details necessary to implement basic research paradigms in clinical settings. As a result, a constant feature of all the announcements that were developed was a formal requirement that at least one basic and one clinical scientist demonstrate a significant involvement in the research plan. Although some talented clinical scientists found this requirement somewhat procrustean, it nevertheless reflected the basic goal of the program—which was not simply to encourage more clinical research in general, but in particular to increase the amount of basic behavioral research which was being applied to clinical issues. Another facet of the program was recognition that, due to the extant barriers, it was unrealistic to expect that many research groups would have collaborative efforts that were sufficiently advanced to compete successfully for large-scale center grant applications. Rather, mechanisms for bootstrapping the development of investigative teams needed to be developed. Accordingly, a program announcement (PA) entitled “Building Translational Research in Behavioral Science” was developed, which incorporated two relatively unique mechanisms of support. The first was a pilot/exploratory mechanism (the R21, in NIH parlance) that provided support for the formation of research groups, including dedicated time on the project and small-scale feasibility studies. The second,
Chapter 12.qxd
3/28/2007
1:00 PM
Page 331
12. TRANSLATIONAL RESEARCH AND THE FUTURE
331
an R24 mechanism, was designed to support larger scale work that would provide the resources necessary to conduct translational research, such as devising new instruments or measures, adapting basic-science paradigms for use with clinical patients, and running the large-scale pilot studies that are prerequisite for R01 or center grant applications. These types of support mechanisms are relatively unusual, and reflect NIMH’s awareness of the need to encourage nascent collaborations between basic and clinical investigators. NIMH also provides support for larger-scale work in translational research. An additional PA, “Translational Research Grants in Behavioral Science,” provides support for R01 grants; although this PA includes the minimum requirement of one basic and one clinical investigator, larger research groups are preferable in this case. Finally, at the high end of the funding scale, there is a mechanism for research centers in a recently reissued PA, “Translational Research Centers in Behavioral Science” ; this center involves the P50 mechanism, and is comparable to the Conte neuroscience centers in permitting direct costs of up to $1.5 million per year. These descriptions are mentioned at some length for several reasons. First, they exemplify the emphasis that NIMH has placed on translational research as a vehicle for providing support in the area of behavioral science research. That is to say, although NIMH continues to support a considerable amount of basic behavioral science research, the emphasis clearly has shifted to research that can demonstrate potential clinical utility in as near a term as possible. Second, the PAs continue to be active, with recently expired announcements slated for reissuance in the near future; thus, these are optimal mechanisms for investigators who wish to develop projects in the area of psychological clinical science. As is always the case, the best procedure for investigators who are potentially interested in submitting an application is to contact a program officer whose area of coverage appears to be appropriate. This suggestion holds true not only for NIMH, but for all of the NIH institutes. More broadly, however, one of the strongest trends in the future is likely to be an ever-increasing emphasis on cross-disciplinary efforts— both between basic and clinical scientists, and among investigators from a wide variety of disciplines. The NIMH programs for translational research in behavioral science thus exemplify, while explicitly supporting, what the future is going to look like. A statement about increasing multi-disciplinary research seems like a cliché. However, although this
Chapter 12.qxd
332
3/28/2007
1:00 PM
Page 332
CUTHBERT
topic receives much verbiage, much of the machinery of the NIH—from the organization of study sections to deeply ingrained cultural mind sets—still implicitly is conceived in terms of single scientists submitting investigator-initiated R01 applications; veteran investigators know that an R01 application that includes multiple disciplinary elements is a highrisk proposal. Within the NIH, any discussion about innovative approaches to funding multidisciplinary work inevitably provokes a concern in response about the effects on the R01 payline. One senses, however, that the necessity for cross-disciplinary science is slowly overcoming the stationary inertia of the NIH system. A particularly salient bellwether of this change is the NIH Roadmap program. Briefly, the NIH Roadmap (“the Roadmap”) is a collection of funding initiatives designed to facilitate interdisciplinary research across the research areas of interest that are typically funded by individual NIH institutes. Notably, one major component of the Roadmap is named “Research Teams of the Future,” with one of its major units called “Interdisciplinary Research.” In the molecular and cellular domain, some examples of Roadmap activities include molecular and cellular libraries, nanotechnology, and genomics initiatives. In behavioral research, funding RFAs have included supplements for methodological developments, and behavioral research components have been included in a number of interdisciplinary announcements. The Roadmap is perhaps not the most aptly named program, as the initial startup activities have had perhaps more the flavor of the Oklahoma Land Rush, with settlers dashing out in search of uncharted territory. However, there is no doubt that the Roadmap marks the way of the future, and that it will be important for psychological clinical science to be part of this enterprise. The explosion of knowledge in so many areas demands no less. The sections that follow indicate some areas in which the intersection of basic and clinical science may play a particularly noteworthy role over the years to come. PSYCHOLOGICAL CLINICAL SCIENCE AND THE DIAGNOSTIC AND STATISTICAL MANUAL FOR MENTAL DISORDERS (DSM) The current scheme of the DSM came into being in 1980 with the DSM–III (with evolutionary changes eventuating in the current DSM–IV), and incorporated the approach developed by Robins and Guze (1970) in which categorical disease entities were inferred on the basis of
Chapter 12.qxd
3/28/2007
1:00 PM
Page 333
12. TRANSLATIONAL RESEARCH AND THE FUTURE
333
clinical symptoms, course, and familial aggregation (American Psychiatric Association, 1980, 1994). An important feature of the DSM was that investigators could provide clear definitions as to how samples of patients had been ascertained, overcoming the individual clinician or site biases that were often present in earlier clinical studies. However, the categories that were provided were based largely on clinical experience and expert consensus; further, the “Chinese menu” scheme, in which patients could be given a diagnosis for meeting a certain number of symptoms from a longer list, meant that patients with the same diagnosis might have quite different symptom patterns. This situation prompted a recent director of the NIMH to comment, “If a relative strength of DSM is its focus on reliability, a fundamental weakness lies in problems related to validity” (Hyman, 2003, p. xii). In spite of this concern, the structure and authority of the DSM have meant that nearly all the studies proposed in grant applications (and in published reports) follow with little questioning the categories used in the DSM. One unfortunate effect has been to reify these categories and, to no small extent, to calcify them—stultifying efforts to transcend them in more systematic ways. However, accumulating experience with the DSM has led to growing concerns with whether the categories are fundamentally sound. These problems include the high degree of comorbidity, which would not be expected if the categories were truly independent disorders; the proliferation of subtypes, calling into question the validity of the main diagnosis; the apparent incorrect placement of some disorders; and the confusion occasioned by the multiaxial system, especially for Axes I and II (see e.g., an excellent set of papers by Clark, 2005; Krueger, Markon, Patrick, & Iacono, 2005; Watson, 2005; Widiger & Samuel, 2005). It has been somewhat surprising, from the viewpoint of an NIMH program official, that so few applications take on the challenge of redefining the fundamental nature of the categories themselves—but the authority of the DSM, in concert with the hegemony of study sections oriented to DSM definitions, exerts a powerful suasion against heterodoxy. Over the next few decades, however, increasing recognition of the problems with the current DSM can be expected to foster research that incorporates a more iconoclastic view of how mental disorders are defined, and what the appropriate categories or dimensions are. As of this writing, the American Psychiatric Association is conducting a series of conferences to establish what is known and still needs to be established
Chapter 12.qxd
334
3/28/2007
1:00 PM
Page 334
CUTHBERT
as a foundation for the next iteration of the DSM, namely, the DSM–V. It is clear from the meetings held to date that there is considerable openness to recognizing the problems with the current system, and to implementing change. This is an area in which clinical science can, and should, maintain a leading role in the future. Psychological scientists have been at the forefront of those adapting new developments in statistical models and measurement theory to the study of psychopathology, such as item response theory, taxonometric analyses, hierarchical models, and dimensional approaches to psychopathology. Most of the current scales and instruments for the assessment of psychopathology have also emerged from psychological research, and this is a trend that can be expected to continue in the future. However, considerably more work needs to be done in these areas. Knotty theoretical issues need to be resolved, such as the current series of debates as to whether dimensional models are more appropriate to the phenomena in various domains of disorder as compared to categorical models (e.g., Phillips, First, & Pincus, 2003). However, an equally demanding task is to devise instruments that are appropriate in time and burden for clinical use, that can measure functional impairment accurately, and, where dimensional approaches are incorporated, that provide ready cutpoints for clinicians to generate the kinds of categorical information necessary for such quotidian needs as billing and insurance reports. This is an area where progress will be responsive to both clinicians’ and neuroscientists’ perspectives about psychological clinical science. As noted earlier, a number of clinicians and clinical researchers fault clinical scientists for attempting to establish lists of empirically supported treatments for specific DSM disorders, typically stemming from single-disorder RCTs; in contrast, real-world patients present with a multitude of problems, some of which may encompass multiple DSM diagnoses, and others that fail to fit neatly into any particular DSM category. To a large extent, this indicates a problem with the diagnostic system itself; as diagnostic systems evolve to reflect more accurately the fundamental biobehavioral nature of mental disorders, we can expect that many of these concerns will be largely transcended. As Beutler and Malik noted recently, “Imagine what would happen if diagnostic decisions actually did direct us to qualities of treatment that improved outcomes … we would be positioned to develop a set of markers that would direct the selection of a specific treatment” (2002, p. 267).
Chapter 12.qxd
3/28/2007
1:00 PM
Page 335
12. TRANSLATIONAL RESEARCH AND THE FUTURE
335
At the other end of the spectrum, neuroscience, in spite of the rapid advances in so many areas, also has as its Achilles heel what some have dubbed the phenotype problem: For all the sophisticated neuroimaging, genetic, and molecular/cellular technologies, neuroscience is ultimately dependent on the clinical definitions of the disorders, or behavioral and psychological measurements of characteristics that represent risk factors for disorders. For instance, without the personality constructs and associated measurement instruments that have been developed over the last 20 years, it would have been impossible to produce the research demonstrating the relationship, described earlier in this chapter, between negative affect and the short form of the serotonin transporter gene. As the clinical impairments of mental disorders occur in behavioral and psychological domains, refinements in measurement are an obvious vital component of an integrative neuroscience research program. Again, this is an area in which comments about what is needed may seem obvious, but the needs involve steps that are not easy to accomplish in the real world. It remains the case that it is very difficult to get grant applications funded, or manuscripts published, if patient samples have not been ascertained according to DSM criteria. This is an understandable situation, for the problems of scientific communication and diagnostic reliability remain. It would appear that, for some time to come, it will be necessary to provide both DSM diagnoses for patient samples and data regarding measures that are intended to transcend the current categories. Only in this way can the exigencies of the current diagnostic criteria still be met, while forging ahead with new conceptualizations and measures. It is also the case that a large majority of current studies, even those that acknowledge the problems of comorbidity, confine themselves to a single DSM diagnosis to avoid the analytical problems introduced by heterogeneity in number and type of comorbid (or co-occurring, to use the more theoretically neutral term) conditions. Obviously, new diagnostic schemes, particularly crosscutting dimensional approaches, must of necessity incorporate multiple entries from the current DSM system. At the least, it is apparent that larger samples than are necessary for single-diagnosis studies will be required in order to provide adequate statistical power, and variance within each of the old DSM categories, for adequate analyses. It may be that one or more conferences would be useful in order to develop consensus approaches for dealing with the practical and conceptual problems of comorbidity in research studies.
Chapter 12.qxd
3/28/2007
1:00 PM
Page 336
336
CUTHBERT
Psychological Clinical Science and Neuroscience As noted earlier, the exponentially accelerating progress of research in genetics and neuroscience is proceeding at a pace that carries negative implications for the long-term viability of psychological clinical science as an independent discipline. As with some other comments in this chapter, this is a strongly worded statement, but one stemming from direct experience in the NIH community. It is possible to respond to this kind of view in several different ways. One of the most obvious, of course, is that clinical science in psychology is far from excluded from work in the neurosciences. Many of the investigators producing some of the most exciting work in brain sciences were trained in psychology, and/or continue to work in psychology departments. At another level, however, the point is not that psychology has a number of outstanding neuroscientists, but that many scientists in the biomedical community regard a science of psychological concepts, or of behavior, to be rapidly falling behind as a significant contributor to progress in understanding mental disorders. It is possible to be defensive about this state of affairs. From a public health point of view, however, it is simply the case that approaches that appear to have the greatest payoff will inevitably attract the largest share of scientific interest and research dollars. Accordingly, the burden will continue to fall on the psychological clinical enterprise over the next few decades to demonstrate its value as a science. (This is, one suspects, a position with which Dr. McFall would agree.) It is clear that in the future, developments in psychological theory risk being ignored by the wider research community if they are not framed within current findings in brain science. A good example is the Pezawas et al. (2005) article mentioned earlier. The amygdala is well established as a major center for the processing of emotionally salient stimuli, and the ACC has received increasing attention for its role in integrating and regulating affective as well as cognitive processes (e.g., Bush, Luu, & Posner, 2000; Posner & Rothbart, 2000). The findings that these two structures appear to have their morphology jointly regulated by the serotonin transporter gene, and that a polymorphism in this gene strongly associated with vulnerability or risk tends to uncouple functional interactions between the two locations, appear to have profound implications for how we think about affective processes and their control. In this light, future psychological theories about emotion regulation would be much stronger to the extent that they were developed in the context of findings such as the Pezawas et al article. (The example is provided as a general illustration,
Chapter 12.qxd
3/28/2007
1:00 PM
Page 337
12. TRANSLATIONAL RESEARCH AND THE FUTURE
337
and no implicit criticism of any current emotional regulation theory is intended nor should be inferred.) At the same time, however, it remains true, as B. F. Skinner (1972) stated, that “A comprehensive set of causal relations stated with the greatest possible precision is the best contribution which we, as students of behavior, can make in the co-operative venture of giving a full account of the organism as a biological system” (p. 270). In this sense, as discussed earlier, a rigorous psychological science is likely to retain an indispensable role in defining the phenomena relevant to mental disorders. A stellar example of the ways in which translational research in behavioral science and relevant neuroscience can intersect to produce significant advances is provided by recent decades of research in child development, personality, and neuroimaging. In broad strokes, years of research on child development have shown that the superordinate dimensions of temperament are best considered in terms of three factors of positive affectivity, negative affectivity, and what might be called effortful control (e.g., Rothbart & Bates, 1998). In turn, the collaboration between Rothbart, a developmental psychologist, and Posner, a cognitive psychologist and pioneer in neuroimaging, led to the initial findings that the ACC is a critical component of the effortful control factor, and is involved not only in cognitive regulation but also in affective regulation (Posner & Rothbart, 2000); subsequent behavioral tasks were developed by this group expressly to serve as indicants of this anterior network activity. Relevant behavioral genetics work with child temperament suggested that these factors are strongly genetically determined, particularly for negative affectivity (e.g., Goldsmith, Lemery, Buss, & Campos, 1999). In a different domain, investigators pursuing independent lines of research in personality theory also arrived at the importance of superordinate personality dimensions of which positive and negative emotionality were of primary importance (e.g., Costa & McCrae, 1992; Saucier & Goldberg, 1998; Tellegen, Watson, & Clark, 1999); relatedly, investigators pursuing new statistical models have been able to explicate the hierarchical nature of symptoms and disorders in ways that relate more fine-grained and more global conceptions of disorders (e.g., Markon, Krueger, & Watson, 2005). Finally, Clark (2005), in a masterful summary and integration of multiple literatures, has made a persuasive case for the primacy of the three temperament factors just indicated, discussed how they serve as the fundament for both normal personality and mental disorders, and concisely argued that Axis I and Axis II disorders fail to show sufficient differentiation in time of onset or clinical course to justify
Chapter 12.qxd
338
3/28/2007
1:00 PM
Page 338
CUTHBERT
keeping them on separate axes (i.e., categories). The ramifications of all the work summarized here are likely to guide and stimulate research for decades to come in both the brain sciences and psychological clinical science, and to have a major influence on future versions of the DSM. One of the major areas in which psychological clinical science can make a substantial contribution vis-à-vis the brain sciences lies in the area of brain plasticity. One of the attractive, and indeed seductive, aspects of the recent exciting findings in the genetics of brain activity concerns the apparent potency and near-inevitability of the genetic effects. For instance, despite appropriate caveats in most of the reports, it is not difficult to conclude that persons with two short copies of the promoter allele in the serotonin transporter gene are virtually doomed to suffer from affective disorders, or at least significant distress, and that there is little that they can do to avoid this fate. It also seems to be a short leap to the conclusion that such biological bases of mental disorders can therefore only be treated with biological treatments, that is, pharmacologic approaches. However, an equally exciting trend in modern brain science involves new discoveries about the remarkable plasticity and adaptability of the brain, and this is an area that is virtually untouched with respect to its potential for mental disorders. To date, much of the work on systems-level plasticity has involved sensorimotor systems. For instance, there is now abundant evidence that primary sensory receiving areas in the cortex are highly plastic, and that the size and organization of these areas change after sustained stimulation of appropriate surface areas of the body (e.g., Braun, Schweizer, Elbert, Birbaumer, & Taub, 2000; Elbert, Pantev, Wienbruch, Rockstroh, & Taub, 1995). Further, it is now well established that this same kind of cortical reorganization is responsible for phantom limb pain following amputation (e.g., Flor et al., 1995). The potential of such plasticity for therapeutic use has been ably demonstrated in a series of studies by Michael Merzenich and colleagues with respect to children who have language impairments due to deficits in the processing of phonemes. This work involved an analysis of phoneme processing deficits as revealed by event-related potential measurement, and subsequently the creation of a motivationally entertaining computer game in which performance depends on correct processing of different consonant sounds. The outcomes were highly successful in demonstrating improved language processing following several sessions of treatment (e.g., Merzenich et al., 1996), and the approach has been replicated and extended in a series of
Chapter 12.qxd
3/28/2007
1:00 PM
Page 339
12. TRANSLATIONAL RESEARCH AND THE FUTURE
339
studies by Kraus and colleagues (e.g., Hayes, Warrier, Nicol, Zecker, & Kraus, 2003). There have been fewer examples to date of plasticity in areas of the brain relevant to mental disorders, particularly for late childhood and adulthood. (Research on stress-induced changes in the brain during early development is of course a very productive research area [e.g., Plotsky & Meaney, 1993]; the clinical implications of early childhood stress on adult clinical disorders have also begun to be explored [e.g., Heim, Plotsky, & Nemeroff, 2004].) However, some intriguing examples have begun to emerge of the changes that can result from systematic environmental manipulations. For example, in a recent study, dopamine D2 receptor activity in the ventral striatum was measured in cynomolgus macaques in isolation and then after a period of social housing with four-animal groups. D2 levels did not differ in isolation, but after the period of social housing, dominant animals exhibited significant increases in D2 activity whereas subordinate monkeys showed no change. More important, in a subsequent behavioral test, subordinate animals self-administered significantly greater amounts of cocaine than dominant monkeys (Morgan et al., 2002). These results are notable for demonstrating the ways in which the effects of behavioral interventions can be assessed with central nervous system (CNS) measures as well as by important behavioral outcomes. To date, there have been few studies that involve the measurement of changes in CNS activity in patients with mental disorders following behavioral interventions. However, one interesting report, from an innovative series of positron emission tomography (PET) studies being conducted by Helen Mayberg and associates with respect to major depression, involved the effects of cognitive-behavioral therapy (CBT; Goldapple et al., 2004). In this study, the investigators reported a pattern of changes in the dorsal cingulate gyrus and the hippocampus that was distinct from the pattern of brain changes previously shown to be associated with positive response to antidepressant medications—obviously suggesting the appealing conclusion that CBT works through a different mechanism in the brain than medications do. This study represents an outstanding and necessary first step of examining an extant therapy in the context of a clinical series, and augurs the feasibility of examining changes in brain circuitry following various kinds of behavioral therapies and interventions—not necessarily as the ultimate dependent variable, but as part of an overall assessment that provides valuable information about mechanisms of response. In the future,
Chapter 12.qxd
3/28/2007
340
1:00 PM
Page 340
CUTHBERT
it is possible that targeted behavioral interventions could produce equally specific responses in relevant circuits of the brain, much in the way that the Merzenich group manipulations affected particular aspects of language processing. Because plasticity in the brain evolved to support flexible responses to environmental demands and to produce many different kinds of behavior, it only makes sense that interventions that alter the environment could be highly precise in changing specific parts of the brain in specific ways. As yet it is difficult to know exactly what behavioral interventions might be effective, what the time course of treatment would be, and what kinds of paradigms would be appropriate to demonstrate successful plasticity in the relevant areas. However, one of the strengths of psychological science lies in the development of these kinds of paradigms and measures. To date, most of the studies of brain plasticity have necessarily concentrated on the negative changes that occur due to stress and other deleterious events. A virtually untouched area concerns what kinds of environmental manipulations could produce positive or protective types of changes. This promise leads to one of the major goals of the NIMH, but one that has tended to languish due to a lack of appropriate paradigms: prevention. Given a sufficient understanding of the changes involved in mental disorders, it should be possible not only to develop superior treatments, but to introduce appropriate interventions before disorders develop or when they are at a mild level. This issue is directly relevant to current controversies in the literature regarding the importance of mild mental disorders, and whether or not they should be treated (see e.g., Kessler et al., 2003). There is an understandable sentiment in many areas of the mental health community that scarce treatment resources should be reserved for serious illness. However, from a public health perspective, it makes sense to pay attention to mild disorders, because treating them represents essentially a prevention intervention that can avert more serious disorder subsequently. Although this entire research domain is as yet almost completely undeveloped, it nevertheless thus represents an area of tremendous potential. TRAINING IN PSYCHOLOGICAL CLINICAL SCIENCE Preparing students for future research in psychological science is one of the most important missions for clinical area groups. Fortunately, this issue has received considerable support in recent years, fostered by academic
Chapter 12.qxd
3/28/2007
1:00 PM
Page 341
12. TRANSLATIONAL RESEARCH AND THE FUTURE
341
associations such as the Association for Psychological Science (previously the American Psychological Society) and the APA. This area is of interest to the NIMH in terms of ensuring a pool of well-trained investigators for the future. Reflecting this interest, in January 2004 a joint meeting was held between NIMH and APCS to address specifically the topic of clinical science training. Nearly 30 eminent investigators from the field joined NIMH and other NIH staff in discussing future directions for clinical science programs and the characteristics that comprise successful training programs. Although space prohibits a thorough discussion of the extensive deliberations of this meeting, a few salient points may be noted. First, there is growing support for developing a clinical science training model that transcends the Boulder Model and practitioner-scientist models. (The Indiana program is one notable example, in which translational research is unusually well developed by virtue of the extensive integration of the clinical training program with other areas of the department.) Second, extensive discussions were held about how to train students in an era of neuroscience and genetics, and how to train them for “Big Science.” Third, it is also vitally important to maintain strong communication between clinical science and services delivery in clinical psychology. The need to train students to function in an era of genomics, brain science, and “Big Science” is particularly clear from the vantage point of an NIH institute. One easy prediction is that it will be imperative to provide students with training in genetics, neuroimaging, and brain sciences that is, at a minimum, adequate to prepare them to collaborate with investigators in various areas of the neurosciences. This does not mean that they need to be independent experts, capable of conducting such research themselves (although of course some students will choose this route). Rather, students need to be in possession of the core principles, findings, assumptions, and vocabularies of these other disciplines so that they can participate effectively and constructively in collaborative projects. As the demands of core curricula become ever more stretched, it is likely that such basic training in brain sciences will be an indispensable component for any psychological clinical scientist. For a similar reason, the undergraduate background of students who wish to enter clinical science programs needs to be given careful consideration. Students interested in clinical psychology frequently take primarily a number of psychology courses in abnormal, developmental, cognitive, and so forth, both in preparation for graduate work and as a
Chapter 12.qxd
342
3/28/2007
1:00 PM
Page 342
CUTHBERT
reflection of their interests. However, they will in many cases be much better served by taking a minimum of psychology courses, and instead preparing themselves with a strong background in relevant sciences and methodology areas—such as physiology, genetics, mathematics and statistics, chemistry, and physics. This will provide them with a much more rigorous preparation for the kinds of work that they will need to do in graduate and postgraduate work, and also will give them an enhanced foundation for collaborations with investigators in other disciplines. Beyond coursework per se, there are other, more abstract considerations about the nature and direction of clinical training. There has long been a tradition in psychology that in order to demonstrate independence from one’s mentor, young investigators must develop a unique field or subfield of study, preferably developing a new theory to account for findings. This proliferation of one-investigator, one-theory approaches contributes to a balkanization of efforts in the discipline. One’s students may carry on tests of the theory in graduate school, and for a few years after moving on to their first academic position. However, to demonstrate their own independence, students must in turn develop their own subareas, and so on. Further, such a proliferation of theories too often results in progress being hampered by spats over the merits of individual theories, rather than investigators working as a community to advance a common knowledge base. Of course, science is always about healthy intellectual disputes, but too often in psychology the intellectual energy seems to be dissipated in ways that, from a broader perspective, do not move us ahead as fast as we might like. As opposed to the solipsistic attractions of one’s own personal theory, it appears that the demands of multidisciplinary environments will place a greater premium on the skills of young investigators to work effectively in research teams with scientists from many different disciplines. Young clinical scientists must be able to merge their own expertise and competencies effectively into integrative research projects that include genomics, molecular and cellular neuroscience, neuroimaging, and bioengineers for the neurosciences; and clinicians from various disciplines, primary care physicians, health systems professionals, and industry representatives for clinical trials work. Training students in such skills is a relatively new endeavor, and it is neither easy nor clear as to what the appropriate balance between individual laboratory work and interdisciplinary efforts would be. However, one might expect that in the later years of graduate school and through postdoctoral opportunities, the percentage of time spent in collaborative efforts might
Chapter 12.qxd
3/28/2007
1:01 PM
Page 343
12. TRANSLATIONAL RESEARCH AND THE FUTURE
343
well approach or exceed 50% of the student’s time. Such a shift could represent a seismic change for many programs, and would of course not be appropriate or necessary for all students. However, over the course of the coming decades, it is imperative that clinical science training programs in psychology prepare their students to function in the context of the coming generations of brain sciences. Finally, again from the viewpoint of an NIMH program official, it is somewhat surprising as to how little coaching students seem to have received about the mechanics and culture of the NIH grant system and other granting agencies. It is not uncommon for students even from top programs to be unfamiliar with the need to contact program staff about the appropriateness of their research ideas, to fail to appreciate the highly selective (and often narrow) nature of the research priorities at the various NIH institutes, and to be unaware of the need to couch grant applications in such a way as to fit the granting agencies’ priorities. There is also a lack of familiarity with the basic grant mechanisms of the NIH and other granting agencies. The demands of research and clinical training may obviate any formal attention to these issues in the graduate curriculum. However, training programs might well consider whether a series of brown-bag luncheons, mock grant-writing assignments, and other similar activities would be well worth the time in preparing their students for the challenges of participating in the real-world research endeavor. CONCLUSIONS It is perhaps appropriate to close by restating the cardinal principle from the McFall Manifesto: “Scientific clinical psychology is the only legitimate and acceptable form of clinical psychology” (McFall, 1991, p. 76). This statement will be even more apropos in the next two decades than in the last two. With respect to clinical practice in psychology, the evertightening demands of health services economics will undoubtedly place increasingly stringent mandates for the use of empirically supported treatments. With pharmacological treatments for mental disorders growing in use due to concerns about cost and provider availability, the pressure to provide demonstrably effective behavioral and psychosocial therapies will be even stronger. However, this does not mean that the burden is all on the side of the practice community to fall in line. It is the responsibility of clinical science in psychology to contribute to new and more valid conceptualizations of mental disorders, which in turn generate
Chapter 12.qxd
3/28/2007
1:01 PM
Page 344
344
CUTHBERT
diagnostic systems that more accurately capture the nature of mental illness as it occurs in the real world; and, given that, to develop and validate treatments with demonstrable efficacy in treating and preventing these disorders. McFall’s Manifesto is thus not antithetical to clinical practice, but represents an ideal of what could happen when clinical science and clinical practice are organic parts of the same enterprise. Vis-à-vis brain science, the strength of psychological clinical science will be dependent on its ability to demonstrate, in some social Darwinian sense, its continuing contributions to the science of mental disorders during what is likely to be one of the most exciting times in the history of the field. Fortunately, psychological clinical science appears to be very well poised to do so. First, mental illness consists of disorders whose primary impairments and burdens are measured in behavioral and psychological terms. Thus, a discipline with a strong science base in rigorous construct development and measurement must always be a vital component of research advances. Second, clinical science has as a potential resource the vast scope and sophistication of modern basic psychological science, and other related disciplines, to draw on in explicating the dysfunctions of mental illness. This is the promise of translational research. Finally, one of the most potentially exciting areas of discovery, which as yet remains almost totally uncharted, is an exacting specification of the ways in which brain functioning may be changed by the systematic manipulation of an organism’s environment and its behavioral contingencies. Because the adaptability of the human brain is perhaps its most remarkable feature, this control over brain plasticity is likely to rival, at the least, the promise inherent in the genetics revolution. It will indeed be exciting to see what the coming two decades of psychological clinical science will bring. ACKNOWLEDGMENTS The views expressed are the private views of the author and do not reflect the official position of the National Institute of Mental Health, the National Institutes of Health, or the federal government. REFERENCES American Psychiatric Association. (1980). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, DC: Author. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author.
