sports medicine Volume 39 Issue 9 September 2009

Sports Med 2009; 39 (9): 697-708 0112-1642/09/0009-0697/$49.95/0

LEADING ARTICLE

ª 2009 Adis Data Information BV. All rights reserved.

Cervical Spine Injuries in American Football Jeffrey A. Rihn,1 David T. Anderson,1 Kathleen Lamb,2 Peter F. Deluca,1 Ahmed Bata,1 Paul A. Marchetto,1 Nuno Neves3 and Alexander R. Vaccaro1 1 The Rothman Institute, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania, USA 2 Jefferson Medical College, Philadelphia, Pennsylvania, USA 3 Orthopaedic Department, Hospital Sa˜o Joa˜o-Porto Medical School, Alameda Prof. Hernaˆni Monteiro, Porto, Portugal

Abstract

American football is a high-energy contact sport that places players at risk for cervical spine injuries with potential neurological deficits. Advances in tackling and blocking techniques, rules of the game and medical care of the athlete have been made throughout the past few decades to minimize the risk of cervical injury and improve the management of injuries that do occur. Nonetheless, cervical spine injuries remain a serious concern in the game of American football. Injuries have a wide spectrum of severity. The relatively common ‘stinger’ is a neuropraxia of a cervical nerve root(s) or brachial plexus and represents a reversible peripheral nerve injury. Less common and more serious an injury, cervical cord neuropraxia is the clinical manifestation of neuropraxia of the cervical spinal cord due to hyperextension, hyperflexion or axial loading. Recent data on American football suggest that approximately 0.2 per 100 000 participants at the high school level and 2 per 100 000 participants at the collegiate level are diagnosed with cervical cord neuropraxia. Characterized by temporary pain, paraesthesias and/or motor weakness in more than one extremity, there is a rapid and complete resolution of symptoms and a normal physical examination within 10 minutes to 48 hours after the initial injury. Stenosis of the spinal canal, whether congenital or acquired, is thought to predispose the athlete to cervical cord neuropraxia. Although quite rare, catastrophic neurological injury is a devastating entity referring to permanent neurological injury or death. The mechanism is most often a forced hyperflexion injury, as occurs when ‘spear tackling’. The mean incidence of catastrophic neurological injury over the past 30 years has been approximately 0.5 per 100 000 participants at high school level and 1.5 per 100 000 at the collegiate level. This incidence has decreased significantly when compared with the incidence in the early 1970s. This decrease in the incidence of catastrophic injury is felt to be the result of changes in the rules in the mid1970s that prohibited the use of the head as the initial contact point when blocking and tackling. Evaluation of patients with suspected cervical spine injury includes a complete neurological examination while on the field or the sidelines. Immobilization on a hard board may also be necessary. The decision to obtain radiographs can be made on the basis of the history and

Rihn et al.

698

physical examination. Treatment depends on severity of diagnosed injury and can range from an individualized cervical spine rehabilitation programme for a ‘stinger’ to cervical spine decompression and fusion for more serious bony or ligamentous injury. Still under constant debate is the decision to return to play for the athlete.

American football has long been recognized as an activity that places players at risk for cervical spine injuries with potential for neurological deficits. Cervical spine injuries affect players of all positions and levels of play, and can range in severity from a cervical strain with no neurological deficit to a catastrophic cervical fracture dislocation with a complete neurological deficit. Significant efforts have been made throughout the past few decades to minimize the risk of cervical injury and improve the management of injuries that do occur. These changes include changes in technique, changes in the rules of the game and improvements of the on- and off-field care of the injured athletes. Despite these efforts, however, cervical spine injuries remain a serious concern in the game of American football. The focus of this article is to review the demographics, mechanism, evaluation and treatment of cervical spine injuries with associated neurological deficit that occur in American football players. The injuries that will be covered include the ‘stinger’, cervical cord neuropraxia and catastrophic neurological injury. A literature search was performed using PubMed. Our search period dated back to 1970. Searches included combinations of the following terms: ‘cervical spine injury’, ‘American football’, ‘burners’, ‘stingers’, ‘neuropraxia’, ‘transient quadriplegia’, ‘spinal cord injury’, ‘catastrophic spine injury’, ‘diagnosis’, ‘management’, ‘treatment’ and ‘return to play’. Articles written about rugby and soccer players were excluded. Also excluded were articles written about the paediatric population. Athletes included high school through to professional level. Case reports were also excluded. The literature we used was chosen based on its utility in providing detailed explanations of the spectrum of injuries we wished to highlight. ª 2009 Adis Data Information BV. All rights reserved.

1. Root/Brachial Plexus Neuropraxia One of the most commonly occurring injuries in American football players is neuropraxia of the cervical nerve root(s) or brachial plexus, which is commonly referred to as a ‘stinger’. This injury represents a reversible peripheral nerve injury of the upper extremity that results from a temporary physiological block in nerve conduction. It has been reported to occur in 50–65% of players over a 4-year collegiate career, and it most commonly occurs in linemen, defensive ends and linebackers.[1] These ‘stinger’ or ‘burner’ injuries are characterized by unilateral burning pain radiating from the neck, down the arm to the hand. If the pain occurs in the bilateral upper extremities, further evaluation should be performed to rule out spinal cord injury.[2] The pain usually lasts seconds to hours, and rarely beyond a 24-hour period. Players may experience associated weakness of the deltoid and/or spanati that typically resolves within 24 hours to 6 weeks following the injury.[1] On physical examination, patients who have a stinger usually demonstrate pain-free active and passive range of motion of the neck and have no tenderness to palpation of the cervical spine or surrounding soft tissues. Three mechanisms of injury have been described for the stinger. Injury can occur by compression of the cervical nerve root due to hyperextension, often with lateral flexion of the neck and an axial load. This can result in compression of the nerve root by narrowing of the intervertebral foramen. In addition, Penning[3] described the ‘pincer mechanism’, in which there is some degree of spinal cord compression by the posterior-inferior margin of the superior vertebral body and the anterior-superior portion of the lamina of the vertebra below. The mechanism of compression of the cervical nerve root is often Sports Med 2009; 39 (9)

Cervical Spine Injuries in American Football

suggested by a positive Spurling’s test on physical examination.[1] Meyer et al.[4] reported that 85% of players with stingers had a mechanism of extension-compression. Levitz et al.[5] echoed these results as 83% of stinger patients in their study suffered extension-compression injuries. These patients tend to have more severe symptoms and tend to be older, with radiographic evidence of cervical disk degeneration and spondylosis. From his study of 55 patients with recurrent burners, Levitz et al.[5] reported that 53% had a narrowed cervical canal and 87% had evidence of degenerative disk disease. Another commonly reported mechanism of injury is that of traction or stretching of the brachial plexus due to depression of the ipsilateral arm and lateral bending of the head toward the unaffected side. As opposed to the pincer mechanism, these patients are often younger without prior cervical spine injury or cervical spondylosis.[1] The Spurling’s test is usually negative in these patients as the injury is not a result of nerve compression. Loss of cervical range of motion and neck pain is not characteristic of this mechanism of injury.[6] The third mechanism of injury is that of direct trauma to the brachial plexus that results from the shoulder pads impinging on Erb’s point, where the brachial plexus is most superficial.[7] Initial evaluation of patients with suspected nerve root or brachial plexus neuropraxia includes a complete neurological examination while on the field or the sidelines. Patients who have loss of consciousness, neck pain or any evidence of a neurological deficit on the field should have their cervical spine stabilized prior to transportation off the field. Once the player is removed from the field, a more thorough history and physical examination can be conducted, with attention given to the direction and mechanism of injury. Although cervical radiographs are often obtained in the setting of a stinger injury, the radiographs are usually normal. Five to 10% of players with stinger symptoms have more serious injuries with prolonged neurological deficits.[1] Patients with symptoms persisting beyond 2 weeks should be evaluated for the possibility of injury to the spinal cord. Anteriorposterior and flexion-extension radiographs ª 2009 Adis Data Information BV. All rights reserved.

699

should be evaluated for fracture or gross instability. MRI evaluation serves to rule out significant ligamentous injury, disk herniation or cervical stenosis. There is some controversy as to the usefulness of electromyography (EMG), although this study may show a mild conduction block demonstrated by positive sharp waves and fibrillation potentials, demonstrating membrane instability due to axonal damage.[8] The origin of injury can be identified with EMG as either the nerve root or the brachial plexus itself. Although Hershman et al.[9] noted its limited utility, more contemporary management guidelines suggest EMG may have a role in determining return to play, as fibrillation potentials or moderate sharp waves in the presence of objective weakness indicate the need to refrain from participation in play.[8] Researchers have noticed a relationship between stingers and cervical stenosis. Meyer et al.[4] reported that 47% of university football players with known stinger injuries had concomitant cervical stenosis. Patients with a congenitally narrow spinal canal have shortened pedicles and narrow intervertebral foramen. This foraminal narrowing is felt to increase the likelihood of nerve compression and cause a stinger injury.[6] Cervical stenosis of the spinal canal can be identified using lateral radiographs. Measurement of the distance from the mid-point of the posterior aspect of the vertebral body to the nearest point on the spinolaminar line is considered the diameter of the spinal canal. Normal spinal canal diameter at the levels of C3–C7 ranges from 14 mm to 23 mm.[10] Cervical stenosis is defined by a canal diameter of £13 mm.[10,11] In 1986 Torg et al.[12] and in 1987 Pavlov et al.[13] suggested the ratio method as a more accurate analysis of evaluation for canal stenosis that eliminates the variation and errors in measurement associated with simply measuring the diameter of the spinal canal. The ratio method measures the distance from the midpoint of the posterior aspect of the vertebral body to the nearest point on the spinolaminar line, divided by the sagittal diameter of the vertebral body. The normal ratio in asymptomatic controls was 1, while a ratio of £0.8 indicated cervical stenosis Sports Med 2009; 39 (9)

Rihn et al.

700

A B

Fig. 1. A lateral cervical spine radiograph depicting the measurements for cervical spinal stenosis. Line ‘A’ indicates the anteroposterior dimension of the cervical spinal canal, measured from the posterior vertebral body to the spinolaminar line. Line ‘B’ indicates the anteroposterior width of the vertebral body. The Torg ratio is defined as A/B. Cervical spinal stenosis has been defined by a cervical spinal canal diameter of £13 mm or a Torg ratio of <0.8.

(figure 1).[12,13] A reported limitation of the ratio method is that football players tend to have increased vertebral body size compared with the general population. This leads to a reduced Torg ratio despite an adequate canal diameter, thus over-reporting the incidence of cervical stenosis in this population.[12-15] Treatment of stinger injuries involves an individualized and comprehensive cervical spine rehabilitation programme that is sufficiently aggressive without causing further injury to the patient. The step-wise goals of therapy are to provide protection to the neck and injured nerves, control pain, correct imbalances in strength and flexibility, correct posture, re-condition the patient and prevent further or recurrent injury prior to return to play.[16] 2. Cervical Cord Neuropraxia Cervical cord neuropraxia, also referred to as transient quadriplegia, was first described in 1986 ª 2009 Adis Data Information BV. All rights reserved.

by Torg et al.[12] as the clinical manifestation of neuropraxia of the cervical spinal cord due to hyperextension, hyperflexion or axial loading. It is characterized by temporary pain, paraesthesias and/or motor weakness in more than one extremity with a rapid and complete resolution of symptoms and a normal physical examination within 10 minutes to 48 hours after the initial injury. Torg et al.[12] reported an incidence of cervical cord neuropraxia of 1.3 per 10 000 1984 National Collegiate Athletic Association (NCAA) football players with transient weakness with paraesthesias and an additional 6 per 10 000 players with transient paresthesias, totaling an estimated 7 per 10 000 players with signs and symptoms of cervical cord neuropraxia per year in this athletic population. Boden et al.[17] more recently reported on 23 high school and 20 collegiate football players who sustained cervical cord neuropraxia between the years of 1989 and 2002. This translated into 3.3 cervical cord neuropraxia injuries per year and 0.17 cervical cord neuropraxia injuries per 100 000 participants at the high school level, and 1.5 per year and 2.05 per 100 000 participants at the collegiate level. The rate of cervical cord neuropraxia reported by Boden et al.[17] is considerably lower than that documented by Torg et al.,[12] which is likely reflective of rule changes that have been implemented in high school and collegiate football. The mechanism of injury that typically results in cervical cord neuropraxia includes hyperextension or hyperflexion of the cervical spine resulting in a temporary physiological conduction block within the spinal cord. Torg et al.[18] explain the temporary disruption of cervical spinal cord function as the result of local anoxia and an increase in intracellular calcium that results from cord compression at the time of the injury. Extension-compression injuries (i.e. Penning’s[3] ‘pincer mechanism’) are felt to be responsible for the most severe spinal cord compression in the anteroposterior direction. Hyperextension injuries may be further exacerbated by infolding of the ligamentum flavum, which can cause an additional 30% decrease in the anteroposterior diameter of the spinal canal.[6,19] Hyperflexion injury is felt to result from impingement of the spinal Sports Med 2009; 39 (9)

Cervical Spine Injuries in American Football

cord by the superior vertebra and the anterior superior lamina of the subjacent vertebra.[2] The degree of compression depends on the sagittal diameter of the spinal canal, presence of any spondylotic changes and the degree of soft tissue hypertrophy and infolding.[3,20] Symptoms of neuropraxia include burning pain, tingling and loss of sensation in more than one extremity, with no associated neck pain other than a burning sensation. The motor changes range from no or partial weakness to complete paralysis.[2,20] Frequently, however, motor and sensory symptoms will coexist in patients. Symptoms are completely transient and resolve within 10 minutes to 48 hours following the injury.[6,12,17] The duration of symptoms in the study of Boden et al.[17] was documented in 12 of the 43 athletes with cervical cord neuropraxia and included the following: <15 minutes in five patients, between 15 minutes and 24 hours in five patients and >24 hours in two patients. The evaluation of patients with cervical cord neuropraxia includes a complete history and physical examination with particular attention given to the neurological examination and the mechanism of injury, including the direction of the traumatic force. Anteroposterior, lateral and open mouth odontoid radiographs are performed initially to rule out any obvious fracture or dislocation. In the absence of neck pain, flexion and extension cervical radiographs are obtained to rule out instability. If the patient does have neck pain, flexion and extension radiographs should not be performed until a cervical MRI rules out any ligamentous injury. Radiographic findings of fracture or dislocation are usually absent in cases of cervical cord neuropraxia. Often seen on radiographs, however, are signs of congenital cervical stenosis and congenital abnormalities such as Klippel-Feil syndrome.[2] Furthermore, intervertebral disk disease, acquired cervical stenosis and/or cervical instability (defined as >3.4 mm of anterior-posterior translation or angulation of 11 between adjacent vertebrae on a lateral view) are relatively common radiographic findings in patients with neuropraxia.[20,21] The mechanism of injury that causes cervical cord neuropraxia is usually the ‘pincer mechanism’, in ª 2009 Adis Data Information BV. All rights reserved.

701

which hyperextension with an associated axial load to the cervical spine results in spinal cord compression. MRI allows a more thorough evaluation of the discoligamentous complex, the spinal cord itself and the cerebrospinal fluid surrounding the spinal cord. The cerebrospinal fluid space has been referred to as the functional reserve of the spinal canal, and the absence of this space is felt to place the spinal cord at risk for injury in the event of a trauma.[10,22,23] MRI or CT myelogram must be performed in any patient demonstrating an abnormal neurological examination, signs or symptoms of neuropraxia, or translation >3 mm on x-ray. These findings are suggestive of intrinsic or extrinsic cord compression, nerve root compression and/or ligamentous injury. Stenosis of the spinal canal, whether congenital or acquired, is thought to predispose the athlete to cervical cord neuropraxia.[12,14] In 1996, Torg et al.[14] published an analysis of 45 football players who suffered from cervical cord neuropraxia. This analysis included cohorts of asymptomatic football players, asymptomatic non-athlete males, football players who had an episode of cervical cord neuropraxia and football players who sustained cervical injuries that resulted in permanent quadriplegia. These authors reported that 93% of football players with cervical cord neuropraxia had a Torg ratio of <0.8.[14] The notion that football players have increased vertebral body size compared with the general population, however, suggests that the Torg ratio may overestimate the presence of significant spinal stenosis in this population.[14] This notion is supported by Odor et al.[24] and Herzog et al.,[25] who reported that 34% and 49% of asymptomatic professional football players, respectively, had a Torg ratio of <0.8. Torg et al.[14] reported that 48% of asymptomatic football players and only 12% of asymptomatic nonathletes had Torg ratios of <0.8 and that the positive predictive value of football players with a Torg ratio of <0.8 for sustaining cervical cord neuropraxia is only 0.2%. It is, therefore, not recommended that the Torg ratio be used as a factor when determining whether or not an athlete should be allowed to participate in football.[14] Sports Med 2009; 39 (9)

Rihn et al.

702

N

XIO

R FLE

Fig. 2. A flexion lateral cervical radiograph demonstrating the mechanism of cervical injury that occurs with spear tackling. When the neck is flexed, as is the case when a player spear tackles, contact is initiated with the crown of the helmet. With the neck in a flexed position, the alignment of the cervical spine is straight to slightly kyphotic, as demonstrated in the lateral cervical radiograph. Almost the entire applied load (black arrow) is axially applied to the cervical spinal column. When this load is excessive, the cervical spinal column fails in a flexion and compression (see figure 3).

manent quadriplegia, they are not as common as injuries of the subaxial cervical spine.[17,28,33,34] Combined injuries of the upper cervical and subaxial cervical spine have also been reported.[17] This must be considered when evaluating the patient radiographically. The majority of cervical cord injuries occur during actual games and are most common in defensive players.[14,17,28,32,34] Over the last 30 years, defensive players have sustained 70.3% of the 269 permanent neurological injuries reported.[28,34] Defensive back is the position at highest risk of injury, representing 35.3% of all spinal cord lesions reported, followed by linebackers (9.7%) and special teams players (8.2%). The overwhelming majority of the injuries occur while tackling or blocking.[28,34] Since the first intercollegiate American football game between Princeton and Rutgers in 1869, repeated attempts have been made to decrease

3. Catastrophic Neurological Injury It has been estimated that each year approximately 11 000 neck injuries that are sustained while playing American football present to emergency departments in the US.[26] Fortunately, permanent neurological injury and death as a result of these injuries are infrequent.[27,28] Athletic activity is the fourth most common overall cause of spinal cord injuries after motor vehicle accidents, violent crime and falls, and the second most common cause in the first three decades of life.[29-31] The level of risk is particularly high in contact sports like American football and hockey or high-energy sports such as gymnastics.[8] In American football, the mechanism of catastrophic cervical injury is most often a forced hyperflexion injury, as occurs when ‘spear tackling’.[17,28,32] Spear tackling refers to a technique of tackling or blocking in which the player initiates contact with the crown of the helmet, with the neck in a slightly flexed position (figure 2). Axial force applied to the helmet is transmitted to the cervical spine, and the subaxial spine fails in flexion in the form of a fracture and/or subluxation or dislocation (figure 3). Although upper cervical spine injuries (i.e. occiput to C2) have been reported to occur in football and result in perª 2009 Adis Data Information BV. All rights reserved.

Fig. 3. Lateral radiograph of a 21-year-old collegiate defensive end who sustained a cervical injury when making a tackle. This represents a flexion compression injury at the C4-5 level, with compression of the C5 vertebral body and a bilateral C4-5 facet dislocation. He underwent an anterior C5 corpectomy and fusion and a posterior spinal fusion from C4 to C6 within 24 hours of the injury.

Sports Med 2009; 39 (9)

Cervical Spine Injuries in American Football

the frequency of death and catastrophic injury sustained in the course of play.[35] In 1964, Schneider[36] reported that 56 cervical fractures or dislocations occurred during the years 1959–63 (1.4/100 000), 30 of which resulted in permanent quadriplegia (0.73/100 000). The National Collegiate Athletic Association and the National Federation of State High School Associations have collected injury data since the 1970s, which has been reported as part of the Annual Survey of Football Injury Research since 1980.[30] The primary purpose of these efforts was to make the game of football safer for the participants. In 1975, the National Football Head and Neck Injury Registry was established in order to document the extent of head and neck injuries. In 1979, Torg et al.[37] retrospectively gathered information dating back to 1975 and found an increase in the number of players rendered quadriplegic.[37] The cause for this increase was attributed to better protective capabilities of modern helmets. These helmets decreased the incidence of head injuries, but also led to the use of the helmet as a primary point of contact in blocking, tackling and head butting, placing the cervical spine at risk (i.e. spear tackling).[37] Fifty-two percent of the cervical injuries with associated permanent quadriplegia were the result of spear tackling or direct force applied to the helmet.[37] Although hyperflexion[36,38,39] and hyperextension[40,41] have traditionally been implicated in the development of cervical spine injuries, pathological, biomechanical and cinematographic analyses have determined that axial loading is the primary mechanism of football related injuries.[42,27] As a consequence of these findings, the National Collegiate Athletic Association and the National Federation of State High School Associations implemented major rule changes in 1976 prohibiting the use of the head as the initial contact point when blocking and tackling; these rules were later supported by the American Football Coaches Association Ethics Committee.[43] These changes had a dramatic effect on the incidence of permanent cervical quadriplegia. According to the National Football Head and Neck Injury Registry[27,37,44] in 1975, prior to the ª 2009 Adis Data Information BV. All rights reserved.

703

rules change, there were a total of 89 cervical fractures or dislocations (6.65/100 000 and 29.3/100 000 at the high school and college levels, respectively) resulting in 28 cases of quadriparesis (2.1/100 000 and 8.0/100 000 at the high school and college levels, respectively). By 1987, the number of cervical fractures and dislocations had fallen to 22 (2.31/100 000) at high school and eight (10.66/100 000) at college level. The number of cases of permanent quadriplegia had fallen to eight (0.73/100 000) at the high school level and none at the collegiate level.[44,27] The most recent report from The National Center for Catastrophic Sport Injury Research and the Annual Survey of Catastrophic Football Injuries show that football has been responsible for 269 permanent cervical spinal cord injuries since 1977.[28,34] High school players represent the vast majority of these injuries accounting for 222 cases, college players for 33, recreational players for 5, and professional players for only 9.[28,34] When per-participant incidence is considered, however, there is an increase in incidence with the level of player. These data also showed a dramatic decrease in the number of injuries since the reports from the 1960s and early 1970s.[28,34] During the 1990s, the average figure was 6.9 injuries per year and a mean incidence over the last 30 years of 0.52/100 000 participants at high school level and 1.47/100 000 at college level.[28,34] Nevertheless, in four of the last eight seasons, the number of permanent spinal cord lesions has surpassed ten cases.[28,34] The incidence in 2006 is above average, especially at the collegiate level (i.e. 2.66/100 000).[28,34] Recent changes in the rules have been implemented to hopefully decrease the incidence of catastrophic injury. In 2005 the word ‘intentional’ was dropped from the spearing rule, so that now no forms of spearing, intentional or not, are permitted. Additional rule changes effective from 2006 and 2007 include the requirement of at least a 4-point chinstrap to secure the helmet, coloured mouth guards and revision of illegal helmet contact rules.[28,34] Continuous surveillance of sustained injuries will help us to realize the impact of these recent rule changes on the incidence of catastrophic cervical spine injuries. Sports Med 2009; 39 (9)

Rihn et al.

704

4. Early Evaluation and Management Evaluation and management of a football player with a cervical spine injury begins on the field. In fact, there should be extensive pre-event planning and preparations for dealing with these types of injuries. A standard protocol for prehospital care, proper organization of equipment and a trained group of personnel are essential to ensure the best care for the injured athlete. Basic sideline equipment should always include a spine-board, stretcher, tools necessary to remove protective gear and maintain cervical spine immobilization, and items for airway management and cardiopulmonary resuscitation. In addition, efficient communication with the emergency room will optimize care.[45] As with any traumatic injury, the primary objective in early management is to address any life-threatening conditions and to prevent further injury. The ABCDE approach should be utilized. Airway is first assessed while maintaining cervical spine stability. Breathing and ventilation are next assessed, following by circulation and disability (neurological status). Lastly, the athlete should be exposed for the secondary survey (although this usually doesn’t occur until the player is in the emergency room). Findings during the primary survey should be addressed and appropriate resuscitation efforts should be made. During the on-field evaluation, the athlete’s helmet and shoulder pads should remain in place with immobilization of the cervical spine.[46] In fact, the helmet and shoulder pads serve to provide support and alignment to the injured cervical spine.[45] It may be necessary to remove the face mask for airway control. This should be accomplished using the proper face mask shears or bolt cutters. A quick on-field history and examination should reveal if the athlete is experiencing unilateral or bilateral arm pain, neck pain, weakness or paraesthesias. Transportation to a medical facility is required for the player with an altered mental status, neck pain or tenderness, limited cervical motion or any neurological symptoms suggesting a spinal cord injury.[2] The helmet and shoulder pads should only be removed once the patient is in a controlled setting ª 2009 Adis Data Information BV. All rights reserved.

and in the hands of personnel trained in such procedures. Once the athlete’s care has been transferred from the on-field team to the emergency room medical staff, the primary and secondary survey will be repeated, including a full neurological examination. At this time, cervical anteroposterior, lateral and odontoid radiographs should be obtained. If plain radiographs are inadequate (i.e. the C7-T1 level cannot be visualized), than a cervical CT should be performed. If warranted, a consultation with a spine surgeon should be quickly obtained. Any signs or symptoms suggestive of a spinal cord injury warrant a cervical MRI. Treatment of cervical spine injury depends on the spectrum of symptoms and the presence of fracture, dislocation, ligamentous injury or spinal cord injury. Although athletes with unstable cervical spine injuries require surgical treatment, the majority of athletes suffer from stable cervical injuries that can be treated conservatively, with rehabilitation. A thorough rehabilitation programme allows for restoration of motion, posture, strength and the prevention of further injury. Most programmes follow a protocol starting with isometric exercises. Forces applied to the head without motion are followed by concentric resistive programmes, gradually allowing greater degrees of neck motion. Progression should be slow and extra care should be taken to avoid the return of pain. Stretching exercises should be avoided in the early inflammatory phase, as they may exacerbate muscle spasm and stiffness. Only after a full painless arc of motion is attained should the athlete partake in eccentric muscle strengthening.[2] Long a controversial topic, the use of glucocorticosteroids in the treatment of acute spinal cord injury is guided by the National Acute Spinal Cord Injury Study (NASCIS) I and II in 1985 and 1990, respectively.[47,48] Although NASCIS I showed no improvement, analysis of NASCIS II using a higher dose of methylprednisolone (i.e. 30 mg/kg IV over 15 minutes, followed by a 45-minute rest and then 5.4 mg/kg/h for 23 hours) showed that the group receiving methylprednisolone within 8 hours of injury had Sports Med 2009; 39 (9)

Cervical Spine Injuries in American Football

significant neurological score improvement. Interestingly, methylprednisolone given after 8 hours of injury was actually found to cause significant worsening of neurological score. The authors also conceded that the improvements in neurological score did not equate with improvements in functional status. Additionally, patients with coexisting life-threatening injuries and spinal cord injuries secondary to gun-shot wounds were excluded. NASCIS I showed increased rate of infection with corticosteroids whereas NASCIS II did not. Bracken et al.,[49] in a more recent study, showed that patients who receive high-dose corticosteroids between 3 and 8 hours after injury had improved neurological outcomes when treated with 48 hours of corticosteroids rather than 24 hours. This group of patients had higher rates of severe sepsis and severe pneumonia, but otherwise did not show increased complications or mortality. The authors concluded that patients with acute spinal cord injury who present within 3 hours of their injury should be treated with the high-dose methylprednisolone regimen (i.e. 30 mg/kg loading dose followed by 5.4 mg/kg/h) for 24 hours. Patients who present between 3 and 8 hours after their spinal cord injury should be maintained on the high-dose corticosteroid regimen for a total of 48 hours. Administration of high-dose corticosteroids is not indicated in the patient with life-threatening co-morbidity and may be harmful if given after 8 hours of injury.[49] The results of the NASCIS studies and the efficacy of methyprednisolone as a neuroprotective agent following acute spinal cord injury have been heavily debated over the years. Many authors feel as though high-dose methylprednisolone causes more harm than good and that it should not be administered in the setting of acute spinal cord injury.[50-52] This remains a controversial topic, although many spinal cord injury centres continue to treat acute spinal cord injury patients according to the above protocol.[53] 5. Return to Play Despite significant efforts to develop guidelines for return to play for the spectrum of cerviª 2009 Adis Data Information BV. All rights reserved.

705

cal spine injury, the issue of return to play remains controversial. Currently, no set of guidelines for return to play is agreed upon. This issue is often complicated by extrinsic pressures placed on the physician from coaches, players, families and other involved parties. Players with injuries desiring to return to play must be evaluated thoroughly to minimize the risk of recurrent injury. Evaluation includes a detailed history and physical examination and a complete neurological examination. The patient must be able to demonstrate a full, painless cervical range of motion and have no evidence of neurological deficit prior to returning to play.[20] The decision of when, if at all, to return an athlete to contact sports is often difficult. This decision should be based on the mechanism of injury, objective findings by clinical examination and radiographic evaluation, extent of treatment required (i.e. surgical vs nonsurgical) and the ability of the athlete to successfully complete a comprehensive rehabilitation programme.[54] Traditionally, the athlete sustaining a stinger may return to play when the paraesthesias resolve and full strength and painless full range of neck motion are appreciated.[2] It is essential that there is no pain in the neck with motion. If this criterion is not met, the athlete should be immobilized and excluded from activity until radiographs and possibly an MRI are obtained and a definitive diagnosis is reached. After return to play, properly fitting protective equipment can help prevent recurrent injury.[42] The risk of sustaining a recurrent compressive injury can be minimized by wearing a thermoplastic total contact neckshoulder-chest orthosis under properly fitting shoulder pads. Although it doesn’t protect against a compressive injury, a U-shaped neck roll can protect the neck by limiting extreme rangeof-motion and preventing hyperextension and excessive lateral bending.[2] Return to contact sports following an episode of cervical cord neuropraxia is a highly debated issue. Boden et al.[17] reported on 76 athletes who sustained permanent quadriplegia playing football. Forty-six were available for questioning, 38 of whom reported no previous injury and eight of whom reported a previous stinger. None of the Sports Med 2009; 39 (9)

706

athletes reported a previous episode of cervical cord neuropraxia. Of the 43 players in the series of Boden et al.[17] who sustained cervical cord neuropraxia, two reported a prior episode of cervical cord neuropraxia and two reported a prior stinger. Torg et al.[14] surveyed 77 athletes who had sustained permanent quadriplegia playing football, none of whom reported a previous episode of cervical cord neuropraxia. These authors also studied 45 athletes who had sustained an episode of cervical cord neuropraxia, none of whom had a subsequent injury resulting in permanent neurological deficit. In another study, Torg et al.[55] reported on 110 athletes who sustained cervical cord neuropraxia, 87% of which occurred playing football, with 105 of these athletes available for an average follow-up of 3.3 years. No permanent neurological deficit occurred as a result of their episode of cervical cord neuropraxia, and no permanent neurological deficit occurred subsequently in those who returned to contact sports. Of those that returned to contact sports, however, 56% had a recurrent episode of cervical cord neuropraxia. Decreased space for the cervical spinal cord (i.e. cervical stenosis) strongly correlated with the risk of experiencing a recurrent episode of cervical cord neuropraxia. These data suggest that an episode of cervical cord neuropraxia does not increase the risk of sustaining subsequent permanent quadriplegia, but does increase the risk of having subsequent episodes of cervical cord neuropraxia with return to contract sports, particularly in athletes with cervical stenosis. For this reason, players at an elevated level of play (professional) who have an episode of cervical cord neuropraxia are often permitted to return to contact sports, but warned of the increased risk of recurrence. Although return to play following cervical cord neuropraxia is controversial, many agree that relative contraindications to return to play include recurrent episodes of cervical cord neuropraxia, an episode of cervical cord neuropraxia with symptoms lasting >24 hours and cervical cord neuropraxia with associated congenital or acquired spinal stenosis.[20,54,56-58] An absolute contraindication to return to play is an episode of cervical cord neuropraxia that is associated with congenital spinal anomalies (e.g. Klippel-Feil), spinal inª 2009 Adis Data Information BV. All rights reserved.