Chapter 12.qxd
3/28/2007
1:01 PM
Page 345
12. TRANSLATIONAL RESEARCH AND THE FUTURE
345
Ansorge, M., Zhou, M., Lira, A., Hen, R., & Gingrich, J. (2004). Early-life blockade of the 5-HT transporter alters emotional behavior in adult mice. Science, 306, 879–881. Beutler, L. (2004). The empirically supported treatments movement: A scientistpractitioner’s response. Clinical Psychology: Science and Practice, 11, 225–229. Beutler, L. E., & Malik, M. M. (2002). Diagnosis and treatment guidelines: The example of depression. In L. E. Beutler & M. M. Malik (Eds.), Rethinking the DSM: A psychological perspective (pp. 251–277). Washington, DC: American Psychological Association. Beutler, L. E., Moleiro, C., & Talebi, H. (2002). How practitioners can systematically use empirical evidence in treatment selection. Journal of Clinical Psychology, 58, 1199–1212. Braun, C., Schweizer, R., Elbert, T., Birbaumer, N., & Taub, E. (2000) Differential activation in somatosensory cortex for different discrimination tasks. Journal of Neuroscience, 20, 446–450. Bush, G., Luu, P., & Posner, M. (2000). Cognitive and emotional influences in anterior cingulate cortex. Trends in Cognitive Sciences, 4, 215–222. Caspi, A., Sugden, K., Moffitt, T., Taylor, A., Craig, I., Harrington, H., et al. (2003). Influence of life stress on depression: Moderation by a polymorphism in the 5-HTT gene. Science, 301, 386–389. Chambless, D., Baker, M., Baucom, D., Beutler, L, Calhoun, K., Crits-Christoph, P., et al. (1998). Update on empirically validated therapies: II. Clinical Psychologist, 51, 3–16. Clark, L. A. (2005). Temperament as a unifying concept in the study of personality and psychopathology. Journal of Abnormal Psychology, 114, 505–521. Costa, P. T., Jr., & McCrae, R. R. (1992). Normal personality assessment in clinical practice: The NEO Personality Inventory. Psychological Assessment, 4, 5–13. Crits-Cristoph, P., Wilson, G. T., & Hollon, S. D. (2005). Empirically supported psychotherapies: Comment on Westen, Novotny, and Thompson-Brenner (2004). Psychological Bulletin, 131, 412–417. Eisch, A. J., & Nestler, E. J. (2002). To be or not to be: Adult neurogenesis and psychiatry. Clinical Neuroscience Research, 2, 93–108. Elbert, T., Pantev C., Wienbruch, C., Rockstroh, B., & Taub E. (1995). Increased cortical representation of the fingers of the left hand in string players. Science, 270, 305–307. Flor, H., Elbert, T., Knecht, S., Wienbruch, C., Pantev, C., Birbaumer, N., et al. (1995). Phantom-limb pain as a perceptual correlate of cortical reorganization following arm amputation. Nature, 375, 482–484. Goldapple, K., Segal, Z., Garson, C., Lau, M., Bieling, P., Kennedy, S., et al. (2004). Modulation of cortical-limbic pathways in major depression. Archives of General Psychiatry, 61, 34–41. Goldsmith, H. H., Lemery, K. S., Buss, K. A., & Campos, J. J. (1999). Genetic analyses of focal aspects of infant temperament. Developmental Psychology, 35, 972–985. Greenberg, B., Li, Q., Lucas, F., Hu, S., Sirota, L., Benjamin, J., et al. (2000). Association between the serotonin transporter promoter polymorphism and personality traits in a primarily female sample. American Journal of Medical Genetics–Neuropsychiatric Genetics, 96, 202–216. Hayes, E. A., Warrier, C. M., Nicol, T. G., Zecker, S. G., & Kraus, N. (2003). Neural plasticity following auditory training in children with learning problems. Clinical Neurophysiology, 114, 673–684.
Chapter 12.qxd
346
3/28/2007
1:01 PM
Page 346
CUTHBERT
Heim, C., Plotsky, P. M., & Nemeroff, C. B. (2004). Importance of studying the contributions of early adverse experience to neurobiological findings in depression. Neuropsychopharmacology, 29, 641–648. Hyman, S. E. (2003). Foreword. In K. A. Phillips, M. B. First, & H. A. Pincus (Eds.), Advancing DSM: Dilemmas in psychiatric diagnosis (pp. xi–xix). Washington, DC: American Psychiatric Association. International HapMap Consortium. (2003). The International HapMap Project. Nature, 426, 789–796. Kendler, K. S., Kuhn, J. W., Vittum, J., Prescott, C. A., & Riley, B. (2005). The interaction of stressful life events and a serotonin transporter polymorphism in the prediction of episodes of major depression. Archives of General Psychiatry, 62, 529–535. Kessler, R. C., Merikangas, K. R., Berglund, P., Eaton, W. W., Koretz, D. S., & Walters, E. E. (2003). Mild disorders should not be eliminated from the DSM–V. Archives of General Psychiatry, 60, 1117–1122. Krueger, R. F., Markon, K. E., Patrick, C. P., & Iacono, W. G. (2005). Externalizing psychopathology in adulthood: A dimensional-spectrum conceptualization and its implications for DSM–V. Journal of Abnormal Psychology, 114, 537–550. Levant, R. F. (2004). The empirically-validated treatments movement: A practitioner perspective. Clinical Psychology: Science and Practice, 11, 219–224. Markon, K., Krueger, R. F., & Watson, D. (2005). Delineating the structure of normal and abnormal personality: An integrative hierarchical approach. Journal of Personality and Social Psychology, 88, 139–157. McFall, R. M. (1991). Manifesto for a science of clinical psychology. Clinical Psychologist, 44, 75–88. Merzenich, M., Jenkins, W., Johnston, P., S., Schreiner, C., Miller, S. L., & Tallal, P. (1996). Temporal processing deficits of language-learning impaired children ameliorated by training. Science, 271, 77–81. Morgan, D., Grant, K. A., Gage, H. D., Mach, R. H., Kaplan, J. R., Prioleau, O., et al. (2002). Social dominance in monkeys: Dopamine D2 receptors and cocaine selfadministration. Nature Neuroscience, 5, 169–174. National Institute of Mental Health. (2000). Translating behavioral science into action: Report of the National Advisory Mental Health Council Advisory Workgroup. Bethesda, MD: Author. Pezawas, L., Meyer-Lindenberg, A., Drabant, E., Verchinski, B., Munoz, K., Kolachana, B., et al. (2005). 5-HTTLPR polymorphism impacts human cingulated-amygdala interactions: A genetic susceptibility mechanism for depression. Nature Neuroscience, 8, 828–834. Phillips, K. A., First, M. B., & Pincus, H. A. (Eds.). (2003). Advancing DSM: Dilemmas in psychiatric diagnosis. Washington, DC: American Psychiatric Association. Pitman, R. K., Sanders, K. M., Zusman, R. M., Healy, A. R., Cheema, F., Lasko, N. B., et al. (2002). Pilot study of secondary prevention of posttraumatic stress disorder with propranolol. Biological Psychiatry, 51, 189–192. Plotsky, P. M., & Meaney, M. J. (1993). Early, postnatal experience alters hypothalamic corticotropin-releasing factor (CRF) mRNA, median eminence CRF content and stress-induced release in adult rats. Molecular Brain Research, 18, 195–200.
Chapter 12.qxd
3/28/2007
1:01 PM
Page 347
12. TRANSLATIONAL RESEARCH AND THE FUTURE
347
Posner, M., & Rothbart, M. (2000). Developing mechanisms of self-regulation. Development and Psychopathology, 12, 427–441. Rakic, P. (2002). Neurogenesis in adult primate neocortex: an evaluation of the evidence. Nature Reviews, 3, 65–71. Robins, S., & Guze, Y. (1970). Psychiatric diagnosis. American Journal of Psychiatry, 13, 45–99. Rosen, G. M., & Davison, G. C. (2003). Psychology should list empirically supported principles of change (ESPs) and not credential trademarked therapies or other treatment packages. Behavior Modification, 27, 300–312. Rothbart, M.K., & Bates, J.E. (1998). Temperament. In W. Damon & N. Eisenberg (Eds.), Handbook of child psychology: Vol. 3. Social, emotional, and personality development (5th ed., pp. 105–176). Hoboken, NJ: Wiley. Santarelli, L., Saxe, M., Gross, C., Surget, A., Battaglia, F. Dulawa, S., et al. (2003). Requirement of hippocampal neurogenesis for the behavioral effects of antidepressants. Science, 301, 805–809. Saucier, G., & Goldberg, L. R. (1998). What is beyond the Big Five? Journal of Personality, 66, 495–524. Sen, S., Burmeister, M., & Ghosh, D. (2004). Meta-analysis of the association between a serotonin transporter promoter polymorphism (5-HTTLPR) and anxiety-related personality traits. American Journal of Medical Genetics–Neuropsychiatric Genetics, 127, 85–89. Skinner, B. F. (1972). Cumulative record: A selection of papers (3rd ed.). New York: Appleton-Century-Crofts. Stirman, S., Crits-Christoph, P., & DeRubeis, R. J. (2004). Achieving successful dissemination of empirically supported psychotherapies: A synthesis of dissemination theory. Clinical Psychology: Science and Practice, 11, 343–359. Tanenbaum, S. J. (2003). Evidence-based practice in mental health: Practical weaknesses meet political strengths. Journal of Evaluation in Clinical Practice, 9, 287–301. Tanenbaum, S. J. (2005). Evidence-based practice as mental health policy: Three controversies and a caveat. Health Affairs, 24, 163–173. Task Force on Promotion and Dissemination of Psychological Procedures. (1995). Training in and dissemination of empirically-validated psychological treatments. Clinical Psychologist, 48, 3–23. Tellegen, A., Watson, D., & Clark, L. A. (1999). On the dimensional and hierarchical nature of affect. Psychological Science, 10, 297–303. Watson, D. (2005). Rethinking the mood and anxiety disorders: A quantitative hierarchical model for DSM-V. Journal of Abnormal Psychology, 114, 522–536. Weisz, J. R., Weersing, V. R., & Henggeler, S. W. (2005). Jousting with straw men: Comment on Westen, Novotny, and Thompson-Brenner (2004). Psychological Bulletin, 131, 418–426. Westen, D., Novotny, C. M., & Thompson-Brenner, H. (2004). The empirical status of empirically supported psychotherapies: Assumptions, findings, and reporting in controlled clinical trials. Psychological Bulletin, 130, 1–33. Widiger, T. A., & Samuel, D. B. (2005). Diagnostic categories or dimensions: A question for DSM–V. Journal of Abnormal Psychology, 114, 494–504.
Chapter 12.qxd
3/28/2007
1:01 PM
Page 348
Chapter 13.qxd
3/28/2007
1:01 PM
Page 349
13 The Future of the Clinical Science Movement: Challenges, Issues, and Opportunities Robert W. Levenson University of California, Berkeley
In the 14 years since Richard McFall wrote his Manifesto (MCFall 1991), the clinical science movement has gained a great deal of momentum, and the term clinical science has achieved enviable brand name recognition. This can be seen in numerous ways, ranging from the subtle (a substantial number of university-based graduate programs in clinical psychology now refer to themselves as clinical science programs) to the profound (the clinical science model is now recognized by the field’s primary accreditation body as being distinct from the older and more generic scientist-practitioner and practitioner-scientist models). The establishment of the Academy of Psychological Clinical Science (the Academy) in 1995, which largely grew out of McFall’s vision, provided an umbrella organization for certifying that member predoctoral and internship training programs in clinical and health psychology embraced and adhered to the principles of the clinical science model. Currently, 45 graduate and nine internship programs have gained membership in the Academy, representing a veritable who’s who of the top training programs in the United States and Canada. Increasingly, undergraduates considering 349
Chapter 13.qxd
3/28/2007
350
1:01 PM
Page 350
LEVENSON
careers in scientific clinical psychology make use of the list of Academy programs in deciding which graduate schools to apply for, and graduate students do the same when applying for internship training. Clearly, clinical science has won the initial round in the battle for the hearts and minds of scientifically oriented clinical psychologists. Still, the war goes on and the ultimate fate of the clinical science movement is yet to be determined. Throughout its history, clinical psychology has focused on a holy trinity of diagnosis, treatment, and etiology. In this chapter, I examine some of the challenges, issues, and opportunities the clinical science movement faces in each of these areas as well as in the general area of training. Many of these represent unsettled issues inherited from earlier times, but all have been shaped and made even more complex by the rapidly changing worlds of clinical science and practice. DIAGNOSIS: IS THE DSM THE WRONG BLUEPRINT FOR CLINICAL SCIENCE? The fruits of decades of work on diagnosis in clinical psychology and psychiatry are reflected in the various incarnations of the Diagnostic and Statistical Manual (DSM; American Psychiatric Association, 1994). Depending on one’s viewpoint, the DSM can be situated somewhere on a bipolar scale anchored on the one end by “an evolving, increasingly reliable, progressively more objective basis for parsing pathology” and on the other by “a highly uneven, hopelessly politicized patchwork of dubious descriptions that identify some ‘natural kinds’ and many hopelessly heterogeneous diagnostic categories.” Many clinical scientists view the DSM as a major impediment for building a true science of psychopathology (for thoughtful discussions of some of these issues, see Follette & Houts, 1996; Widiger & Clark, 2000). In this view, the heterogeneity of some diagnostic categories means that different investigators studying patients with the same diagnosis are, in fact, studying very different disease processes. This, of course, is a formula for disaster, especially when attempting to aggregate findings from different laboratories on what is ostensibly the same disorder. Clinical scientists have been actively involved for decades in attempts to perfect (or at least improve) the DSM in a number of different ways (Widiger, Frances, Pincus, Davis, & First, 1991). One thrust of these efforts has been to modify criteria and clarify descriptions so that the reliability of diagnosis is improved (Nathan & Langenbucher, 1999; Williams et al., 1992). These clearly noble efforts
Chapter 13.qxd
3/28/2007
1:01 PM
Page 351
13. THE FUTURE OF THE CLINICAL SCIENCE MOVEMENT
351
are reflected in the most recent versions of the DSM requiring much less inference and containing far fewer unoperationalized constructs than did earlier versions. However, as we all learned in introductory psychometrics, reliability establishes a ceiling on validity but it does not in any way ensure validity. The unsettling question remains: Are we getting better and better at assigning individuals to categories that are ultimately going to prove to be not very useful for understanding etiology, course, and treatment? A second thrust (Hinshaw, 1987) has been in identifying subtypes of some DSM disorders (where observed heterogeneity is thought to reflect the existence of more that one disease process) and clusters of others (where putative distinctions between disorders are more apparent than real). In many ways, this has been an area in which clinical scientists have really shone, applying their considerable observational, psychometric, and empirical skills in the service of these efforts. Unfortunately, the application of findings from these efforts back into the DSM has been both slow and uneven. Thus, two parallel diagnostic universes now exist, one in which the DSM is the bible and the other in which research criteria are. This state of affairs is understandable, but provides a poor model for integration of science and practice. Outside of the world of the DSM, an increasing number of clinical scientists have explicitly or implicitly abandoned the DSM entirely and are gravitating toward a transdiagnostic, symptom-oriented approach (Harvey, 2001; Harvey, Watkins, Mansell, & Shafran, 2004). The symptom-oriented approach to diagnosis is certainly not new, finding considerable favor during the heyday of behaviorism in clinical psychology (Wolpe, 1958). The transdiagnostic approach, however, is an important extension. The behaviorists focused on symptoms as an endpoint for diagnosis and treatment. In contrast, the transdiagnosticians attempt to leverage the observation that certain symptoms are manifest in a range of disorders (e.g., affect dysregulation, sleep disruption) into a deeper understanding of these symptoms as the basic building blocks of psychopathology and as promising loci for intervention. Clinical science is at a decision point regarding the DSM. It can keep its scientific portfolio diverse by straddling the fence on these issues, or it can make a choice and invest its resources in ways that are most likely to lead to significant scientific discoveries and applied payoffs. For me, there are a number of reasons that argue in favor of investing most heavily in the transdiagnostic approach. The transdiagnostic approach allows clinical science to take advantage of the massive body of knowledge in nonclinical psychological science concerning measurement of, functions served by,
Chapter 13.qxd
352
3/28/2007
1:01 PM
Page 352
LEVENSON
and underlying biological substrates for basic behavioral, cognitive, and social processes such as thinking, feeling, developing, remembering, deciding, attaching, individuating, affiliating, and so on. These are the precisely the processes that are most vulnerable to disruption in psychological disorders. Moreover, these disruptions often constitute the most debilitating aspects of mental illness for the patient, the patient’s family, and society. Importantly, this kind of analysis is at the core of the translational research movement (National Institute of Mental Health, 2000), which envisions the application of advances in the most basic behavioral and social science to reducing the burden of mental illness. The clinical science movement has a real opportunity to embrace these issues and take a bold leadership position in the debate over the future of the DSM in clinical research. Moreover, these issues have enormous implications for whether we will train clinical scientists for obsolescence or for producing and participating in significant, groundbreaking clinical research in the coming years. TREATMENT: ARE EMPIRICALLY SUPPORTED TREATMENTS THE FUTURE? Clinical science confronts several striking ironies in the realm of treatment. First there is the issue of empirically supported treatments. There is a long tradition of findings that treatment type doesn’t really matter (tracing back to Smith & Glass, 1977). Complementing this are findings that the common, nonspecific aspects of treatment (placebo, expectancy, therapeutic alliance) account for most of the variance in outcome (Ahn & Wampold, 2001; Frank & Frank, 1991). However, there is a remarkable amount of effort currently being devoted to manualizing specific treatment protocols and amassing support for their inclusion in approved lists of empirically supported treatments (Chambless & Hollon, 1998). Among the most avid consumers of these empirically supported treatments are mental health care providers and insurers—who understandably want to use the most effective and efficient treatments—and clinical science training programs—who are committed to training their students in scientifically based clinical practice. A second irony revolves around the question of who most needs psychological services and how to best deliver these services to them. A great deal of contemporary effort in treatment development and treatment evaluation in clinical science has focused on mounting well-controlled clinical trials of treatments conducted in university clinics. Implicit in
Chapter 13.qxd
3/28/2007
1:01 PM
Page 353
13. THE FUTURE OF THE CLINICAL SCIENCE MOVEMENT
353
these efforts has been the assumption that those treatments that are established as efficacious in the research clinic could then be exported into real-world community settings where they would be enthusiastically embraced by practitioners and reduce impairment in patients. However, when researchers began to investigate the relative effectiveness of these treatments administered in different settings, they found a dramatic reduction in effectiveness as treatments moved from the university research clinic into the community (Weisz, Donenberg, Han, & Weiss, 1995). There are many explanations for why this may be. For example, patients and therapists in university trials may not be representative of those in community settings (Weisz, Doss, & Hawley, 2005), and manualized treatments that are found to be effective in university-based trials may be too cumbersome for real-world use. The bottom line for clinical science is that we need to embrace the goal of developing treatments that work and can be administered not only in the rarefied atmosphere of university-based randomized clinical trials but also in real-world settings where the great majority of clinical services are delivered. ETIOLOGY: WHAT IS THE APPROPRIATE LEVEL OF ANALYSIS? Clinical psychology has lived through a number of paradigm shifts in psychology. Each model along the way—psychodynamic, behavioral, cognitive, and brain/neuroscience—has left its mark on our theories, methods, measures, and treatments. To an extent, nothing in clinical psychology is ever lost. Practitioners of psychodynamic, behavioral, cognitive, and pharmacologic treatments all still compete for patients in many urban centers. Similarly, in many hospital settings, projective tests and functional brain imaging are both used. Nonetheless, the heyday of discovery in the older paradigms has passed. In keeping with this view, when NIMH recently retargeted their grant portfolio toward research most likely to lead to breakthroughs in the treatment of severe psychopathology, biological and genetic models were enthusiastically endorsed while psychodynamic and classic behavioral paradigms were given short shrift (Levenson, 2005). Although a great deal of the current excitement in mental health research centers on the brain, its neurotransmitters, and the neural circuits that can be traced through the methods of contemporary neuroscience, it is likely that the next big thing in psychopathology research will be genetics. The application of new molecular genetics methodologies
Chapter 13.qxd
354
3/28/2007
1:01 PM
Page 354
LEVENSON
to the study of mental illness focuses on particular proteins that are involved in the basic processes thought to underlie normal and abnormal functioning. Unlike the quantitative genetics of the past, which utilized twin studies to establish the relative power of heredity and environment in the transmission of a number of DSM disorders, these new molecular methods show great promise for uncovering some of the mechanisms that underlie disease and health. Already, exciting findings are emerging using these new methods. For example, links have been established between genotypical variation related to catecholamine metabolism and deficits in performance on working memory tasks in schizophrenia (Goldenberg et al., 2003). Promising gene by environment interactions have also been elucidated such as one linking a polymorphism in a serotonin transporter gene to variation in the likelihood that life stress will lead to depressive symptomatology and suicide (Caspi et al., 2003). The implications of these kinds of findings for identifying risk for psychopathology and for designing future interventions are enormous. What is the optimal level of analysis for clinical scientists wishing to work with modern functional brain imaging and molecular genetics? It seems clear to me that these new methods are better suited for building links to basic processes (e.g., attention, social interest, affect regulation) than to full-blown, often-heterogeneous DSM diagnostic categories (e.g., autism, hyperactivity, bipolar disease). If this view is correct, clinical scientists are going to need to become much more expert in basic psychological processes. It makes little sense to marry first-rate neuroscience and first-class genetics with second-rate psychology. For this reason, clinical science should be actively courting the participation of researchers from the nonclinical areas of psychology where there already is impressive expertise in conceptualization and measurement of basic behavioral, affective, social, and cognitive processes (but sadly, often little familiarity with psychopathology). THE FUTURE OF THE CLINICAL SCIENCE MOVEMENT Judgments of the ultimate importance of the clinical science movement will be based on the issues it raises, the ideas it promulgates, the science it produces, the scientists it trains, and its influence on the field. Since its inception, the clinical science movement has articulated a clear vision of a scientifically grounded clinical and health psychology and has established a standard for member graduate and internship programs that goes
Chapter 13.qxd
3/28/2007
1:01 PM
Page 355
13. THE FUTURE OF THE CLINICAL SCIENCE MOVEMENT
355
far beyond merely talking the talk. These are its clearest victories to date. Some movements do not go beyond this point and, certainly, the clinical science movement could decide to rest on its laurels content in the knowledge that it spawned dialog, organizations, and had a significant influence on the field’s consciousness, accreditation procedures, and market for its services and products. Hopefully the energy behind the clinical science movement has not been spent, for there are still many areas in which its potential has not yet been met. In this final section, I turn to some of the areas where the final judgment on the clinical science movement is still out and where future challenges and opportunities abound. Producing Science Has the clinical science movement fundamentally changed the nature of research produced by clinical psychologists and has it spawned the kinds of breakthroughs in diagnosis, treatment, and etiology that the field, the funding agencies, and the public crave? These kinds of questions are always difficult, if not impossible, to answer. What we can say is that there are significant obstacles that stand in the path of faculty and students in clinical science programs who wish to produce significant science. Many of the top psychology programs in the country either do not have clinical science programs or have relatively small programs. Students and faculty in existing clinical science programs have to devote significant effort to clinical psychology coursework and applied clinical training that divert valuable time away from research. Gaining accreditation by APA and being competitive for internships are classic tail-wagsdog situations that can further dilute scientific efforts, reduce degrees of freedom in program design and individual curriculum choices, and add huge administrative overhead that further consumes time, energy, and resources. Clinical science is highly time-consuming. Intervention studies take time, recruiting patient populations takes time, coding behavior takes time, and using paradigms that don’t lend themselves to group testing takes time. Many clinical scientists end up moving toward using analog populations rather than “real” patients and mount relatively brief interventions with no or only brief follow-up. Other clinical scientists gravitate away from the study of clinical populations and interventions completely, instead studying basic psychological processes such as emotion, affiliation, and cognition in normal populations.
Chapter 13.qxd
3/28/2007
356
1:01 PM
Page 356
LEVENSON
Another recent development worthy of note is that some of the most exciting work in clinical science is now being done by psychologists in nonclinical areas. Thus, there are a growing number of neuroscientists who are studying autism, dementia, and other developmental disorders (Amaral, Bauman, & Schumann, 2003) and a number of social and personality psychologists who are studying gene by environment interactions in depression, addictions, and other forms of mental illness (Caspi et al., 2003). From the vantage point of clinical science, it is encouraging to see researchers from nonclinical areas of psychology taking on these important problems; however, it is also important that clinical scientists stay involved in this kind of cutting-edge research. Clinical scientists who drift away from research on clinical populations and nonclinical scientists who move into these areas are not problematical in and of themselves, especially if good science comes out of these efforts. However, to the extent that research and training choices of clinical scientists are being dictated in significant part by the competing demands associated with the mounting of traditional clinical psychology programs, it may be time to break this tie to the past. Clinical science programs may want to start anew and design themselves from the ground up in ways that conform to the ideals of the clinical science movement and the priority of producing important science that is relevant to mental health and illness. Training Clinical psychology in the coming decades is likely to see rapid changes in the realms of diagnosis, treatment, and etiology. It is difficult to argue against the general principle that we should be training clinical scientists who are master problem solvers rather than masters of the status quo. However, this still means striking a balance between training in the existing body of theories, methods, and techniques and training in how to develop new theories, new methods, and new techniques. Earlier sections of this chapter argue in favor of training in the new transdiagnostic, basic process-oriented approaches to diagnosis, in treatment development and evaluation, and in new neuroscience and genetics methodologies. Introducing this kind of training would clearly mean a reduction in emphasis on older theories, the DSM diagnostic system, and the current catalog of empirically supported treatments. As noted earlier, clinical science programs often find their training options constrained by past practices, by demands of accreditation, and by perceived demands of
Chapter 13.qxd
3/28/2007
1:01 PM
Page 357
13. THE FUTURE OF THE CLINICAL SCIENCE MOVEMENT
357
internship programs. In an ideal world, clinical science programs would be freed from these constraints and able to make pedagogical decisions solely in terms of what they believe would produce the best science and the best scientists. One aspect of pre-internship training that was once prominent and now seems to be in danger of being lost is exposure to a range of patients, especially to those with severe psychopathology. I recently conducted an informal poll of students in our clinical science program and learned that few had any significant exposure to schizophrenic patients prior to internship. Reasons for this are myriad, including sharp reductions in inpatient facilities, increasing training of clinical science students inhouse or in facilities primarily devoted to outpatient care, increased use of pharmacologic treatments, and the disappearance of live patient case conferences. Exposure to patients suffering from severe psychopathology early in students’ training can be a rich source of research hypotheses and can stimulate subsequent research. Moreover, students can learn how to work with families and mental health system gatekeepers to locate and recruit patients and can develop a comfort level with interacting with patients, both critical for patient-centric research. For many students, if the door into the world of patient research is not opened early, it will never be opened. In fact, I have argued recently that this kind of early exposure to patients should be part of the training of both clinical and nonclinical students (Levenson, 2004a, 2004b). If psychological science in general and clinical science in particular are going to stay relevant (and funded) in the years ahead, it is important that our students be trained in ways that makes them capable of and inclined toward conducting research with clinical populations (including those with the most severe psychopathologics). Science and Practice: Bridging the Gap or Increasing the Schism? One of the most sobering crises facing the clinical science movement is its relationship with clinical practice. In defining its own identity, the clinical science movement has often adopted harsh rhetoric criticizing current clinical practice, training in practice-oriented graduate programs, and the priorities of professional organizations that are devoted to the interests of practitioners. Practitioners have adopted their own harsh rhetoric about failures of clinical science to help real-world patients, the impracticality of manualized treatments, the superiority of clinical over
Chapter 13.qxd
358
3/28/2007
1:01 PM
Page 358
LEVENSON
empirical evidence, and the priorities of professional organizations devoted to the interests of clinical scientists. Clinical scientists and clinical practitioners have taken opposing positions on a number of critical issues, including accreditation, prescription privileges, licensing, and the direction and leadership of professional organizations. In reality, clinical practitioners and clinical scientists occupy overlapping niches, share common concerns, and work with many of the same issues. Clinical psychology will be severely diminished as a field and the mental health of the public will suffer if most of those who produce the science become estranged from most of those who proffer the treatments. Clearly there is a great deal of benefit to both groups if clinical practitioners respect and utilize the available science and if clinical scientists respect and utilize the experience that practitioners have working with clinical problems. There will always be tensions between scientists and practitioners, but in many other public health fields these groups are able to work together in much less adversarial ways. We need more collaboration between the groups rather than more isolation. There are ample opportunities to work together in research, in the application of research to practice, and in the training of scientists and practitioners. I think it is time for the clinical science movement to soften its rhetoric and take steps toward pursuing a rapprochement between these warring factions.