Rihn et al.

stability, ligamentous injury, persistent neck pain or loss of motion and/or oedema in the spinal cord.[20,54,56-58] In general, congenital anomalies of the upper cervical spine, including os odontoideum, odontoid hypoplasia or aplasia and atlantooccipital fusion, are an absolute contraindication to participation in contact sports. Bailes et al.[59] created three prognostic categories based on characteristics of the cervical spine injury. Type I injuries included those with permanent spinal cord injury, cord haemorrhage, cord contusion or swelling on MRI. These players are not to return to contact sports. Type II injuries included transient neurological symptoms referable to the cervical cord. Neurological examination and radiographs are normal. Diagnoses in this group include brachial plexopathy, burning hands syndrome and cervical cord neuropraxia. As long as there are no neurological deficits or radiographic abnormalities, players may return to play. Type III injuries include those vertebral column injuries only demonstrated on radiographs in a patient with a normal neurological examination. Unstable fractures or dislocations requiring bracing or surgery should not return to play. Any player with an injury requiring atlantoaxial fusion is restricted from play. White et al.[21] devised a set of principles to follow for subaxial injuries. If any of the following criteria are met, the player’s cervical spine can be considered unstable and may require surgical stabilization: combined disruption of the anterior and posterior elements, 3.5 mm of horizontal segment displacement or more than 11 of angulation between adjacent levels. Patients with non-tender, healed, stable compression fractures or spinous process fractures may return to play. As with all injuries, any neurological examination abnormalities or painful motion should restrict an athlete from return to play. An athlete who sustains a disk herniation may require an anterior disectomy with interbody fusion. If limited to one or two levels, this operation is not a contraindication to eventual return to play given that the segments above and below the fusion are normal.[58] The athlete can return to play when symptoms have resolved, the graft is radiographically well incorporated and the athlete exhibits a full painless range of neck motion Sports Med 2009; 39 (9)

Cervical Spine Injuries in American Football

and full strength. Although surgical options provide the opportunity to return to sport in patients with focal disease, Maroon et al.[23] have reported a 25% increase in adjacent segment disk herniation within 10 years following a single level fusion. 6. Summary American football is a high-energy contact sport that places participants at risk of a variety of injuries, including those of the cervical spine. Injuries to the cervical spine range from a simple sprain to a catastrophic injury with permanent quadriplegia. Over the last three decades, improved understanding of injury mechanism and subsequent rule changes have led to an overall decrease in the number of annual catastrophic cervical spine injuries sustained playing football. Advances in on- and off-field evaluation and management, rehabilitation strategies and return-to-play guidelines have improved the care of athletes that sustain cervical injuries. Continued surveillance of cervical injuries in football and other contact sports will hopefully lead to further improvements in preventative strategies. Acknowledgements No sources of funding were used in the preparation of this review, and the authors have no conflicts of interest that are directly relevant to the content of this review.

References 1. Shannon B, Klimkiewicz JJ. Cervical burners in the athlete. Clin Sports Med 2002 Jan; 21 (1): 29-35, vi 2. Thomas BE, McCullen GM, Yuan HA. Cervical spine injuries in football players. J Am Acad Orthop Surg 1999 Sep-Oct; 7 (5): 338-47 3. Penning L. Some aspects of plain radiography of the cervical spine in chronic myelopathy. Neurology 1962 Aug; 12: 513-9 4. Meyer SA, Schulte KR, Callaghan JJ, et al. Cervical spinal stenosis and stingers in collegiate football players. Am J Sports Med 1994 Mar-Apr; 22 (2): 158-66 5. Levitz CL, Reilly PJ, Torg JS. The pathomechanics of chronic, recurrent cervical nerve root neuropraxia: the chronic burner syndrome. Am J Sports Med 1997 Jan-Feb; 25 (1): 73-6 6. Page S, Guy JA. Neuropraxia, ‘‘stingers’’, and spinal stenosis in athletes. South Med J 2004 Aug; 97 (8): 766-9 7. Markey KL, Di Benedetto M, Curl WW. Upper trunk brachial plexopathy: the stinger syndrome. Am J Sports Med 1993 Sep-Oct; 21 (5): 650-5

ª 2009 Adis Data Information BV. All rights reserved.

707

8. Kim DH, Vaccaro AR, Berta SC. Acute sports-related spinal cord injury: contemporary management principles. Clin Sports Med 2003 Jul; 22 (3): 501-12 9. Hershman EB, Wilbourn AJ, Bergfeld JA. Acute brachial neuropathy in athletes. Am J Sports Med 1989 Sep-Oct; 17 (5): 655-9 10. Cantu RC. The cervical spinal stenosis controversy. Clin Sports Med 1998 Jan; 17 (1): 121-6 11. Eismont FJ, Clifford S, Goldberg M, et al. Cervical sagittal spinal canal size in spine injury. Spine 1984 Oct; 9 (7): 663-6 12. Torg JS, Pavlov H, Genuario SE, et al. Neuropraxia of the cervical spinal cord with transient quadriplegia. J Bone Joint Surg Am 1986 Dec; 68 (9): 1354-70 13. Pavlov H, Torg JS, Robie B, et al. Cervical spinal stenosis: determination with vertebral body ratio method. Radiology 1987 Sep; 164 (3): 771-5 14. Torg JS, Naranja Jr RJ, Pavlov H, et al. The relationship of developmental narrowing of the cervical spinal canal to reversible and irreversible injury of the cervical spinal cord in football players. J Bone Joint Surg Am 1996 Sep; 78 (9): 1308-14 15. Torg JS. Cervical spinal stenosis with cord neuropraxia: evaluations and decisions regarding participation in athletics. Curr Sports Med Rep 2002 Feb; 1 (1): 43-6 16. Weinstein SM. Assessment and rehabilitation of the athlete with a ‘‘stinger’’: a model for the management of noncatastrophic athletic cervical spine injury. Clin Sports Med 1998 Jan; 17 (1): 127-35 17. Boden BP, Tacchetti RL, Cantu RC, et al. Catastrophic cervical spine injuries in high school and college football players. Am J Sports Med 2006 Aug; 34 (8): 1223-32 18. Torg JS, Thibault L, Sennett B, et al. The Nicolas Andry Award: the pathomechanics and pathophysiology of cervical spinal cord injury. Clin Orthop Relat Res 1995 Dec; (321): 259-69 19. Taylor AR. The mechanism of injury to the spinal cord in the neck without damage to vertebral column. J Bone Joint Surg Br 1951 Nov; 33-B (4): 543-7 20. Allen CR, Kang JD. Transient quadriparesis in the athlete. Clin Sports Med 2002 Jan; 21 (1): 15-27 21. White 3rd AA, Johnson RM, Panjabi MM, et al. Biomechanical analysis of clinical stability in the cervical spine. Clin Orthop Relat Res 1975; (109): 85-96 22. Cantu RC. Functional cervical spinal stenosis: a contraindication to participation in contact sports. Med Sci Sports Exerc 1993 Mar; 25 (3): 316-7 23. Maroon JC, El-Kadi H, Abla AA, et al. Cervical neuropraxia in elite athletes: evaluation and surgical treatment: report of five cases. J Neurosurg Spine 2007 Apr; 6 (4): 356-63 24. Odor JM, Watkins RG, Dillin WH, et al. Incidence of cervical spinal stenosis in professional and rookie football players. Am J Sports Med 1990 Sep-Oct; 18 (5): 507-9 25. Herzog RJ, Wiens JJ, Dillingham MF, et al. Normal cervical spine morphometry and cervical spinal stenosis in asymptomatic professional football players: plain film radiography, multiplanar computed tomography, and magnetic resonance imaging. Spine 1991 Jun; 16 (6 Suppl.): S178-86 26. Delaney JS, Al-Kashmiri A. Neck injuries presenting to emergency departments in the United States from 1990 to

Sports Med 2009; 39 (9)

Rihn et al.

708

27.

28.

29. 30.

31.

32.

33.

34.

35.

36. 37.

38. 39. 40. 41.

42.

43.

44.

1999 for ice hockey, soccer, and American football. Br J Sports Med 2005 Apr; 39 (4): 1-5 Torg JS, Vegso JJ, O’Neill MJ, et al. The epidemiologic, pathologic, biomechanical, and cinematographic analysis of football-induced cervical spine trauma. Am J Sports Med 1990 Jan-Feb; 18 (1): 50-7 Mueller FO, Cantu RC. Annual Survey of Catastrophic Football Injuries 1977 – 2006 [online]. Available from URL: http://www.unc.edu/depts/nccsi/FootballCatastrophic.pdf [Accessed 2007 May 1] Wilson JB, Zarzour R, Moorman 3rd CT, et al. Spinal injuries in contact sports. Curr Sports Med Rep 2006 Feb; 5 (1): 50-5 Cantu RC, Mueller FO. Catastrophic spine injuries in American football, 1977-2001. Neurosurgery 2003 Aug; 53 (2): 358-62; discussion 353-62 Nobunaga AI, Go BK, Karunas RB. Recent demographic and injury trends in people served by the Model Spinal Cord Injury Care Systems. Arch Phys Med Rehabil 1999 Nov; 80 (11): 1372-82 Banerjee R, Palumbo MA, Fadale PD. Catastrophic cervical spine injuries in the collision sport athlete, part 1: epidemiology, functional anatomy, and diagnosis. Am J Sports Med 2004 Jun; 32 (4): 1077-87 Torg JS, Sennett B, Vegso JJ, et al. Axial loading injuries to the middle cervical spine segment: an analysis and classification of twenty-five cases. Am J Sports Med 1991 Jan-Feb; 19 (1): 6-20 Mueller FO, Cantu RC. National Center for Catastrophic Sports Injury Research: twenty-sixth annual report. Fall of 1982-spring of 2006 [online]. Available from URL: http:// www.unc.edu/depts/nccsi/AllSport.pdf [Accessed 2007 May 1] Mueller FO. Fatalities from head and cervical spine injuries occurring in tackle football: 50 years’ experience. Clin Sports Med 1998 Jan; 17 (1): 169-82 Schneider RC. Serious and fatal neurosurgical football injuries. Clin Neurosurg 1964; 12: 226-36 Torg JS, Truex Jr R, Quedenfeld TC, et al. The National Football Head and Neck Injury Registry: report and conclusions 1978. JAMA 1979 Apr 6; 241 (14): 1477-9 Dolan KD, Feldick HG, Albright JP, et al. Neck injuries in football players. Am Fam Physician 1975 Dec; 12 (6): 86-91 Funk FF, Wells RE. Injuries of the cervical spine in football. Clin Orthop Relat Res 1975; (109): 50-8 Burke DC. Hyperextension injuries of the spine. J Bone Joint Surg Br 1971 Feb; 53 (1): 3-12 Marar BC. Hyperextension injuries of the cervical spine: the pathogenesis of damage to the spinal cord. J Bone Joint Surg Am 1974 Dec; 56 (8): 1655-62 Torg JS. Epidemiology, pathomechanics, and prevention of athletic injuries to the cervical spine. Med Sci Sports Exerc 1985 Jun; 17 (3): 295-303 Cantu RC, Mueller FO. Catastrophic football injuries: 1977-1998. Neurosurgery 2000 Sep; 47 (3): 673-5; discussion 675-7 Torg JS, Vegso JJ, Sennett B. The National Football Head and Neck Injury Registry: 14-year report on cervical quadriplegia (1971-1984). Clin Sports Med 1987 Jan; 6 (1): 61-72

ª 2009 Adis Data Information BV. All rights reserved.

45. Banerjee R, Palumbo MA, Fadale PD. Catastrophic cervical spine injuries in the collision sport athlete, part 2: principles of emergency care. Am J Sports Med 2004 Oct-Nov; 32 (7): 1760-4 46. Waninger KN. Management of the helmeted athlete with suspected cervical spine injury. Am J Sports Med 2004 Jul-Aug; 32 (5): 1331-50 47. Bracken MB, Shepard MJ, Hellenbrand KG, et al. Methylprednisolone and neurological function 1 year after spinal cord injury: results of the National Acute Spinal Cord Injury Study. J Neurosurg 1985 Nov; 63 (5): 704-13 48. Bracken MB. Methylprednisolone in the management of acute spinal cord injuries [letter]. Med J Aust 1990 Sep 17; 153 (6): 368 49. Bracken MB, Shepard MJ, Holford TR, et al. Administration of methylprednisolone for 24 or 48 hours or tirilazad mesylate for 48 hours in the treatment of acute spinal cord injury: results of the Third National Acute Spinal Cord Injury Randomized Controlled Trial. National Acute Spinal Cord Injury Study. JAMA 1997 May 28; 277 (20): 1597-604 50. Hurlbert RJ. The role of steroids in acute spinal cord injury: an evidence-based analysis. Spine 2001 Dec 15; 26 (24 Suppl.): S39-46 51. Hurlbert RJ, Moulton R. Why do you prescribe methylprednisolone for acute spinal cord injury? A Canadian perspective and a position statement. Can J Neurol Sci 2002 Aug; 29 (3): 236-9 52. Hurlbert RJ, Hamilton MG. Methylprednisolone for acute spinal cord injury: 5-year practice reversal. Can J Neurol Sci 2008 Mar; 35 (1): 41-5 53. Eck JC, Nachtigall D, Humphreys SC, et al. Questionnaire survey of spine surgeons on the use of methylprednisolone for acute spinal cord injury. Spine 2006 Apr 20; 31 (9): E250-3 54. Vaccaro AR, Klein GR, Ciccoti M, et al. Return to play criteria for the athlete with cervical spine injuries resulting in stinger and transient quadriplegia/paresis. Spine J 2002 Sep-Oct; 2 (5): 351-6 55. Torg JS, Corcoran TA, Thibault LE, et al. Cervical cord neuropraxia: classification, pathomechanics, morbidity, and management guidelines. J Neurosurg 1997 Dec; 87 (6): 843-50 56. Cantu RC. Stingers, transient quadriplegia, and cervical spinal stenosis: return to play criteria. Med Sci Sports Exerc 1997 Jul; 29 (7 Suppl.): S233-5 57. Cantu RC. Cervical spine injuries in the athlete. Semin Neurol 2000; 20 (2): 173-8 58. Torg JS, Ramsey-Emrhein JA. Management guidelines for participation in collision activities with congenital, developmental, or postinjury lesions involving the cervical spine. Clin J Sport Med 1997 Oct; 7 (4): 273-91 59. Bailes JE, Hadley MN, Quigley MR, et al. Management of athletic injuries of the cervical spine and spinal cord. Neurosurgery 1991 Oct; 29 (4): 491-7

Correspondence: Dr Jeffrey A. Rihn, The Rothman Institute, Thomas Jefferson University Hospital, 925 Chestnut St, Fifth Floor, Philadelphia, PA 19107, USA.

Sports Med 2009; 39 (9)

Sports Med 2009; 39 (9): 709-721 0112-1642/09/0009-0709/$49.95/0

CURRENT OPINION

ª 2009 Adis Data Information BV. All rights reserved.

Is it Time to Retire the ‘Central Governor’? Roy J. Shephard Faculty of Physical Education and Health, University of Toronto, Toronto, Ontario, Canada

Abstract

Over the past 13 years, Noakes and his colleagues have argued repeatedly for the existence of a ‘Central Governor’, a specific brain centre that provides a feed-forward regulation of the intensity of vigorous effort in order to conserve homeostasis, protecting vital organs such as the brain, heart and skeletal muscle against damage from hyperthermia, ischaemia and other manifestations of catastrophic failure. This brief article reviews evidence concerning important corollaries of the hypothesis, examining the extent of evolutionary pressures for the development of such a mechanism, the effectiveness of protection against hyperthermia and ischaemia during exhausting exercise, the absence of peripheral factors limiting peak performance (particularly a plateauing of cardiac output and oxygen consumption) and proof that electromyographic activity is limiting exhausting effort. As yet, there is a lack of convincing experimental evidence to support these corollaries of the hypothesis; furthermore, some findings, such as the rather consistent demonstration of an oxygen consumption plateau in young adults, argue strongly against the limiting role of a ‘Central Governor’.

1. Concept of a ‘Central Governor’ An early graduate textbook[1] made a systematic review of possible factors limiting physical performance. Depending on duration of the bout of effort, critical physiological factors were, in turn, maximal anaerobic power, maximal anaerobic capacity, maximal aerobic power, maximal aerobic capacity, and reserves of food (glycogen and fat), water and minerals. In some situations, environmental constraints such as difficulty in eliminating metabolic heat could also intervene, and with any period of exercise, psychological factors (motivation, arousal and release of cerebral inhibition) could sustain or enhance an individual’s performance. However, this comprehensive evaluation found no need to postulate that a specific subconscious feed-forward control mechanism was setting an upper limit to the intensity of exercise.

A more recent systematic review of the causes of fatigue during prolonged endurance cycling[2] made brief mention of the idea of such a control system, tracing its origin to Ulmer.[3] Ulmer had envisaged a feedback of information on force generation, displacement and metabolism to a central regulator, where these data were matched against motor learning and desired objectives to ensure a teleo-anticipatory optimization of the individual’s metabolic rate. In developing the concept of a ‘Central Governor’, Noakes[4] reached back to the classic studies of Hill et al.,[5] who had postulated some mechanism (either peripherally, in cardiac muscle or in the CNS) that restricted cardiac output as the arterial oxygen saturation began to fall. Noakes envisaged the central control mechanism as limiting motoneuron output in order to conserve homeostasis and prevent tissue damage from such threats as ischaemia or hyperthermia.[4]

Shephard

710

A long list of papers promoting the concept of a ‘Central Governor’ has been written by Noakes, his immediate colleagues and collaborators,[4,6-24] with a varying degree of support from some other laboratories.[3,25-40] However, most sports scientists have yet to accept the hypothesis.[41-46] In a recent debate conducted by the Journal of Applied Physiology,[22] seven commentators offered arguments against the hypothesis, against one brief and tepid endorsement from Carl Foster:[47] ‘‘y as anyone who has performed an incremental exercise test knows, this leads to the compelling ‘I don’t want to continue’ sensation. So, yes, there must be a command coming from the CNS that tells the exerciser that homeostasis is becoming disturbed and that it would be advisable to stop. But, many, if not most, of these ‘stop’ signals are reasonably attributable to limitations in central O2 transport and aerobic ATP generation.’’ Under the title ‘The implausible governor,’ a recent editorial[48] in the journal Sportscience (p. 9) summarized the issue thus: ‘‘When you’re trying to outrun a predator or cross the line first in an Olympic final, is your physiology truly pushed to the limit? Most exercise scientists assume it is, but for the last 10 years or so one of our number has been promoting a different view. Tim Noakes believes that our bodies are capable of exercise intense enough to cause catastrophic failure of physiology somewhere in the body. He argues that the demand for oxygen or the release of metabolites or heat by exercising muscles could result directly or indirectly in development of catastrophic malfunction in the muscles, heart, lungs or brain. Noakes believes that our brains therefore have a ‘governor’ that caps the drive from the brain to skeletal muscles before these catastrophes occur.’’ Reasoned discussion of the hypothesis has been hampered by the absence of a systematic and clearly enunciated listing of its inherent correlates. Sometimes, enthusiasm has also led to an unfortunate misrepresentation of opposing viewpoints, and/or too quick a dismissal of critical questions as ‘irrelevant’. ª 2009 Adis Data Information BV. All rights reserved.

2. Key Correlates of the ‘Central Governor Hypothesis’ Review of papers written to promote the idea of a ‘Central Governor’ suggests several key correlates of the concept: 1. The governor has evolved in response to evolutionary pressures. Although Noakes currently argues that the hypothesis applies to all forms of exercise with the specific exception of running on a treadmill where speed is controlled by the investigator, most discussions of possible evolution have focused on species that engage in prolonged endurance activity when hunting their prey.[14,17,49] 2. The controller limits the drive to the working muscles in order to maintain homeostasis throughout the body. Thus, the brain, heart, skeletal muscle and other body organs are protected against ischaemic and thermal injury, and the lungs against pulmonary oedema.[6,7,9,49] 3. Cardiac output does not reach a ceiling value that limits the maximal intensity of exercise when measuring maximal oxygen intake or engaging in other forms of intense endurance activity.[22] 4. Oxygen consumption does not reach a plateau value that limits maximal exercise intensity. If there is neither a ceiling of cardiac output nor a plateauing of oxygen consumption, then some alternative mechanism must be limiting maximal performance, and this could be the postulated ‘Central Governor’.[50,51] 5. Feed-forward regulation limits central drive to the muscles, so that electromyographic (EMG) activity is submaximal when an individual is making a peak voluntary effort.[10,52] These several issues are examined critically below. 2.1 Potential for the Evolution of a ‘Central Governor’

How strong are evolutionary pressures to develop a mechanism such as the postulated Central Governor? Many years ago, Mosso[53] noted that 3 years of learning was needed to perfect migration patterns in quails; he argued that the birds were then able to regulate their energy Sports Med 2009; 39 (9)

Is it Time to Retire the ‘Central Governor’?

expenditures over prolonged flights in such a manner as to conserve a final store of metabolites against easy capture. Marino[49] saw this observation as evidence for the evolution of a feed-forward regulator that enabled a bird to calculate its physiological requirements ahead of time, and to regulate potentially fatiguing exercise patterns accordingly. Marino[54] argued that in similar fashion, evolutionary pressures had led to the development of central regulatory mechanisms in both hunting dogs and humans, allowing prey to be run into the ground without endangering the homeostasis of the hunter. However, Hopkins[48] has recently pointed out that this scenario could be argued in several different ways – perhaps with predators risking a loss of homeostasis in order to avoid death from starvation, or the prey risking loss of homeostasis in order to avoid capture. Animals plainly have adapted to hostile environments, given sufficient time and selective pressures. The potential for human genetic adaptation to challenging habitats was thus a major concern of the International Biological Programme (IBP).[55] Particular attention was focused on the ability of various isolated populations to undertake fatiguing work of various durations. However, the IBP found surprisingly little evidence that humans had developed unusual physiological characteristics in response to prolonged residence in extreme environments. The available findings of human anthropology, the extent of selective pressures and molecular genetics all argued against significant modification of human exercise tolerance in this way, even in small and circumscribed populations that had been exposed to rigorous conditions for many generations. Anthropological evidence emphasized that most human groups lived at the junction of several habitats, which were exploited at different seasons. Thus, an adaptation favouring survival in one habitat would often prove a disadvantage in a second habitat exploited at other times during the year. Experimental data showed not diversity, but rather remarkably similar physiological working capacity in populations that occupied strikingly different environments.[55] A partial explanation may lie in differences in hunting techniques between animals and humans. ª 2009 Adis Data Information BV. All rights reserved.

711

Although some animals capture their prey by ‘running them into the ground’, field studies of traditional neolithic populations suggest that their success in hunting trips, and thus any selective pressures, depends much more on intellect than brute force, running ability or tolerance of physical fatigue.[56] Even in the savannah that Marino envisaged, !Kung bushmen can survive by working only 2.2 days per week.[57] Thus, if individuals are handicapped by a poor physique, they can compensate for this by working a slightly longer week. Moreover, hunting skills such as a knowledge of the habits of game, and the ability to fashion and use weapons are acquired progressively over the adult life-span.[56] In contrast, reproduction occurs as a young adult, and this limits the potential for genetic transmission of hunting skills. Further research may yet uncover genetically acquired characteristics favouring sustained exercise in small and isolated populations, but even if they were to be demonstrated, they would be unlikely to persist in the genetic melting pot that is our 21st century. If natural selection were a major determinant of such traits as maximal oxygen intake and heat tolerance, one would anticipate the ready identification of molecular markers of these characteristics. Further research on the human genome may yet uncover consistent genetic markers of exercise tolerance and fatigue resistance, but positive findings have to date been meagre.[58] Some 50 genes can be modified by heat exposure, but to date most interindividual differences in both exercise tolerance and thermoregulation seem phenotypic rather than genotypic.[59] Thus, at present we lack any strong evidence for the existence of selective pressures favouring the evolution of a thermally protective ‘Central Governor’ in humans. 2.2 Ability of a Putative ‘Central Governor’ to Protect against Hyperthermia and Ischaemia

If there is indeed a ‘Central Governor’ with a teleo-anticipatory mechanism designed to conserve homeostasis and protect the body against such hazards as hyperthermia and ischaemia, it seemingly has a limited effectiveness. In the Sports Med 2009; 39 (9)

712

African hunting dog, where natural selection is suggested as having acted strongly to develop such a governor,[49] rectal temperatures climb several degrees higher than those seen in domestic dogs. This observation has been constrained to the ‘Central Governor’ hypothesis by drawing upon the concept[60] that the African hunting dog focuses on conservation of water rather than regulation of its core temperature. Exercising humans also face a significant toll of deaths and dangerous episodes of hyperthermia in hot weather, particularly during American football games[61] and endurance runs.[62,63] Available data suggest that there is a far from perfect evolution of mechanisms to ensure thermal homeostasis. Rectal temperatures of those undertaking prolonged exercise commonly reach a level of 42C (compared with the prudent ceiling of 39.5C imposed by most Committees on Human Experimentation), and 21 US football players are known to have died of heat stroke between 1995 and 2001.[64] Nevertheless, the number of individuals who die is small relative to those at risk, and some mechanism, whether peripheral or central, at least protected the majority of the population against such fatalities. The putative system also seems to have a less than perfect ability to protect the exerciser against myocardial ischaemia and its fatal consequences. A bout of vigorous exercise can increase the risk of sudden cardiac death as much as 50-fold.[65] Possibly, many of those who succumb to a bout of exercise had some predisposing atherosclerosis. However, even in young adults, participation in an endurance or ultra-endurance event leads to signs usually interpreted as indicative of minor myocardial damage, such as the release of cardiac troponins, and an associated depression of myocardial function of variable duration.[66,67] Myocardial fibrosis may also develop with repeated participation in such events.[68] In older adults, episodes of severe exercise-induced myocardial ischaemia are commonplace.[69] In similar fashion, there is incomplete protection against skeletal muscle damage following not only eccentric contractions but also prolonged bouts of endurance exercise.[70] About a ª 2009 Adis Data Information BV. All rights reserved.

Shephard

half of ultramarathoners show elevated levels of circulating myoglobin following a race,[71] and sometimes this progresses to acute renal failure; 19 such cases were observed in the Comrades Marathon between 1969 and 1986.[72] Recent reviews by Gonza´lez-Alonso[73] and Cheung and Sleivert[74] comment specifically on the potential risks of impaired brain, heart and muscle function during and immediately following marathon running. There is also limited protection against pulmonary hypertension, and some endurance competitors show marked right ventricular dysfunction[75] and pulmonary oedema following a prolonged bout of exercise.[76,77] Zavorsky[77] reviewed 11 studies performed in 137 subjects; approximately 65% of those who had performed a prolonged bout of maximal exercise showed radiographic evidence of pulmonary oedema. Exercise plainly leads to large increases in demand on many homeostatic systems. The limited long-term morbidity and mortality that follows a sustained bout of vigorous physical activity could perhaps be construed as evidence for some process that limits physical activity sufficiently to avoid catastrophe in most people, although there are various well known mechanisms that could account for this, without invoking the action of a specific feed-forward ‘Central Governor’. Thus, Hopkins[48] suggests that the fall in local reserves of adenosine triphosphate (ATP) inevitably limits the activity of both skeletal and cardiac muscle as fatigue develops, and the negative effect of insipient pulmonary oedema upon oxygen transfer likewise serves to brake the intensity of exercise. 2.3 Absence of a Ceiling of Cardiac Output during Maximal Effort

If a ceiling of cardiac output were to be reached during maximal effort, this would point strongly to a limitation of performance by oxygen transport rather than to a ‘Central Governor’. The oxygen conductance equation[78] looks at the gradient of oxygen concentration from the inspired air to the active muscles. During vigorous exercise, the main concentration gradient is Sports Med 2009; 39 (9)

Is it Time to Retire the ‘Central Governor’?

between the pulmonary and the tissue capillaries, emphasizing that the individual’s peak cardiac output is the primary factor limiting oxygen transport. Furthermore, since the oxygen concentration in the tissue capillaries is low, the implication seems to be that the limitation is imposed by the pumping ability of the heart rather than a restriction of peripheral demand, as would occur with a feed-forward control of motoneuron activity. The standard understanding of the circulation has been that as the intensity of effort is increased, cardiac output reaches a plateau, mainly because a ceiling of heart rate has been reached, but also (particularly in older adults) because there is no further increase in cardiac stroke volume. The peripheral vasculature nevertheless retains the capacity to accept a larger blood flow.[79,80] Details of the cardiovascular response depend somewhat on the individual’s posture, other details of the exercise protocol such as continuous versus discontinuous testing, and the steepness of the ramp function if a progressive exercise test is used. However, typically, the characteristics observed in a person who is approaching maximal oxygen intake (an ashen grey vasoconstriction of the skin, impaired coordination of the muscles, and a clouding of consciousness progressing to collapse) strongly support a cardiac limitation of performance. Possible underlying factors include local or general myocardial ischaemia, a failure of venous return, and restrictions imposed by the pericardium.[81,82] It has sometimes been objected that a pericardial limitation is unlikely, since the stroke volume is smaller when a person is exercising in a normal, upright position than when supine or semi-supine. This is certainly true of light and moderate exercise, where most of the experimental observations have been made, but in severe and maximal exercise, the cardiac output in the upright position is equal to or slightly greater than that seen when supine.[83] A second alternative possibility, suggested by Noakes and associates, is that blood flow, venous return and cardiac output are all ultimately limited by a central restriction of muscle recruitment; this possibility is addressed below. ª 2009 Adis Data Information BV. All rights reserved.

713

Noakes and Marino[22] have argued against a plateauing of stroke volume, citing as support for their view a paper that applied the acetylene rebreathing technique to ten ordinary young men, ten runners and five elite runners;[84] this investigation used a rather steep ramp function, but nevertheless it demonstrated a clear plateauing of stroke volume in 20 of the 25 subjects. A second small study of seven normal young men and seven competitive cyclists also used the acetylene rebreathing technique; again a plateauing of stroke volume was seen in normal individuals, but values for the cyclists continued to increase up to the highest heart rate tested (190 beats/min, which did not coincide with attainment of an oxygen consumption plateau).[85] Other more recent studies have pushed equally fit subjects to an oxygen consumption plateau. These observations have demonstrated convincingly that the stroke volume does indeed plateau; indeed, it tends to decline as effort is increased, this decline being exacerbated by heat stress.[42,86-89] The second plank in this argument[22] was that if cardiac output plateaued, myocardial ischaemia would necessarily develop.[13] Myocardial ischaemia was denied on the basis of two Scandinavian papers that had suggested a reserve of coronary oxygen supply during vigorous effort.[90,91] However, the proof of this point is less than convincing. In the first of these papers,[90] the level of exercise was only moderate (a heart rate of 130–140 beats/min). In the second paper, maximal effort was said to have been reached,[91] but it is unlikely that subjects attained a true maximal effort, since they had cardiac catheters inserted into their coronary sinuses. In the first study, a switch from air to the breathing of a hypoxic gas mixture (4500 m altitude equivalent) led to a small net production of lactate by the heart muscle, indicating that coronary vasodilatation was not able to compensate completely for the 27% decrease in arterial oxygen content. In the second study, hypoxia was less severe (2300 m altitude equivalent), but nevertheless some of the subjects again showed a net release of lactate from the myocardium. The myocardial ischaemia inferred from lactate release also seems likely on theoretical Sports Med 2009; 39 (9)

Shephard

714

grounds. Maximal effort leads to a 6-fold increase in cardiac work rate.[1] However, oxygen extraction from the coronary vessels is relatively complete even under resting conditions, so that maintenance of an adequate oxygen supply during vigorous exercise depends almost entirely upon coronary vasodilatation. There is a potential for a 5-fold increase in coronary blood flow if the cardiac output increases sufficiently to sustain systemic blood pressure.[92] During maximal effort, any overall margin of coronary oxygen transport is very slender even in a young adult, and there is plainly potential to develop at least local pockets of myocardial ischaemia. It is worth underlining that myocardial ischaemia is not the only possible reason for the observed plateauing of cardiac output. Autonomic function, the mechanics of diastolic filling and even the restraints of the pericardium could play a role. Nevertheless, there is considerable empirical evidence that some degree of myocardial ischaemia does develop during vigorous exercise. In younger adults, this is manifested by minor changes in the electrocardiogram and a release of cardiac troponins, and in many middleaged and older adults the electrocardiogram indicates substantial myocardial ischaemia.[69] The progressive fall in peak heart rate with both aging[93] and exposure to hypoxic environments[94] provides some support to the concept that cardiac output is restricted by local ischaemia in the cardiac pacemaker. However, it remains puzzling why the breathing of oxygen-rich mixtures does not restore the maximal heart rate of the older individual to the values seen in a young adult. Possibly, the small increase in arterial oxygen content induced by inhaling oxygen-rich mixtures is insufficient to compensate for substantial deteriorations in regional myocardial blood flow. 2.4 Absence of an Oxygen Consumption Plateau

Noakes and associates have argued that an oxygen consumption plateau is rarely seen during incremental exercise testing.[4,95] If effort is limited by a feed-forward mechanism rather than a ª 2009 Adis Data Information BV. All rights reserved.

peaking of cardiac output, this would seem a logical corollary. In a review of studies worldwide that included children and the elderly but omitted key papers such as the definitive International Biological Programme Working Party Study,[96] Noakes and St Clair Gibson[13] found that a plateau had been demonstrated in about 50% of subjects. This in itself would seem to argue strongly against a ‘Central Governor’ in at least half of the population. Moreover, review of the papers cited suggests that failure to demonstrate a plateau in many of the remaining subjects reflected a failure to adopt the methodology recommended by the International Working Party.[96] Problems included: (i) failure to commence a definitive ramp test in the recommended manner (close to the individual’s maximal oxygen intake[78]); (ii) use of a metabolic cart with an inappropriately brief gas sampling time;[97] (iii) use of a test mode differing from that of an athlete’s specialty (e.g. testing runners or rowers on a cycle ergometer);[78] (iv) activation of an inadequate muscle mass (e.g. use of a cycle ergometer in people with weak quadriceps muscles);[42,78,98] and (v) poor motivation of the individual or limitations imposed by excessive caution of the observer (particularly in young children, the elderly and those with chronic disease).[78,99] The use of a steep ramp function was perhaps the most common problem. Two examples will suffice. One such study exercised five men and one woman on a treadmill, increasing the individual’s oxygen intake by as much as 7.0 mL/kg/min per stage;[100] only three of the six subjects in this report reached a typical plateau. A second study was even less successful, finding a plateau in only 12 of the 71 individuals;[101] it used a cycle ergometer rather than a treadmill as the source of exercise, and it adopted a ramp function that began with zero loading of the ergometer and increased continuously over a 12-minute protocol. However, many investigators using a more appropriate methodology have had no problems in demonstrating a clear ceiling of oxygen consumption when testing healthy adults.[41,42,44,78,89,102-105] We noted that the few elderly individuals who failed to reach an oxygen consumption plateau Sports Med 2009; 39 (9)

Is it Time to Retire the ‘Central Governor’?

could be identified by an a priori rating of their motivation before and during the test.[78,93] 2.5 Neuromuscular Drive and the Electromyogram

If there were a ‘Central Governor’, this should limit neuromuscular drive and thus EMG activity as fatigue is approached in order to conserve homeostasis and avoid catastrophic biological failure.[10,52] In effect, the peak level of oxygen consumption would be set by this centrally determined drive. Given such a system, it becomes hard to explain why the maximal oxygen intake is augmented by blood transfusion or hyperoxia; it is easy to see how such measures could boost a peripheral limitation of oxygen transport, but much harder to explain how such treatment could modify a central drive.[48] If there were a central limitation of drive, Noakes has argued that a subject would be likely to terminate exercise before reaching full activation of available motor units in the exercising limbs.[51] A magnetic resonance imaging study estimated that between 40% and 80% of potential muscle power was used during treadmill running to exhaustion.[106] Other information on this question is limited, much being based on surface or needle EMG recordings from one or two muscles. There are many limitations to interpreting such data in terms of motor unit recruitment and the generation of force and power during relatively complicated dynamic tasks. Kayser and associates[107] compared maximal exercise to exhaustion at sea level and after 1 month of acclimatization to an altitude of 5050 m (where the sea level maximal oxygen intake was reduced by some 20%). At sea level, exhaustion was associated with an increase rather than a decrease of the integrated EMG, with a sizeable increase in arterial lactate concentration and a decrease in arterial pH, a pattern the authors considered consistent with a peripheral limitation of effort. At altitude, the changes in lactate and pH were smaller and there was no increase of the EMG signal at fatigue, leading them to suggest that in this specific situation there could have been a central limitation of muscle ª 2009 Adis Data Information BV. All rights reserved.