A CLOSING AND MORE PERSONAL THOUGHT It is a great honor to be able to contribute to this Festschrift volume, which recognizes the far-reaching contributions of Dick McFall to clinical science. Our paths crossed early in my career when Dick was courted from the University of Wisconsin to come to Indiana University and direct its clinical psychology program. Of course, he did far more than that, building what has arguably become the national model for a clinical science training program. During the years when we were colleagues at Indiana, we collaborated on research, tackled clinical program issues, and talked often about lives and careers. After I left for Berkeley, I found myself on many occasions attending meetings and serving on committees with Dick. Initially, many of these revolved around his tireless efforts to build the clinical science movement and the Academy of Psychological Clinical Science. Later, we often served together as representatives of the movement he had built. In the past decade, there have been significant changes in the field of clinical psychology, numerous
Chapter 13.qxd
3/28/2007
1:01 PM
Page 359
13. THE FUTURE OF THE CLINICAL SCIENCE MOVEMENT
359
crises have been averted, challenges met, and opportunities seized. Without Dick McFall’s personal integrity, force of intellect, clarity of vision, and unbounded energy it is hard to imagine any of this coming to pass. For me, it has been a great privilege and joy to know and work with Dick McFall over these many years and to be able to consider him a colleague and friend. REFERENCES Ahn, H., & Wampold, B. E. (2001). Where oh where are the specific ingredients? A meta-analysis of component studies in counseling and psychotherapy. Journal of Counseling Psychology, 48(3), 251–257. Amaral, D. G., Bauman, M. D., & Schumann, C. M. (2003). The amygdala and autism: implications from non-human primate studies. Genes , Brain, and Behavior, 2, 295–302. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders. Fourth edition (DSM–IV). Washington, DC: American Psychiatric Association. Caspi, A., Sugden, K., Moffitt, T. E., Taylor, A., Craig, I. W., Harrington, H., et al. (2003). Influence of life stress on depression: Moderation by a polymorphism in the 5-HTT gene. Science, 301(5631), 386–389. Chambless, D. L., & Hollon, S. D. (1998). Defining empirically supported therapies. Journal of Consulting & Clinical Psychology, 66(1), 7–18. Follette, W. C., & Houts, A. C. (1996). Models of scientific progress and the role of theory in taxonomy development: A case study of the DSM. Journal of Consulting & Clinical Psychology, 64(6), 1120–1132. Frank, J. D., & Frank, J. B. (1991). Persuasion and healing: A comparative study of psychotherapy (3rd ed.). Baltimore, MD: John Hopkins University Press. Goldberg, T. E., Egan, M. F., Gscheidle, T., Coppola, R., Weickert, T., Kolachana, B. S., et al. (2003). Executive subprocesses in working memory: Relationship to catecholO-methyltransferase Val158Met genotype and schizophrenia. Archives of General Psychiatry, 60(9), 889–896. Harvey, A. G. (2001). Insomnia: Symptom or diagnosis? Clinical Psychology Review, 21(7), 1037–1059. Harvey, A. G., Watkins, E., Mansell, W., & Shafran, R. (2004). Cognitive behavioural processes across psychological disorders: A transdiagnostic approach to research and treatment. New York: Oxford University Press. Hinshaw, S. P. (1987). On the distinction between attentional deficits/hyperactivity and conduct problems/aggression in child psychopathology. Psychological Bulletin, 101(3), 443–463. Levenson, R. W. (2004a). Patients and impatience. APS Observer, pp. 5, 44. Levenson, R. W. (2004b). Patients and impatience (Part II). APS Observer, pp. 5, 42. Levenson, R. W. (2005). Basic research funding: An exercise in NIH-ilism. APS Observer, 18(2). McFall, R. M. (1991). Manifesto for a science of clinical psychology. Clinical Psychologist, 44(6), 75–88.
Chapter 13.qxd
360
3/28/2007
1:01 PM
Page 360
LEVENSON
Nathan, P. E., & Langenbucher, J.W. (1999). Psychopathology: description and classification. Annual Review of Psychology, 50, 79–107. National Institute of Mental Health. (2000). Report of the National Advisory Mental Health Council’s Behavioral Science Workgroup: Translating behavioral science into action. Bethesda, MD: Author. Smith, M. L., & Glass, G. V. (1977). Meta-analysis of psychotherapy outcome studies. American Psychologist, 32(9), 752–760. Weisz, J. R., Donenberg, G. R., Han, S. S., & Weiss, B. (1995). Bridging the gap between laboratory and clinic in child and adolescent psychotherapy. Journal of Consulting and Clinical Psychology, 63, 688–701. Weisz, J. R., Doss, A. J., & Hawley, K. M. (2005). Youth psychotherapy outcome research: a review and critique of the evidence base. Annual Review of Psychology, 56, 337–363. Widiger, T. A., & Clark, L. A. (2000). Toward DSM–V and the classification of psychopathology. Psychological Bulletin, 126(6), 946–963. Widiger, T. A., Frances, A. J., Pincus, H. A., Davis, W. W., & First, M. B. (1991). Toward an empirical classification for the DSM–IV. Journal of Abnormal Psychology, 100(3), 280–288. Williams, J. B., Gibbon, M., First, M. B., Spitzer, R.L., Davies, M., Borus, J., et al. (1992). The Structured Clinical Interview for DSM–III–R (SCID). II. Multisite test–retest reliability. Archives of General Psychiatry, 49(8), 630–636. Wolpe, J. (1958). Psychotherapy by reciprocal inhibition. Stanford, CA: Stanford University Press.
Chapter 14.qxd
3/28/2007
1:01 PM
Page 361
V PERSPECTIVES ON PSYCHOLOGICAL CLINICAL SCIENCE
Chapter 14.qxd
3/28/2007
1:01 PM
Page 362
Chapter 14.qxd
3/28/2007
1:01 PM
Page 363
14 On Psychological Clinical Science Richard M. McFall Indiana University–Bloomington
Writing this festschrift chapter was a special opportunity, not to be squandered. I might have offered an overview of my research, but the chapters by Joe Steinmetz and Teresa Treat already touched on some of the latest work. So instead I’ve written a conceptual chapter focused on the principles and passions at the core of my professional journey thus far. My greatest rewards as a researcher have come not from collecting data that proved I had guessed right, but from following up on unexpected results that forced me to think more deeply about things, and that led to discoveries beyond the horizons of anything I had imagined. Similarly, my most gratifying moments as a professor have come not from teaching students facts, but from sharing conceptual perspectives that helped the students look at problems more critically, seeing things anew in a more exciting and productive light. My greatest satisfactions as a clinician have come less from taking responsibility for the way people changed than from knowing that I had acted responsibly, tempering my eagerness to be helpful with a discipline that forced me to rely on the scientific literature, to follow the evidence, and to move cautiously to avoid falling prey to the illusions that can lead to harming people instead of helping them. As an administrator, my greatest joys have come not from wielding influence, but from teaming with others to develop a vision and then facilitating efforts to advance clinical psychology as a science. In all these spheres, 363
Chapter 14.qxd
3/28/2007
1:01 PM
Page 364
364
MCFALL
my goal has been to perform in a manner, and at a level, that would allow my family, students, colleagues, and mentors—the most influential being Julian B. Rotter, George A. Kelly, and Peter J. Lang—to conclude that I deserve the label psychological clinical scientist. This chapter’s title is ambiguous: On could mean either about or onward. Both are intended here. The chapter is both an exploration of psychological clinical science and a call to action (as in the fight song, “On Wisconsin”). The chapter looks at psychological clinical science in three time frames: present, past, and future. The Academy of Psychological Clinical Science (APCS, or the Academy) is the centerpiece of the first section. As the Academy’s founding president, I am proud of my association with the organization; however, it is featured here because it is a champion for the ideals and principles of psychological clinical science. The second section looks at the history of how we got to our present situation. History is not science; it is storytelling, no matter how many references are attached. I tell the story as I know and experienced it, keeping references to a minimum, but providing pointers for any who might want to pursue things on their own. The final section involves a bit of crystal ball gazing. I sketch what I see as the two most challenging frontiers facing psychological clinical science and discuss their implications for the future. TODAY In May 2005, the Academy celebrated the 10th anniversary of its founding. At this significant milestone, the Academy, which is a voluntary association of U.S. and Canadian doctoral training programs in clinical and health psychology, had 54 members—45 university-based PhD training programs and 9 predoctoral internship training programs. All member programs share a commitment to training clinical scientists in psychology, as proclaimed in the Academy’s formal mission statement (APCS, n.d.; for more information, including a list of APCS member programs, go to http://www.psychclinicalscience.org): The Academy’s broad mission is to advance clinical science. “Clinical science” is defined as a psychological science directed at the promotion of adaptive functioning; at the assessment, understanding, amelioration, and prevention of human problems in behavior, affect, cognition or health; and at the application of knowledge in ways consistent with scientific evidence. The Academy’s emphasis on the term “science” underscores its commitment to empirical
Chapter 14.qxd
3/28/2007
1:01 PM
Page 365
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
365
approaches to evaluating the validity and utility of testable hypotheses and to advancing knowledge by this method. The Academy sees the development and application of clinical science as ongoing and dynamic processes, and is committed to facilitating the evolution of clinical science. Toward that end, it has established five specific goals: Training: To foster the training of students for careers in clinical science research, who skillfully will produce and apply scientific knowledge. Research and Theory: To advance the full range of clinical science research and theory and their integration with other relevant sciences. Resources and Opportunities: To foster the development of, and access to, resources and opportunities for training, research, funding, and careers in clinical science. Application: To foster the broad application of clinical science to human problems in responsible and innovative ways. Dissemination: To foster the timely dissemination of clinical science to policymaking groups, psychologists and other scientists, practitioners, and consumers.
Academy members are programs, not individuals. Membership is open to any interested research-oriented, university-based doctoral program or internship training program that shares the Academy’s values, mission, and goals. Applications from nontraditional clinical programs, such as those with an emphasis on health, child, personality, or experimental psychopathology, not only are welcome, but are encouraged. To become an Academy member, interested applicant programs must go through an anonymous peer review similar to a study section review. Applicants provide a narrative description of their principles and philosophy, faculty, students, curriculum, resources, achievements, and future directions. This is supplemented by documentation: the program’s brochure, curriculum vitae of the faculty, lists of student publications and presentations, lists of program graduates and job placements, and syllabi for core courses. Because there are multiple ways to train scientists successfully, applicants are given latitude to develop their own system of training. Rather than judging programs against a fixed checklist of expected features, reviewers evaluate programs on the basis of their overall quality, coherence, integrity, truth in advertising, and documented results.
Chapter 14.qxd
366
3/28/2007
1:01 PM
Page 366
MCFALL
Reviewers focus on a program’s record of training productive clinical scientists, and its commitment to critical self-examination and continuous quality improvement. Academy membership is not a form of accreditation; nor is it an avenue to credentialing, licensure, or practice. Instead, acceptance into membership is like earning the Good Housekeeping Seal of Approval. Membership status does not mean that Academy programs have no room for improvement. On the contrary, psychological clinical science is an abstract ideal, a lofty aspiration, which is represented only imperfectly by any specific program. Moreover, clinical science is a moving target— dynamic and evolving—so that even the best training programs must evolve to keep pace in their pursuit of this ideal. Thus, Academy programs are expected to engage in critical self-assessment and continuous quality improvement, and are expected to undergo periodic re-review with the bar raised over time. Of course, Academy membership is voluntary. For whatever reason, some high-quality programs have not yet applied. As a result, it is not reasonable to infer that the training offered by a nonmember program necessarily is inferior to that of member programs; however, because Academy members have gone through a peer review process, this provides some assurance that the training offered by these programs is consistent with the clinical science model and that it is reasonably effective at preparing students for careers as clinical scientists. Academy membership represents more than a seal of approval for the individual programs, however. It also represents a pact among member programs to work together to advance clinical science and to improve the training of clinical scientists. The Academy’s use of the labels clinical science and clinical scientist understandably has provoked a number of questions. These labels rarely were encountered prior to the Academy’s founding. Now they are used regularly by Academy-type programs, by the National Institute of Mental Health (NIMH) and other granting agencies, and by the field generally. For example, the Committee on Accreditation (CoA), the accrediting body for training programs in professional psychology, now recognizes clinical science as one of the legitimate training models for applicant programs. But what exactly does the clinical science label signify? How do training programs using this label differ, if at all, from the vast majority of other clinical training programs using other labels? The Boulder Model of clinical training, for instance, which has dominated clinical training for
Chapter 14.qxd
3/28/2007
1:01 PM
Page 367
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
367
over 50 years (see special section, “Boulder at 50,” edited by Benjamin & Baker, 2000), always has held that doctoral students should be trained both as scientists and as practitioners, with research training being an essential component. Many Boulder Model programs can point with pride to graduates with productive research careers. So what is the difference between the clinical science model and the Boulder Model? Does the clinical science label reflect a meaningful distinction, or is it little more than marketing hype, or perhaps elitist one-upmanship? In truth, the clinical science label is used, in part, for marketing purposes, although not in a misleading or self-serving way. The label helps draw attention to, reinforce, and promote the scientific values that have served for a century as the keystone for doctoral training in all areas of psychology. These values were acknowledged in the Boulder Model, following World War II, and still are represented, to one degree or another, in most PhD clinical training programs. However, the Academy believes that these values have been eroded and compromised in clinical training over the years by a variety of forces, as will be discussed shortly, and that it is critical to the future of clinical psychology that they be restored to preeminence. By shining a spotlight on them, with this label, the Academy is advocating a rededication, without compromise, to the ideal of building clinical psychology as a science. Programs that use the clinical science label are committing themselves to make the training of scientific researchers their primary and overriding mission; in addition, for reasons that are explained soon, clinical science programs also have deemphasized practitioner training. Thus, clinical science and scientistpractitioner programs now differ in their training focus. The Academy’s role in all of this is to provide the research-training programs with an association through which to work cooperatively to improve the quality of training, to create a nurturing environment for research and theory, to foster the dissemination and application of knowledge, and to exert a positive influence on the future of the discipline. Thus, the Academy is not a self-serving, elitist club. Naturally, it works to enhance the training, resources, and impact of its member programs, but its overarching mission is to elevate the quality of training, research, and practice across the entire discipline. The application process, for example, is not designed to keep programs out; instead, it is designed to encourage all interested programs to join in striving for excellence in research training. If an applicant program falls short in the peer review process, the program is not merely rejected and sent away; rather, it is
Chapter 14.qxd
368
3/28/2007
1:01 PM
Page 368
MCFALL
given clear, constructive feedback about how it might work to gain membership. The Academy has no interest in denigrating nonmember programs or other training models. However, because it aspires to promote the highest ideals of scientific training in clinical psychology and is dissatisfied with anything less, it often takes positions that challenge the status quo. These principled stances sometimes offend others, but this is an unintended and unfortunate by-product of advocating for psychological clinical science. What, specifically, is psychological clinical science? To understand this concatenated concept, it helps to examine the individual parts and then reassemble them into a coherent whole. Psychological qualifies clinical, which, in turn, qualifies science. In other words, the concept refers to taking a scientific approach to the investigation of psychological phenomena encountered in a clinical context. As proclaimed in the Academy’s mission statement, the science is “directed at the promotion of adaptive functioning; at the assessment, understanding, amelioration, and prevention of human problems in behavior, affect, cognition or health; and at the application of knowledge in ways consistent with scientific evidence” (APCS, n.d.). Problems are viewed from a psychological perspective, broadly defined, which includes biological, cognitive, developmental, genetic, learning, neural, sensory, social, and all other facets that contribute to variance in human behavior, adaptation, and health. Ideally, any barriers that have been separating these sub-areas will be bridged, and all relevant scientific knowledge from psychology and beyond will be incorporated as well. The anchor for the full concept, of course, is the noun science. Science is only one of many epistemological systems—ways of knowing—that humans have devised and used throughout history (Boorstin, 1983). Some others are authority, revelation, insight, intuition, inspiration, inductive and deductive logical reasoning, innate knowledge, sensory or extra-sensory experience, and magic. Perhaps the most common approach historically has been the use of force (might makes right; truth is on the winner’s side). Although science is not the oldest or most common epistemological approach in the history of human thought, its productivity and influence, particularly over the past two hundred years, has been profound (Gigerenzer et al., 1989). It has led to revolutionary advances in virtually every field of inquiry where it has been applied— from A (agronomy) to Z (zoology). Science now is embraced throughout much of the world as the gold standard for evaluating and choosing
Chapter 14.qxd
3/28/2007
1:01 PM
Page 369
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
369
among conflicting claims to knowledge, and for generating new insights, discoveries, solutions, and inventions. What specific features define a scientific epistemology, distinguishing it from pseudo-science or other approaches to knowing? Philosophers of science have written volumes on this topic, yet after centuries of analysis and debate, there still is no consensus on the definitive criteria for science. Philosophers generally can agree that certain exemplars represent either science or nonscience, and among the exemplars of science, they can distinguish between good and bad science. Nevertheless, they cannot agree on the specific features that allow them to distinguish among these classes. There seem to be exceptions to every rule. Thus, although some philosophical analyses have been influential (e.g., Feyerabend, 1981; Kuhn, 1970; Popper, 1962), debates about the exact nature of science have been of limited value to scientific progress. Interestingly, philosophical debates about science are not, themselves, science. Speculations about the essence of science rarely lend themselves to scientific tests, which mean that the debates never may be resolved. Fortunately, scientists have not been deterred by the debates; they simply have proceeded with the business of making fresh inroads into the unknown, discovering new solutions to old puzzles, and creating revolutionary inventions. Scientists understand that equating science with a finite set of rules, methods, or techniques is like equating art with paintby-number kits. They also understand that science is a way of thinking, not limited to any particular content area, problem, or theory; that science will take different forms, depending on the particular content area and problem being addressed; and that the category of science represents a fuzzy set—a multidimensional space populated by clear-cut prototypes at its center, with ambiguous and less prototypic examples at its margins. Science may be abstract and fuzzy, but history has shown that it works. Scientists have used this way of thinking to advance knowledge, despite its intrinsic ambiguities. In my experience on grant panels and editorial boards, scientists generally are able to judge the scientific merits of grant applications and journal manuscripts with high levels of interjudge agreement. This suggests that scientists know good science when they see it. Technically, it would be more accurate to say that they know bad science when they see it. Essentially, science takes a negative approach to finding truth. This is because, in humans, the unaided processing of information simply is too fallible, too prone to error and bias (Kahneman, Slovic, & Tversky, 1982).
Chapter 14.qxd
370
3/28/2007
1:01 PM
Page 370
MCFALL
The scientific approach, with its emphasis on caution, skepticism, and critical thinking, is an orthotic for fallible information processing, created specifically to protect humans—especially scientists—from fooling themselves. Good science involves the systematic collection, examination, and interpretation of empirical evidence in ways that are as error-free as possible. To minimize errors, science takes an indirect approach to advancing knowledge. Rather than setting out to prove that ideas or hypotheses are true—as humans, with their confirmation bias, naturally are inclined to do—a scientific approach seeks instead to disprove ideas. The goal is to eliminate plausible rival hypotheses from further consideration by exposing their errors (Campbell & Stanley, 1963). Science starts by entertaining all plausible possibilities, then systematically seeks to ferret out and eliminate the erroneous possibilities, gradually whittling the initial list to those remaining candidates that have resisted elimination. Only those ideas that have survived rigorous efforts to disprove them are credited as having validity, or as possibly being true. Logically, no idea ever can be declared the absolute, final truth, however, because no matter how often an idea has survived critical tests, it always is possible that future tests may reveal it to be erroneous and invalid. Viewed from this perspective, bad science is any search for truth that is flawed by being insufficiently cautious, skeptical, disciplined, and critical to protect us from fooling ourselves (e.g., see Feynman, 1985, on cargo cult science). Scientific ideas (explanations, hypotheses, theories) can come from any source, not just from scientists. However, if there is no conceivable way that an idea can be subjected to critical tests capable of revealing its errors and disproving it, then to take such an idea seriously is an example of bad science. For a theory to be taken seriously in science, its predictions must be tied to events that are empirically verifiable; the events should be publicly observable, measurable, reliable, and replicable. Furthermore, a scientific theory should make risky predictions. That is, the predictions should extend beyond the events that gave rise to the theory in the first place; they should have utility beyond what could have been achieved without the theory; and their predictive power should exceed that of other, simpler theories. In short, theories that offer post hoc explanations, lack incremental validity, or are not parsimonious are considered bad scientific theories (Meehl, 1978; Sechrest, 1963). The ultimate goal of all science is to inform. Information, in turn, is defined as the reduction of uncertainty (Pierce, 1980; Shannon & Weaver, 1949).
Chapter 14.qxd
3/28/2007
1:01 PM
Page 371
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
371
These key concepts of information and uncertainty can be measured. Thus, the scientific value of theories can be assessed quantitatively, as well as qualitatively (see Macmillan & Creelman, 2005; McFall, 2006; McFall & Treat, 1999). You may be wondering, at this point, whether this elementary discussion of the general characteristics of a scientific approach really is necessary. Aren’t such basics taught routinely in high school and college science classes? At the graduate level, for example, haven’t these ways of knowing served as the cornerstone for psychological training, in general, and for the Boulder Model of training in clinical psychology, in particular, for half a century or more? Don’t professional clinical psychologists, most of whom consider themselves to be scientists as well as practitioners, appreciate and embrace these very basic empirical foundations of psychological clinical science? Unfortunately, the centrality of science in clinical psychological training and practice cannot be taken for granted. The preeminent role of science in training has been eroding slowly for many years, and in the last two decades it has come under direct attack. Indeed, the Academy of Psychological Clinical Science was created, in large part, to counteract and reverse these alarming historical trends. YESTERDAY The seeds for these developments were planted following World War II when doctoral training in clinical psychology was transformed from its long-standing focus on training researchers into a new enterprise focused on training scientist-practitioners (see Benjamin, 2005). This historic shift led to the gradual growth of all the trappings that go with training health care professionals—the preoccupation with accumulating practicum and internship experiences; the introduction of accreditation and licensing, with their homogenizing influence and their standardized, litigationavoidant, detailed requirements and procedures; the proliferation of specialties and subspecialties, with boards, certificates, and turf battles; the imposition of continuing education requirements; and guild-related battles over reimbursement parity, managed care, and prescription privileges. Even before World War II, Woodworth (1937) had warned the field against starting down the slippery slope of professional training in clinical psychology. In retrospect, his concerns were prescient (Sechrest, 1992). Translating the Boulder Model’s idealistic vision of dually trained scientist-practitioners into viable training programs required that psychology
Chapter 14.qxd
372
3/28/2007
1:01 PM
Page 372
MCFALL
departments make a number of unforeseen and unfortunate compromises. Even the most research-oriented clinical training programs soon found that the demands of preparing their students for professional practice undermined their efforts to train these students as research scientists. At nearly every turn, the demands of professional training seemed to trump the competing, less glamorous, more challenging demands of science training. But once the Boulder Model train had left the station, there was no turning back, regardless of the unexpected costs and consequences. There were too many seductive forces at work. The PhD in clinical psychology, with its professional aura, quickly became enormously popular, attracting the very best and brightest undergraduate applicants. Who could resist working with some of the best doctoral students in the university, even if they weren’t committed to research careers? And the resources thrown into practitioner training were compelling, too. Who could resist harvesting the cornucopia of funding for practitioner training that flowed from the Veterans Administration, the U.S. Public Health Service, and in-house clinics and externships? The job market for applied clinical psychologists seemed inexhaustible, as well. Who could turn their backs on the national campaign to train practitioners to serve the needs of the mentally ill, especially in light of the federal government’s campaign, started in the 1960s, to build a national network of community mental health centers? Indeed, this golden era of clinical psychology— extending from the late 1940s into the early 1980s—was a heady time. The future of applied clinical training seemed unlimited. The number of clinical training programs mushroomed as universities rushed to get a piece of the action. Extrapolating from the growth rate during the 1960s in the number of PhDs from clinical psychology training programs, it was projected, in jest, that by the middle of the 21st century everyone in the United States would have a PhD in clinical psychology. Beneath the surface of this rosy picture, all was not well, however. In the midst of the euphoric expansion of practitioner training, basic science training in clinical psychology was suffering—although the signs were subtle. For instance, even as clinical graduate students continued to take their substantive science courses, to learn applied statistics, and to satisfy their PhD research requirements, the modal number of research publications by clinical PhDs fell to zero! Entering clinical students increasingly were more interested in private practice careers than in research careers. It was not uncommon for academic counselors to advise undergraduate students to apply to graduate programs in experimental psychology if they
Chapter 14.qxd
3/28/2007
1:01 PM
Page 373
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
373
liked research, but to apply to clinical programs if they didn’t. Given the highly competitive admissions process, clinical applicants were smart enough to feign research interests on their applications, but once in graduate school, they tended to treat the research requirements as speed bumps on the road to earning the credentials for practice careers. The occasional students who initially may have been interested in research careers often found their aspirations swamped by the professional training requirements. The tensions between science and practice were exacerbated by the fact that the science and practice components of clinical training typically were poorly integrated. Too often, students encountered faculty role models who talked the talk of integration, but failed to walk the walk. Some faculty members were two-headed clinicians, modeling scientific rigor in their laboratories and classrooms, but checking their scientific rigor and critical thinking at the therapy room door. Other faculty members served as models of high-powered researchers, but tended to consider practicum training beneath them, happily farming out practicum supervision to clinicians at externships and internship sites, or to adjunct faculty members. These supervisors, in turn, often tended to sympathize with students’ distaste for the science requirements, and to undermine in various ways the program’s emphasis on science training. Students found the softer, more intuitive epistemology modeled by their clinical supervisors to be less demanding and stressful than the epistemology modeled by their research advisors. Such experiences convinced many students that the hyphen in the scientist-practitioner model—a hyphen initially meant to symbolize integration—actually was a separator, symbolizing the schism between science and practice that the students had encountered personally. The disconnect between science and practice probably was unavoidable in the early 1950s. Although the Boulder Model’s blueprint called for an applied profession built on a rock-solid foundation of scientific evidence, no such foundation had been laid. Researchers could offer practitioners few solutions to the puzzles they faced daily in the clinic. Yet, to be credible as professionals, clinicians on the firing line felt that they could not sit on their hands waiting for the slow progress of science to deliver the knowledge and tools they needed. Real people with real problems expected action—now!—even if it meant improvising. Stop-gap clinical practices, born of this necessity, soon took on lives of their own. These practices became functionally autonomous of their origins (Astin, 1961).
Chapter 14.qxd
374
3/28/2007
1:01 PM
Page 374
MCFALL
As they were passed from one generation of clinicians to the next, they became gospel. Skeptics who questioned clinicians’ reliance on unproven assessments (e.g., Meehl, 1954, 1973) and interventions (e.g., Schofield, 1964) simply were dismissed as ivory tower academics, out of touch with reality. The gulf between science and practice widened. Early practitioners might have been excused for dismissing researchers’ concerns about the validity of their methods, given that the skeptics offered few positive alternatives. This picture began to change, however, as the profession approached the 1960s. In 1957, Carl Rogers announced a revolutionary research program in which he would violate the sanctity of the psychotherapy session by recording and analyzing the interactions of therapists and clients in client-centered psychotherapy, thus daring to test empirically his theories about the necessary and sufficient conditions for psychotherapeutic change. In 1960, Peter Lang and his colleagues developed the first controlled, laboratory paradigm for evaluating the effectiveness of systematic desensitization therapy (Lazovik & Lang, 1960). In 1961, Albert Bandura published a Psychological Bulletin review of the extant research on social learning approaches to psychotherapy; it filled 17 pages. From there, things took off. In the mid 1960s, the American Psychological Association published a three-volume series on psychotherapy research. In 1968, two groundbreaking broadsides—one by Walter Mischel, the other by Donald Peterson—challenged traditional psychological assessment. In 1969, Cyril Franks edited an influential book on behavior therapy, which included a seminal paper by Gordon Paul (1969). Paul criticized the common tendency to regard all psychotherapies, patients, problems, therapists, and measures as though they were interchangeable (i.e., the logical error that Kiesler, 1966, had dubbed the uniformity myth). Paul urged researchers to ask specific questions: “What treatment, by whom, is most effective for this individual with that specific problem, under which set of circumstances, and how does it come about?” (p. 44). By the end of the decade, there was enough material for Bandura (1969) to fill a 677-page book on Principles of Behavior Modification. If anything, the pace of psychotherapy research accelerated during the 1970s. By the early 1980s, practitioners no longer could blame their use of untested and unsupported methods on a lack of research foundations. Prior to the 1980s, practitioners rarely disparaged the value of science—at least not openly; their identification with the scientist half of the scientist-practitioner model gave them bragging rights over their
Chapter 14.qxd
3/28/2007
1:01 PM
Page 375
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
375
mental health colleagues from psychiatry and social work. However, as the body of research challenging their methods grew, practitioners faced a dilemma. Either they had to bring their practices into line with the research evidence or they had to find some way to discount the evidence. Many did not want to change what they were doing. It was at this point that the broad consensus among psychologists started to unravel. A growing number of practice-oriented clinical psychologists openly suggested that the conventional scientific epistemology of psychological researchers was flawed, irrelevant, even bankrupt. They argued that the evidence from the personal experiences of practicing psychologists was more valid and relevant than the evidence from laboratory experiments and controlled clinical trials. Essentially, they were advocating that clinical psychologists—and the programs that trained them—adopt a new epistemology! By challenging the relevance of the scientific evidence, in this way, they were free to continue with business as usual. At the same time, the hegemony of the scientist-practitioner model was being challenged on another front. The doctor of psychology (PsyD) movement, which had begun as a mere trickle, was becoming a significant new force in clinical psychology. It began with the University of Illinois’ experimental offering of a PsyD degree designed to train graduate students exclusively for careers as clinical practitioners. The idea was that if students were freed from the traditional research requirements of the PhD degree, they could devote more time to acquiring the knowledge and skills that would make them better practitioners. PsyD students were expected to be informed consumers of psychological research, but not to produce research. Although Illinois soon terminated its short-lived PsyD program, it had provided a precedent that other universities quickly followed. Before long, PsyD programs were being offered both by universities and by free-standing professional schools. Compared to the traditional, university-based scientist-practitioner programs represented by the Council of University Directors of Clinical Psychology (CUDCP), these new PsyD programs seemed to have problems on a number of dimensions. For example, on average, they admitted much larger classes of students, had lower admission standards, had fewer full-time faculty, had higher student–faculty ratios, placed a lower percentage of their graduates, left their graduates with larger student-loan debts, and produced students who earned lower scores on the national licensing exam for psychologists (e.g., Cherry, Messenger, & Jacoby, 2000; Maher, 1999; Norcross, Castle, Sayette, & Mayne, 2004; Peterson, 2003).