715

drive; however, it is difficult to assess the influence of other factors such as respiratory distress, weight loss, dehydration and altered acid-base balance. Further evidence to support the expectations of the ‘Central Governor’ hypothesis was sought in 100 km cycling time trials.[108] The peak effort over the first 4 km high-intensity epoch of the trial averaged 318 W, with a mean integrated rectus femoris EMG that was 16% of that recorded during a maximal voluntary isometric contraction. In contrast, over the final 4 km high intensity epoch, power output had dropped to 278 W, and the integrated EMG showed a similar decrease, to 11% of that seen in the maximal voluntary contraction.[108] The authors concluded that from an early stage in the trial there had been a protective limitation of muscle activation, and that this protection had increased over the course of the trial. However, the decrease in the EMG signal could also reflect in part technical factors such as changes in temperature, conductivity and displacement of the electrodes, and a close relationship between the signal and power output seems unlikely. Unfortunately, this investigation did not examine changes in the maximal voluntary contraction at the end of the trial, which might have served to show how far the decrease in EMG reflected a protective decrease of central drive. Nicol and associates[109] demonstrated a decrease in maximal voluntary contraction following participation in a marathon run, although they wisely concluded that this could reflect either a decrease of motivation or a change in central recruitment tactics. Stronger evidence against the ‘Central Governor’ came from a study of soleus fatigue induced by electrical stimulation of the peripheral nerves.[110] At fatigue, maximal plantar flexion was reduced, and a testing of the Hoffman reflex demonstrated a peripheral reflex inhibition of the alpha motoneurons innervating the muscle.[110] Electrical stimulation can induce a more powerful contraction than is possible by conscious effort but, nevertheless, the end-result of prolonged stimulation was fatigue and not catastrophic injury of the muscle; in this situation, there was no centrally imposed ceiling acting to prevent Sports Med 2009; 39 (9)

Shephard

716

catastrophe.[111] Again, in brief ramp function tests, motor unit activation does not reach a centrally defined ceiling at a power output corresponding to the individual’s maximal oxygen intake; the rectified EMG increases substantially if subjects are persuaded to exercise at an intensity 5–15% above that corresponding to the oxygen consumption plateau.[42,89] Terminal intramuscular concentrations of lactate are high during both ‘supramaximal’ and sustained exhausting exercise,[78] and any decrease of EMG activity at fatigue seems likely to reflect an inadequate local blood flow, and an inhibition of phosphagen regeneration as lactate and/or

hydrogen ions accumulate, rather than the action of a ‘Central Governor’. 3. Factors Limiting Endurance Performance From the foregoing discussion, we may conclude that, in humans, evidence concerning the potential for evolution of a ‘Central Governor’ and resulting protection against hyperthermia and ischaemia is equivocal. However, a person can increase his or her neuromuscular drive substantially above the level associated with a plateauing of oxygen transport, and the

Higher centres

Motor cortex

+

Mesencephalon

Hypothalamus

− + Cardiac centre −

− + −

Vasoconstrictor centre

−

− Sensors

+

Sensors

+

Central blood pressure

Venous return

+

Respiratory centres +

+

− + − − +

− +

+

Vasodilator centre

− +

+

Sensors

− + Heart

− +

Humoral factors

Blood vessels −

Local tissue conditions

Fig. 1. Potential feedback loops regulating vigorous exercise.[1] + indicates positive feedback; – indicates negative feedback.

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (9)

Is it Time to Retire the ‘Central Governor’?

demonstrated ceilings of cardiac output and oxygen transport argue strongly against the ‘Central Governor’ hypothesis. If there is no ‘Central Governor’, how is fatiguing exercise regulated? The limiting factors undoubtedly vary with the duration of activity. Noakes currently argues that the ‘Central Governor’ limits all forms of exercise, but much of the thinking of ‘Central Governor’ proponents concerns events such as a marathon or a supermarathon run, whereas critics of the hypothesis have often focused on much shorter bouts of exercise. A variety of standard texts have illustrated the many mutually redundant feedback loops that limit exercise (see for example, figure 1[1]). Among these, there is good empirical evidence for loops signalling peripheral ischaemia, hypoglycaemia and hyperthermia.[112] Plainly, there is also a potential for input from higher centres, including the motor cortex, but the effect is probably other than envisaged by Noakes and colleagues. The higher centres of an endurance athlete who is competing over a 1-mile track event call forth an initial effort to gain the desired position on the field; for most of the remaining distance, a combination of personal experience and coaching instruction hold oxygen demand just below the individual’s maximal oxygen intake, at a level where a minimal accumulation of lactate in the peripheral muscles is sensed. A final sprint is begun at a distance set by coaching instruction, accumulated experience or a signal from a friend who is helping with pacing; the rate of energy expenditure is then increased so that the competitor’s anaerobic capacity is fully exploited.[1,78,113] Various areas of the brain contribute to the choice of pace over an event, but the statement ‘‘only the Central Governor can explain this’’[20] (p. 376) is plainly incorrect. Likewise, when determining maximal oxygen intake in a moderately fit adult, personal motivation or verbal stimulation from the monitoring physician can call forth a final extreme effort that overcomes most of the constraints imposed by feedback. Moreover, sustained effort can be limited by mental fatigue, and thus a greater perception of the required effort[114] – with a decreased cortical glucose supply contributing to this ª 2009 Adis Data Information BV. All rights reserved.

717

process.[112] Particularly in a hot environment, effort may be limited by both a failure of venous return, and an increase in ratings of perceived exertion at a given rate of working.[115] However, the supposed ‘proof’[116] that cardiovascular factors do not limit exercise in the heat is based on moderate rather than maximal effort,[117] and it is hardly necessary to class an early reluctance[118] to exercise as hard in a hot environment as the action of an ‘anticipatory Central Governor’.[116] In an older person the sensations arising from peripheral cardiac or skeletal muscle ischaemia (angina and intermittent claudication, respectively) can cause a termination of exercise. Over prolonged exercise, glycogen depletion is also undoubtedly a factor. Fatigue is postponed by dietary manipulations that increase intramuscular glycogen, and although muscle biopsies may show a residue of glycogen in some fibres at exhaustion, if the average reserves are decreased by 75%, depletion is likely complete in the most active fibres. Furthermore, the apparent absence of a depletion of Krebs cycle intermediaries in some biopsy specimens probably reflects regeneration of these substances in the interval between biopsy and freezing of the muscle sample.[119] 4. Conclusions Over the past 13 years, a small group of investigators has argued repeatedly for the existence of a ‘Central Governor’ – an anticipatory feed-forward mechanism that regulates the intensity of vigorous effort with the intent of conserving homeostasis, thus protecting vital organs such as the brain, heart and skeletal muscle against hyperthermia and ischaemia. There seems mounting evidence against several key correlates of this hypothesis. It is difficult to discern evolutionary pressures that would favour the development of such a mechanism in humans. Protection of the exerciser against problems of hyperthermia and ischaemia is incomplete. Cardiac output generally reaches a plateau, perhaps in response to local oxygen lack. Most laboratories also have no difficulty in demonstrating an oxygen consumption plateau in well motivated young adults, and at least in some forms of Sports Med 2009; 39 (9)

Shephard

718

exercise EMG activity can increase after oxygen consumption has plateaued. Until there is convincing experimental evidence of an underlying physiological mechanism, most sports scientists will continue to express scepticism concerning the existence of a ‘Central Governor’.[120] Acknowledgements No funding was received for the preparation of this article, and the author has no conflicts of interest that are directly relevant to the content of this article.

References 1. Shephard RJ. Physiology and biochemistry of exercise. New York (NY): Praeger Publications, 1982 2. Abbis CR, Laursen PB. Models to explain fatigue during prolonged endurance cycling. Sports Med 2005; 35: 865-98 3. Ulmer CV. Concept of an extracellular regulation of muscular metabolic rate during heavy exercise in humans by psychophysiological feedback. Experientia 1996; 52: 416-20 4. Noakes TD. How did A.V. Hill understand the VO2max and the ‘‘plateau phenomenon’’? Still no clarity? Br J Sports Med 2008; 42: 574-80 5. Hill AV, Long CHN, Lupton H. Muscular exercise, lactic acid and the supply and utilisation of oxygen: parts VII-VIII. Proc Roy Soc B 1924; 97: 155-76 6. Noakes TD. Physiological models to understand exercise fatigue and the adaptations that predict or enhance athletic performance. Scand J Med Sci Sports 2000; 10: 123-45 7. Hampson DB, St Clair Gibson A, Lambert MI, et al. The influence of sensory cues on the perception of exertion during exercise and central regulation of exercise performance. Sports Med 2001; 31: 935-52 8. Kay D, Marino FE, Cannon J, et al. Evidence for neuromuscular fatigue during high intensity cycling in warm, humid conditions. Eur J Appl Physiol 2001; 84: 115-21 9. Noakes TD, Peltonen JE, Rusko HK. Evidence that a central governor regulates exercise performance during acute hypoxia and hyperoxia. J Exp Biol 2001; 204 Pt 18: 3225-34 10. Noakes TD. Linear relationship between the perception of effort and the duration of constant load exercise that remains. J Appl Physiol 2004; 96: 1571-2 11. Ansley L, Schabort EJ, St Clair Gibson A, et al. Regulation of pacing strategies during successive 4 km time trials. Med Sci Sports Exerc 2004; 38: 1819-25 12. Noakes TD, St Clair Gibson A, Lambert EV. From catastrophe to complexity: a novel model of integrative central neural regulation of effort and fatigue during exercise in humans. Br J Sports Med 2004; 38: 511-4

ª 2009 Adis Data Information BV. All rights reserved.

13. Noakes TD, St Clair Gibson A. Logical limitations to the catastrophe models of fatigue during exercise in humans. Br J Sports Med 2004; 38: 648-9 14. St Clair Gibson A, Noakes TD. Evidence for complex system regulation and dynamic neural regulation if skeletal muscle recruitment during exercise in humans. Br J Sports Med 2004; 38: 797-806 15. Noakes TD, Calbet JA, Boshel R, et al. Central regulation of skeletal muscle recruitment explains the reduced maximal cardiac output during hypoxia. Am J Physiol 2004; 287: R996-9 16. Tucker R, Rauch I, Harley YXR, et al. Impaired exercise performance in the heat is associated with an anticipatory reduction in skeletal muscle recruitment. Pflu¨g Archiv 2004; 448: 422-30 17. Lambert M, St Clair Gibson A, Noakes TD. Complex systems model of fatigue: integrative homeostatic control of peripheral physiological systems during exercise in humans. Br J Sports Med 2005; 39: 52-62 18. Tucker R, Marle T, Lambert EV. The rate of heat storage mediates an anticipatory reduction in exercise intensity during cycling at a fixed rating of perceived exertion. J Physiol 2006; 574: 905-15 19. Noakes TD. Determining the extent of neural activation during maximal effort [abstract]. Med Sci Sports Exerc 2007; 39: 2092 20. Noakes TD. The central governor of exercise regulation applied to the marathon. Sports Med 2007; 37: 374-77 21. Noakes TD. Testing for maximum oxygen consumption has produced a brainless model of human exercise performance. Br J Sports Med 2008; 42: 551-5 22. Noakes TD, Marino FE. Maximal oxygen uptake is limited by a central nervous system governor. J Appl Physiol 2009; 106: 338-9 23. Crewe H, Tucker R, Noakes TD. The rate of increase of perceived exertion predicts the duration of exercise to fatigue at a fixed power output in different environmental conditions. Eur J Appl Physiol 2008; 103: 569-77 24. Noakes TD, Tucker R. Do we really need a central governor to explain brain regulation of exercise performance? A response to the letter of Dr Marcora. Eur J Appl Physiol 2008; 104: 933-5 25. Kayser B. Exercise starts and ends in the brain. Eur J Appl Physiol 2003; 90: 411-9 26. Abbis CR, Laursen PB. Describing and understanding pacing strategies during athletic competition. Sports Med 2008; 38: 239-52 27. Baron B, Noakes TD, Dekerle J, et al. Why does exercise terminate at the maximal lactate steady state intensity. Br J Sports Med 2008; 42: 528-33 28. Baron B, Deruelle F, Moullan F, et al. The eccentric muscle loading influences the pacing strategies during repeated downhill sprint intervals. Eur J Appl Physiol 2009; 105: 749-57 29. Castle PC, Macdonald AL, Philp A, et al. Precooling leg muscle improves intermittent sprint exercise performance in hot, humid conditions. J Appl Physiol 2006; 100: 1377-84

Sports Med 2009; 39 (9)

Is it Time to Retire the ‘Central Governor’?

30. Clark SA, Bourdon PC, Schmidt W, et al. The effect of acute simulated altitude on power, performance and pacing strategies in well-trained cyclists. Eur J Appl Physiol 2007; 102: 45-55 31. Edwards AM, Mann ME, Marfell-Jones MJ, et al. Influence of moderate hydration on soccer performance: physiological responses to 45 min of outdoor match-play and the immediate subsequent performance of sport-specific and mental concentration tests. Br J Sports Med 2007; 41: 385-91 32. Eston R, Faulkner J, St Clair Gibson A, et al. The effect of antecedent fatiguing activity on the relationship between perceived exertion and physiological activity during a constant load exercise task. Psychophysiol 2007; 44: 779-86 33. Flouris AD, Cheung SS. Human conscious response to thermal input is adjusted to changes in mean body temperature. Br J Sports Med 2009; 43: 199-203 34. Joseph T, Johnson B, Battista RC, et al. Perception of fatigue during simulated competition. Med Sci Sports Exerc 2008; 40: 381-6 35. Kabitz H-J, Walker D, Schwoerer A, et al. New physiological insights into exercise-induced diaphragmatic fatigue. Resp Physiol Neurobiol 2007; 158: 88-96 36. Morante SM, Brotherhood JR. Autonomic and behavioural thermoregulation in tennis. Br J Sports Med 2008; 42: 679-85 37. Nummela AT, Heath KA, Paavolainen LM, et al. Fatigue during a 5-km running time trial. Int J Sports Med 2008; 29: 738-45 38. Racinais S, Bringard A, Puchaux K, et al. Modulation in voluntary neural drive in relation to muscle soreness. Eur J Appl Physiol 2008; 102: 439-46 39. Ross EZ, Middleton N, Shave R, et al. Corticomotor excitability contributes to neuromuscular fatigue following marathon running in man. Exp Physiol 2007; 92: 417-26 40. Thomas R, Stephens P. Prefrontal cortex oxygenation and neuromuscular responses to exhaustive exercise. Eur J Appl Physiol 2008; 102: 153-63 41. Bergh U, Ekblom B, A˚strand PO. Maximal oxygen uptake ‘‘classical’’ versus ‘‘contemporary’’ viewpoints. Med Sci Sports Exerc 2000; 32: 85-8 42. Brink-Elfegoun T, Holmberg H-C, Ekblom MN, et al. Neuromuscular and circulatory adaptations during combined arm and leg exercise with different maximal work loads. Eur J Appl Physiol 2007; 101: 603-11 43. di Prampero PE, Capelli C, Ferretti G. Positive effects of intermittent hypoxia (live high:train low) on exercise performance are/are not mediated primarily by augmented red cell volume [comments on point:counterpoint]. J Appl Physiol 2005; 99: 2453-62 44. Snell PG, Stray-Gundersen J, Levine BD, et al. Maximal oxygen uptake as a parametric measure of cardiorespiratory capacity. Med Sci Sports Exerc 2007; 39: 103-7 45. Shephard RJ. Is the measurement of maximal oxygen intake passe´? Br J Sports Med 2009; 43: 83-5 46. Shephard RJ. Hard evidence for a central governor is still lacking. J Appl Physiol 2009; 106: 343 47. Foster C. Untitled. J Appl Physiol 2009; 106: 343

ª 2009 Adis Data Information BV. All rights reserved.