Chapter 14.qxd
376
3/28/2007
1:01 PM
Page 376
MCFALL
Nevertheless, these programs soon were producing a disproportionate share of all the doctorates in clinical psychology, and their graduates increasingly were exerting political influence on APA and licensing boards, affecting the directions being taken in the field. Because the new PsyD programs were training pure practitioners, as opposed to scientist-practitioners, the programs were free to challenge the conventional scientific epistemology of Boulder Model programs, and to advocate for a new epistemology that might support many of their practices. Some of their criticisms were reminiscent of philosophical arguments designed to persuade unwary listeners, through faulty logic, that an arrow shot at a target never will reach that target if the remaining distances repeatedly are divided into infinitely smaller halves. For example, the critics pointed to unsolved clinical problems as proof that conventional science is incapable of solving such problems and, therefore, that a new approach is required. To prove that the analytic methods of conventional science are inferior to a more intuitive, idiographic approach to clinical problems, critics pointed to comorbidity and other multidimensional problems, and asserted that such complexity defied the powers of conventional science. To prove that evidence from randomized clinical trials is irrelevant to clinical practice, critics pointed to contextual differences between clinical trials and real-world clinical settings, and declared that the findings from efficacy research (i.e., clinical trials using traditional methods) cannot be generalized to practice settings; only the findings of effectiveness research (i.e., naturalistic studies relying on a new epistemology) are valid. Essentially, critics concluded that conventional science had not covered the entire distance to its target, so it never would get there. Moreover, they argued that scientific efforts aimed at establishing empirically supported treatments miss the real target entirely. (For a recent example of this kind of perspective, see Westen, Novotny, & Thompson-Brenner, 2004; also see replies by Crits-Christoph, Wilson, & Hollon, 2005; Weisz, Weersing, & Henggeler, 2005; Westen, Novotny, & Thompson-Brenner, 2005.) Critics also offered some pragmatic warnings about the potential dangers of relying on conventional science in psychology (e.g., see Peterson, 1996). For instance, they argued that scientific skepticism about the validity of common clinical practices undermines public confidence in the profession, thereby diminishing the effectiveness of the services. (Translation: Science harms patients.) In addition, if clinical psychologists were limited to empirically supported procedures, they would have
Chapter 14.qxd
3/28/2007
1:01 PM
Page 377
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
377
little to offer; other professions would fill the void, leaving psychology out in the cold. (Translation: Science is not competitive.) If clinical practice were restricted to methods that meet scientific standards, this would put practitioners—and the programs that train them—out of business. (Translation: Science dooms psychology.) One of the favorite candidates to replace the traditional science model is the local clinical science (LCS) model (sometimes referred to more generically as the scholar-practitioner model), which also is recognized by CoA as a legitimate training model for accreditation.1 The LCS model’s reference to science appears to be a case of scientism (Christian Science and Scientology are examples of scientism in another realm), where the term is used to evoke all of the positive associations normally linked to science, but actually does not refer to anything close to ordinary science. The LCS movement in psychology was inspired, in part, by Donald A. Schön’s (1983) book, The Reflective Practitioner: How Professionals Think in Action. Schön was highly critical of contemporary science, in general. He focused on psychotherapy, in one chapter, to illustrate his points. He argued that every patient is unique, a universe of one, which means that the skillful practitioner cannot apply predetermined nomothetic analyses and solutions, but must treat each case as an unique local problem requiring an idiographic solution. A number of PsyD training programs have incorporated Schön’s critical analysis and epistemological philosophy as a cornerstone for their training and practice. (For more on these programs—their model, history, rationale, and national association—see, for example, Peterson, Peterson, Abrams, & Stricker, 1997; Peterson & Trierweiler, 1995; Stricker & Trierweiler, 1999.) A detailed summary and critique of the LCS model is beyond the scope of this chapter. However, in a nutshell, the approach teaches clinicians to start from scratch, with no preconceptions, when formulating and treating each case. One might wonder, if clinicians were to do this, what special attributes they would bring to a case that would justify their charging for their services. They claim to employ an N = 1 approach to formulating and testing their unique theories for individual cases. This abstract idea may have emotional appeal, but it does not address the concerns 1 Even as this chapter was being finalized, a Summit Meeting on Accreditation was held at Snowbird, UT, July 23-25, 2005. The meeting, which I attended, was devoted to discussions among representatives from all interested constituent groups regarding the future structure and process of accreditation in professional psychology. Although it seems virtually certain that the accreditation system will be revised somehow, the final structure still is unclear.
Chapter 14.qxd
378
3/28/2007
1:01 PM
Page 378
MCFALL
that gnaw at skeptics. For example, if a clinician starts from scratch, where do the clinician’s unique theories come from? Schön suggested, metaphorically, that the theories emerge from a reflective conversation with the situation. Without relying on nomothetic generalizations, clinicians draw reflectively from their experience, skill, and expertise to grasp the truth of the unique case. But what is experience, after all, if not abstracting nomothetic generalizations from unique events? A major problem with relying on reflections and personal experience, of course, is that if different experts arrive at different truths for a given case, based on their different reflections and personal experiences, then we are back to square one. That is, how do we decide among these competing truths? Even when experts agree, these consensual conclusions may not be true. History is full of examples of consensual views that proved false or even harmful. On closer inspection, the LCS model appears to be old wine in new bottles, relying on ways of knowing, such as personal experience and intuitive judgment, that have been tried and found wanting in the past (e.g., Chapman & Chapman, 1969; Dawes, Faust, & Meehl, 1989; Garb, 2005; Grove & Meehl, 1996). When LCS practitioners formulate idiographic theories for individual patients, we need objective methods of determining whether these theories are valid. At a minimum, reliability is a necessary-but-not-sufficient condition for validity, meaning that multiple experts should arrive independently at the same formulation for a given case. What is the reliability of LCS clinicians’ judgments and formulations? Of course, reliability is not validity. Even if LCS clinicians agree, we still need to rule out the possibility that they are making common errors in judgment, falling prey to pervasive errors in human information processing. For clinicians to avoid fooling themselves and others, they must put their theories and methods to the kinds of rigorous, objective, empirical tests required by a conventional scientific epistemology. If LCS advocates reject conventional scientific standards of reliability and validity, then what substitute methods can they offer instead for deciding which ideas or decisions are valid? Essentially, the LCS model makes implicit and explicit claims about its ability to outperform traditional scientific approaches in the assessment and treatment of clinical cases. Such claims can be evaluated empirically by traditional scientific methods. For example, by keeping outcome records for individual therapists and clients, it is possible to compare the effectiveness of services based on the LCS model with the effectiveness
Chapter 14.qxd
3/28/2007
1:01 PM
Page 379
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
379
of other services. For any claims to be taken seriously, we need empirical evidence of this sort. Until LCS advocates provide compelling support for their model, their claims of superiority remain unsupported claims. Ordinarily, the onus for establishing the validity of such claims falls on the shoulders of the advocates; skeptics are free to ignore unsupported claims. However, given the growing number of graduates from PsyD programs trained in the LCS model who now are engaging in clinical practice, the model’s potential impact on people’s lives should cause advocates and skeptics alike to be concerned about the model’s validity and safety. The fundamental ethical question is this: Should clinical psychologists be providing services that have not been shown by empirical scientific evidence to be safe and effective? In 1971, Julian Rotter proposed that clinical psychology adopt Food and Drug Administration (FDA)-like standards, limiting psychological services to those that are safe and effective. He predicted that if psychologists failed to police themselves in this way, they eventually would find themselves regulated by outside agencies. Although Rotter’s proposal fell on deaf ears, by the early 1980s his warning about external policing proved to be prophetic. The U.S. health care system, staggering under the weight of escalating costs and public concerns about malpractice and quality control, began moving away from the traditional fee-for-service model toward a new managed-care model. This move was aimed primarily at controlling the costs of medical services; however, psychological services, which accounted for a relatively small share of the total health care costs, got caught in the same net as the bigger fish. The health care industry sought to establish guidelines for determining which clinicians should be reimbursed at what rates for providing what services to what patients with what problems. This move to managed care had several implications for clinical psychologists: First, it meant that some services might not be reimbursed. Second, it meant that there would be limits on the total amount of billable services. Third, because there was no good evidence that the treatment outcomes of social workers and psychologists differed, and because social workers provided services at a lower cost, it meant that psychologists would start to lose their market share and experience a decline in income. In short, the future for psychologists as providers of mental health services did not look bright. At the start of the 1980s, the federal government saw another way to cut costs. Given the proliferation of clinical training programs and the
Chapter 14.qxd
380
3/28/2007
1:01 PM
Page 380
MCFALL
superabundance of and maldistribution of practicing psychologists, the government decided that it no longer made sense to provide grant support for the predoctoral training of clinical psychologists going into practice careers. Suddenly, the sun seemed to be setting on the quarter century of psychology’s golden era of clinical training and practice. Clinical psychologists reacted to these changes in two general ways. Some psychologists, typically those engaged in practice, saw these changes as threats to their professional survival and began scrambling to higher ground, not only engaging in rear-guard attacks on managed care, but also looking for ways to reinvent clinical psychology. Among other things, they launched the American Psychological Association Practice Organization, a lobbying organization technically independent from APA but funded by special assessments on all clinical psychologist members of APA. They also campaigned for legislation to grant prescription privileges to psychologists, pushed for higher level credentialing of practice specialties in psychology aimed at holding competitors at bay, and explored a wide range of potentially lucrative new roles for psychologists. Other psychologists, typically those involved primarily in research and training, were more inclined to welcome the changes in health care delivery, seeing them as long overdue and as providing an opportunity for psychology to strip away the professional trappings and other impediments to scientific progress that had weighed down the field since World War II. In their view, it was a chance for psychology to return to its basicscience roots, to get back on track. Thus, the field was being pulled in two very different directions, one concerned with income, the other with outcome (thanks to Russell Bauer for this framing, personal communication, August 28, 2004). Essentially, the erosion of resources exposed the underlying fault lines in psychology. After several minor rumbles in the early 1980s, a major rupture occurred in 1988 along those lines. A large block of academic and research psychologists, led by a group of clinical psychologists, broke away from the American Psychological Association (APA) to form the American Psychological Society (APS), an organization dedicated exclusively to building psychology as a science.2 APA and APS often have been characterized as symbolizing, respectively, the two poles of a practice versus science dichotomy in psychology. On closer inspection, this characterization is too simplistic and misleading. 2 As of January 1, 2006, APS officially changed its name to the Association for Psychological Science, in part as a way of emphasizing this dedication.
Chapter 14.qxd
3/28/2007
1:01 PM
Page 381
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
381
On one hand, increasing numbers of practitioners affiliated with APA are attending to the scientific evidence regarding the effectiveness of specific treatments for specific problems. So it is inaccurate to imply that all clinical practitioners are anti-science. In addition, offices within APA, especially the Science Directorate, have made and continue to make significant contributions to the advancement of science. Many of the leading science journals in psychology are published by APA. Unfortunately, some of these valuable contributions tend to be diluted by other APA activities. On the other hand, the psychologists who joined APS are not insensitive to the suffering of clinical patients; nor are they opposed to the delivery of effective clinical services to deal with such clinical problems. Members of the Academy, for instance, have an explicit commitment to “the assessment, understanding, amelioration, and prevention of human problems in behavior, affect, cognition or health” (APCS, n.d.). So it is inaccurate to characterize as anti-practice all psychologists who are pushing for clinical practices to be consistent with the best scientific evidence. The relationship between practice and science—or between APA and APS—cannot be represented accurately in a simple onedimensional diagram, with basic science anchoring one end of a continuum and applied practice anchoring the other end. The clinical science perspective advocated by the Academy requires at least a two-dimensional representation, with the vertical dimension anchored by science at the top and nonscience/pseudoscience at the bottom, and with the orthogonal horizon dimension anchored by basic at one end and applied at the other. For clinical scientists, the vertical dimension in this model (i.e., the science vs. nonscience dimension) is the crucial one. Psychological activities are legitimate as long as they are consistent with the available scientific evidence and contribute to the advancement of science. In principle, this means that all activities in the top half of this diagram— ranging from the most basic to the most applied—are examples of legitimate clinical science. We return to this issue in the next section. The preeminence of science was the central theme of my paper, “Manifesto for a Science of Clinical Psychology” (McFall, 1991; also see McFall, 1996, 2000; Peterson, 1996). The paper set forth one cardinal principle: “Scientific clinical psychology is the only legitimate and acceptable form of clinical psychology.” Then it outlined the implications of this principle in two corollaries. The first focused on clinical practice, asserting that psychological services should not be administered
Chapter 14.qxd
382
3/28/2007
1:01 PM
Page 382
MCFALL
to the public (except under strict experimental control) until their exact nature and claimed benefits had been made explicit and until they had been shown by scientific evidence to be both safe and effective. The second corollary focused on clinical training, asserting that the primary and overriding aim of doctoral training in clinical psychology is to train competent clinical scientists. Although the manifesto essentially reiterated and elaborated Rotter’s message from 20 years earlier, the message had an unexpected impact this time, arriving just as the winds of change were starting to blow through psychology. At about that time, for example, David Barlow, president of Division 12 (Clinical Psychology) of APA, had appointed a Task Force on the Promotion and Dissemination of Psychological Procedures, chaired by Dianne Chambless (1995), to review the research literature on clinical interventions and identify the treatments for which there was a reasonable body of empirical support (e.g., Chamblesset al., 1996, 1998; Chambless & Ollendick, 2001). Accountability was in the air and the manifesto added to the electrical charge in this atmosphere. Meanwhile, other events also were helping to focus attention on the second issue, as well—that is, on the aims and structures of doctoral training programs. Among other things, chairpersons of psychology departments were becoming frustrated by the control that the Committee on Accreditation (CoA) of the American Psychological Association was exerting over the design of the graduate curriculum in their departments. To be accredited, clinical programs were being forced to tailor their curriculum and training experiences to conform to the CoA’s detailed checklist of expectations and requirements. Chairs saw this external control as an infringement on academic freedom and departmental prerogatives. Clinical faculties also felt hamstrung by the requirements, many of which lacked good scientific foundations and tended to undermine efforts to offer high-quality research training. These frustrations led to a 1992 Summit on Accreditation in Chicago, sponsored by the Council of Graduate Departments of Psychology (COGDOP), APS, and NIMH. The summit was aimed at helping departments and clinical programs regain control over their curriculum. Two possibilities were explored: One possibility was to convince CoA to become more flexible; the other was to create an alternative to the CoA’s accreditation system. The summit led to the formation of the Alternative Accreditation Steering Committee (AASC), which was charged with designing a fully independent, alternative accreditation system. At the
Chapter 14.qxd
3/28/2007
1:01 PM
Page 383
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
383
same time, ongoing negotiations were taking place on a separate front, aimed at persuading the powers that be to modify CoA’s guidelines and procedures. By early 1994, as AASC was completing its plans for an alternative system, the parallel negotiations also were paying off, yielding significant changes in the CoA’s accreditation system. Thus, AASC decided to delay implementing its alternative system, giving the new CoA system a chance. AASC did not disband; rather, it went into recess, monitoring the success of the newly revised CoA system, standing ready to reconvene if the revised CoA system proved unsatisfactory. The summit had stimulated dialogue, attention to the issues, and excitement among the faculties of research-oriented clinical training programs. Capitalizing on this momentum, a conference on “Clinical Science in the 21st Century” was convened in April 1994 at Indiana University, with support from NIMH and APS. There was general consensus among the conference participants—representatives from 25 leading graduate programs in clinical or health psychology, plus one representative each from NIMH and APS—that the time was ripe for PhD programs to increase their emphasis on research training, and that the programs needed a mechanism for working together to strengthen this aspect of training. The conference formed a steering committee and charged it with creating a new organization of research-oriented graduate training programs in clinical and health psychology. By January 1995, this committee had created the Academy of Psychological Clinical Science, and had launched the bootstrap process by which applications for membership in the new organization would be solicited and reviewed. In July 1995, the first official meeting of APCS was held in New York City, in conjunction with the annual APS convention. In the Academy’s 10 years, it has made good progress toward the five goals set forth in its mission statement. Member programs have shared syllabi and exchanged information on the training of clinical scientists. Academy programs now are able to apply for CoA accreditation under the clinical science model. Since 1995, at least one representative from a clinical science program has served on CoA under the umbrella of CUDCP. Academy programs are able to select accreditation site visitors from peer programs, increasing the odds that the visitors will understand their training model and decreasing the pressures to conform to inappropriate criteria. In April 2002, the Academy joined with APS and NIMH in cosponsoring a conference at Indiana University on integrative psychological science, which explored integrative models of training.
Chapter 14.qxd
3/28/2007
1:01 PM
Page 384
384
MCFALL
In January 2004, NIMH and the Academy cosponsored a conference in Washington, DC, aimed at promoting integrative research training for clinical scientists. In general, the Academy seems to be serving as a magnet, attracting attention to the cause of rigorous scientific training in clinical psychology. The Academy has grown steadily, attracting new members from like-minded training programs. For several years now, nine internship programs have been Academy members. This has helped foster a continuity and integration of science training as students move from universities to internships to postdocs. It also has helped decrease students’ anxieties about accumulating practicum hours in order to get into a good internship. The Academy also has taken responsibility for helping to organize the clinical track of the APS annual convention program. Academy faculty members have been mutually supportive in their roles as reviewers on grant panels, journal editors, consultants on editorial boards, textbook authors, and so on. Although Academy programs represent only about one quarter of the clinical training programs in CUDCP, the faculties and students from Academy programs account for a significant share of the research published in the top psychological clinical science journals. The Academy has become an influential voice in the dissemination of scientific knowledge through its interactions with granting agencies, policymakers, prospective students, colleagues, and the public at large. The Academy has lived up to its founders’ dreams, having become a major force on behalf of science training in clinical and health psychology. As it enters its second decade, it is poised to assume an even larger role as the leading advocate for a new vision of clinical science. In the next section, I present what I see as the two most pressing issues facing the Academy over the next decade, and explore their implications for the future of psychological clinical science. TOMORROW
Integrity. Our first priority must be to do all that we can to strengthen the integrity of psychological clinical science. Of course, we all are familiar with the basic ground rules of integrity: Research must not be biased. Data collection, analysis, and interpretation must be done with rigor and care. Knowingly distorting, corrupting, misrepresenting, or misusing scientific evidence is unacceptable. We must tell what we don’t know, as well as what we think we do know. We must not make things
Chapter 14.qxd
3/28/2007
1:01 PM
Page 385
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
385
look more coherent and successful than they are, or sacrifice accuracy for the sake of telling a more interesting or impressive story. We must not confuse our hopes and speculations with evidence. We must follow the data, wherever they take us. Above all, we must avoid doing harm. In short, all actions should be consistent with, and guided by, the best scientific evidence and knowledge, and should satisfy the highest ethical standards. If we take such principles seriously, we no longer can justify continuing to do things the same old way. We must take a long, hard, skeptical look at all aspects of our training and practice to ensure that, at a minimum, these are consistent with the best scientific evidence. Where things are not consistent, integrity requires change. To illustrate this point, I’ll focus on just one of the articles that imply change. Bickman (1999) has challenged six widely accepted beliefs about clinical services, beliefs that drive much of our current training and practice. Specifically, he reviewed the empirical evidence for and against the beliefs that “effective services are assured by (a) more experienced clinicians, (b) degree programs, (c) continuing education, (d) licensing, (e) accreditation, and (f) clinical supervision” (p. 965). He concluded that these beliefs are myths, clearly unsupported by the empirical evidence. If we are serious about training with integrity, Bickman’s analysis requires that we make major changes in the designs of our clinical science training programs, bringing them into line with the evidence. Consider, for example, the current custom of using the quantity of practicum hours as an index of a student’s readiness to provide clinical services. This custom always has seemed illogical, akin to deciding when to grant someone a degree based on the number of hours the person sat in the library. But now, in light of the evidence reviewed by Bickman, this index simply is indefensible. If such a “readiness index” were needed (which is debatable, for reasons to be discussed in a moment), then at least the index should be valid, most likely based on performance samples. But Bickman’s evidence also indicates that we must drop the illusion that degree matters. We must abandon the idea that having a PhD degree makes someone a more effective service provider. Data contradicting this myth have been available for a good while (e.g., Berman & Norton, 1985). If persons with less training, say with an MA or BA, are as effective as PhD psychologists at providing services, and are able to do it at a lower cost, then shouldn’t we be championing these providers as the best choice for serving the interests of patients and society—and science—
Chapter 14.qxd
386
3/28/2007
1:01 PM
Page 386
MCFALL
rather than acting to protect psychology’s guild interests? Bickman’s review also exposes the lack of validity behind such professional trappings as continuing education, licensing, accreditation, and clinical supervision. Again, although the evidence has been available, psychologists have hesitated to pull back the scrim curtain and expose the myths. As a matter of integrity, however, we must have the courage to align our professional practices with the empirical evidence. Bickman’s analysis has implications for our choice of training models as well. For example, what is the justification for investing scarce resources in the training of PhD-level service providers, as in the scientist-practitioner training model, when the evidence indicates that such training has no incremental validity, and that it increases the cost of services? The implication seems clear: The PhD in clinical psychology should be a research degree, as it is in most other basic sciences. Research training should become our central focus, as it is in the clinical science model. Invariably, this line of reasoning leads to anxious questions: “Would this mean that clinical science students no longer have contact with clinical patients, conduct assessments, or learn to carry out interventions?” Again, the answer is dictated by integrity: “It depends.” What is the rationale for the specific training component? Is there evidence that it actually advances the training mission? Undoubtedly, investigators need to have a solid grasp of their research problems in order to make valuable scientific contributions to solving them. The Academy’s mission statement explicitly says that clinical science is “directed at the promotion of adaptive functioning; at the assessment, understanding, amelioration, and prevention of human problems in behavior, affect, cognition or health; and at the application of knowledge in ways consistent with scientific evidence” (APCS, n.d.). This suggests that clinical scientists need firsthand knowledge of the clinical patients and disorders that they are studying, and that they need to understand the conceptual, methodological, and practical aspects of clinical assessment, treatment, and prevention for these same clinical problems. What is not clear, however, is how best to arrange for students to acquire this vital knowledge. The evidence indicates that expertise is not a function of the amount of time a student devotes to patient contact, or to administering psychological tests and treatments. So what is the most efficient and effective way to give students this training? One-size-fits-all program requirements surely are not the answer. More likely, the old approaches should be replaced by training systems tailored to the needs, research interests, career goals, and
Chapter 14.qxd
3/28/2007
1:01 PM
Page 387
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
387
measurable achievements of individual students. We should not cling to cookie-cutter approaches to training simply out of fear that individualized training might lead to decreased standardization and loss of quality control. Individualized training should be embraced as a way of capitalizing on each student’s interests and strengths with the aim of producing the most powerful next generation of scientists possible. Some Academy training programs already have adopted such an approach, demonstrating not only that it is feasible, but also how it might be done efficiently and effectively while ensuring quality control. In the last few years, there has been movement in the opposite direction, with increased talk of developing a standard list of competencies that should become the focus of all doctoral training programs in clinical (e.g., Kaslow, 2004). I attended the 2002 Competencies Conference in Scottsdale, AZ, at which such proposals were discussed and promoted. I was struck by how confident everyone seemed to be about what a competency was, which competencies were important enough to be put on the master list, and how to assess these competencies. Almost no one worried about the empirical support for the validity of their favorite competencies or their favorite measures. My skepticism was aroused, as well, by the enthusiastic support for this competency movement among psychologists who advocated for replacing the traditional scientific epistemology with the local clinical science epistemology. I have devoted a good deal of my research career to studying behavioral competence and its assessment. Although I generally favor a competence-based approach to assessment, I know that the first step always must be a careful, empirically grounded definition of specific competencies, the careful development of targeted measures, and a clear demonstration that the competence measures are relevant to some important outcomes. Not all hypothesized competencies, and not all methods of assessing them, are equally valid. In redesigning our training programs, we should be wary of outsiders’ proposals to impose lists of requirements on trainees without first providing evidence that these are relevant to specific training goals, or showing that the competence measures actually have incremental validity for predicting critical outcomes. Ultimately, each training program must be responsible for identifying its own training goals, for designing the best way of achieving those goals, and for implementing quality control systems to evaluate and improve the training outcomes. It is time to redesign our doctoral training programs, starting with a blank sheet of paper, fixing our sights on the specific goals, and basing
Chapter 14.qxd
388
3/28/2007
1:01 PM
Page 388
MCFALL
each part of the new design on the best evidence. After we’ve made changes, we need to engage in an ongoing process of quality improvement, continuously gathering data to assess the wisdom of our choices, and working to refine the rough edges of our design. Every aspect of training should contribute meaningfully and efficiently to our overriding mission of training clinical scientists. There is no room for fluff, just for the sake of tradition. We cannot impose requirements on students simply because “we’ve always done things this way.” Decisions about whether a student should take a course, engage in a practicum, go on an internship, teach a class, or learn a particular research tool or clinical skill should be dictated by the student’s needs, research interests, and career goals. We cannot assume that a given training element serves its intended purpose; we need to gather evidence regarding its value. Again the anxious questions: “Will clinical science students be able to get internships? Will they be able to get licensed? Will they be able to see patients?” Setting aside for the moment our concerns about the validity of such things as licensing, let’s examine these questions: Obviously, they are based on an assumption that it is reasonable to expect clinical training and practice to continue as we’ve known it. This assumption is false. Whether we like it or not, the new economic and workforce realities wrought by managed care are forcing radical changes in training and practice in clinical psychology (e.g., APA, 2004; Cummings, 1995; Pion, Kohout, & Wicherski, 2000; Robiner & Crew, 2000). For instance, in 1992, under the fee-for-services system, social workers provided only about 5% of mental health services; by 1997, under the managed care system, social workers were providing 56% of these services, and their market share was expanding rapidly (Clay, 1998). Career opportunities for psychologists in private practice simply have begun to evaporate (Fyfe, 1995). Indeed, radical changes in the mental health care system seem inevitable over the next decade, not only in terms of who delivers what services to whom and how these services are paid for, but also in terms of the fundamental structure of the system. Just as medical care increasingly is becoming hierarchically organized and increasingly specialized, with consequent improvements in reliability, accountability, efficiency, and effectiveness (e.g., Gawande, 1998), parallel changes are almost certain in mental health care. Faced with such changes, there is little justification for continuing to train doctoral students for careers as primary service providers. We must take these new realities into account when redesigning
Chapter 14.qxd
3/28/2007
1:01 PM
Page 389
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
389
our programs, rather than letting concerns about licensing and accreditation dictate our training programs’ designs. PhD psychologists’ future career choices increasingly will be narrowed to positions requiring specialized research training and scientific knowledge. These include not only traditional research and academic positions, but also positions involving the development and evaluation of interventions and assessments, supervision and training, consultation, education, administration, and public policy and public health systems research. Integrity requires that we refocus our training programs on these objectives, all of which require first-rate science training, not training focused on preparing students for service delivery roles. Program brochures should state these new realities explicitly. Programs also should consider reducing the number of predoctoral students admitted, should admit only students appropriately suited for the careers available in this new system, and should tailor all aspects of training to serve these specific goals. Earlier, when describing the two-dimensional model of clinical science, I stated that any activity in the top (science) half of the diagram— from basic to the applied—was legitimate, as long as it was consistent with the scientific evidence and contributed to the advancement of the science. One implication of the current discussion of integrity, however, is that training PhD psychologists for careers as service providers no longer makes sense, given the evidence and new realities. Although the range of reasonable career choices for doctoral trainees has been narrowed, this is not meant to denigrate or devalue in any way the importance of promoting the delivery of empirically supported clinical services. It simply means that such services should be delivered by others while clinical scientists focus on roles that capitalize on their research training, such as the development and evaluation of services, training and supervision of the providers, and investigations into the etiology and prevention of clinical disorders.