719

48. Hopkins WG. The implausible governor. Sportscience 2009; 13: 9-11 49. Marino FE. The evolutionary basis of thermoregulation and exercise performance. In: Marino FE, editor. Thermoregulation and human performance: physiological and biological aspects. Basel: Karger Publications, 2008: 1-13 50. Noakes TD. Implications of exercise testing for prediction of athletic performance: a contemporary perspective. Med Sci Sports Exerc 1988; 20: 319-30 51. Noakes TD. Maximal oxygen uptake: ‘‘classical’’ versus ‘‘contemporary’’ viewpoints – a rebuttal. Med Sci Sports Exerc 1998; 30: 1381-98 52. St Clair Gibson A, Schabort EJ, Noakes TD. Reduced neuromuscular activity and force generation during prolonged cycling. Am J Physiol 2001; 281: R187-96 53. Mosso A. Fatigue. London: George Allen & Unwin, 1915 54. Marino FE. Comparative thermoregulation and the quest for athletic supremacy. In: Marino FE, editor. Thermoregulation and human performance: physiological and biological aspects. Basel: Karger Publications, 2008: 14-25 55. Shephard RJ. Human physiological work capacity. London: Cambridge University Press, 1978 56. Godin G, Shephard RJ. Activity patterns of the Canadian Eskimo. In: Edholm O, Gunderson EK, editors. Human polar biology. Cambridge: Heinemann, 1973 57. Lee RB. Kung bushmen subsistence: an input-output analysis. In: Vayda AP, editor. Environment and cultural behavior. New York (NY): Natural History Press, 1969 58. Wolfarth B, Rankinen T, Muhlbauer S, et al. Association between a beta2-adrenergic receptor polymorphism and elite endurance performance. Metab Clin Exper 2007; 56: 1649-51 59. Lambert MI, Mann T, Dugas JP. Ethnicity and temperature regulation. In: Marino FE, editor. Thermoregulation and human performance: physiological and biological aspects. Basel: Karger Publications, 2008: 104-20 60. Taylor CR, Schmidt-Nielsen K, Dmi’el R, et al. Effect of hyperthermia on heat balance during running in the African hunting dog. Am J Physiol 1971; 220: 823-7 61. Luke AC, Bergeron MF, Roberts WO. Heat injury prevention practices in high school football. Clin J Sports Med 2007; 17: 488-93 62. Pugh LGCE, Corbett JL, Johnson RH. Rectal temperatures, weight losses and sweat rates in marathon running. J Appl Physiol 1967; 23: 347-52 63. Sutton JR, Coleman MJ, Millar AP, et al. The medical problems of mass participation in athletic competition: the ‘‘City to Surf’’ race. Med J Aust 1972; 2: 127-33 64. Mueller FO, Cantu RC. Twentieth annual report: Fall 1982-Spring 2002, National Center for Catastrophic Sport Injury Research. Chapel Hill (NC): University of North Carolina, National Center for Catastrophic Sport Injury Research, 2004 65. Vuori I. Exercise and sudden cardiac death: effects of age and type of activity. Sports Sci Rev 1995; 4 (2): 46-84 66. Neilan TG, Yoerger D, Douglas P, et al. Persistent and reversible cardiac dysfunction among amateur marathon runners. Eur Heart J 2006; 27: 1079-84

Sports Med 2009; 39 (9)

Shephard

720

67. Neilan TG, Jnauzzi JL, Lee-Lewandrowski E, et al. Myocardial injury and ventricular dysfunction related to training levels among nonelite participants in the Boston marathon. Circulation 2006; 114: 2325-33 68. Whyte GP. Clinical significance of cardiac damage and changes in function after exercise. Med Sci Sports Exerc 2008; 40: 1416-23 69. Shephard RJ. Ischemic heart disease and exercise. London: Croom Helm, 1981 70. Clarkson PM. Exertional rhabdomyolysis and acute renal failure in marathon runners. Sports Med 2007; 27: 361-3 71. Schiff HB, MacSearraigh ET, Kallmeyer JC. Myoglobinuria, rhabdomyolysis and marathon running. Quart J Med 1978; 47: 463-72 72. Seedat YK, Aboo N, Naicker S, et al. Acute renal failure in the ‘‘Comrades Marathon’’ runners. Renal Failure 1989; 11: 209-12 73. Gonza´lez-Alonso J. Hyperthermia impairs brain, heart and muscle function in exercising humans. Sports Med 2007; 37: 371-3 74. Cheung SS, Sleivert GG. Multiple triggers for hyperthermic fatigue and exhaustion. Ex Sport Sci Rev 2004; 32: 100-6 75. Lagerche A, Prior D. Exercise: is it possible to have too much of a good thing? Heart Lung Circ 2007; 16 Suppl. 3: S102-4 76. McKechnie JK, Lear WP, Noakes TD, et al. Acute pulmonary edema in two athletes during a 90 km running race. S Afr Med J 1979; 56: 261-5 77. Zavorsky GS. Evidence of pulmonary oedema triggered by exercise in healthy humans and detected with various imaging techniques. Acta Physiol 2007; 189: 305-17 78. Shephard RJ. Aerobic fitness and health. Champaign (IL): Human Kinetics, 1994 79. Andersen P, Saltin B. Maximal perfusion of skeletal muscle in man. J Physiol 1985; 366: 233-49 80. Rowell LB. Human cardiovascular control. Oxford: Oxford University Press, 1993 81. Stray-Gundersen J, Musch TI, Haidet GC, et al. The effect of pericardiectomy on maximal oxygen consumption and maximal cardiac output in untrained dogs. Circulation Res 1986; 58: 523-30 82. Hammond HK, White FC, Bhargava V, et al. Heart size and maximal cardiac output are limited by the pericardium. Am J Physiol 1992; 263: H1675-81 83. Leyk D, Egfeld D, Hoffmann U, et al. Postural effect on cardiac output, oxygen uptake and lactate during cycle exercise of varying intensity. Eur J Appl Physiol 1994; 68: 30-5 84. Zhou B, Conlee RK, Jensen R, et al. Stroke volume does not plateau during graded exercise in elite male distance runners. Med Sci Sports Exerc 2001; 33: 1849-54 85. Gledhill N, Cox D, Jamnik R. Endurance athletes’ stroke volume does not plateau: major advantage is diastolic function. Med Sci Sports Exerc 1994; 26: 1116-21 86. Gonza´lez-Alonso J, Calbet J. Reductions in systemic and skeletal muscle blood flow and oxygen delivery limit

ª 2009 Adis Data Information BV. All rights reserved.

87.

88.

89.

90.

91.

92.

93. 94.

95.

96.

97.

98.

99.

100.

101.

102.

103.

104.

maximal aerobic capacity in humans. Circulation 2003; 107: 824-30 Mortensen SP, Dawson EA, Yoshiga CC, et al. Limitations to systemic and locomotor limb muscle oxygen delivery and uptake during maximal exercise in humans. J Physiol 2005; 566: 273-85 Mortensen SP, Damsgaard R, Dawson EA, et al. Restrictions in systemic and locomotor skeletal muscle perfusion, oxygen supply and VO2 during high-intensity whole-body exercise in humans. J Physiol 2008; 586: 2621-35 Ekblom B. Point/counterpoint: maximal oxygen intake is not limited by a central nervous system governor. J Appl Physiol 2009; 106: 341-2 Kaijser L, Grubbstrom J, Berglund B. Myocardial lactate release during prolonged exercise under hypoxemia. Acta Physiol Scand 1993; 149: 427-33 Grubbstrom J, Berglund B, Kaijser L. Myocardial blood flow and lactate metabolism at rest and during exercise with reduced arterial oxygen content. Acta Physiol Scand 1991; 142: 567-74 Vatner SF, Higgins CB, Franklin D, et al. Role of tachycardia in mediating the coronary hemodynamic response to severe exercise. J Appl Physiol 1972; 32: 380-9 Shephard RJ. Aging, physical activity and health. Champaign (IL): Human Kinetics, 1997 Pugh LGCE. Physiological and medical aspects of the Himalayan Scientific and Mountaineering Expedition, 1960-1961. BMJ 1962; 2: 621-7 Doherty M, Nobbs I, Noakes TD. Low frequency of the ‘‘plateau phenomenon’’ during maximal exercise in elite British athletes. Eur J Appl Physiol 2003; 89: 619-23 Shephard RJ, Allen C, Benade AJS, et al. The maximum oxygen intake: an international reference standard of cardio-respiratory fitness. Bull WHO 1968; 38: 757-64 Myers J, Walsh D, Sullivan M, et al. Effect of sampling on variability and plateau in oxygen uptake. J Appl Physiol 1990; 68: 404-10 Shephard RJ, Bouhlel E, Vandewalle H, et al. Muscle mass as a factor limiting physical work. Eur J Appl Physiol 1988; 64: 1472-9 Astorino TA, Willey J, Kinnahan J, et al. Elucidating determinants of the plateau in oxygen consumption at VO2max. Br J Sports Med 2005; 39: 655-60 Myers J, Walsh D, Buchanan N, et al. Can maximal cardiopulmonary capacity be recognized by a plateau in oxygen uptake? Chest 1989; 96: 1312-6 Day JR, Rossiter HB, Coats EM, et al. The maximally attainable VO2 during exercise in humans: the peak vs maximum issue. J Appl Physiol 2003; 95: 1901-7 A˚strand PO, Saltin B. Maximal oxygen uptake and heart rate in various types of muscular activity. J Appl Physiol 1961; 16: 977-81 Hill AV, Lupton H. Muscular exercise, lactic acid, and the supply and utilisation of oxygen. Quart J Med 1923; 16: 135-71 Taylor HL, Buskirk ER, Henschel A. Maximal oxygen intake as an objective measure of cardio-respiratory performance. J Appl Physiol 1955; 8: 73-80

Sports Med 2009; 39 (9)

Is it Time to Retire the ‘Central Governor’?

105. Wagner PD. New ideas on limitations to VO2max. Exerc Sport Sci Rev 2000; 28: 10-4 106. Sloniger MA, Cureton KJ, Prior BM, et al. Lower extremity muscle activation during horizontal and uphill running. J Appl Physiol 1997; 83: 2073-9 107. Kayser B, Narici M, Binzoni T, et al. Fatigue and exhaustion in chronic hypobaric hypoxia: influence of exercising muscle mass. J Appl Physiol 1994; 76: 634-40 108. St Clair Gibson A, Lambert MI, Noakes TD. Neural control of force output during maximal and submaximal exercise. Sports Med 2001; 31: 637-50 109. Nicol C, Komi PV, Marconnet P. Fatigue effects of marathon running on neuromuscular performance. Scand J Sci Sports 1991; 1: 18-24 110. Garland SJ, McComas A. Reflex inhibition of human soleus muscle during fatigue. J Physiol 1990; 429: 17-27 111. Spriet LL, Soderlund KK, Bergstrom M, et al. Anaerobic energy release in skeletal muscle during electrical stimulation in man. J Appl Physiol 1987; 62: 611-5 112. Hargreaves M. Fatigue mechanisms determining exercise performance: integrative physiology is systems biology. J Appl Physiol 2008; 104: 1541-2 113. Tucker R, Lambert MI, Noakes TD. An analysis of pacing strategies during men’s world performances in track athletics. Int J Sports Physiol Perf 2006; 1: 233-45 114. Marcora S, Staiano W, Manning V. Mental fatigue impairs physical performance in humans. J Appl Physiol 2009; 106: 857-64

ª 2009 Adis Data Information BV. All rights reserved.

721

115. Galloway SDR, Maughan RJ. Effect of ambient temperature on the capacity to perform prolonged cycle exercise in man. Med Sci Sports Exerc 1997; 29: 1240-9 116. Tucker R. Thermoregulation, fatigue, and exercise modality. In: Marino FE, editor. Thermoregulation and human performance: physiological and biological aspects. Basel: Karger Publications, 2008: 26-38 117. Savard GK, Nielsen B, Laszczynska J, et al. Muscle blood flow is not reduced in humans during moderate exercise and heat stress. J Appl Physiol 1988; 64: 649-57 118. Tatterson AJ, Hahn AG, Martin DT, et al. Effect of heat and humidity on time trial performance in Australian national team road cyclists. J Sci Med Sport 2000; 3: 186-93 119. Baldwin J, Snow RJ, Gibala MJ, et al. Glycogen availability does not affect the TCA cycle or TAN pools during prolonged, fatiguing exercise. J Appl Physiol 2003; 94: 2181-7 120. Weir JP, Beck TW, Cramer JT, et al. Is fatigue all in your head? A critical review of the central governor model. Br J Sports Med 2006; 40: 573-86

Correspondence: Dr Roy J. Shephard, PO Box 521, Brackendale, BC V0N 1H0, Canada. E-mail: [email protected]

Sports Med 2009; 39 (9)

Sports Med 2009; 39 (9): 723-741 0112-1642/09/0009-0723/$49.95/0

REVIEW ARTICLE

ª 2009 Adis Data Information BV. All rights reserved.

Competitive Elite Golf A Review of the Relationships between Playing Results, Technique and Physique John Hellstro¨m ¨ rebro University, O ¨ rebro, Sweden School of Health and Medical Sciences, O

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Playing Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Total Score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Partial Scores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.1 Putting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.2 Short Game . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.3 Long Game . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.4 Shot by Shot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Ball Displacement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Impact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Clubface, Clubhead and Shaft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 Clubface Angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2 Clubhead Velocity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.3 Shaft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Hands on the Shaft, Wrists and Arms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1 Pull and Centripetal Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.2 Force Couple . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.3 Wrist Torque Generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.4 Work Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Torso, Legs and Feet on the Ground . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.1 Ground Reaction Force and Centre of Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.2 X-Factor and Stretch-Shortening Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.3 The Kinematic and Kinetic Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Physique. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Elite Players’ Physique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Physique, Clubhead Speed and Ball Flight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

723 724 724 725 725 725 727 727 728 729 729 729 729 729 730 730 730 731 731 731 732 732 732 734 735 735 735 738

Elite golfers commonly use fitness and technical training to become more competitive. The aim of this paper was to review the literature regarding the relationships between elite golfers’ playing results, technique and physique. The competitive outcome is a direct function of the score. The three golf statistical measures that show the strongest correlations to scoring average are greens in regulation (GIR), scrambling, and putts per GIR. However,

Hellstro¨m

724

more detailed game statistics are needed where the distances to the targets are known before and after the strokes. Players affect ball displacement by controlling clubhead velocity and clubface angle during club and ball impact. X-factor studies have produced ambiguous results, possibly caused by different definitions of upper torso, rotation and top of backswing. Higher clubhead speed is generally associated with larger spinal rotation and shoulder girdle protraction at the top of the backswing. It is also associated with higher ground reaction forces and torques, a bottom-up and sequential increase of body segment angular velocities, a rapid increase of spinal rotation and a late adduction of the wrists during the downswing. Players can increase the clubhead speed generated by a swinging motion by actively adding a force couple. Wrist, elbow and shoulder force couple strategies should be differentiated when investigating the technique. Physical parameters such as anthropometrics, strength and flexibility are associated with skill level and clubhead speed. Current studies have investigated the linear correlation between arm and shaft lengths and clubhead speed, but a quadratic relationship may be stronger due to changes in moment of inertia. Fitness training can increase and perhaps decrease the clubhead speed and striking distance, depending on training methods and the player’s fitness and level of skill. Future studies may focus on individual training needs and the relationship between physique, execution and its relation to accuracy of impact and ball displacement.