Integration. Our second priority should be to work toward integrating psychological clinical science with the rest of psychology, as well as with relevant sciences beyond psychology. Our best hope for making rapid and enduring progress toward solving clinical problems is to draw from the very best scientific theories and methods, regardless of where we find them. Of course, to do this, we first must become aware of what is happening outside our particular area. Unfortunately, the whole field of psychology has become increasingly Balkanized over the last half century,
Chapter 14.qxd
390
3/28/2007
1:01 PM
Page 390
MCFALL
with psychologists in each area knowing little or nothing about the advances in neighboring areas. Clinical psychology may be the most isolated of all—physically (housed in the clinic wing) as well as conceptually. Although clinical psychologists pay lip service to the biopsychosocial model of abnormal behavior, for instance, they may be unaware of the advances actually taking place in biological or social psychology. Although clinical psychologists often consider themselves to be cognitive-behaviorists, their theoretical notions about cognitive processes and methods of assessing cognitive processes may have little in common with the powerful theories and methods of contemporary cognitive science. Similarly, it seems odd that clinical neuropsychologists continue to rely on 5- to 8-hour test batteries to get indirect and imprecise answers to diagnostic questions that neuroscientists now can answer much more efficiently, directly, and accurately with imaging methods. Parochialism impedes scientific progress. It is time to build bridges to the other areas of psychological science. There has been an explosion of basic knowledge, for example, in cognitive science and neuroscience. Clinical scientists need to put themselves in a position to exploit this information in their search for new solutions to the important puzzles of clinical psychology. Ideally, clinical science should not differ from other areas of psychology in terms of the quality or scope of its science; it should be distinguishable primarily in terms of its problem focus. Indeed, clinical science is focused on some of the most interesting and challenging problems in all of psychology. Integration not only should help clinical science solve these problems, but, if integration truly becomes a two-way street, clinical science should contribute reciprocally to advances in other areas, as well. We have developed specialized knowledge about the situational specificity of behavior, the conceptualization and measurement of individual differences, classification systems, and change and plasticity. Sharing this knowledge with other areas while drawing on their specialized knowledge would be consistent with the synergistic spirit behind the current emphasis on translational research at NIMH and other granting agencies. How might we redesign clinical science training programs to make them more integrative? The first step is to break down the artificial barriers that have kept us from capitalizing on scientific advances in neighboring areas. Not only must clinical scientists become more aggressive about acquiring the relevant knowledge and tools in these other areas, but we also should entice scientists from these other areas to work on clinically relevant
Chapter 14.qxd
3/28/2007
1:01 PM
Page 391
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
391
problems. At Indiana University, to cite one example, our clinical science program has adopted a hybrid training model, in which doctoral students are expected to become experts not only in clinical science, but also in at least one other substantive area of psychology—typically, cognitive science, neuroscience, or developmental. Some ambitious students even have combined three areas (e.g., clinical–cognitive–neuroscience). Faculty members from outside the traditional clinical area have begun to apply their expertise to clinical problems in collaboration with the clinical faculty and students, have become co-mentors of clinical students, and even have joined the clinical science training program as core faculty members. We not only have welcomed these faculty colleagues, but have encouraged students from other areas to take our clinical courses and to tackle clinical problems. Thanks to the leadership of our recent chair, Joe Steinmetz, the entire department has joined in this effort to break down old barriers. All students are encouraged to develop individualized training programs and to bridge areas, often working with multiple mentors. With the advice and approval of an advisory committee of faculty members from the relevant areas, each student is able to design a plan of study tailored specifically to that student’s theoretical and methodological expertise, programmatic research interests, and career goals. Clearly, this kind of individualized, problem-focused, hybrid, integrative training differs from the conventional approach to clinical training. Although no two students’ training looks exactly alike, by the end of their training these hybrid clinical students have gone far beyond the superficial knowledge provided by the usual survey courses. They have achieved a breadth and depth of integrated training that allows them to hold their own in interactions with faculty and students from multiple areas of psychology. In effect, this training goes beyond translational research. In the translational model, a basic researcher and an applied clinical researcher are expected to collaborate in the hope that this will foster synergistic integration. In the hybrid model, the basic and clinical research perspectives are combined and integrated within a single investigator!
Concluding Thoughts. It is one thing to talk abstractly about striving for integrity and integration in clinical science; it is quite another matter to take the concrete steps required to realize these goals. Clinical faculty members and students immediately start to worry—not just about internships, licensing, and accreditation, as noted before, but also about
Chapter 14.qxd
392
3/28/2007
1:01 PM
Page 392
MCFALL
effects on the existing program: “How would this affect our applicant pool?” “If we reduce admissions, how will we fill our classes?” “What happens to students who decide that they don’t like research?” “Won’t integrative training take longer?” “How can I supervise students’ integrative research when I’m not an expert in other areas myself?” I’ve heard all of these concerns and many more. As I’ve noted, such questions reflect an assumption that continuing to operate our training programs as we have in the past remains a viable option, that the field is not going to change. But the field is changing…rapidly. It is unlikely that psychology departments as we have known them the last 100+ years—or that clinical psychology programs as we’ve known them for the last 50+ years—will continue to exist unchanged. It also seems likely that the health care system will continue its rapid evolution. We should not design our programs based on fear or nostalgia. We should design them for the future with courage, daring to do what should be done, what is best for the future of our science. This is what it means to act with integrity. Striving for integration requires that we stop defending our turf and become more open-minded and adventurous, willing to adopt promising ideas or methods wherever we find them. There currently is distress among some clinical scientists, for example, over the biologizing of psychology, as though biological approaches to our problems somehow diminish the value of psychological theories and methods. Although it is critical that we think carefully about how we attempt to integrate the models offered by different conceptual frameworks, and avoid making logical errors about causal relationships among these various frameworks, we should become less defensive about considering potentially valuable information simply because it does not come from our particular theoretical perspective. In fact, it has been my experience, in collaborating with colleagues from cognitive science and neuroscience, that the payoff from tackling a problem from multiple perspectives far exceeds any costs that might be involved in working to integrate the different perspectives. Integrative science may not be easiest road, but, being the road less traveled, it offers the greatest promise of important new discoveries. If anyone doubts the feasibility and power of integrative clinical science, they need look no farther than the chapters in this festschrift volume! It has been one my great joys over the years to have known and worked in one way or another with most of these extraordinary scholars and scientists. I feel doubly fortunate to be able to claim each of them as a friend. I am deeply honored by their generosity in agreeing to contribute
Chapter 14.qxd
3/28/2007
1:01 PM
Page 393
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
393
to this collection. It provides an inspiring and exciting array of work. Each contributor exemplifies what I believe it means to be a psychological clinical scientist. The fact that some were not even trained as clinicians makes it all the better! Several generations of scientists are represented in this volume. Peter Lang, who is one of the senior architects of the clinical science model, was a mentor to me in my first faculty job at the University of Wisconsin; has built an incredible record of training outstanding clinical scientists; and, as his chapter shows, remains a highly productive and creative researcher. Teresa Treat represents the most recent generation. I had the privilege of serving as her co-mentor when she was a doctoral student at Indiana, and we continue to collaborate. Her exceptional integrative work proves that the future of clinical science is in very capable hands, indeed. Taken together, the broad range of research presented in these chapters by multiple generations of clinical scientists illustrates precisely the kind of cutting-edge science that we all should strive to emulate. The papers by these pioneers not only define the current frontiers of psychological clinical science, but also inspire us to dream about the exciting possibilities that await us in the future. REFERENCES Academy of Psychological Clinical Science. (n.d.). Mission and specific goals. Retrieved August 3, 2005, from http://www.psychclinicalscience.org. APA Task Force on Workforce Analysis. (2004, November). Final report. Washington, DC: APA. Astin, A. W. (1961). The functional autonomy of psychotherapy. American Psychologist, 16, 75–78. Bandura, A. (1961). Psychotherapy as a learning process. Psychological Bulletin, 58, 143–159. Bandura, A. (1969). Principles of behavior modification. New York: Holt, Rinehart and Winston. Benjamin, L. T., Jr. (2005). A history of clinical psychology as a profession in America (and a glimpse at its future). Annual Review of Clinical Psychology, 1, 1–30. Benjamin, L. T., Jr., & Baker, D. B. (Eds.). (2000). Special section: Boulder at 50. American Psychologist, 55, 233–254. Berman, J. S., & Norton, N. C. (1985). Does professional training make a therapist more effective? Psychological Bulletin, 98, 401–407. Bickman, L. (1999). Practice makes perfect and other myths about mental health services. American Psychologist, 54, 965–978. Boorstin, D. J. (1983). The discoverers: A history of man’s search to know his world and himself. New York: Random House.
Chapter 14.qxd
394
3/28/2007
1:01 PM
Page 394
MCFALL
Campbell, D. T., & Stanley, J. (1963). Experimental and quasi-experimental designs for research. Chicago: Rand-McNally. Chambless, D. L., Chair. Task Force on Promotion and Dissemination of Psychological Procedures. (1995). Training in and dissemination of empiricallyvalidated psychological treatments: Report and recommendations. Clinical Psychologist, 48, 2–23. Chambless, D. L., Baker, M. L., Baucom, D. H., Beutler, L. E., Calhoun, K. S., Crits-Christoph, P., Daiuto, A., DeRubeis, R., Detweiler, J., Haaga, D. A. F., Bennett Johnson, S., McCurry, S., Mueser, K. T., Pope, K. S., Sanderson, W. C., Shoham, V., Stickle, T., Williams, D. A., & Woody, S. R. (1998). Update on empirically validated therapies: II. Clinical Psychologist, 51, 3–16. Chambless, D. L., Sanderson, W. C., Shoham, V., Bennett Johnson, S., Pope, K. S., Crits-Christoph, P., Baker, M., Johnson, B., Woody, S. R., Sue, S., Beutler, L., Williams, D. A., & McCurry, S. (1996). An update on empirically validated therapies. Clinical Psychologist, 49, 5–18. Chambless, D. L., & Ollendick, T. H. (2001). Empirically supported psychological interventions: Controversies and evidence. Annual Review of Psychology, 52, 685–716. Chapman, L. J., & Chapman, J. P. (1969). Illusory correlation as an obstacle to the use of valid psychodiagnostic signs. Journal of Abnormal Psychology, 74, 271–280. Cherry, D. K., Messenger, L. C., & Jacoby, A. M. (2000). An examination of training model outcomes in clinical psychology programs. Professional Psychology: Research and Practice, 31, 562–568. Clay, R. (1998, September). Mental health professions vie for position in the next decade. APA Monitor, 29, 20–21. Crits-Christoph, P., Wilson, G. T., & Hollon, S. D. (2005). Empirically supported psychotherapies: Comment on Westen, Novotny, and Thompson-Brenner (2004). Psychological Bulletin, 131, 412–417. Cummings, N. A. (1995). Impact of managed care on employment and training: A primer for survival. Professional Psychology: Research and Practice, 26, 10–15. Dawes, R., Faust, D., & Meehl, P. E. (1989). Clinical versus actuarial judgment. Science, 243, 1668–1674. Feyerabend, P. (1981). How to defend society against science. In I. Hacking (Ed.), Scientific revolutions (pp. 156–167). New York: Oxford University Press. Feynman, R. (1985). Surely you’re joking, Mr. Feynman! Adventures of a curious character. New York: Norton. Fyfe, B. (1995, December). A field in transition. APA Monitor, 26, 5. Garb, H. N. (2005). Clinical judgment and decision making. Annual Review of Clinical Psychology, 1, 67–89. Gawande, A. (1998, March). No mistake. The future of medical care: Machines that act like doctors, and doctors who act like machines. Newsweek, pp. 74–81. Gigerenzer, G., Swijtink, Z., Porter, T., Daston, L., Beatty, J., & Krüger, L. (1989). The empire of chance: How probability changed science and everyday life. New York: Cambridge University Press. Grove, W. M., & Meehl, P. E. (1996). Comparative efficiency of informal (subjective, impressionistic) and formal (mechanical, algorithmic) prediction procedures: The clinical–statistical controversy. Psychology, Public Policy, and Law, 2, 293–323.
Chapter 14.qxd
3/28/2007
1:01 PM
Page 395
14. ON PSYCHOLOGICAL CLINICAL SCIENCE
395
Kahneman, D., Slovic, P., & Tversky, A. (1982). Judgment under uncertainty: Heuristics and biases. Cambridge, England: Cambridge University Press. Kaslow, N. J. (2004). Competencies in professional psychology. American Psychologist, 59, 774–781. Kiesler, D. J. (1966). Some myths of psychotherapy research and the search for a paradigm. Psychological Bulletin, 65, 110–136. Kuhn, T. S. (1970). The structure of scientific revolutions (2nd ed.). Chicago: Chicago University Press. Lazovik, A. D., & Lang, P. J. (1960). A laboratory demonstration of systematic desensitization psychotherapy. Journal of Psychological Studies, 11, 238–247. Macmillan, N. A., & Creelman, C. D. (2005). Detection theory: A user’s guide (2nd ed.). Mahwah, NJ: Lawrence Erlbaum Associates. Maher, B. A. (1999). Changing trends in doctoral training programs in psychology: A comparative analysis of research-oriented versus professional-applied programs. Psychological Science, 10, 475–481. McFall, R. M. (1991). Manifesto for a science of clinical psychology. Clinical Psychologist, 44, 75–88. McFall, R. M. (1996). Making psychology incorruptible. Applied & Preventive Psychology, 5, 9–15. McFall, R. M. (2000). Elaborate reflections on a simple manifesto. Applied & Preventive Psychology, 9, 5–21. McFall, R. M. (2006). Theory and utility: Key themes in evidence-based assessment. Psychological Assessment, 17, 312–323. McFall, R. M., & Treat, T. A. (1999). Quantifying the information value of clinical assessments with signal detection theory. Annual Review of Psychology, 50, 215–241. Meehl, P. E. (1954). Clinical versus statistical prediction: A theoretical analysis and a review of the evidence. Minneapolis: University of Minnesota Press. Meehl, P. E. (1973). Psychodiagnosis: Selected papers. Minneapolis: University of Minnesota Press. Meehl, P. E. (1978). Theoretical risks and tabular asterisks. Sir Karl, Sir Ronald, and the slow progress of soft psychology. Journal of Consulting and Clinical Psychology, 46, 806–834. Mischel, W. (1968). Personality and assessment. New York: Wiley. Norcross, J. C., Castle, P. H., Sayette, M. A., & Mayne, T. J. (2004). The PsyD: Heterogeneity in practitioner training. Professional Psychology: Research and Practice, 35, 412–419. Paul, G. L. (1969). Behavior modification research: Design and tactics. In C. M. Franks (Ed.), Behavior therapy: Appraisal and status (pp. 29–62). New York: McGraw-Hill. Peterson, D. R. (1968). The clinical study of social behavior. New York: AppletonCentury-Crofts. Peterson, D. R. (1996). Making psychology indispensable. Applied & Preventive Psychology, 5, 1-8. Peterson, D. R. (2003). Unintended consequences: Ventures and misadventures in the education of professional psychologists. American Psychologist, 58, 791–800.
Chapter 14.qxd
396
3/28/2007
1:01 PM
Page 396
MCFALL
Peterson, R. L., Peterson, D. R., Abrams, J. C., & Stricker, G. (1997). The National Council of Schools and Programs of Professional Psychology educational model. Professional Psychology: Research and Practice, 28, 373–386. Peterson, R. L., & Trierweiler, S. J. (1999). Scholarship in psychology: The advantages of an expanded vision. American Psychologist, 54, 350–355. Pierce, J. R. (1980). An introduction to information theory: Symbols, signals and noise. (2nd rev. ed.) New York: Dover. Pion, G., Kohout, J., & Wicherski, M. (2000). “Rightsizing” the workforce through training reductions: A good idea? Professional Psychology: Research and Practice, 31, 266–271. Popper, K. (1962). Conjectures & refutations. New York: Basic Books. Robiner, W. N., & Crew, D. P. (2000). Rightsizing the workforce of psychologists in health care: Trends from licensing boards, training programs, and managed care. Professional Psychology: Research and Practice, 31, 245–263. Rogers, C. R. (1957). The necessary and sufficient conditions of therapeutic personality change. Journal of Consulting Psychology, 21, 95–103. Rotter, J. B. (1971). On the evaluation of methods of intervening in other people’s lives. Clinical Psychologist, 24, 1–2. Sechrest, L. (1963). Incremental validity: A recommendation. Educational and Psychological Measurement, 23, 153–158. Sechrest, L. (1992). The past future of clinical psychology: A reflection on Woodworth (1937). Journal of Consulting and Clinical Psychology, 60, 18–23. Schofield, W. (1964). Psychotherapy: The purchase of friendship. Englewood Cliffs, NJ: Prentice-Hall. Shannon, C. E., & Weaver, W. (1949). The mathematical theory of communication. Urbana: University of Illinois Press. Schön, D. A. (1983). The reflective practitioner: How professionals think in action. New York: Basic Books. Stricker, G., & Trierweiler, S. J. (1995). The local clinical scientist: A bridge between science and practice. American Psychologist, 50, 995–1002. Weisz, J. R., Weersing, V. R., & Henggeler, S. W. (2005). Jousting with straw men: Comment on Westen, Novotny, and Thompson-Brenner (2004). Psychological Bulletin, 131, 418–426. Westen, D., Novotny, C. M., & Thompson-Brenner, H. (2004). The empirical status of empirically supported psychotherapies: Assumptions, findings, and reporting in controlled clinical trials. Psychological Bulletin, 130, 631–663. Westen, D., Novotny, C. M., & Thompson-Brenner, H. (2005). EBP ≠ EST: Reply to Crits-Christoph et al. (2005) and Weisz et al. (2005). Psychological Bulletin, 131, 427–433. Woodworth, R. S. (1937). The future of clinical psychology. Journal of Consulting Psychology, 1, 4–5. Reprinted (1992) Journal of Consulting and Clinical Psychology, 60, 16–17.
Bibiography.qxd
4/6/2007
2:41 PM
Page 397
Bibliography of Richard M. McFall 1960s McFall, R. M., & Saxman, J. H. (1968). Verbal communication as a mediator of expectancy effects: Methodological artifact? Psychological Reports, 23, 1223–1228. Gottman, J. M., McFall, R. M., & Barnett, J. T. (1969). Design and analysis of research using time series. Psychological Bulletin, 72, 299–306. 1970s McFall, R. M., & Schenkein, D. (1970). Experimenter expectancy effects, need for achievement, and field dependence. Journal of Experimental Research in Personality, 4, 122–128. McFall, R. M. (1970). Effects of self-monitoring on normal smokingbehavior. Journal of Consulting and Clinical Psychology, 35, 135–142. McFall, R. M., & Marston, A. R. (1970). An experimental investigation of behavior rehearsal in assertive training. Journal of Abnormal Psychology, 76, 295–303. Marston, A. R., & McFall, R. M. (1971). Comparison of behavior modification approaches to smoking reduction. Journal of Consulting and Clinical Psychology, 36, 153–162. McFall, R. M., & Hammen, C. L. (1971). Motivation, structure, and selfmonitoring: The role of nonspecific factors in smoking reduction. Journal of Consulting and Clinical Psychology, 37, 80–86. McFall, R. M., & Lillesand, D. B. (1971). Behavior rehearsal with modeling and coaching in assertion training. Journal of Abnormal Psychology, 77, 313–323. Gottman, J. M., & McFall, R. M. (1972). Self-monitoring effects in a program for potential high school dropouts: A time series analysis. Journal of Consulting and Clinical Psychology, 39, 273–281.
397
Bibiography.qxd
398
4/6/2007
2:41 PM
Page 398
BIBLIOGRAPHY OF RICHARD M. McFALL
McFall, R. M., & Twentyman, C. T. (1973). Four experiments on the relative contribution of rehearsal, modeling, and coaching to assertion training. Journal of Abnormal Psychology, 81, 199–218. Goldsmith, J. B., & McFall, R. M. (1975). Development and evaluation of an interpersonal skill-training program for psychiatric inpatients. Journal of Abnormal Psychology, 84, 51–58. Twentyman, C. T., & McFall, R. M. (1975). Behavioral training of social skills in shy males. Journal of Consulting and Clinical Psychology, 43, 384–395. Sieck, W. A., & McFall, R. M. (1976). Some determinants of self-monitoring effects. Journal of Consulting and Clinical Psychology, 44, 958–965. McFall, R. M. (1976). Behavioral training: Skill acquisition approaches to clinical problems. Morristown, NJ: General Learning Press (monograph). McFall, R. M. (1977). Analogue methods in behavioral assessment. In J. D. Cone & R. P. Hawkins (Eds.), Behavioral assessment: New directions in clinical psychology (pp. 152–177). New York: Brunner/Mazel. McFall, R. M. (1977). Parameters of self-monitoring. In R. B. Stuart (Ed.), Behavioral self-management: Strategies, techniques, and outcomes (pp. 196–214). New York: Brunner/Mazel. McFall, R. M. (1978). Smoking cessation research. Journal of Consulting and Clinical Psychology, 46, 703–712. (Special methodological issue.) Freedman, B. J., Rosenthal, L., Donahoe, C. P., Jr., Schlundt, D. G., & McFall, R. M. (1978). A social-behavioral analysis of skill deficits in delinquent and nondelinquent adolescent boys. Journal of Consulting and Clinical Psychology, 46, 1448–1462. Haberman, M. C., Chapman, L. J., Numbers, J. S., & McFall, R. M. (1979). Relation of social competence to scores on two scales of psychosis proneness. Journal of Abnormal Psychology, 88, 675–677. 1980s Twentyman, C. T., Boland, T., & McFall, R. M. (1981). Heterosocial avoidance in college males: Four studies. Behavior Modification, 5, 523–552. Muehlenhard, C. L., & McFall, R. M. (1981). Dating initiation from a woman’s perspective. Behavior Therapy, 12, 682–691. Gaffney, L. R., & McFall, R. M. (1981). A comparison of social skills in delinquent and nondelinquent adolescent girls using a behavioral
Bibiography.qxd
4/6/2007
2:41 PM
Page 399
BIBLIOGRAPHY OF RICHARD M. McFALL
399
role-playing inventory. Journal of Consulting and Clinical Psychology, 49, 959–967. McFall, R. M. (1982). A review and reformulation of the concept of social skills. Behavioral Assessment, 4, 1–33. McFall, R. M., & Dodge, K. A. (1982). Self-management and interpersonal skills learning. In P. Karoly & F. Kanfer (Eds.), Self-management and behavior change: From theory to practice (pp. 353–392). New York: Pergamon. Fisher-Beckfield, D., & McFall, R. M. (1982). The development of a competence inventory for college men and evaluation of relationships between competence and depression. Journal of Consulting and Clinical Psychology, 50, 697–705. Muehlenhard, C. L., & McFall, R. M. (1983). Automated assertion training: A feasibility study. Journal of Social and Clinical Psychology, 1, 246–258. Corty, E. C., & McFall, R. M. (1984). Response prevention in the treatment of cigarette smoking. Addictive Behaviors, 9, 405–408. Schlundt, D. G., & McFall, R. M. (1985). New directions in the assessment of social competence and social skills. In M. A. Milan & L. L. L’Abate (Eds.), Handbook of social skills training and research (pp. 22–49). New York: Wiley. McFall, R. M. (1985). Nonbehavioral training for behavioral clinicians. Behavior Therapist, 8, 27–30. McFall, R. M., & McDonel, E. C. (1986). The continuing search for units of analysis in psychology: Beyond persons, situations, and their interactions. In R. O. Nelson & S. C. Hayes (Eds.), Conceptual foundations of behavioral assessment (pp. 201–241). New York: Guilford. McFall, R. M. (1986). Theory and method in assessment: The vital link. Behavioral Assessment, 8, 3–10. Ward, C. I., & McFall, R. M. (1986). Further validation of the Problem Inventory for Adolescent Girls: Comparing Caucasian and Black delinquents and nondelinquents. Journal of Consulting and Clinical Psychology, 54, 732–733. Lipton, D. N., McDonel, E. C., & McFall, R. M. (1987). Heterosocial perception in rapists. Journal of Consulting and Clinical Psychology, 55, 17–21. Schlundt, D. G., & McFall, R. M. (1987). Classifying social situations: A comparison of five methods. Behavioral Assessment, 9, 21–42.
Bibiography.qxd
400
4/6/2007
2:41 PM
Page 400
BIBLIOGRAPHY OF RICHARD M. McFALL
1990s McFall, R. M. (1990). The enhancement of social skills: An informationprocessing analysis . In W. L. Marshall, D.R. Laws, & H. E. Barbaree (Eds.), Handbook of sexual assault: Issues, theories, and treatment of the offender (pp. 311–330). New York: Plenum. McGrew, J. H., & McFall, R. M. (1990). A scientific inquiry into the validity of astrology. Journal of Scientific Exploration, 4, 75–83. McFall, R. M. (1991). Selecting scientific models to revere. Interviewed by S. C. Hayes. Scientist Practitioner, 1(2), 4–11. McDonel, E. C., & McFall, R. M. (1991). Construct validity of two heterosocial perception skill measures for assessing rape proclivity. Violence and Victims, 6, 17–30. McFall, R. M. (1991). Manifesto for a science of clinical psychology. Clinical Psychologist, 44(6), 75–88. Goddard, P., & McFall, R. M. (1992). Decision-making skills and heterosocial competence in college women: An information-processing analysis. Journal of Social & Clinical Psychology, 11, 401–425. McGrew, J. H., & McFall, R. M. (1992). A collaborative Vernon Clark experiment. Correlation, 11, 2–10. McFall, R. M. (1993). The essential role of theory in psychological assessment. In R. L. Glueckauf, L. B. Sechrest, G. R. Bond, & E. C. McDonel (Eds.), Improving assessment in rehabilitation and health (pp. 11–32). Newberry Park, CA: Sage. Viken, R. J., & McFall, R. M. (1994). Paradox lost: Implications of contemporary reinforcement theory for behavior therapy. Current Directions in Psychological Science, 3, 121–125. McFall, R. M. (1994). Invited column: Clinical science in the 21st century. APS Observer, 7(2), 28–29. McFall, R. M. (1995). Models of training and standards of care. In S. C. Hayes, V. M. Follette, R. M. Dawes, & K. E. Grady (Eds.), Scientific standards of psychological practice: Issues and recommendations (pp. 125–137). Reno, NV: Context Press. McFall, R. M. (1995). The future of mental-health care. Chronicle of Higher Education, XLI(45), B1–B3. McFall, R. M., Townsend, J. T., & Viken, R. J. (1995). Diathesis-stress model or “just-so” story? Commentary on Mealey’s “The sociobiology of sociopathy.” Behavioral & Brain Sciences, 18(3), 565–566. McFall, R. M. (1996). Making psychology incorruptible. Applied & Preventive Psychology, 5, 9–15.
Bibiography.qxd
4/6/2007
2:41 PM
Page 401
BIBLIOGRAPHY OF RICHARD M. McFALL
401
McFall, R. M., Treat, T. A., & Viken, R. J. (1997). Contributions of cognitive theory to new behavioral treatments. Psychological Science, 8, 174–176. McFall, R. M. (1998). Implications of prescription privileges for psychological research and training. In S. C. Hayes & E. M. Heiby (Eds.), Prescription privileges for psychologists: A critical appraisal (pp. 101–114). Reno, NV: Context Press. McFall, R. M., Treat, T. A., & Viken, R. J. (1998). Contemporary cognitive approaches to studying clinical problems. In D. K. Routh & R. J. DeRubeis (Eds.), The science of clinical psychology: Accomplishments and future directions (pp. 163–197). Washington, DC: American Psychological Association. McFall, R. M., & Townsend, J. T. (1998). Foundations of psychological assessment: Implications for cognitive assessment in clinical science. Psychological Assessment, 10, 316–330. McFall, R. M., & Treat, T. A. (1999). Quantifying the information value of clinical assessments with signal detection theory. Annual Review of Psychology, 50, 215–241. Tracy, J. A., Ghose, S. S., Stecher, T., McFall, R. M., & Steinmetz, J. E. (1999). Classical conditioning in a non-clinical obsessive compulsive population. Psychological Science, 10, 9–12. McFall, R. M., Eason, B. J., Edmondson, C. B., & Treat, T. A. (1999). Social competence and eating disorders: Development and validation of the Anorexia and Bulimia Problem Inventory. Journal of Psychopathology and Behavioral Assessment, 21, 365–394. 2000s McFall, R. M., Tracy, J. A., Ghose, S. S., & Steinmetz, J. E. (2000). Can eyeblink classical conditioning provide a foundation for integrating clinical science and cognitive-neuroscience in the study of psychopathology? In D. S. Woodruff-Pak & J. E. Steinmetz (Eds.), Eyeblink classical conditioning: Human (pp. 253–274). Norwell, MA: Kluwer Academic. McFall, R. M. (2000). Elaborate reflections on a simple manifesto. Applied and Preventive Psychology, 9, 5–21. Charles, S. T., Carstensen, L. L., & McFall, R. M. (2001). Problem-solving in the nursing home environment: Age and experience differences in emotional reactions and responses. Journal of Clinical Geropsychology, 7, 319–330.