Competition is extremely tough in world-class sports, forcing athletes to continuously find means of improving their performance. It is therefore important for both athletes and experts to help them to gather relevant information and accurately assess strengths and weaknesses before making interventions.[1] Golf coaches, fitness trainers and other experts help elite players toward lower scores. Research may help them to achieve that goal more effectively. The purpose of this paper is therefore to review the research concerning elite golfers (i.e. professionals and amateurs with a handicap of <5) relevant to the competitive outcome, with focus on playing results, technique and physique. One review article, 44 original research articles, 28 proceedings from the World Scientific Congress of Golf, two other conference proceedings, 14 books and ten dissertations were reviewed. The databases searched were PubMed, SPORTDiscus and ProQuest, with the key words ‘golf’ and ‘elite’ or ‘professional’. Additional manual searches were made through article reference lists. For the purpose of ª 2009 Adis Data Information BV. All rights reserved.

this review, all swing-related activity refers to a right-handed golf swing. 1. Playing Results 1.1 Total Score

The total number of strokes taken is the ultimate measure of performance when evaluating professional golfers. There are ongoing attempts on most professional tours to measure the players’ game. Several variables are used, such as ranking points, money earned, top finishes and scores. Several statistical studies have used earnings[2-9] or scoring average[2,4,5,9-12] as the dependent variable to examine the relative importance of various parts of the game such as driving accuracy, greens in regulation, sand saves and putts per round. Various ranking systems such as top finishes only give the relative performance compared with the results of the other competitors, and these are based on an arbitrary figure. Top finishes give Sports Med 2009; 39 (9)

Competitive Elite Golf

information on whether a player is better or worse than the other competitors (ordinal scale), but not by how much (interval or ratio scale). When comparing ranking points, one has to consider several variables such as the strength of the field and previous results. The system depends on the players’ rankings in each competition. A player can in theory achieve a personal best score and still not be awarded ranking points. Money won depends on the sponsor’s purse and the value of the currency, which can vary considerably. Using prize money as the measure also makes comparison between professionals and amateurs difficult. The total score can be measured in many ways. It is common to analyse the total score of a round and tournament, and the average score for several rounds. The average score is called scoring average or scoring average (actual).[9,13] The condition of the course and the weather may change however, making the scoring average (actual) imperfect as a single evaluator of performance. The compared players do not all play in the same tournaments. Scoring average (adjusted) has recently been used to compensate for the difference in course difficulties and weather conditions.[9,13] It is a weighted scoring average, which takes the stroke average of the field into account. Scoring average (adjusted) has shown a stronger correlation to earnings than scoring average (actual) [-0.77 vs -0.68], and it may be a better measure than scoring average (actual) when comparing tour players over a year.[9] Thus, there is no single foolproof variable for evaluating performance.[14] The total score, however, determines the competitive outcome and reflects the idea of the game. 1.2 Partial Scores

It is important to evaluate the partial scores that make up the total score to optimize the training. The parts of the game are often categorized by the different techniques, such as driving, bunker shots and putting. Statistical information can be gathered and presented in numerous ways,

725

such as round-by-round, hole-by-hole or shotby-shot. Most professional golf tours measure the statistics round by round, for example, the number of putts per round, greens hit in regulation, sand saves, etc.[8,15] They also present hole-by-hole statistics such as the score in relation to par on par three, par four and par five holes. Research indicates that some of the tour statistics can be used as predictive tools for low scoring average.[2,10] Better skills in putting, short game and long game are related to lower scores.[2,5,10,12] 1.2.1 Putting

The ability to putt well is important for the total score. The best players on the Professional Golfers’ Association (PGA) Tour differentiate themselves by better putting than the rest of the field.[5] The average number of putts made on the greens (putts per round) have shown strong correlations to scoring average (actual and adjusted) [table I].[2,4,9-11] However, the number of putts per round may increase if the players hit more greens in regulation (GIR1).[2,11] Players that miss many greens may hit a chip, pitch or bunker shot and thereby hit their first putt from a closer distance to the hole than those hitting many GIR. Putts per GIR was created to show a putting statistic that reduces the effect of the short game. It is calculated as the average number of putts per greens hit in regulation. Tour statistics show that putts per GIR became lower between 1968 and 1993[16] and between 1990 and 2004.[12] Better green qualities[16] and thereby an increased ability to read the greens[17] may explain that improvement. Putts per GIR has shown a stronger correlation to scoring average (adjusted) than putts per round (0.31 vs 0.63),[13] and is therefore a relevant measure for putting skill.[5,12,13] 1.2.2 Short Game

The best players on the PGA Tour distinguish themselves from the rest of the players by a more accurate short game.[5,18] It is important for the score, as chipping, pitching and bunker shots will have a large effect on the number of

1 Greens in regulation (GIR) is the number of greens reached on two shots (or less) than par for the hole.

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (9)

Hellstro¨m

726

Table I. Investigations into the relationship between total scores and partial scores: scoring average (actual or adjusted) and part scores and skills Study

Tour yearsa

No. of players

Findings

Davidson and Templin[2]

1983

119 of top 125

GIR was found most important for scoring averageb, then putts per round and driving distance. These skills explained 86% of the variance in scoring average and 59% of money earned. Sand save did not have appreciable effect on scoring average. Scoring average correlated strongly (-0.77) to earnings

Jones[4]

1988

99 of top 100

Score before cut had the strongest correlation with scoring average (0.78) and money earned (0.60). Putts per round (0.36), GIR (-0.29), driving distance (-0.21) and sand save (-0.20) correlated significantly* with scoring average

Nix and Koslow[11]

1987

Top 100

Lower scoring average correlated strongly to increased earning (-0.74). GIR and putts per round explained the scoring average to 83%. Adding driving distance and sand saves improved the explained variation only marginally (87%)

Larkey and Smith[20]

1997

195

The authors recommended putting, sand save, driving distance, driving accuracy, scrambling and GIR as all-round measures on Tour players’ skills. Normalization was suggested for combining these skills instead of using rank summary computations. Scoring is strongly related to earnings (-0.93) and top 10 (0.83)

Belkin et al.[10]

1986–88

180, 182, 186

GIR were most strongly related to low scoring average during all years (-0.64 to -0.82), followed by driving accuracy (-0.53 to -0.58), putts per round (0.32 to 0.48), sand saves (-0.20 to-0.39), and average driving distance (-0.11 to -0.26). These skills accounted for 84% of the variance in scoring average over the 3-year period

Dorsel and Rotunda[5]

1990

130 and top 42

GIR and putts per GIR significantly* correlated with the scoring average (-0.45 and 0.24), top 10 finishes (0.26 and -0.22) and earnings (0.24 and -0.28). Driving accuracy was only significantly associated with scoring average. Driving distance was not significantly related to any performance variable, and showed weaker correlations than driving accuracy in all three categories. Putts per GIR was the top predictor for the best 42 players. Scoring average correlated strongly with both top 10 finishes (-0.79) and earnings (-0.74)

Finley and Halsey[9]

2002

196

Proposed scoring average (adjusted) as better measure of performance than scoring average (actual). The former correlated stronger (-0.77) with earnings than the latter (-0.68). Scrambling (-0.68 and -0.67), GIR (-0.73 and -0.63), bounce back (-0.40 and -0.41), putts per round (0.34 and 0.36), driving accuracy (-0.43 and -0.36) and sand saves (-0.23 and -0.30) were significantly* related to scoring average (actual and adjusted), but driving distance (-0.18 and -0.15) was not. The best three-way model for explaining the variance in scoring average (actual) was GIR, putts per round, and scrambling (R2 = 0.93)

Wiseman and Chatterjee[12]

1990–2004

179–202

GIR and putts per GIR were the skills explaining the most variance in scoring average over all years. Total driving (driving distance [yards] driving accuracy [%]) correlated stronger to scoring average than either driving distance or driving accuracy. The effect of total driving on score seemed to decrease (r = -0.38 to -0.50 between 1990 and 2002, and -0.24 and -0.29 during 2003 and 2004, respectively). Driving accuracy correlated stronger with scoring average than driving distance during all years, except 1985




Quinn[13]

2004

196

Scoring average (adjusted) correlated stronger to putts per GIR (0.63) and GIR (-0.62), than putts per round (0.31), driving accuracy (-0.15) and driving distance (-0.05)

a

The statistics are based on the men’s US Professional Golfers’ Association Tour data.

b

Scoring average is an abbreviation of scoring average (actual).

GIR = greens in regulation; * p < 0.05.

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (9)

Competitive Elite Golf

one-putts.[18,19] Previous investigations of the short game use data that is a combination of short game and putting skills. Sand saves are calculated as the percentage the hole is reached with two strokes or less from a greenside bunker. It has shown a significant correlation to the scoring average in some,[4,9] but not all,[2] studies. Scrambling, the ability to make par when missing GIR, is now used by the PGA Tour, as previously suggested by Larkey and Smith.[20] It has revealed strong correlations to scoring average (actual and adjusted) [-0.68 and -0.67].[9] Scrambling is currently the best general measure of the combination of short game and putting skills. Future studies should separate short game skills from the putting skill, and investigate chipping and pitching skills too. 1.2.3 Long Game

The ability to hit the ball a long way and to be accurate from tee to fairway affects the final score. The field average and longest driving distance on tour increases each decade.[16,21] A trend was found between 1986 and 1988 toward increased importance of driving distance for the scoring average.[10] The PGA Tour increased the difficulty of the hole placement in 1988. Longer drives may have increased ability to aim for the pins, as the course was set up then.[10] However, driving accuracy2 has been shown to be a better predictor of the scoring average than driving distance.[9,10,12,13] Driving accuracy had stronger correlation to scoring average than driving distance between 1990 and 2004, except for 1995.[12] The explanation for the change that year may be the introduction of titanium clubheads, which resulted in an increased elasticity of the clubface and higher initial ball speed after impact.[22,23] Longer shots are more likely to miss the fairway when the fairways are set up to be very narrow, which may in turn reduce the importance of driving.[12] Course design and weather conditions will affect scores, so game statistics should be related to the environmental conditions. The combination of driving distance and driving accuracy is called total driving. It has

727

been measured in several ways,[2,6,8,12,21] and the results should be interpreted with caution. The US PGA Tour adds the players’ ranking in driving accuracy and driving distance to obtain a measure of total driving, which has been used in investigations.[6-8] This solution is problematic because non-numerical, ordinal-scaled data are introduced that cannot be meaningfully summarized. Currently, the most effective method of calculating total driving is probably to use standardized z-scores, which takes into account the magnitude of the differences between players on driving accuracy and driving distance.[21] However, driving distance is the average distance recorded only on two holes per round (in opposite directions to counteract the effect of wind) where the players are likely to use their drivers. Driving accuracy, on the other hand, is measured about 14 times per round (tee shots on all par 4 and 5 holes), and the players may use irons and fairway woods on several of those holes. Thus, there is a validity problem in combining these two skills into one parameter. Driving accuracy should only be measured on holes where the driver is used. The skill previously named driving accuracy should be named ‘tee to fairway accuracy’, to better communicate the skill measured. Long game accuracy of approach shots has not been uniquely investigated. GIR is a combination of shots from the tee, fairway and other positions toward the green. GIR has shown the strongest correlation to scoring average in several investigations.[2,9,10,12,13] It is currently the best measure for the long game skill used in the scientific literature. The difference in GIR when hitting from on or off the fairway should be investigated in the future. 1.2.4 Shot by Shot

Most score-related research has been based on statistics published by the PGA Tour, which have been somewhat imprecise with much nominal data (table I). Nominal data, such as on or off the fairway or green, do not say how far off target the player was, or how severe the penalty was. It might be tactically wise to aim toward the rough

2 Driving accuracy is defined as the percentage of tee shots on par 4 and par 5 holes that end up on the fairway.

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (9)

Hellstro¨m

728

1.3 Ball Displacement

The most used pedagogical model in golf instruction is Laws, Principles and Preferences.[25,26] This model considers five parameters that affect the ball flight: clubhead speed, centredness of contact, clubhead path, clubface position and angle of approach. However, these definitions do not adequately explain all ball flights. For example, the clubface position is defined as ‘‘the degree at which the leading edge of the clubface is at right angles to the swing path’’ (p.5),[25] which only considers the closed/open position and not the loft angle nor the tilt (toe up/ down) angle. The model can therefore not explain why the five parameters can be constant and the players can still achieve different launch angles, or why the launch angle and sidespin of a ball will change when struck from a higher or lower position than the feet. Furthermore, the angle of approach is defined as ‘‘the angle formed by the descending or ascending arc of the clubhead on the downswing in relation to the slope of the ground’’ (p.5).[25] A relation to the slope instead of the horizontal plane means that two identical ball flights can have different attack angles if the ground conditions differ, which can lead to confusion. ª 2009 Adis Data Information BV. All rights reserved.

Linear velocity

Clubfa

ce norm

lar gu y An locit ve

or away from the pin, depending on the course design. The Percent Error Index (PEI) is the distance that the ball finishes from the target divided by the distance to the target before the impact.[18] The advantage of using the PEI is that the relative precision between clubs, shots and distances can be compared. Pelz[18] found that the PGA Tour players had a PEI between 5% and 9% for all fullshot distances and clubs, a putting PEI of 5–10%, and a short game PEI (less than full shots) of 13–26%. The long shots are generally missed more in the lateral (sideways) direction, whereas short game and putting errors tend to be related to distance (too short or too long).[18,24] The US PGA Tour have started to publish some detailed shot-by-shot statistics. Investigations of such detailed game statistics are valuable to increase the understanding of the current standard and development of elite golf.

al

D-plane ity

loc

e dv

a

he

ub

Cl

Target Fig. 1. The D-plane.

Jorgensen[27] presented a model where initial ball trajectory and rotation after impact follow the so-called D-plane (see figure 1), which lies between the clubhead velocity vector and the clubface normal vector. Since ball rotation is along the D-plane, the lift force will initially lie in the D-plane. This model considers the directions in all three dimensions of the clubhead velocity and the clubface normal. Thus, the concept of the D-plane[27-29] is more accurate in describing ball flight than the model in Laws, Principles and Preferences.[25,26] However, there is a difference between measuring clubhead velocity and clubface angle at the instant before impact with the ball, during the contact phase or at the instant of release. The D-plane model does not say exactly when to measure the clubhead speed and clubface angle and therefore needs to be further developed. In summary, it is helpful to analyse the game statistics to set goals and plan the training toward lower scores. Players can record if they hit the GIR and the number of putts on each green at the same time they write the score on each hole. Scrambling and putts per GIR can then be calculated. Statistics are then available on the long game, short game and putting, which have shown strong correlation to the scoring average. Motivated players can record the distance to the target before the swing, club selection, ball flight, where the ball carried and the distance to the target sideways and distance when it stopped. This information makes comparing of the relative precision between clubs and distances possible, and helps players to understand what circumstances create the highest and lowest PEI. Furthermore, Sports Med 2009; 39 (9)

Competitive Elite Golf

players can use this information to plot the shot variation over an image of the next tournament course to improve the game strategic decisions. 2. Technique 2.1 Impact

Initial ball flight and change in the momentum of the ball depend on the impulse, i.e. the time that the normal force and a tangential force act on it during impact. The total momentum of the ball and clubhead is almost conserved through impact (some energy is lost due to heat dissipation and permanent deformation[30]). The largest transfer of momentum from clubhead to ball occurs when the resulting force passes through the centre of mass of the clubhead and ball.[31] A greater decrease in linear clubhead velocity during impact is a sign of effective transfer of momentum, which has been shown to result in longer drives.[30,31] However, due to the need for backspin, which increases the lift force during flight, the resultant force passes under the centre of the ball, decreasing the ball’s linear momentum and increasing its angular momentum.[32] The D-plane model does not consider that the impact occurs during a time interval. The ball travels with the clubhead for about half a millisecond and 2 cm for a driver. During that time, it is affected by accelerations of over 50 000 m/s2.[33,34] Two forces affect the ball at impact: the normal force and the tangential force parallel to the clubface. The tangential force is the result of friction with the clubface.[30] The resultant force vector gives the ball a ‘line of compression’.[35] What normally is called compression is a deflection under load, which is measured and reported in different ways by different golf ball manufacturers.[36] If the impact on the clubface is offcentre, some momentum ‘leaks’ into rotating the clubface. The clubface angle can change considerably after impact, and thus decrease the ball flight distance and increase the dispersion from the intended flight direction.[34,37] A golfer can, however, only strive to achieve the desired ball displacement by creating the suitable combinaª 2009 Adis Data Information BV. All rights reserved.