Bibiography.qxd
402
4/6/2007
2:41 PM
Page 402
BIBLIOGRAPHY OF RICHARD M. McFALL
Treat, T. A., McFall, R. M., Viken, R. J., & Kruschke, J. K. (2001). Using cognitive science methods to assess the role of social information processing in sexually coercive behavior. Psychological Assessment, 13, 549–565. McFall, R. M. (2002). Training for prescriptions vs. prescriptions for training: Where are we now? Where should we be? How do we get there? Journal of Clinical Psychology, 58, 659–676. Treat, T. A., McFall, R. M., Viken, R. J., Nosofsky, R. M., MacKay, D. B., & Kruschke, J. K. (2002). Assessing clinically relevant perceptual organization with multidimensional scaling techniques. Psychological Assessment, 14, 239–252. Viken, R. J., Treat, T. A., Nosofsky, R. M., McFall, R. M., & Palmeri, T. (2002). Modeling individual differences in perceptual and attentional processes related to bulimic symptoms. Journal of Abnormal Psychology, 111, 598–609. Lease, A. M., McFall, R. M., Treat, T. A., & Viken, R. J. (2003). Assessing children’s representations of their peer group using a multidimensional scaling technique. Journal of Social and Personal Relationships, 20, 707–728. Lease, A. M., McFall, R. M., & Viken, R. J. (2003). Distance from peers in the group’s perceived organizational structure: Relation to individual characteristics. Journal of Early Adolescence, 23, 194–217. Viken, R. J., Treat, T. A., Bloom, S. L., & McFall, R. M. (2005). Illusory correlation for body type and happiness: Covariation bias and its relationship to eating disorder symptoms. International Journal of Eating Disorders, 38, 65–72. McFall, R. M. (2005). Theory and utility—Key themes in evidence-based assessment: Comment on the special issue. Psychological Assessment, 17, 312–323. Tracy, J. A., McFall, R. M., & Steinmetz, J. E. (2005). Effects of emotional valence and arousal manipulation of eyeblink classical conditioning and autonomic measures. Integrative Physiological and Behavioral Science, 40, 45–54. McFall, R. M. (2006). Doctoral training in clinical psychology. Annual Review of Clinical Psychology, 2, 21–49. Farris, C., Viken, R. J., Treat, T. A., & McFall, R. M. (2006). Heterosocial perceptual organization: A cognitive science application to sexual coercion. Psychological Science, 17, 869–875. Treat, T. A., McFall, R. M., Viken, R. J., Kruschke, J. K., Nosofsky, R. M., & Wang, S. S. (in press). Clinical-cognitive science: Applying quantitative methods of cognitive processing to examine cognitive aspects
Bibiography.qxd
4/6/2007
2:41 PM
Page 403
BIBLIOGRAPHY OF RICHARD M. McFALL
403
of psychopathology. In R. W. J. Neufeld (Ed.), Advances in clinical cognitive science: Formal modeling and assessment of processes and symptoms. Washington, DC: APA Books. McFall, R. M. (in press). On psychological clinical science. In T. A. Treat, R. R. Bootzin, & T. B. Baker (Eds.), Psychological clinical science: Papers in honor of Richard M. McFall. London: Psychology Press. Yeater, E. A., Viken, R. J., Wagner, L. R., & McFall, R. M. (in press). Sexual attitudes and instructional set affect estimates of risk and response effectiveness. Journal of Psychopathology and Behavioral Assessment.
Bibiography.qxd
4/6/2007
2:41 PM
Page 404
Bibiography.qxd
4/6/2007
2:41 PM
Page 405
About the Editors Teresa A. Treat is an Assistant Professor in the Departments of Psychology and Cognitive Science at Yale University. She received a joint PhD in clinical psychology and cognitive science from Indiana University–Bloomington in 2000 and completed her clinical internship and postdoctoral training at Western Psychiatric Institute and Clinic in Pittsburgh. Her primary research interests reside at the conjunction of clinical science and quantitative cognitive science. Her awards include Yale’s Graduate Mentor Award, a Junior Faculty Fellowship from Yale, and the Arthur Greer Memorial Prize from Yale in recognition of outstanding teaching and research. She currently is secretary of the Academy of Psychological Clinical Science. Richard R. Bootzin is a Professor of Psychology at the University of Arizona. He received his bachelor’s degree from the University of Wisconsin and his PhD in clinical psychology from Purdue University. His research interests include the behavioral treatment of insomnia, sleep and cognition, research methodology, and training in clinical psychology. He is a past president of the Society for a Science of Clinical Psychology and the Academy of Psychological Clinical Science, a past member of the board of directors of the Sleep Research Society, and a current member of the board of directors of the Association for Psychological Science. Timothy B. Baker is a Professor of Medicine at the University of Wisconsin–Madison School of Medicine and Public Health. Baker received his bachelor’s degree from the University of California and his master’s degree and PhD in clinical psychology from the University of Utah. Baker has recently published studies on smoking withdrawal, gender differences in smoking cessation, helping women to quit smoking, and treating tobacco use and dependence. He is particularly interested in motivational factors in addiction and has published multiple papers on that subject. He has been the Director of Research of the University of Wisconsin’s Center for Tobacco Research and Intervention (CTRI) since its inception in 1992. Baker recently received the James McKeen Cattell 405
Bibiography.qxd
406
4/6/2007
2:41 PM
Page 406
ABOUT THE EDITORS
Fellow Award from the Association for Psychological Science in recognition of a lifetime of outstanding applied research contributions. Baker also recently served as the editor for the Journal of Abnormal Psychology.
Treat - Author Index.qxd
3/28/2007
1:01 PM
Page 407
Author Index A Abramowitz, J., 23, 26 Abramowitz, J. S., 94, 99, 116, 128 Abrams, J. C., 377, 396 Abramson, L. Y., 85, 97 Academy of Psychological Clinical Science, 364, 368, 381, 386, 393 Ackermann, H., 267, 284 Addis, M., 108, 128 Addis, M. E., 105, 108, 109, 120, 126, 129, 130 Adler, L. E., 57, 60, 68, 70, 73, 74 Agras, W. S., 79, 89, 97, 109, 112, 117, 120, 122, 124, 127, 128, 130, 132 Ahissar, E., 210, 222 Ahn, H., 359 Aiken, L. S., 140, 142, 143, 163 Aine, C., 58, 68, 71 Aine, C. J., 58, 74 Albee, G. W., 14, 25 Alberti, R. E., 114, 127 Alberts, S. C., 84, 102 Alfonso-Reese, L. A., 305, 314 Alkon, D. L., 262, 287 Allen, J. J. B., 61, 71 Allen, L. B., 88, 97, 123, 127 Allen, R., 83, 97 Alloy, L. B., 85, 97, 224, 255 Altman, J., 84, 102 Amaral, D. G., 175, 190, 359 Ameli, R., 175, 192 American Psychiatric Association, 272, 283, 333, 344, 350, 359
American Psychological Association, 12, 13, 15, 18, 26 Amsterdam, J. D., 107, 127, 129 Amundsen, M. J., 95, 98 Anderer, P., 277, 285 Anderson, T., 106, 131 Andreasen, N. C., 276, 277, 283, 287 Andrews, G., 83, 97, 101 Anen, C., 55, 71 Angelone, L. M., 59, 68 Ansorge, M., 327, 345 APA Task Force on Workforce Analysis, 388, 393 Arensdorf, A., 95, 98 Arnow, B. A., 83, 100, 300, 315 Asarnow, R. F., 223, 255 Ashby, F. G., 223, 231, 235, 238, 244, 245, 246, 247, 249, 257, 293, 305, 313, 314 Assenheimer, J. S., 170, 171, 194 Astin, A. W., 373, 393 Astrachan, D. I., 189, 192 Atkins, D., 108, 128 Attwell, P. J. E., 262, 284 Audrain, J., 147, 151, 161 B Babiloni, C., 59, 68 Babiloni, F., 59, 68 Baile, W. F., 141, 159 Bailey, G. K., 208, 221 Bailey, W. C., 137, 147, 148, 149, 160 Baker, D. B., 368, 393
407
Treat - Author Index.qxd
3/28/2007
1:01 PM
408 Baker, M., 324, 345, 382, 394 Baker, M. L., 382, 394 Baker, T. B., 138, 142, 148, 149, 150, 160, 161, 162, 163 Balabanis, M. H., 148, 162 Bandura, A., 16, 26, 80, 97, 107, 123, 127, 199, 203, 219, 374, 393 Banich, M. T., 61, 64, 65, 69, 71, 72 Bao, S., 262, 283 Barch, D. M., 253, 256 Barker, P., 38, 39, 50 Barlow, D. H., 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 93, 94, 96, 97, 98, 99, 102, 107, 111, 112, 114, 115, 123, 127, 169, 179, 191 Baron, R. M., 135, 142, 143, 144, 145, 146, 159 Barrett, P. M., 87, 97 Basoglu, M., 90, 101, 212, 221 Bates, J.E., 337, 347 Battaglia, F., 326, 347 Baucom, D., 324, 345 Baucom, D. H., 382, 394 Bauman, M., 272, 283 Bauman, M. D., 359 Beatty, J., 368, 394 Beck, A., 107, 109, 127 Beck, A. T., 80, 81, 89, 97, 169, 170, 183, 191, 299, 300, 314 Becker, E. S., 300, 317 Benjamin, J., 327, 345 Benjamin, L. T., Jr., 11, 15, 26, 368, 371, 393 Benn, K. D., 253, 254 Bennett Johnson, S., 382, 394 Benson, H., 174, 180, 191 Berg, P., 59, 73 Berger, T. W., 265, 283, 286 Berglund, P., 340, 346 Berman, J. S., 385, 393 Bertagnolli, A., 116, 131 Berthier, N. E., 262, 264, 265, 283, 286 Bertrand, O., 58, 73 Best, E., 58, 74 Beutler, L., 324, 345, 382, 394 Beutler, L. E., 123, 124, 127, 325, 340, 345, 382, 394
Page 408
AUTHOR INDEX Bickman, L., 118, 127, 385, 393 Biederman, J. K., 83, 102 Bieling, P., 339, 345 Biemond, R., 93, 100 Bijl, R. V., 187, 195 Birbaumer, N., 267, 284, 338, 345 Bird, A., 38, 39, 49 Birmaher, B., 23, 28 Bixler, M. A., 173, 193 Bjork, R. A., 218, 219 Blaha, L. M., 246, 254 Blalock, J. A., 141, 159 Blanchard, D. C., 174, 191 Blanchard, E., 203, 219 Blanchard, R. J., 174, 191 Blankenship, M. R., 265, 283 Blaszczynski, A., 85, 102 Boksman, K., 225, 226, 239, 244, 256, 313, 317 Bolger, N., 134, 136, 139, 141, 142, 144, 145, 146, 161, 162 Bolt, D. M., 149, 161 Bonthius, D. J., 269, 283 Boorstin, D. J., 367, 393 Bootzin, R. R., 7, 18, 19, 21, 26, 28 Boring, E. G., 4, 26 Borkovec, T. D., 94, 97, 188, 189, 194 Bornfleth, H., 59, 73 Borus, J., 360 Bostrom, A.,116, 131 Bouton, M. E., 81, 97, 199, 201, 202, 207, 208, 211, 214, 217, 218, 219, 220 Boutros, N. N., 57, 73 Bowlby, J., 86, 98 Bradley, A., 30, 34, 51 Bradley, M. M., 57, 61, 69, 71, 73, 169, 175, 176, 177, 178, 186, 189, 191, 192, 194 Braff, D. L., 60, 72 Bramon, E., 57, 59, 69 Brandon, T. H., 214, 220 Braun, C., 338, 345 Brawn, C. M., 267, 286 Brent, D. A., 23, 28 Britton, G. B., 262, 287 Broderson, L., 84, 99 Brody, P. E., 111, 131
Treat - Author Index.qxd
3/28/2007
1:01 PM
AUTHOR INDEX Broga, M. I., 225, 226, 254, 255 Brooks, D. C., 207, 208, 214, 220 Brown, G. K., 170, 183, 191 Brown, L. L., 107, 127 Brown, S. M., 275, 276, 283 Brown, T. A., 83, 85, 98, 102, 169, 179, 191 Brownell, K. D., 299, 313, 318 Bruch, M. A., 87, 98 Bryson, S. E., 272, 286 Bufka, L. F., 111, 112, 127 Bullock, D., 276, 284 Burke, P., 85, 101 Burmeister, M., 327, 347 Busemeyer, J. R., 313, 314 Bush, G., 336, 345 Buss, A. H., 169, 172, 191 Buss, K. A., 337, 345 Butman, J., 58, 71 Bystritsky, A., 90, 98, 117, 118, 131 C Cacioppo, J. T., 61, 73 Cadenhead, K. S., 60, 72 Calhoun, K., 324, 345 Calhoun, K. S., 382, 394 Camerer, C. F., 55, 71 Campbell, D. T., 7, 19, 26, 370, 394 Campbell, L., 295, 310, 311, 314, 316 Campbell, L. A., 169, 191 Campos, J. J., 337, 345 Cañive, J. M., 57, 74 Canterbury, R., 300, 314 Carducci, F., 59, 68 Carmichael, S. T., 175, 190 Carrillo, M. C., 267, 286 Carroll, C. A., 275, 276, 283 Carson, M. A., 61, 73 Carter, B. L., 141, 159 Carter, C. S., 224, 254 Carter, J. C., 105, 128 Carter, J. R., 225, 226, 239, 244, 253, 254, 256, 313, 314, 317 Carter, S. R., 176, 195 Caspi, A., 327, 345, 359 Castle, P. H., 375, 395
Page 409
409 Cawthra, E. M., 60, 68 Celaya, L. J., 224, 254 Cerny, J. A., 81, 93, 94, 97 Chaderjian, M. R., 224, 254 Chambless, D., 324, 345 Chambless, D. L., 22, 26, 81, 89, 98, 99, 113, 114, 127, 359, 382, 394 Champoux, M., 84, 101 Chapman, J., 56, 72 Chapman, J. P., 378, 394 Chapman, L. J., 378, 394 Chapman, T. F., 83, 99 Cheema, F., 329, 346 Chen, H. T., 136, 159 Chen, L., 262, 283 Cherry, D. K., 375, 394 Choate, M. L., 88, 97, 123, 127 Chorpita, B. F., 83, 85, 86, 95, 98, 169, 191 Christensen, A., 111, 114, 129 Christner, R., 58, 68, 71 Christner, R. F., 58, 74 Cicchetti, D. V., 36, 49 Cinciripini, L. G., 141, 159 Cinciripini, P. M., 159 Cinciripini, P. R., 141, 159 Cioffi, F., 41, 49 Clark, D. M., 81, 90, 91, 98, 102, 116, 118, 123, 127, 128 Clark, G. A., 264, 286 Clark, L. A., 83, 98, 169, 170, 171, 187, 191, 194, 333, 337, 345, 347, 360 Clark, R. E., 264, 284 Clarkin, J. F., 113, 130 Claus, E. D., 65, 72 Clay, R., 388, 394 Clementz, B. A., 60, 72 Codispoti, M., 61, 69, 175, 191 Cohen, J. D., 62, 67, 69, 73 Cohen, N. J., 65, 69, 72 Cohen, S. J., 137, 147, 148, 149, 160 Cole, D. A., 85, 86, 98, 103, 140, 141, 142, 143, 160 Collins, B. N., 214, 220 Collins, J. F., 116, 128 Collins, L. M., 136, 145, 147, 160 Colonius, H., 243, 255
Treat - Author Index.qxd
3/28/2007
1:01 PM
410
Page 410
AUTHOR INDEX
Colpaert, F. C., 210, 222 Compton, R. J., 61, 69 Cook, E. W., 61, 73 Cook, E. W., III, 177, 178, 186, 191, 194 Cook, J., 91, 102 Cook, W., III, 57, 69 Cooke, S. F., 262, 284 Coon, H., 57, 70 Cooper, T. B., 84, 98 Cooper, Z., 110, 114, 121, 122, 128 Coplan, J. D., 84, 98 Coppola, R., 359 Costa, P. T., Jr., 337, 345 Courchesne, E., 272, 284 Craig, I., 327, 345 Craig, I. W., 359 Craighead, W. E., 109, 129 Craske, M. G., 81, 85, 87, 89, 90, 91, 93, 94, 97, 98, 102, 115, 117, 118, 127, 131, 200, 202, 203, 204, 205, 206, 207, 209, 210, 211, 212, 215, 216, 217, 220, 221, 222 Creelman, C. D., 301, 316, 371, 395 Crew, D. P., 388, 396 Crick, N. R., 291, 314 Crits-Christoph, P., 111, 115, 116, 117, 118, 123, 126, 127, 131, 324, 325, 345, 347, 376, 382, 394 Crits-Cristoph, P., 325, 345 Cronbach, L. J., 15, 26 Crosby, R. D., 111, 130 Crow, S. J., 111, 112, 127, 130 Cummings, N., 48, 49 Cummings, N. A., 30, 49, 388, 394 Cunningham, C. L., 210, 220 Curry, J. F., 119, 127 Curtin, J. J., 148, 160 Cuthbert, B. N., 57, 61, 69, 71, 73, 169, 175, 177, 178, 186, 189, 191, 192, 194 Czerwinski, M., 246, 254 D Dadds, M. M., 87, 97 Dai, D., 83, 102 Daiuto, A., 382, 394 Dalgleish, T., 300, 314
Dang, Q., 148, 162 Dantendorfer, K., 277, 285 Dar, R., 90, 101 Daston, L., 368, 394 Daum, I., 267, 284 Davidson, R. J., 55, 61, 69, 74 Davies, M., 360 Davis, A., 57, 70 Davis, L., 58, 71 Davis, M., 55, 72, 173, 174, 175, 181, 189, 191, 192 Davis, W. W., 360 Davison, G. C., 107, 132, 325, 347 Dawes, R., 53, 69, 121, 127, 378, 394 Dawes, R. M., 30, 40, 43, 50 Dawson, M. E., 57, 69 de Graaf, R., 187, 195 de Ruiter, C., 91, 99 de Vries, S., 93, 99 DeFries, J. C., 83, 101 Densen, M. E., 275, 284 Derryberry, D., 61, 74 DeRubeis, R., 382, 394 DeRubeis, R. J., 107, 111, 123, 127, 129, 131, 139, 160, 163, 325, 347 Desmond, J. E., 262, 286 Detweiler, J., 382, 394 Deyo, R. A., 265, 286 Dichgans, J., 267, 284 Dichter, G. S., 188, 192 Diederich, A., 243, 254, 255 Diener, H. C., 276, 285 Dike, G. L., 265, 288 Dimidjian, S., 89, 99, 108, 128 DiNardo, P. A., 179, 191 Dishion, T. J., 136, 160 Disterhoft, J. F., 265, 267, 286 Dobson, K., 108, 128 Dobson, K. S., 108, 109, 128, 129 Doby, D., 277, 285 Dodge, K. A., 291, 314 Donegan, N. H., 263, 286 Donenberg, G. R., 360 Donkervoet, J. C., 95, 98 Donovan, K. A., 60, 68 Doss, A. J., 360 Dow, M. G. T., 92, 100
Treat - Author Index.qxd
3/28/2007
1:01 PM
Page 411
411
AUTHOR INDEX Dow, R. S., 276, 285 Drabant, E., 328, 336, 346 Drobes, D., 169, 175, 177, 178, 186, 189, 191 Duffy, M., 116, 118, 128 Dulawa, S., 326, 347 Dunner, D., 108, 128 Dykman, B., 224, 255 Dzhafarov, E., 250, 252, 256 E Eason, B. J., 302, 316 Eaton, W. W., 340, 346 Eaves, L. J., 100 Ebsworthy, G., 295, 311, 316 Echemendia, R. J., 94, 97 Echiverri, A. M., 205, 209, 210, 211, 216, 217, 221 Eddy, J. M., 136, 160 Edgar, C., 57, 74 Edgar, J. C., 57, 58, 61, 64, 70, 71, 72, 74 Edmondson, C. B., 302, 316 Egan, M. F., 359 Egeth, H. E., 232, 242, 255 Ego, V., 210, 222 Ehlers, A., 87, 99 Eifert, G. H., 189, 192, 194 Eilers, A. T., 269, 284 Eisch, A. J., 326, 345 Elbert, T., 58, 73, 338, 345 Elias, M. J., 291, 317 Elkin, I., 116, 128 Elliott, D. S., 223, 255 Emery, G., 80, 81, 97, 107, 127 Emmelkamp, P. M., 93, 99 Emmelkamp, P. M. G., 93, 100 Emmons, M. L., 114, 127 Engberg, J. B., 148, 162 Engelberg, B., 94, 100, 115, 129 Engels, A. S., 61, 64, 70, 72 Epstein, L., 147, 151, 161 Epstein, S., 169, 192 Erbaugh, J., 169, 191 Erickson, M. A., 305, 314 Eulitz, C., 58, 73 Eysenck, H. J., 13, 14, 26
F Fairburn, C. G., 105, 109, 110, 114, 117, 120, 121, 122, 124, 127, 128, 130, 132 Farris, C., 313, 315 Faust, D., 378, 394 Feeley, M., 139, 160 Feldner, M. T., 189, 192, 194 Fenz, W. D., 169, 192 Ferguson, K., 30, 45, 49, 51 Ferguson, K. E., 36, 37, 51 Fernberger, S. W., 11, 26 Feyerabend, P., 369, 394 Feynman, R., 370, 394 Fiala, J. C., 276, 284 Fiegenbaum, W., 118, 129, 130 Fific, M., 225, 257 Filion, D. L., 57, 69 Filoteo, J. V., 313, 315 Finch, A. E., 82, 100 Finger, M. S., 169, 170, 193 Finn, P. R., 272, 273, 287 Fiore, M. C., 137, 138, 142, 147, 148, 149, 150, 160, 161, 162, 163 Fiorito, E., 57, 73 First, M. B., 334, 346, 360 Fisher, J. E., 30, 50, 61, 70, 71 Fisher, R. P., 216, 220 Fitzsimmons, J. R., 175, 176, 194 Flaherty, B. P., 136, 145, 147, 160 Flor, H., 338, 345 Foa, E., 23, 26, 199, 220 Foa, E. B., 80, 93, 94, 99, 116, 128 Fodor, J. A., 55, 70 Follette, W. C., 80, 99, 359 Fouladi, R. T., 141, 159 Fowler, K. A., 34, 50 Frances, A. J., 360 Frangou, S., 57, 59, 69 Frank, J. B., 359 Frank, J. D., 359 Frank, M., 23, 27, 28, 118, 129, 130 Franklin, M., 23, 26 Franklin, M. E., 94, 99, 116, 128 Franks, C. M., 119, 128 Franks, R. D., 57, 60, 68, 70 Freedman, R., 57, 60, 68, 70, 73
Treat - Author Index.qxd
3/28/2007
1:01 PM
412
Page 412
AUTHOR INDEX
Freeman, B. B., 107, 129 Frerk, N., 216, 220 Freud, S., 41, 50 Friedman, B. H., 188, 192 Frishkoff, G., 59, 74 Fujita, S., 58, 71 Fujiwara, S., 58, 71 Funderbunk, F., 177, 188, 192 Fyer, A. J., 83, 99 Fyfe, B., 388, 394 G Gabrieli, J. D. E., 267, 286 Gage, H. D., 339, 346 Gagnon, R., 106, 128 Gallop, R., 107, 127, 129 Garb, H. M., 30, 33, 43, 50, 52 Garb, H. N., 378, 394 Garber, J., 119, 129 Garcia-Gutierrez, A., 199, 217, 218, 220 Gardner, R., 212, 221 Garfield, S. L., 105, 106, 128 Garson, C., 339, 345 Garssen, B., 91, 99 Garvey, C. R., 5, 26 Gaser, C., 275, 287 Gaston, L., 106, 128 Gaudiano, B. A., 30, 50 Gault, J., 57, 72 Gawande, A., 388, 394 Geer, J. H., 188, 193 Geiselman, R. E., 216, 220 Gelder, M., 90, 98 Gelder, M. G., 91, 102, 198, 221 Gelfand, L. A., 107, 127, 139, 160 Gershuny, B. S., 83, 99 Geyer, M. A., 60, 72 Ghaderi, A., 122, 128 Gholson, B., 38, 39, 50 Ghose, S. S., 278, 279, 287 Ghosh, D., 327, 347 Gibbon, M., 360 Gigerenzer, G., 368, 394 Gillespie, K., 116, 118, 128 Gingrich, J., 327, 345 Giorgini, M., 250, 252, 256
Girgus, J. S., 85, 101 Gladis, M. M., 107, 127 Glass, G. V., 359 Gleitman, H., 198, 221 Glickstein, M., 264, 288 Globisch, J., 175, 192 Gloor, P., 173, 192 Gnys, M., 148, 162 Goddard, P., 291, 315 Gold, J. M., 253, 256 Goldapple, K., 339, 345 Goldberg, L. R., 337, 347 Goldberg, T. E., 253, 256, 359 Goldfried, M. R., 113, 128, 139, 160 Goldsmith, H. H., 337, 345 Goldstein, A. J., 81, 99 Goldstone, R., 311, 315 Goldstone, R. L., 246, 255, 311, 315 Golinelli, D., 117, 118, 131 Gollan, J. K., 108, 109, 128, 129 Gonzalez, S., 59, 72 Goodlett, C. R., 269, 270, 271, 284, 286 Gorman, J. M., 84, 90, 93, 97, 98, 111, 130 Gormezano, I., 260, 284 Gormican, S., 246, 257 Gorsuch, R. L., 170, 183, 194 Gortner, E., 108, 109, 129 Gortner, E. T., 108, 128 Gotlib, I. H., 83, 100, 300, 315 Gould, R. A., 89, 99 Gould, T. J., 262, 265, 284 Grace, A. A., 63, 64, 70 Graham, F. K., 57, 70, 73 Graham, J. W., 136, 145, 147, 160 Granger, D. A., 84, 99 Granholm, E., 223, 255 Grant, K. A., 339, 346 Grave de Peralta, R., 59, 72 Gray, J. A., 189, 192, 194 Green, J. T., 270, 271, 284 Green, M. F., 253, 256 Greenberg, B., 327, 345 Greenberg, L. S., 139, 160 Greenberg, R. L., 80, 81, 97 Greenwald, M. K., 175, 192 Griffith, C. R., 9, 26 Griffith, J. M., 60, 73
Treat - Author Index.qxd
3/28/2007
1:01 PM
AUTHOR INDEX Grillon, C., 175, 192 Grilo, C. M., 112, 129 Grisham, J. R., 169, 191 Gross, C., 326, 347 Gross, J. J., 188, 192 Grossberg, S., 276, 284 Grove, W. M., 33, 50, 378, 394 Gscheidle, T., 359 Guerri, C., 269, 284 Guinther, P. M., 56, 57, 71 Gunnar, M., 84, 101 Gunnar, M. R., 84, 99 Guze, Y., 332, 347 Gwaltney, C. J., 148, 162 H Haaga, D. A. F., 382, 394 Hackman, A., 90, 98 Hackmann, A., 116, 118, 128 Hadzi-Pavlovic, D., 91, 102 Hager, F., 275, 287 Hahleg, K., 118, 130 Hahlweg, K., 23, 27, 118, 129 Hairdarliu, S., 210, 222 Hall, S. M., 148, 160 Halmi, K. A., 112, 127 Halsey, L., 36, 41, 51 Haman, K. L., 107, 129 Hamilton, D. A., 57, 70 Hamm, A. O., 175, 192, 208, 221 Hammen, C. L., 133, 139, 142, 162 Hammond, T. R., 43, 51 Hampson, S., 58, 73 Hamre, K. M., 269, 285 Han, S. S., 360 Hanlon, F. M., 57, 58, 70, 71, 74 Hanna, H. H., 93, 94, 102 Hannigan, J. H., 269, 285 Haque, W., 159 Hardiman, M. J., 264, 288 Harmon-Jones, E., 61, 71 Harrington, H., 327, 345, 359 Harrington, P. J., 90, 102 Harris, J. A., 208, 221 Harris, J. G., 60, 68 Harris, L. M., 84, 99
Page 413
413 Harrison, P. J., 66, 71 Hartlage, S., 224, 255 Hartz, D. T., 148, 160 Harvey, A. G., 359 Hasselt, V. B., 79, 99 Hastings, J. E., 203, 221 Hatgis, C., 105, 120, 126 Havik, O. D., 113, 129 Hawk, L., 147, 151, 161 Hawley, K. M., 360 Hayaki, J., 117, 120, 130 Hayes, E. A., 339, 345 Hayes, S., 96, 97 Hayes, S. C., 80, 82, 96, 99, 188, 192 Hayhurst, H., 108, 109, 132 Haynes, R. D., 111, 131 Haynes, S. N., 135, 160 Hazelrigg, M. D., 115, 116, 131 Hea, R. A., 60, 73 Healy, A. R., 329, 346 Heath, A. C., 83, 100, 101 Heaton, R. K., 253, 256 Heim, C., 339, 346 Heimberg, R. G., 87, 94, 98, 100, 115, 129 Heinrichs, N., 90, 99 Heinrichs, R. W., 253, 255 Heller, W., 61, 63, 71, 73 Hembree, T. L., 264, 285 Hen, R., 327, 345 Henderson, A. S., 83, 97 Henderson, D., 30, 35, 44, 51 Henderson, S. H., 85, 102 Henggeler, S. W., 111, 117, 120, 132, 325, 347, 376, 396 Herbert, J. D., 30, 40, 50 Herman, S., 92, 101 Herrington, J., 61, 69 Herrington, J. D., 61, 63, 64, 70, 71, 72, 73 Hersen, M., 79, 97, 99 Hershenson, M., 223, 255 Herting, J. R., 143, 160 Hertsgaard, L., 84, 99 Hesselink, J. R., 272, 284 Hetherington, E. M., 85, 102 Hetrick, W. P., 56, 71, 275, 276, 283
Treat - Author Index.qxd
3/28/2007
1:01 PM
Page 414
414
AUTHOR INDEX
Hickcox, M., 148, 162 Highgate, S., 244, 256 Hilgard, E. R., 12, 26 Hinshaw, S. P., 359 Hirsch, C., 111, 116, 129 Hladek, D., 200, 203, 204, 205, 206, 207, 208, 209, 210, 217, 221, 222 Ho, M. R., 61, 70 Hobbs, N., 14, 27 Hodges, L., 93, 102 Hodgson, R., 198, 222 Hoehn-Saric, R., 177, 188, 192 Hofer, E., 277, 285 Hoffman, J. M., 142, 143, 146, 150, 155, 161 Hofmann, S. G., 90, 99 Holcombe, L., 216, 220 Holker, L., 295, 311, 316 Holland, H. L., 216, 220 Holland, P. W., 140, 160 Hollon, S., 108, 111, 126, 127, 128 Hollon, S. D., 89, 98, 107, 108, 113, 114, 119, 127, 129, 325, 345, 359, 376, 394 Holmbeck, G., 145, 160 Holtzworth-Munroe, A., 291, 315 Hopkins, J., 57, 72 Hothersall, D., 4, 5, 6, 7, 27 Houts, A. C., 359 Howard, K. I., 139, 160 Howard, R. C., 118, 129 Howe, G. W., 85, 102 Hoyle, R. G., 142, 144, 160 Hoyningen-Huene, P., 38, 39, 50 Hu, S., 327, 345 Huang, M. X., 57, 58, 68, 70, 71, 72, 74 Huckfeldt, R., 265, 283 Hudson, S. M., 296, 318 Hulsbosch, A. M., 93, 99 Humfleet, G .L., 148, 160 Hyman, S. E., 62, 67, 73, 333, 346 I Iacono, W. G., 333, 346 Iedema, J., 187, 195 Ilardi, S. S., 109, 129
Ille, N., 59, 73 Imber, S. D., 116, 128 Institute of Medicine, 45, 50 International HapMap Consortium, 327, 346 Irwin, J., 57, 70, 74 Ito, T., 61, 73 Ivry, R. B., 276, 285 Iyengar, S., 23, 28 J Jackson, D. C., 61, 69 Jacobs, G. A., 170, 183, 194 Jacobson, B. L., 65, 69 Jacobson, E., 177, 193 Jacobson, N. S., 89, 99, 108, 109, 111, 114, 119, 128, 129, 130 Jacoby, A. M., 375, 394 Jaimcz, T. S., 93, 94, 102 Jaimez, T. L., 90, 102 Jamerson, B., 138, 162 James, P., 148, 162 Jamner, L. D., 148, 162 Jang, K. L., 83, 100 Jansson, L., 89, 100 Jaquin, K. M., 90, 102 Jardine, R., 83, 101 Jenkins, W., 338, 346 Jernigan, T. L., 272, 284 Jetté, J., 225, 226, 238, 239, 244, 256, 313, 317 Jiang, Y., 293, 317 Johansen, M. K., 295, 305, 308, 309, 310, 315 Johnson, B., 382, 394 Johnson, J. E., 275, 285 Johnson, L. D., 82, 100 Johnson, S., 216, 220 Johnson, S. L., 300, 315 Johnson, T. B., 271, 284 Johnston, D. W., 92, 100, 198, 221 Johnston, J. A., 138, 162 Johnston, L., 296, 318 Johnston, P. S., 338, 346 Jolley, S., 111, 116, 129 Jones, A., 57, 74
Treat - Author Index.qxd
3/28/2007
1:01 PM
AUTHOR INDEX Jones, A. P., 57, 70 Jones, M. C., 78, 100 Jones, M. L., 208, 221 Joormann, J., 300, 315 Jorenby, D. E., 138, 150, 162, 150, 163 Jorm, A. F., 115, 116, 131 Judd, C. M., 136, 139, 142, 143, 144, 145, 146, 160 Juster, H. R., 94, 100, 115, 129 K Kadlec, H., 313, 315 Kahneman, D., 33, 52, 247, 255, 369, 395 Kalin, N. H., 61, 69 Kanno, A., 58, 71 Kaplan, J. R., 339, 346 Kapp, B. S., 173, 193 Karkowski, L. M., 83, 100 Kasch, K. L., 83, 100, 300, 315 Kashy, D. A., 134, 136, 139, 141, 142, 144, 146, 161 Kaslow, N. J., 387, 395 Kass, E., 189, 192 Kassel, J. D., 148, 162 Katkin, E., 17, 27 Katon, W., 117, 118, 131 Katz, D. B., 263, 264, 285 Kaufmann, V., 147, 151, 161 Kauneckis, D., 84, 99 Kawamura, T., 58, 71 Kazdin, A. E., 18, 27, 113, 130, 135, 138, 139, 140, 161 Keele, S. W., 276, 285 Kehoe, E. J., 260, 284 Keller, J., 55, 73 Kelly, E. L., 12, 26 Kelly, G. A., 16, 27, 290, 299, 315 Kemper, T., 272, 283 Kenardy, J., 92, 101 Kenardy, J. A., 92, 100 Kendall, K. C., 139, 161 Kendler, K. S., 83, 100, 327, 346 Kennedy, S., 339, 345 Kenney, F. A., 211, 220 Kenny, D. A., 134, 135, 136, 139, 141, 142, 143, 144, 145, 146, 159, 160, 161
Page 415
415 Kessler, R. C., 100, 340, 346 Kettner, R. R., 264, 286 Kieffaber, P. D., 275, 276, 283 Kiesler, D. J., 374, 395 Kihlstrom, J., 84, 101 Kihlstrom, J. F., 8, 27 Kihlstrom, L. C., 8, 27 Kim, J., 300, 316 King, D. A., 201, 220 King-Casas, B., 55, 71 Kirby, M., 90, 101 Kirk, J., 111, 116, 132 Kirsch, J. R., 188, 193 Kisley, M. A., 56, 57, 71 Klein, D. F, 83, 92, 99, 100 Klorman, R., 203, 221 Klosko, J. S., 81, 93, 94, 97 Knecht, S., 338, 345 Knight, R. A., 223, 255 Knopke, A. J., 111, 130 Koenig, O., 55, 56, 71 Koerner, K., 108, 109, 129 Kohlenberg, R., 108, 128 Kohout, J., 388, 396 Kolachana, B., 328, 336, 346 Kolachana, B. S., 359 Kolko, D. J., 23, 28 Konorski, J., 172, 193 Koretz, D. S., 340, 346 Koselka, M., 91, 102 Kosslyn, S. M., 55, 56, 71 Koven, N. S., 61, 63, 71, 73 Kowalski, P., 177, 188, 192 Kozak, M., 23, 26, 199, 220 Kozak, M. J., 55, 60, 61, 71, 73, 80, 93, 94, 99, 116, 128, 177, 178, 193, 194 Kraemer, H., 112, 127 Kraemer, H. C., 109, 122, 124, 127, 130 Kratochwill, T. R., 18, 27 Kraus, N., 339, 345 Kremen, I., 172, 193 Krijn, M., 93, 99, 100 Krueger, R. F., 169, 170, 187, 193, 333, 346 Krüger, L., 368, 394 Krupa, D. J., 264, 285
Treat - Author Index.qxd
3/28/2007
1:01 PM
416
Page 416
AUTHOR INDEX
Kruschke, J. K., 224, 257, 295, 296, 298, 299, 300, 302, 304, 305, 306, 308, 309, 310, 311, 313, 314, 315, 317, 318 Kuch, K., 90, 101, 212, 221 Kuhn, J. W., 327, 346 Kuhn, T. S., 37, 38, 39, 50, 369, 395 Kunzel, R., 124, 131 L Labouvie, E., 117, 120, 130 Labus, J. S., 205, 216, 217, 221 Laird, N. M., 83, 102 Lalich, J., 30, 52 Lambert, M. J., 82, 100 Lang, P. J., 55, 57, 61, 69, 71, 72, 73, 80, 100, 107, 114, 130, 169, 172, 174, 175, 176, 177, 178, 181, 186, 189, 191, 192, 193, 194, 195, 203, 208, 221, 374, 395 Langenbucher, J.W., 359 Lantz, G., 59, 72 Larson, M., 84, 99 Lasko, N. B., 61, 73, 329, 346 Lau, M., 339, 345 Laudan, L., 39, 50 Lavond, D. G., 261, 262, 264, 285, 286, 287 Lawrence, D. L., 149, 161 Lazarus, A. A., 106, 119, 130, 132 Lazovik, A. D., 374, 395 Leahy, R. M., 58, 72 Lease, A. M., 299, 315 LeDoux, J. E., 55, 72, 173, 193 Lee, J. H., 93, 102 Lee, M. D., 293, 315 Lee, R. R., 57, 58, 70, 71 Lee, W. W., 293, 314 Lehman, C. L., 169, 191 Leiner, A. L., 276, 285 Leiner, H. C., 276, 285 Leischow, S. J., 138, 162 Leitenberg, H., 79, 89, 97 Lelliott, P. T., 90, 101 Lemery, K. S., 337, 345 Lemieux, S. K., 267, 285 Leonard, S., 57, 70, 72 Lerman, C., 147, 151, 161
Leshner, A., 66, 72 Levant, R. F., 22, 27, 324, 346 Levenson, R. W., 188, 192, 359 Levin, D. N., 61, 73, 177, 178, 193, 194 Levis, D. J., 80, 102 Levitt, J., 23, 26 Levitt, J. T., 94, 99, 111, 112, 116, 127, 128 Levy, L., 225, 226, 239, 244, 256 Lewin, K., 21, 27 Lewin, M., 91, 98 Lewine, J. D., 57, 58, 72 Ley, R., 91, 92, 100 Li, Q., 327, 345 Light, G. A., 60, 72 Lightfoot, N., 246, 254 Lilienfeld, S. O., 30, 34, 36, 40, 50, 52 Lillesand, D. B., 133, 162 Lim, S., 300, 316 Lincoln, J. S., 264, 285 Lincoln, T., 23, 27 Lincoln, T. M., 118, 130 Lippa, Y., 311, 315 Lipsedge, M. S., 212, 221 Lipsey, M. W., 18, 27, 136, 161 Lipton, D. N., 291, 316 Lira, A., 327, 345 Liu, K. S., 148, 162 Liu, W., 147, 151, 161 Livesley, W. J., 83, 100 Lloyd, A., 37, 43, 51 Lockwood, C. M., 139, 142, 143, 146, 150, 155, 161 Loeb, K. L., 117, 120, 130 Logel, J., 57, 72 Logue, S. F., 264, 287 Lohr, J. M., 30, 34, 36, 40, 50 Longabaugh, R., 137, 138, 140, 145, 162 Losee, J., 38, 50 Lovett, M. L., 107, 127, 129 Lu, J., 293, 317 Lucas, F., 327, 345 Lucas, J. A., 93, 94, 102 Lucas, R. A., 93, 94, 102 Luckey, B., 12, 26 Ludgate, J., 90, 98 Lueger, R. J., 139, 160 Lushene, P. R., 170, 183, 194
Treat - Author Index.qxd
3/28/2007
1:01 PM
AUTHOR INDEX Luten, A., 85, 100 Lutzenberger, W., 267, 284 Luu, P., 59, 61, 74, 336, 345 Lynn, S. J., 30, 34, 36, 40, 50 Lyons, L. C., 115, 116, 131 M Ma, S. H., 109, 130 MacCorquodale, K., 143, 161 Mach, R. H., 339, 346 MacKay, D. B., 224, 257, 298, 299, 313, 318 MacKinnon, D. P., 136, 138, 139, 140, 142, 143, 145, 146, 150, 155, 161, 216, 220 Macklin, M. L., 61, 73 MacLeod, C., 189, 194, 295, 299, 300, 310, 311, 314, 316, 318 Macmillan, N. A., 301, 316, 371, 395 Madden, J., 262, 285 Maddox, W. T., 293, 313, 314, 315 Magaro, P. A., 223, 255 Maher, B. A., 375, 395 Mahoney, J. C., 269, 286 Maidenberg, E., 90, 98 Malik, M. M., 340, 345 Maling, M. S., 139, 160 Mamounas, L. A., 262, 285 Mancil, R. B., 169, 191 Mandler, G., 172, 193 Mandler, J. M., 172, 193 Manicavasgar, V., 91, 102 Mannuzza, S., 83, 99 Manoach, D. S., 223, 255 Mansell, W., 359 Marcus, M. D., 110, 114, 124, 128 Marcussan, B. L., 269, 286 Marder, S. R., 223, 255 Markley, D., 108, 128 Markon, K., 337, 346 Markon, K. E., 333, 346 Markowitsch, H. J., 173, 194 Marks, I. M., 80, 83, 89, 90, 100, 101, 169, 193, 212, 221 Marmar, C., 139, 160 Marr, D., 55, 67, 72 Marshall, B. S., 260, 284 Marshall, W. L., 296, 318
Page 417
417 Marston, A. R., 133, 136, 140, 161 Martell, C. R., 89, 99, 108, 130 Marter, S. R., 223, 255 Martin, J. M., 85, 98 Martin, L. Y., 83, 99 Martin, N. G., 83, 101 Martinovich, Z., 139, 160 Masterman, M., 38, 51 Mathews, A., 295, 299, 300, 311, 316, 318 Mathews, A. M., 169, 193, 198, 221 Matt, G. E., 115, 116, 131 Matthews, A., 189, 194 Mattson, S. N., 269, 286, 287 Maude-Griffin, R., 148, 160 Mauk, M. D., 262, 263, 264, 265, 286 Maxwell, S. E., 140, 141, 142, 143, 160 Mayne, T. J., 375, 395 McCann, J. T., 43, 51 McCarthy, D. E., 142, 148, 149, 160, 161, 162 McCarty, T. S., 225, 256 McCauley, E., 85, 101 McClearn, G. E., 83, 101 McCormick, D. A., 262, 264, 285, 286 McCormick, R. A., 170, 171, 194 McCrae, R. R., 337, 345 McCurry, S., 382, 394 McDonel, E. C., 291, 316 McFall, R. M., 14, 15, 18, 20, 22, 27, 29, 32, 33, 34, 36, 44, 53, 72, 115, 130, 133, 135, 136, 137, 139, 140, 142, 145, 146, 161, 162, 168, 194, 224, 253, 255, 257, 278, 279, 281, 287, 289, 290, 291, 293, 296, 297, 298, 299, 300, 301, 302, 304, 305, 306, 313, 310, 311, 313, 315, 316, 317, 318, 322, 343, 346, 359, 371, 381, 395 McGee, C., 95, 98 McGhie, A., 56, 72 McGlashan, T. H., 112, 129 McGlinchey, J., 108, 128 McGlinchey-Berroth, R., 267, 286 McKinley, S. C., 305, 317 McLaughlin, J., 265, 286 McLean, A., 61, 73 McLean, A., Jr., 177, 194 McLeod, D. R., 177, 188, 192
Treat - Author Index.qxd
3/28/2007
1:01 PM
Page 418
418 McNally, R. J., 43, 51, 170, 194 McNamee, G., 90, 101 McNaughton, N., 189, 194 McNeil, D. W., 178, 186, 191, 194 McRae, K. A., 60, 68, 73 McTeague, L. M., 186, 194 Meadows, E. A., 89, 98 Meaney, M. J., 339, 346 Medina, J. F., 262, 286 Meehl, P., 53, 72 Meehl, P. E., 16, 27, 29, 33, 42, 43, 50, 51, 53, 60, 72, 137, 143, 161, 162, 370, 374, 378, 394, 395 Melamed, B. G., 178, 186, 191, 194, 203, 221 Menard, S., 152, 162 Mendelsohn, M., 169, 191 Mercer, J., 30, 51 Merikangas, K., 175, 192 Merikangas, K. R., 340, 346 Merrill, K. A., 23, 27, 94, 101, 111, 116, 117, 118, 130 Merzenich, M., 338, 346 Messenger, L. C., 375, 394 Metalsky, G. I., 85, 97 Meuret, A. E., 91, 92, 101 Meyer-Lindenberg, A., 328, 336, 346 Michel, C. M., 59, 72 Miklowitz, D. J., 113, 130 Milham, M. P., 61, 65, 69, 72 Miller, D. P., 265, 288 Miller, G. A., 53, 55, 57, 60, 61, 64, 69, 71, 72, 73, 74, 177, 178, 193, 194 Miller, J., 242, 255 Miller, S. L., 338, 346 Milner, J. S., 291, 316 Mineka, S., 8, 9, 27, 81, 83, 84, 85, 97, 98, 100, 101, 169, 194, 197, 198, 199, 200, 203, 204, 205, 206, 207, 208, 209, 210, 212, 213, 217, 221, 222 Mintz, J., 117, 118, 127 Mischel, W., 16, 27, 374, 395 Mitchell, J. E., 111, 112, 127, 130 Mitchell, J. R., 85, 101 Mock, J., 169, 191 Moffitt, T., 327, 345 Moffitt, T. E., 359 Mohanty, A., 61, 63, 64, 69, 70, 71, 72, 73
AUTHOR INDEX Moleiro, C., 325, 345 Montague, P. R., 55, 62, 67, 71, 73 Montgomery, R. W., 40, 50 Moody, E.W., 199, 217, 218, 220 Moore, H., 63, 64, 70 Moore, J. W., 262, 264, 265, 283, 286 Moore, R. G., 108, 109, 132 Moradi, A., 300, 314 Morgan, D., 339, 346 Morgan-Lopez, A. A., 136, 138, 140, 161 Morgenstern, J., 137, 138, 140, 145, 162 Morris, S. E., 57, 74 Moses, S. N., 57, 58, 70, 71, 74 Mosher, J. C., 58, 72 Moss, S., 85, 101 Moulder, B., 175, 191 Mowrer, O. H, 79, 101, 278, 286 Moyer, J. R., 265, 286 Mueser, K. T., 382, 394 Munoz, K., 328, 336, 346 Munoz, R. F., 148, 160 Murias, M., 59, 73 Murray, M. M., 59, 72 Murray, R. M., 57, 59, 69 Mussell, M. P., 111, 130 Muthen, B. O., 153, 154, 162 Muthen, L. K., 153, 154, 162 Myles-Worsley, M., 57, 70 Mystkowski, J., 204, 205, 206, 207, 208, 209, 210, 217, 221 Mystkowski, J. L., 205, 209, 210, 211, 212, 213, 216, 217, 221 N Nachmias, J., 198, 221 Nagamoto, H. T., 60, 68, 73 Nakasato, N., 58, 71 Napper, R. M. A., 269, 286 Nathan, P. E., 111, 130, 359 National Advisory Mental Health Council Behavioral Science Workgroup, 289, 316 National Institute for Clinical Excellence, 112, 121, 130 National Institute of Mental Health, 329, 346, 359
Treat - Author Index.qxd
3/28/2007
1:01 PM
Page 419
419
AUTHOR INDEX Naugle, A. E., 36, 51 Navarro, A. M., 115, 116, 131 Neale, M. C., 100 Neisser, U., 198, 221 Nelson, J. B., 201, 217, 220 Nelson, R., 96, 97 Nelson, T., 216, 220 Nelson-Gray, R. O., 82, 96, 99 Nemeroff, C. B., 84, 98, 339, 346 Nenadic, I., 275, 287 Neshat-Doost, H., 300, 314 Nestler, E. J., 326, 345 Neufeld, R. W. J., 225, 226, 238, 239, 244, 253, 254, 255, 256, 293, 294, 313, 314, 316, 317 Newman, M. G., 92, 100, 101 Nezworski, M. T., 30, 52 Niaura, R., 147, 151, 161 Nicol, T. G., 339, 345 Nides, M. A., 138, 162 Nietzel, M. T., 115, 116, 131 Nitschke, J. B., 61, 69 Noecker, T. L., 56, 57, 71 Nolen-Hoeksema, S., 85, 101 Norcross, J. C., 375, 395 Nordahl, T. E., 224, 254 Noriega-Dimitri, R., 91, 98 Norton, N. C., 385, 393 Nosfosky, R. M., 224, 257, 293, 295, 296, 297, 298, 299, 301, 302, 305, 310, 311, 313, 317, 318 Noshirvani, H., 90, 101, 212, 221 Novonty, C. M., 94, 103, 106, 112, 117, 120, 126, 132, 324, 347, 376, 396 Nozawa, G., 232, 235, 239, 243, 250, 257 Nuechterlein, K. H., 57, 74, 253, 256 O O’Donnell, B. F., 275, 276, 283 O’Donohue, W., 30, 32, 34, 35, 36, 37, 41, 43, 44, 45, 49, 50, 51, 291, 317 O’Donohue, W. T., 40, 50, 296, 317 O’Leary, D. S., 276, 277, 283, 287 O’Malley, K., 269, 287 O’Reardon, J. P., 107, 127, 129 O’Sullivan, G., 90, 101, 212, 221
Oakley, D. A., 265, 286 Öhman, A., 55, 72, 208, 221 Okiishi, J. C., 82, 100 Olafsson, R. P., 93, 100 Olincy, A., 57, 60, 68, 70 Ollendick, T. H., 22, 26, 382, 394 Oltmanns, T., 8, 9, 27 Ormel, J., 187, 195 Orr, S. P., 61, 73 Orr, W. B., 265, 286 Orrison, W. W., Jr., 57, 72 Orr-Urtreger, A., 57, 70 Öst, L. G., 89, 100, 199, 202, 222 Otto, M. W., 89, 99, 212, 222 Overmier, J. B., 84, 101 Overton, D. A., 210, 222 Owens, M. J., 84, 98 P Pachtman, E., 57, 60, 68, 70 Pagano, M.E., 112, 129 Palmeri, T., 293, 297, 298, 299, 301, 305, 313, 318 Palmeri, T. J., 293, 305, 310, 317 Panksepp, J., 61, 73 Pantev, C., 58, 73, 338, 345 Paradiso, S., 276, 283 Parker, G., 91, 102 Parloff, M. B., 105, 130 Parzen, E., 228, 256 Pascoe, J. P., 173, 193 Patrick, C. J., 169, 175, 177, 178, 186, 189, 191 Patrick, C. P., 333, 346 Paty, J. A., 148, 162 Paul, G. L., 107, 130, 131, 374, 395 Paulson, K. M., 58, 70 Pavlov, I. P., 200, 222 Pecevich, M., 57, 68 Peeke, L. G., 85, 98 Pepping, G., 124, 131 Perlstein, W., 57, 73 Perrett, S. P., 264, 286 Persons, J. B., 111, 116, 131 Perz, W. G., 148, 162 Peterson, C. B., 111, 130
Treat - Author Index.qxd
3/28/2007
1:01 PM
Page 420
420
AUTHOR INDEX
Peterson, D. R., 34, 51, 374, 375, 376, 377, 381, 395, 396 Pettit, Meuret, 91, 102 Petzel, T. P., 275, 285 Pezawas, L., 328, 336, 346 Phillips, K. A., 334, 346 Piasecki, T. M., 142, 148, 149, 161, 162 Pierce, J. R., 370, 396 Pincus, H. A., 334, 346, 360 Pion, G., 388, 396 Piper, M. E., 148, 160 Pitkanen, A., 175, 190 Pitman, R. K., 329, 346 Pizzagalli, D., 61, 69 Plomin, R., 83, 85, 101,102, 169, 191 Plotkin, D., 87, 98 Plotsky, P. M., 339, 346 Pollack, M. H., 89, 99, 212, 222 Pook, M., 23, 28 Pope, C., 112, 131 Pope, K. J., 293, 315 Pope, K. S., 382, 394 Pope, M., 108, 109, 132 Popper, K., 369, 396 Porter, T., 368, 394 Posner, M., 336, 337, 345, 347 Poulsen, C., 59, 74 Poulton, R., 87, 98 Powell, D. A., 265, 286 Powers, W. T., 62, 73 Pratt, E. M., 117, 120, 130 Prescott, C. A., 83, 100, 327, 346 Press, G. A., 272, 284 Price, J. L., 175, 190 Price, L., 93, 102 Prince, M., 8, 27 Prince, S. E., 108, 109, 129 Prioleau, O., 339, 346 Prout, H. T., 115, 116, 131 Putnam, K. M., 61, 69 Q Quartz, S. R., 55, 71 Quine, W. V., 33, 52 Quiring, J., 59, 74
R Rabe-Hesketh, S., 57, 59, 69 Racette, S. R., 83, 102 Rachman, S., 198, 200, 222 Rachman, S. J, 80, 101 Ragusea, S. A., 94, 97 Rakic, P., 326, 347 Ralph, J. A., 85, 100 Rammsayer, T., 275, 287 Ranken, D., 58, 74 Rapee, R. M., 85, 87, 97, 102 Rauch, S. L., 61, 73 Raynor, R., 78, 103, 197, 222 Razran, G, 81, 102 Rebec, G. V., 265, 288 Reichenbach, H., 43, 52 Reiss, D., 85, 102 Reiss, S., 170, 194 Rennard, S. I., 138, 162 Reus, V. I., 148, 160 Richardson, W. S., 111, 131 Richardson-Klavehn, A., 218, 219 Richmond, R., 148, 162 Ricker, D., 57, 74 Rief, W., 23, 27, 118, 130 Riley, B., 327, 346 Riley, E. P., 269, 286, 287 Rinck, M., 300, 317 Rising, C. E., 264, 286 Ritter, B., 203, 219 Ritz, T., 91, 101 Robertson, L. C., 224, 254 Robertson, M. H., 8, 14, 17, 28 Robiner, W. N., 388, 396 Robins, S., 332, 347 Robinson, L., 115, 116, 131 Rockstroh, B., 338, 345 Rodier, P. M., 272, 286 Rodriguez, B., 204, 205, 206, 207, 208, 209, 210, 217, 221 Rodriguez, B. I., 200, 203, 204, 205, 206, 207, 210, 222 Roebuck, T. M., 269, 287 Roemer, L., 188, 189, 194 Rogers, C., 41, 52
Treat - Author Index.qxd
3/28/2007
1:01 PM
AUTHOR INDEX Rogers, C. R., 14, 28, 373, 396 Rogers, R. F., 270, 284 Romani, G. L., 58, 59, 68, 73 Rosa, L., 30, 51 Rosen, G. M., 40, 50, 325, 347 Rosenbaum, J. F., 83, 102 Rosenberg, W., 111, 131 Rosengard, C., 211, 220 Ross. A. O., 198, 222 Rossini, P. M., 59, 68 Roth, D., 147, 151, 161 Roth, W., 91, 102 Roth, W. T., 91, 92, 101 Rothbart, M., 336, 337, 347 Rothbart, M. K., 337, 347 Rothbaum, B. O., 93, 102 Rothman, A., 123, 131 Rothwell, P. M., 113, 131 Rottenberg, J., 83, 100 Rotter, J. B., 16, 17, 18, 28, 85, 102, 379, 396 Rouder, J. N., 293, 317 Routh, D. K.,10, 11, 15, 20, 28 Rowan, V. C., 91, 103 Rowe, M., 91, 98, 202, 220 Roy-Byrne, P. P., 117, 118, 131 Rozeboom, W. W., 143, 162 Rubin, D. B., 140, 162 Rudman, J. C., 291, 317 Ruggill, J. S., 18, 19, 26 Ruiz, B. P., 264, 286 Ruiz, M., 94, 97 Rush, J., 107, 127 Russell, I. S., 265, 286 Rutherford, E., 295, 310, 311, 316 Rutherford, E. M., 295, 310, 311, 314 Rutter, M., 83, 101 Ryan, S. M., 87, 97 S Sabatinelli, D., 175, 176, 191, 194 Sabatino, S. A., 212, 222 Sackett, D. L., 111, 131 Salkovkis, P. M., 90, 98 Salkovskis, P. M., 123, 131 Salomon, R. M., 107, 127, 129 Samuel, D. B., 333, 347
Page 421
421 Sanchez, N., 57, 74 Sanchez-Andres, J. V., 262, 287 Sanders, K. M., 329, 346 Sanderson, W. C., 382, 394 Sanford, R. N., 12, 26 Sanislow, C. A., 112, 129 Santarelli, L., 326, 347 Santiago, H., 91, 102 Sapolsky, R. M., 84, 102 Sarason, S. B., 13, 14, 28 Sarner, L., 30, 51 Sarter, M., 173, 194 Saucier, G., 337, 347 Sauer, H., 275, 287 Saxe, M., 326, 347 Sayama, M., 48, 49 Sayette, M. A., 291, 317, 375, 395 Schell, A. M., 57, 69 Scherg, M., 59, 73 Schewe, P. A., 296, 317 Schmaling, K., 108, 128 Schmidt, N. B., 90, 91, 93, 94, 102 Schneider, D. J., 176, 195 Schneider, W., 245, 246, 256 Schofield, W., 374, 396 Schön, D. A., 377, 396 Schreiner, C., 338, 346 Schreurs, B. G., 262, 287 Schroeder, B., 23, 27, 118, 129, 130 Schuemie, M. J., 93, 99 Schugens, M. M., 267, 284 Schuierer, G., 58, 73 Schulte, D., 124, 131 Schulte-Bahrenberg, D., 124, 131 Schumann, C. M., 359 Schupp, H. T., 61, 73 Schwarz, W., 243, 256 Schwarzkopf, S. B., 57, 73 Schweickert, R., 230, 249, 250, 252, 256 Schweizer, R., 338, 345 Sears, L. L., 272, 273, 277, 287 Sears, R. R., 12, 28 Sechrest, L. B., 7, 19, 28 Sechrest, L., 370, 371, 396 Segal, Z., 339, 345 Segal, Z. V., 108, 109, 131, 132 Seki, K., 58, 71
Treat - Author Index.qxd
3/28/2007
1:01 PM
422 Seligman, M. E. P., 84, 85, 97, 101, 117, 122, 131 Sen, S., 327, 347 Sengun, S., 90, 101 Seroczynski, A. D., 85, 98 Servan-Schreiber, D., 67, 69 Shadish, W. R., 115, 116, 131, 137, 162 Shaffer, L. F., 12, 26 Shafran, R., 110, 114, 121, 122, 128, 132, 359 Shakow, D., 12, 26 Sham, P., 57, 59, 69 Shannon, C. E., 370, 396 Shapiro, D., 148, 162 Shaw, B., 107, 127 Shea, M. T., 112, 116, 128, 129 Shear, M. K., 90, 93, 97 Sheets, V., 142, 143, 146, 150, 155, 161 Shekhar, A., 275, 276, 283 Shelton, R. C., 107, 119, 127, 129, 188, 192 Sher, K. J., 83, 99 Sherbourne, C. D., 117, 118, 131 Sherwood, A., 57, 70 Shiffman, S., 148, 162 Shiffrin, R., 311, 315 Shiffrin, R. M., 224, 245, 246, 256 Shin, L. M., 61, 73 Shindler, K. L., 43, 51 Shoham, V., 382, 394 Sholiton, R., 172, 193 Short, M., 57, 72 Shrout, P. E., 145, 162 Shulz, D. E., 210, 222 Siegle, G., 115, 116, 131 Silove, D., 91, 102 Simmens, S. J., 85, 102 Simons, A. D., 107, 127 Simons, R. F., 57, 73 Singer, M. T., 30, 52 Sirota, L., 327, 345 Skinner, B. F., 337, 347 Skodol, A. E., 112, 129 Slot, L. A., 210, 222 Slovic, P., 369, 395 Smit, F., 187, 195 Smith, D. A., 56, 57, 71, 73 Smith, G. T., 142, 144, 160
Page 422
AUTHOR INDEX Smith, M. L., 115, 116, 131., 359 Smith, S., 93, 102 Smith, S. M., 209, 215, 216, 218, 222 Smith, S. S., 138, 150, 162, 163 Smoller, J. W., 83, 102 Sosnik, R., 210, 222 Sotsky, S. M., 116, 128 Speckman, P., 293, 317 Spencer, M. E., 58, 72 Spiegel, D. A., 90, 99, 102 Spielberger, C. D., 170, 183, 194 Spinelli, L., 59, 72 Spira, A. P., 189, 192, 194 Spitzer, R.L., 360 Squire, L. R., 268, 284 Srinivasan, R., 59, 73 Staats, A. W., 80, 102 Stampfl, T. G., 80, 102 Stanley, J., 370, 394 Stecher, T., 278, 279, 287 Steele, P. M., 262, 286 Steer, R. A., 170, 183, 191 Stein, M. B., 117, 118, 131 Steinmetz, J. E., 260, 261, 262, 263, 264, 265, 266, 267, 270, 271, 272, 273, 275, 276, 278, 279, 281, 283, 284, 285, 287, 288 Steinmetz, J. J., 265, 283 Steinmetz, S. S., 264, 287 Steketee, G., 89, 102 Stephen, J., 58, 68 Stephen, J. M., 58, 74 Sternberg, S., 232, 240, 241, 242, 247, 256, 257 Stewart, G., 83, 97 Stewart, J. L., 61, 71 Steyvers, M., 311, 315 Stice, E., 122, 128 Stickle, T., 382, 394 Stirman, S., 325, 347 Stirman, S. W., 111, 123, 131 Stoolmiller, M., 136, 160 Storey, J., 91, 102 Stout, J. D., 313, 314 Stout, R. L., 112, 129 Strauss, C., 169, 175, 177, 178, 186, 189, 191 Strauss, M. E., 170, 171, 194
Treat - Author Index.qxd
3/28/2007
1:01 PM
AUTHOR INDEX Streissguth, A. P., 269, 287 Stricker, G., 30, 31, 52, 377, 396 Strosahl, K. D., 188, 192 Strupp, H. H., 106, 131 Stuart, G. L., 23, 28, 93, 94, 102, 103, 115, 116, 131, 132 Substance Abuse and Mental Health Administration, 94, 95, 102 Sue, S., 382, 394 Sugden, K., 327, 345, 359 Sullivan, G., 117, 118, 131 Summerskill, W., 111, 131 Summerskill, W. S. M., 112, 131 Sun, D., 293, 317 Sunsay, C., 199, 217, 218, 220 Surget, A., 326, 347 Sutton, S. K., 61, 69, 74, 188, 192 Svartberg, M., 115, 116, 131 Swartzentruber, D., 211, 220 Sweeney, R. B., 137, 162 Swerdlow, N. R., 57, 69 Swijtink, Z., 368, 394 Swinson, R. P., 90, 101, 212, 221 T Taborga, M. P., 136, 138, 140, 161 Taghavi, R., 300, 314 Taitano, K., 55, 74 Talebi, H., 325, 345 Tallal, P., 338, 346 Tanenbaum, S. J., 324, 347 Tang, T. Z., 107, 127, 139, 163 Task Force on Promotion and Dissemination of Psychological Procedures, 324, 347 Taub, E., 338, 345 Taylor, A., 327, 345, 359 Taylor, C. B., 92, 100, 101 Teasdale, J. D., 85, 97, 108, 109, 130, 131, 132 Telch, M. J., 90, 93, 94, 102 Tellegen, A., 337, 347 Thayer, J. F., 188, 192 The British Psychological Society, 96, 103 Thoma, R. J., 57, 58, 70, 71, 74 Thomas, C., 199, 200, 221
Page 423
423 Thomas, J. D., 269, 284 Thompson, J. K., 264, 285 Thompson, R. F., 262, 264, 265, 283, 285, 286 Thompson-Brenner, H., 94, 103, 106, 112, 117, 120, 126, 132, 324, 347, 376, 396 Thomson, A., 92, 100 Tiffany, S., 147, 163 Tolbert, V. E., 23, 27, 94, 101, 111, 116, 117, 118, 130 Tolin, D. F., 40, 50 Tomarken, A. J., 188, 192 Tomlin, D., 55, 71 Torres, F., 57, 74 Torrey, E. F., 276, 287 Townsend, J. T., 223, 225, 226, 227, 230, 231, 232, 234, 235, 237, 238, 239, 242, 243, 244, 245, 246, 247, 249, 250, 253, 254, 255, 256, 257, 258, 289, 291, 313, 314, 315, 316 Tracy, J. A., 262, 275, 276, 278, 279, 281, 283, 287 Traill, S., 300, 315 Trakowski, J., 91, 102 Tran, T., 271, 284 Treat, T. A., 23, 28, 93, 94, 102, 103, 115, 116, 131, 132, 168, 194, 224, 257, 289, 293, 296, 297, 298, 299, 300, 301, 302, 304, 305, 306, 309, 310, 311, 313, 315, 316, 317, 318, 371, 395 Treatment for Adolescents with Depression Study (TADS) Team, 132 Treisman, A., 246, 257 Trierweiler, S. J., 30, 31, 52, 377, 396 Trost, R. C., 84, 98 Truax, P. A., 108, 109, 129 Truglio, R., 85, 98 Tsao, J. C. I., 87, 98 Tucker, D. M., 59, 61, 73, 74 Turken, A. U., 305, 314 Turner, J. E., 86, 98, 103 Turpin, G., 57, 69 Tuschen-Caffier, B., 23, 28 Tversky, A., 33, 52, 369, 395 Twentyman, C. T., 114, 130 Tysk, L., 275, 287
Treat - Author Index.qxd
3/28/2007
1:01 PM
424
Page 424
AUTHOR INDEX V
Vagg, P. R., 170, 183, 194 Vaitl, D., 175, 192, 208, 221 van der Mast, C. A., 93, 99 Van Dyck, R., 91, 99 Van Vunakis, H., 159 Van Zandt, T., 242, 246, 252, 257, 258 VandenBos, G. R., 18, 27, 113, 129 Vázquez, C., 224, 255 Verchinski, B., 328, 336, 346 Verkindt, C., 58, 73 Vernon, L. L., 205, 212, 213, 217, 221 Vernon, P. A., 83, 100 Vianzon, R., 57, 72 Viken, R. J., 168, 194, 224, 257, 289, 293, 296, 297, 298, 299, 300, 301, 302, 305, 309, 310, 311, 313, 315, 316, 317, 318 Viswanathan, R., 212, 221 Vitousek, K. B., 296, 318 Vittum, J., 327, 346 Vohs, J. L., 275, 276, 283 Vollebergh, W. A., 187, 195 Vollick, D., 225, 226, 239, 244, 256, 313, 317 Vollick, D. N., 225, 258 Volz, H.-P., 275, 287 von Witzleben, I., 23, 27, 118, 129, 130 Vrana, S. R., 175, 177, 178, 194 W Wade, W. A., 23, 27, 28, 93, 94, 101, 102, 103, 105, 111, 115, 116, 117, 118, 120, 126, 130, 131, 132 Waldo, M. C., 57, 60, 68, 70 Waldron, E. M., 305, 314 Wallfisch, A., 159 Walsh, B. T., 117, 120, 122, 127, 128, 130 Walters, E. E., 100, 340, 346 Walton, M., 30, 32, 47, 52 Wampold, B. E., 359 Wang, S. S., 296, 310, 311, 313, 318 Wang, S. W., 299, 313, 318 Ward, C. H., 169, 191 Ward, T., 296, 318 Warrier, C. M., 339, 345
Watkins, E., 359 Watkins, J. T., 116, 128 Watson, D., 83, 98, 169, 170, 171, 187, 191, 194, 333, 337, 346, 347 Watson, J. B., 78, 103, 197, 222 Watson, R. I., Sr., 7, 28 Watts, F. N., 299, 300, 318 Weaver, W., 370, 396 Webb, A., 65, 69 Webb, A. G., 61, 63, 73 Webb, M. R., 293, 315 Weber, K., 170, 171, 194 Weersing, V. R., 23, 28, 111, 117, 120, 132, 136, 163, 325, 347, 376, 396 Weerts, T. C., 177, 195, 203, 221 Wegner, D. M., 176, 188, 195 Weickert, T., 359 Weierich, M. R., 313, 318 Weinberger, A., 132 Weinberger, D. R., 66, 71, 276, 287 Weisend, M. P., 57, 58, 70, 71, 74 Weisend, W., 58, 71 Weiss, B., 115, 116, 131, 360 Weiss, J. M, 84, 103 Weisz, J. R., 84, 99, 111, 117, 120, 132, 136, 163, 325, 347, 360, 376, 396 Welch, J. P., 272, 286 Wells, A., 90, 98 Wells, K., 119, 127 Welsch, S. K., 150, 163 Wenger, M. J., 223, 226, 234, 235, 238, 239, 243, 244, 246, 247, 257, 258 Wenzel, E. A., 61, 63, 70, 73 Wenzlaff, R. M., 188, 195 West, J. R., 269, 283, 284, 285, 286 West, S. G., 140, 142, 143, 146, 150, 155, 161, 163 Westbrook, D., 111, 116, 132 Westbrook, F. R., 208, 221 Westen, D., 94, 103, 106, 112, 117, 120, 126, 132, 324, 347, 376, 396 Wetter, D. W., 141, 150, 159, 163 White, I. M., 265, 288 White, J. D., 265, 286 White, P. M., 57, 74 White, T. L., 176, 195 White, W., 265, 288
Treat - Author Index.qxd
3/28/2007
1:01 PM
Page 425
425
AUTHOR INDEX Wicherski, M., 388, 396 Wickwire, K., 90, 101 Widiger, T. A., 333, 347, 360 Wienbruch, C., 338, 345 Wiggins, J. S., 141, 163 Wilhelm, F. H., 91, 92, 101 Williams, D. A., 382, 394 Williams, J. B., 360 Williams, J. M., 299, 300, 318 Williams, M., 108, 131 Williams, R., 111, 116, 129 Williams, S., 108, 109, 132 Williams, S. L., 199, 222 Wilson, E., 295, 310, 311, 316 Wilson, G. T., 105, 107, 109, 110, 111, 112, 113, 114, 117, 118, 119, 120, 121, 122, 124, 125, 126, 127, 128, 130, 132, 291, 317, 318, 325, 345, 376, 394 Wilson, K. G., 188, 192 Wittling, W., 5, 28 Wolfe, B. E., 113, 128 Wolfle, D., 12, 28 Wolpe, J., 79, 91, 103, 106, 119, 132, 169, 195, 197, 222, 360 Wonderlich, S. A., 111, 130 Wood, J. M., 30, 52 Woodruff-Pak, D. S., 266, 267, 285, 288 Woods, A. M., 199, 217, 218, 220 Woods, S. W., 90, 93, 97, 175, 192 Woodworth, R. S., 371, 396 Woody, R. H., 8, 14, 17, 28
Woody, S. R., 382, 394 Woolaway-Bickel, K., 91, 102 Wright, L., 15, 28 Wszalek, T., 65, 69 Wyatt, J. R., 276, 287 Y Yee, C. M., 57, 74 Yen, S., 112, 129 Yeo, C. H., 262, 284, 288 Yeo, R. A., 58, 71 Yeung-Courchesne, R., 272, 284 Yim, L. M., 95, 98 Yokaitis, M. H., 264, 285 Young, P. R., 107, 109, 127, 129 Yule, W., 300, 314 Z Zaki, S. R., 313, 317 Zecker, S. G., 339, 345 Zhang, A. A., 264, 284 Zhou, M., 327, 345 Zimmerli, W. D., 188, 192 Zinbarg, R., 197, 221 Zinbarg, R. E., 205, 212, 213, 217, 221 Zoellner, L. A., 212, 222 Zusman, R. M., 329, 346 Zvolensky, M. J., 189, 192, 194 Zwar, N., 148, 162
Treat - Author Index.qxd
3/28/2007
1:01 PM
Page 426
Treat - subject index.qxd
3/28/2007
1:01 PM
Page 427
Subject Index A Academy of Psychological Clinical Science (APCS), 23–24, 77, 126, 323, 341, 349–350, 358, 364–371, 381–387 Accreditation, 12–23, 126, 349, 354–357, 366, 377, 382–386 Affect regulation, see Emotion regulation Agoraphobia, see Panic disorder Alternative Accreditation Steering Committee (AASC), 20–22, 382–383 American Association of Applied and Preventive Psychology (AAAPP), 19–20 American Association of Applied Psychology (AAAP), 10–12, 20 American Association of Clinical Psychologists (AACP), 9–10 American Biodyne, 48 American Psychiatric Association, 333 American Psychological Association (APA), 5, 9–22, 31, 33, 45, 126, 321, 324, 341, 355, 374–382 American Psychological Society (APS), 19–25, 341, 380–384 Amygdala, 55, 62–66, 173–176, 189, 265, 281, 328, 336 Anhedonia, 170–171 Anorexia nervosa, 110 Anterior cingulate cortex (ACC), 65, 328, 336–337 Antidepressants, 107–108, 119, 326–327, 339 Antipsychotics, 60
Anxiety, 8, 61, 77–96, 107, 114, 118, 123, 167–190, 197–219, 278–281 Anxiety disorders, see Anxiety Anxiety sensitivity, 92, 170–171 Architecture, mental, 223–254 Assertion training, 114 Association for Psychological Science, see American Psychological Society Associative learning, 78–85, 90–92, 172–178, 197–219, 259–283, 304–311 Attention, 61, 65–66, 173, 175, 189, 224, 278, 282, 296–311, 354 Attributional style, 81, 85 Autism, 272–275, 354–355 Automatic processing, 224–225, 240, 244–247 Avoidance, 55, 61, 79–81, 87–90, 123–124, 148, 172, 188–189, 198–219, 278 B Bayesian approaches to model fitting, 226, 293 Behavior therapy, 18, 41, 78–82, 88–96, 197–219, 321, 329, 339–340, 343–344, 351–352 Behavioral activation, 83, 88–89, 108–109 Behavioral activation system, 189 Benchmarking, 22–23, 93–94, 115–116, 118 Benzodiazepines, 211–212 Big Science, 341
427
Treat - subject index.qxd
3/28/2007
1:01 PM
Page 428
428
SUBJECT INDEX
Binge eating disorder (BED), 110 Borderline personality disorder, 112 Boulder model of training, see Clinical training models Brain plasticity, 262–266, 338–344 Breathing training, 89–91 Bulimia, 110, 112, 121–124 Bupropion, 147–157 C Caffeine, 205, 212–213 California School of Professional Psychology, 15 Capacity of processing, 225–227, 235–253 Capnometry, 91–92 Category learning, 304–311 Causality, 134–136, 140, 143–147, 337 Cerebellum, 262–282 Classical conditioning, 78–81, 173–175, 197–219, 259–283 Clinical experience, 18, 31, 33, 43, 118–119, 125, 325, 333, 358, 375, 378, 385 Clinical judgment, 35, 43, 105, 118–125, 387 Clinical science model of training, see Clinical training models Clinical training models clinical science, 10, 17–25, 167, 322–323, 340–343, 349–350, 356–358, 364–393 experimental psychopathology, 17, 365 integrative psychological science (hybrid, interdisciplinary), 259, 289–290, 322–323, 340–343, 383–384, 390–392 local clinical science (scholar-practitioner), 15–17, 377–379, 387 scientist-practitioner (Boulder), 6, 10, 13–17, 321, 341, 349, 366–367, 371–376, 386 Clinical Trials Network Initiative, 95 Coactive model of processing, see Architecture Cognitive avoidance, 188–189 Cognitive restructuring, 81, 89–91, 390
Cognitive science, 24, 56, 223–254, 259, 282–283, 289–314, 351, 390–392 Cognitive therapy (CT), 80–81, 88–91, 107–110, 292, 295, 310–311, 329 Cognitive-behavioral therapy (CBT), 80–96, 106–126, 329, 339 Committee on Accreditation (CoA), 21–22, 366, 377, 382–383 Community mental health, 15, 22–23, 93–94, 112–119, 323, 340, 343, 352–353, 372 Comorbidity, 111–112, 117, 120, 123, 168–169, 179–180, 184–187, 325, 333, 335, 351 Competencies, clinical, 387 Conditioned fear, see Fear conditioning Context specificity, 138, 197–219, 278–282 Continuing education, 385–386 Controlled processing, 224–225, 240 Council of Graduate Departments of Psychology (COGDOP), 21–22, 382 Council of University Directors of Clinical Psychology (CUDCP), 21–22, 375, 383–384 Craving, 148–159, 214 D Defensive motivational system, 61, 83, 172–190 Depression, 61, 84–88, 94, 107–110, 116–119, 123, 138, 168–190, 326–328, 339, 354–356 Diagnosis, clinical, 114–115, 120–123, 168–172, 178–190, 325, 329, 332–335, 343–344, 350–356 Diagnostic and Statistical Manual (DSM) 114, 120–123, 169–170, 325, 329, 332–335, 343–344, 350–356 Dismantling studies, 107–108, 133, 139–140 Dissemination, treatment, 22–24, 89, 92–96, 108, 111–119, 125–126, 138, 324, 329, 365, 367, 382, 384 Dopamine, 339 Dorsolateral prefrontal cortex, 62–66
Treat - subject index.qxd
3/28/2007
1:01 PM
Page 429
429
SUBJECT INDEX E Early response to treatment, 108–110, 122, 124, 139 Eating disorders, 110–115, 121–124, 296–313 Effectiveness of treatment, 22–23, 92–96, 105–126, 324–326, 329, 334, 352–353, 357–358, 376 Efficacy of treatment, 22, 78–96, 105–126, 197–199, 324–326, 334, 352–353, 357–358, 376, 382 Electroencephalography (EEG), 58–59, 64–65 Emotion regulation, 61–65, 78–93, 336–338, 351, 354 Emotional processing theory, 80, 93, 199 Empirically based principles, 118, 120, 123–125 Empirically supported treatments, see Evidence-based treatments Empirically validated treatments, see Evidence-based treatments Engineer, clinical psychologist as behavioral, 30, 46–48 Epistemic duties, 30–35, 44–45 Epistemology, 30, 33–40 Etiology of clinical disorders, 80–89, 198, 323, 351–356 Event-related brain potential (ERP), 57–60, 64, 331 Evidence-based treatments, 22, 31–32, 44–48, 53–54, 77–96, 105–126, 133–139, 147–159, 197–199, 278–283, 322–326, 329, 334, 343–344, 352–353, 357–358, 364, 368, 375–389 Exclusion criteria, 93–94, 111–114 Experimental psychopathology model of training, see Clinical training models Exposure, 79–81, 88–93, 106–107, 114, 124–125, 198–219, 278–279 Eyeblink conditioning, 259–283 F Falsification, 36–37, 40, 43 Fear, 55, 78–93, 167–190, 197–219, 278
Fear conditioning, 78–81, 87, 91–93, 167–190, 197–219, 278, 281 Felix Consent Decree, 95 Fetal alcohol syndrome, 268–272, 275–281 Flooding, 80, 198 Functional magnetic resonance imaging (fMRI), 58, 61–65, 175–176, 226, 267, 327–328 Funding of research, 95, 330–332, 335, 355–357, 365 G Gene by environment interactions, 327, 354–356 Generalizability of treatment-outcome results, see Effectiveness research Generalized anxiety disorder, 86, 168–190 Genetics, 54–55, 66, 82–86, 323, 327–330, 335–338, 341–343, 353–356 H Habituation, 80–81, 93, 197–219 Hippocampus, 58, 62–63, 189, 265–271, 326, 339 Hybrid model of processing, see Architecture Hybrid model of training, see Clinical training models Hypothalamo-pituatary-adrenal (HPA) axis, 84 I Imagery, emotional, 61, 79, 168–190, 199, 215, 278 Inclusion criteria, 94, 111–114 Individualization of treatment, 111–113, 119–125, 377–378 Integrative Psychological Science (IPS), 197–199, 224, 259–260, 289–290, 383–384, 389–392 Integrative psychological science model of training, see Clinical training models
Treat - subject index.qxd
3/28/2007
1:01 PM
Page 430
430
SUBJECT INDEX
Interdisciplinary Model of training, see Clinical training models Interdisciplinary research, 3, 24–25, 167–168, 197–199, 224, 259–260, 281–283, 289–290, 292–295, 329–332, 341–344, 354–356, 365, 383–384, 389–392 International Hapmap Project, 327 Internship, clinical, 12, 24, 323, 349–350, 354–357, 364–365, 371, 373, 384, 388 Interoceptive conditioning, 81, 87–92 L Level of analysis, 67, 277, 282, 353–354 Licensing, clinical, 3, 19, 45, 357, 366, 371, 375–376, 385–389 Local clinical science model of training, see Clinical training models Localization of function, 55–68 Locus of control, 85 M Magnetoencephalography (MEG), 57–59 Managed health care, 379–380, 388 Manifesto for a science of clinical psychology, 20, 29–49, 53–54, 322–324, 343–344, 349, 381–382 Manualized treatment, 35, 48, 82, 89–90, 105–126, 138–139, 324–325, 352–353, 357 Mean interaction contrast, 248–254 Mechanisms of action, 106–126, 133–159, 199–202, 354 Mediational models and analyses, 85–86, 134–159 Memory, 7, 43, 174, 177–178, 188–189, 201, 209, 215–218, 245, 261–262, 266, 268–269, 300–304 Mental-health services delivery, 33–35, 47–49, 53–54, 95–96, 105–126, 138–139, 329, 341, 352–353, 380–381, 389 Metacognitive awareness, 108–109 Meta-science, 29–42 Model-mimicking, 227, 240, 247
Moderator effects, 85–86, 117, 138, 158 Motivational systems, 61, 83, 172–190 Mowrer’s two-factor model, 79, 278–280 N National Child Traumatic Stress Network, 94 National Comorbidity Survey, 187 National Health Service (NHS), 95–96, 116 National Institute for Clinical Excellence (NICE), 110, 121, 124 National Institute of Mental Health (NIMH), 11–13, 21–24, 289, 324, 327–333, 340–343, 366, 382–384, 390 National Institute on Drug Abuse (NIDA), 66 National Institutes of Health (NIH), 289–290, 323, 329–332, 336, 341–343 Negative affect, 65, 82–86, 108, 123, 168–190, 327, 335–337 Neurogenesis, 326–327 Neuroimaging, 54–68, 353–354, 175–176, 226, 267, 272, 275, 327–328, 330, 335–337, 341–342, 390 Neuroscience, 54–68, 168–190, 226, 259–283, 323–343, 353–356, 390–392 Nicotine dependence, 137–138, 147 Nonspecific influences on treatment outcome, 107, 120, 133, 137, 139–140, 325, 352 Normal science, 38–40 Null hypothesis statistical testing, 137 O Obsessive-compulsive disorder, 80, 87, 198, 278–282 Outshining hypothesis, 209 P P50 component, see Event-related brain potential Panic disorder, 80–94, 115–116, 118, 168–190, 198, 215 Paradigmatic science, 38–40 Parallel model of processing, see Architecture
Treat - subject index.qxd
3/28/2007
1:01 PM
Page 431
431
SUBJECT INDEX Parallel processing, 223–254 Perceptual organization, 296–311 Personal construct theory, 290–291 Personality, see also Negative affect, positive affect, 78, 82–83, 335–337 Personality disorder, 112 PhD in clinical psychology, 10–13, 17, 118, 125, 364, 367, 372, 375, 383–389 Phenotype problem, 335 Point predictions, 37, 137 Positive affect, 82–83, 170, 337 Posttraumatic stress disorder, 93–95, 169–170, 178–179, 329 Practice Research Network, 94–95 Practicum, clinical, 371, 373, 384–385 Prescription privileges for psychologists, 19, 357, 371, 380 Prevention, 329, 340, 344 Processing capacity, see Capacity of processing Processing times, 228, 233–241, 248–249 Protocol, treatment, see Manualized treatment Psychodynamic theory and therapy, 321, 353 Psychological space, 296–311 Psychopharmacology, 323, 328–329, 338, 343, 353–357 Psychophysiology, 57–61, 80–84, 91–92, 167–190, 202–210, 260–283, 329–330, 342 PsyD in clinical psychology, 15–17, 375–379 Q Quality improvement, 30–35, 46–49, 366–367, 379, 387–388 R Randomized controlled trials (RCTs), 82, 93–94, 105–126, 135, 324, 353, 376 Rebirthing therapy, 30, 45 Reductionism, 55–68, 323, 328 Relaxation, 79–80, 107, 198 Retrieval cues, 201–202, 214–215 Return of fear, 197–219 Roadmap of NIH, 332
S Schema-focused therapy, 109–110 Schizophrenia, 55–68, 275–282, 354–357 Schizotypy, 63–65 Scholar-practitioner model of training, see Clinical training models Science and practice, 4, 13–20, 29–49, 53–54, 77–96, 105–126, 133–134, 138–9, 167, 197–199, 321–326, 329, 343–344, 351–353, 357–358, 371–377, 380–381, 385–386 Science Directorate, 381 Scientist-practitioner model of training, see Clinical training models Selective serotonin reputake inhibitors (SSRIs), 327 Sensory gating, 56–60 Serial model of processing, see Architecture Serial processing, 223–254 Serotonin transporter (5–HTT) gene, 327–328, 335–338, 354 Sexual aggression, 296, 298, 300, 309–310 Signal detection theory, 301–302 Silvio O. Conte Centers for the Neuroscience of Mental Disorders, 330–331 Small theory, 18 Smoking, 133–159 Social information processing, 290–291 Social phobia, 87–94, 168–190 Society for a Science of Clinical Psychology (SSCP), 20, 321–322 Specific phobia, 78–79, 83, 87, 93, 168–190, 197–219 Startle reflex, 168–190 State-dependent learning, 205, 210–213 Stopping or decision rule, 231–252 Structural magnetic resonance imaging (sMRI), 57–65, 328 Substance Abuse and Mental Health Services Administration (SAMHSA), 94–95 Supervision, clinical, 105, 117–119, 373, 385–386, 389 Systematic desensitization, 106–108, 197–198, 219
Treat - subject index.qxd
3/28/2007
1:01 PM
Page 432
432
SUBJECT INDEX T
Task Force on the Promotion and Dissemination of Psychological Procedures, 22, 324, 382 Technologist, clinical psychologist as, 30, 46–49 Temperament, 337 Therapeutic alliance, 107, 120, 325, 352 Therapist effects on treatment outcome, 117–118 Thought disorder, 55 Timing deficits, 271–277 Transdiagnostic approach to diagnosis, 122–123, 351, 356 Translational research, 79, 96, 168, 197–199, 259–260, 267, 277, 283, 289–290, 294–295, 321–344, 352–356, 390–391 Treatment dissemination, see Dissemination, treatment Treatment manual, see Manualized treatment
Treatment of Adolescents with Major Depression (TADS), 118–119 Tripartite model, 169 Triple vulnerabilities model, 82–89 Two-headed clinician, 15–16, 373 U Uncontrollability, 83–88, 92, 200 Uniformity myth, 374 Unpredictability, 83–84, 87–88 V Veterans Administration (VA), 11–12, 372 Virtual reality exposure therapy, 92–93 W Workload, task, 227, 237–247