729

tion of clubhead velocity and clubface angle during impact, and by choosing the appropriate club for the intended shot. 2.2 Clubface, Clubhead and Shaft 2.2.1 Clubface Angle

The clubface angle is important for initial ball velocity and rotation.[29,31] The clubface angle changes due to the torque applied at the grip. The clubface rotates (around the shaft’s long axis) approximately 90 in the backswing in relation to an imaginary leaning plane between the target line and swing centre between the shoulders. Cochran and Stobbs[38] suggested that approximately 30 arises from the rotation of the forearms, and the other 60 from a combination of arm raising and shoulder turning. The left forearm pronated, on average, 6 during the backswing, and the left shoulder rotated inward 22, in a kinematic study of ten elite golfers.[39] The left forearm starts to supinate when the wrists adduct (un-cock) and the shaft releases into impact, making the clubface rotate in order to become square to the target line.[38,39] Better players have low variability in forearm rotation.[39] The variability of clubface rotation during swings between golfers is relatively high during the backswing, but less at impact.[39] If the clubface is rotated more or less during the backswing, it has to rotate more or less during the downswing in order to return with a square clubface. Many combinations of body rotation, tilt and bend can achieve this clubface rotation during the swing. However, the optimal way of controlling the clubface during different clubhead speeds is yet to be established. The long-axis rotation of the forearms and upper arms contribute a substantial amount of speed of the hitting area in racquet sports,[40] and the importance of this in the kinematic chain, clubhead speed and clubface control, should be further investigated in golf. 2.2.2 Clubhead Velocity

The ability to achieve maximum linear clubhead velocity during impact increases in general with a higher level of skill.[41-44] The linear clubhead velocity (vch) is a function of the grip’s linear velocity (vgrip) and the shaft’s angular velocity Sports Med 2009; 39 (9)

Hellstro¨m

730

(oshaft) multiplied by the length of the shaft (lshaft).3 Choosing a longer shaft extends the lever. There is, however, an optimal shaft length due to the lack of strength of the player to continue to accelerate the club angularly with its increased moment of inertia. There is only a small linear and angular acceleration of the club at the beginning of the downswing. Elite golfers have, in general, lower angular shaft velocity at the start of the downswing and higher velocity later in the downswing than average and low skilled players.[45] Moreover, investigations show that better players also reach maximum grip velocity just before impact,[46,47] and the clubhead does not pass their hands until after impact from a frontal plane view.[38] Some reports suggest that maximum clubhead velocity occurs just before impact instead of at impact.[31,46,48,49] Further investigations are needed to determine if this is due to a technical strategy to control the clubface, a statistical artifact caused by including the clubhead deceleration at impact, or something else. 2.2.3 Shaft

According to Butler and Winfield,[50] players’ shaft loading profiles can be used to categorize their technique. They suggested that most players would fit into one of three different categories. The differences between their suggested categories lay in the amount of shaft bending in the toe/heel direction of the clubhead and the time from the top of the backswing to impact. Category one players had a fast downswing with a distinct ‘single peak’ in shaft bending in the toe direction, which resulted in the greatest shaft bending and highest peak force. Category two had a small ‘double peak’ shaft-bending profile, which indicated that these players loaded the shaft twice during the downswing. Category three had a slower ‘ramp-like’ loading, with about the same shaft deflection in the toe direction and a slightly slower tempo than category two. The shaft changes from backward bending (back strain) in the downswing to forward bending (face strain) at impact for most golfers.[41,51]

The forward bend at impact increases with a longer club. The dynamic loft might change so much for a driver, due to the dynamics of the shaft, that the launch angle of the ball increases.[52] However, a straight shaft at impact is optimal to transfer kinetic energy to the ball.[50] The increase in clubhead speed by the ‘whipping’ effect of the shaft seems to be insignificant, according to some investigations employing computer simulations.[52,53] The correlation between clubhead speed, grip speed and shaft flexibility was investigated with both a swing machine and real golfers.[54] The clubhead speed was then found to change depending on the stiffness of the shaft and grip speed. It was suggested that the relationship between clubhead speed and grip speed is determined partly by shaft flexibility and partly by the player’s ability to adjust to the shaft dynamics. Thus, elite players use the flexible property of the shaft better than average amateurs to maximize striking distance,[41,45] but the effect has to be investigated further. The player’s loading profile on the shaft, especially in the toe/heel direction, is valuable when analysing technique. 2.3 Hands on the Shaft, Wrists and Arms 2.3.1 Pull and Centripetal Force

The ability of the arms and wrists to do work on the shaft determines clubhead velocity and clubface angle. Total work is the combination of linear work (force · linear displacement) and angular work (torque · angular displacement). After torso and hip muscles, the upper extremities constitute the next most important area in producing work on the club, contributing about a quarter of the total work to the shaft, according to results derived from computer models.[55] The resultant force acting on the shaft while swinging can be divided into a centripetal (radial) force and a pull (tangential) force. The centripetal force acts from the grip towards the centre where the hand-arm-shoulder complex rotates, approximately around a central point between the shoulders.[56] It accelerates the club angularly

3 vch » vgrip + oshaft · lshaft; the bending of the shaft decreases the effective lever slightly. ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (9)

Competitive Elite Golf

due to the lever created between the force vector and the club’s centre of gravity. The pull force primarily accelerates the club linearly. However, it can also create an angular acceleration in a direction opposite to the radial component if it is acting on the other side the club’s centre of gravity. The result of the centripetal and pull forces is acceleration of the club in a circular path and rotation of the club relative to its centre of mass. The adduction of the wrists and the ‘release’ (shaft angular acceleration) occur when the torque due to the centripetal force surpasses the opposite torque due to the pull force (i.e. the hands slow down).[56] 2.3.2 Force Couple

Players can apply forces in the opposite direction on the shaft with the left and the right hand. The effect of actively applying a force couple during the full swing is a subject of debate. Some predictive studies suggest that the increase in clubhead speed by such action is at best very limited and requires very accurate timing.[27,48] However, these models were based on two stiff levers (‘shaft’ and ‘left arm’) and did not consider the muscle dynamics. A model with three levers (including ‘left shoulder’) was used later, which also considered the force-velocity curve of human muscles. The increase in clubhead speed by applying wrist torque was then found to be no more than approximately 9%.[57] These predictions were based on two-dimensional models that used a planar downswing.[27,48,57] During real-life swings, however, the clubhead, left wrist and left shoulder do not move in a consistent plane, and three-dimensional models are therefore valuable.[58] Furthermore, the validity might improve if more levers were used,[59] and the shaft is considered as a non-rigid body.[41,54] This has now been achieved. Nesbit and Serrano[45,55] used a three-dimensional, full-body computer model, with 15 body and 16 golf club segments, with representative mass and inertia properties, to investigate the kinematics and kinetics of the full golf swing. They performed energy analysis on 85 players’ swings, and made further comparisons between four players with handicaps between 0 and 18. The wrists were ª 2009 Adis Data Information BV. All rights reserved.

731

found to behave almost like a free hinge during the late downswing phase into impact, but not earlier in the downswing. 2.3.3 Wrist Torque Generators

How a force couple can be generated on the shaft is seldom discussed or differentiated in predictive analyses. The left and right wrist can create a force couple using the forearm adductors. However, a force generated from the right triceps brachii is more effective in creating power, due to a longer lever and larger crosssectional area of muscle. The right elbow does most work of all the upper body joints due to its large angular displacement.[55] The right hand does not need to grip the club to push the shaft when extending at the elbow, thereby decreasing the risk of slowing down the shaft at release. Furthermore, both pectoralis major muscles seem to have an important role in the downswing. Electromyography (EMG) studies indicate high levels of activation on both sides immediately prior to impact.[60,61] The left pectoralis major may be activated eccentrically when trying to decelerate the left arm, while the right is working concentrically trying to accelerate the right arm. The resulting forces can create a force couple on the shaft that may not yet have been considered in predictive analyses. Players who use different wrist torque generators might also have distinctive shaft loading profiles.[50] Future muscle modeling studies are needed to investigate the forces produced and the work done by groups or individual muscles of the upper body. Such studies would provide information on the contributions of the key muscle groups of the upper body to the downswing. 2.3.4 Work Analysis

Data from the players (n = 85) in Nesbit and Serrano’s studies[45,55] showed significant correlations between clubhead speed and total work, as Newton’s laws of motion predict. Total work was calculated as each player’s ability to apply forces and torques in the direction of motion during the downswing. When four players with different playing handicaps within that group Sports Med 2009; 39 (9)

Hellstro¨m

732

were compared, the best player first worked at a slower rate and then faster through impact than the other three less skilled players. The best player achieved the highest total work, and was able to peak his total work closest to impact, by reducing his power output to zero at impact. It was found that better golfers use their arms more than their wrists to do work, with a ratio of about 1.4 : 1 for the best player. The arms were also found to be more important than the wrists in generating power for all 85 subjects. Coming into impact, the wrists could not keep up with the shaft’s angular acceleration but the arms could pull inwards and decrease the distance from the grip to the middle of the upper torso, thereby increasing acceleration more (see also Miura[62]). The linear power therefore peaked after the angular power. Thus, it is important to consider factors such as range of motion, timing and the sustainability and amount of force and torque when analysing the golf swing.[45,55] 2.4 Torso, Legs and Feet on the Ground 2.4.1 Ground Reaction Force and Centre of Pressure

A sequential start of the body segments, commencing from the ground and moving up to the (anatomically) superior segments during the downswing, is beneficial in increasing clubhead speed at impact.[46] This is not necessarily desirable during short shots that do not require high clubhead velocity.[18] The characteristics of ground reaction forces during the full swing may differentiate between players’ skill levels. Elite players transfer their weight toward their right foot during the backswing.[63-65] The lower body initiates the downswing while the upper body and club continue the backswing.[46] Better players may create maximum right foot torque earlier than less skilled players.[44] The right foot pushes in the posterior direction and the left foot pushes in the anterior direction, creating a force couple resulting in a hip slide and anti-clockwise rotation. Some research suggests that better players transfer more weight from their right to left side in a shorter time during the downswing than lesser skilled players.[61,64] ª 2009 Adis Data Information BV. All rights reserved.

Later research questions whether there is one optimal way of transferring weight.[66,67] A cluster analysis of 65 players (ranging from professionals to high handicap amateurs) indicated that there are two major styles of weight transfer away from/toward the target.[66] ‘Front foot’ (left foot) style players continued to move their centre of pressure toward the target through impact, while ‘reversed’ style players then moved their centre of pressure back toward the right foot. When comparing these two groups, it was found that front (left) foot players’ clubhead speeds were associated with a larger range of centre of pressure movements toward the left foot in a shorter time during the downswing.[67] Reversed players’ clubhead speeds were associated with centres of pressure further from the right foot than the other players at the top of the backswing. Reversed players also moved their centre of pressure toward the left foot at the start of the downswing, but the direction then changed. A more rapid transfer of centre of pressure towards the right foot during impact was for this group associated with higher clubhead speeds.[67] The authors suggested that neither style was a technical error since there were elite players in both groups.[66,67] However, a reversed strategy might be an effect – not a cause – of a sub-optimal technique. The reversed players may need to tilt back to adjust from a tilted movement in the backswing to make their clubheads approach to ball at a better angle, which may create more stress on the hip joints and lumbar spine.[68,69] Another possible explanation is that the reversed players do not have enough strength to counteract the forces from the club when it releases into impact. A more rapid movement of the centre of pressure toward the right foot may be explained by an increased need to counteract the higher force from the club when swinging faster. Thus, the different centre of pressure profiles may only be fully understood by analysing them from a broader perspective by considering more variables. 2.4.2 X-Factor and Stretch-Shortening Cycle

Research on skilled players’ pelvic-upper torso separation at the top of their backswings Sports Med 2009; 39 (9)

Competitive Elite Golf

733

Table II. Investigations into elite golfers’ X-factor and X-factor stretch Study

Subjects

Method

Findings

McTeigue et al.[70]

N = 131; 51 US PGA Tour players; 46 US PGA Senior Tour players; 34 amateurs

Rate gyroscope used at 100 Hz. Six medial positioned potentiometers posterior on pelvis and mid-thoracic spine. X-factor was defined as ‘‘the relative rotation of the golfer’s upper body to hips’’ (p.52)

Mean X-factors: 32 (PGA); 29 (senior PGA); 34 (amateurs). 70% of PGA players initiated the downswing with a rotation of pelvis before upper torso. 20% rotated the upper torso before their pelvis and 10% started them simultaneously. The 10 PGA Tour players who ranked top 50 in driving distance had a larger X-factor than the rest of the tour players

Cheetham et al.[71]

N = 19; 10 skilled (8 PGA Tour players and 2 players with hcp £0); 9 less skilled (hcp >15) players

Electromagnetic system used at 30 Hz. Sensors positioned posterior medially on pelvis and at T3. X-factor was defined as the ‘‘difference between the rotational position of the pelvis and T3 sensors’’ (p.194)

The X-factor did not significantly* differ between the groups at the top of the backswing. It increased in the downswing (X-factor stretch) for both groups, and the increase was significantly greater for the skilled group (19%) than the less skilled group (13%)

Zheng et al.[47]

N = 72; 18 PGA Tour players; 18 low (mean hcp 3.2),18 mid (mean hcp 12.5) and 18 high (mean hcp 21.3) hcp golfers

Optoelectronic system used at 240 Hz. Retroreflectiva markers were placed on the greater trochanter, and the upper torso (bilaterally on the tip of acromion). X-factor was defined as the angular difference between the vector of the upper torso and pelvis in the transverse plane

The PGA Tour players had significantly* larger X-factor in the backswing than the high hcp group. In the downswing, PGA Tour players came into impact maintaining a significantly larger X-factor than the high and mid hcp groups

Myers et al.[73]

N = 100; HBV group (n = 14, mean hcp 1.8); MBV group (n = 65, mean hcp 7.8); LBV group (n = 21, mean hcp 8.1)

Optoelectronic system used at 200 Hz. Retroreflective markers measuring the pelvis were positioned on sacrum, and bilaterally on the anterior superior iliac spine. The markers defining the upper torso were positioned at C7 and bilaterally on acromion. X-factor was defined as the angular difference between the vector of the upper torso and pelvis in the transverse plane. Ball velocity was assessed with golf ball radar

Significant* group differences were found between an increase in ball velocity and increased X-factor at the top of the swing (HBV vs LBV and HBV vs MBV), and X-factor stretch (between all groups)

HBV = high ball velocity; hcp = handicap; LBV = low ball velocity; MBV = medium ball velocity; PGA = Professional Golfers’ Association; * p < 0.05.

(so-called X-factor[70-72]) has shown ambiguous results (table II). Studies using medial positioned sensors close to the spine[70,71] have not found significant associations between pelvic-upper torso rotation at the top of the backswing and striking distance or skill level, but studies using laterally positioned markers at the tip of acromion[47,73] have. Thus, the combination of spinal axial rotation and left shoulder girdle protraction at the top of the backswing is ª 2009 Adis Data Information BV. All rights reserved.

important to consider when striving for longer striking distances. Another possible source of confusion when comparing X-factor studies is the different definitions of rotation. The intention behind the X-factor concept is to illustrate the rotation between pelvis and upper torso segment angle, which can be investigated relatively simply in two dimensions. However, the golf swing is performed in three dimensions with rotating, Sports Med 2009; 39 (9)

Hellstro¨m

734

bending and tilting of the spine. Rotations in two different planes are compared when measuring the rotation of pelvis and upper torso, since the segments have different bend and tilt angles during the swing. When projecting both hip and upper torso segment rotation on one plane (such as the global horizontal plane), the bending and tilting motion will affect results. The findings are not reproducible if only one of three angles per segment is presented in three-dimensional studies (only rotation, without bend or tilt). Thus, these methods may give different results when investigating the X-factor. The maximal increase of pelvic-upper torso separation during the downswing (so-called Xfactor stretch[70,72]) seems important regardless of whether the upper torso is measured by medial or lateral sensors.[71,73] The increased muscle stretch during the change of direction has been suggested as creating stretch reflexes and elastic energy, which leads to increased rotational velocities in the more distal segments.[71] However, it is not known if the turn at the top of the backswing elicits a stretch reflex, and the elastic energy is largely a function of the muscle force acting in the muscles and tendons. Parts of the lower body and torso musculature are eccentrically activated before the direction changes in the downswing.[61,74,75] The increase in striking distance associated with X-factor stretch is probably mainly due to the increased force attained through the eccentric action and the increased time which the force can be applied before the downswing, compared with a downswing without an eccentric pre-load.[76] The mechanism and the relative contribution of force from various factors during the rotational stretch between pelvis and upper torso should be investigated further. 2.4.3 The Kinematic and Kinetic Chain

The principle of summation of angular velocity in the golf swing has been discussed for decades.[38] Elite golfers reach their peak angular velocity in the downswing during the full swing in

a bottom-up sequence (i.e. from the segments closest to the ground up through the body and lastly to the clubhead), with an increase in velocity in each segment.[46,47] The kinematic chain of 13 male professional golfers’ full swings was analysed by FujimotoKanatani.[46] Their inter-knee, hip and shoulder segments reached their respective maximum angular velocities (projected on the horizontal plane in the global coordinate system) in the downswing at 93.5%, 94.5% and 97.0% of the swing.4 The maximum linear velocity of the clubhead was achieved just after the wrists. Nesbit and Serrano[55] found that the generation of work during the full swing, i.e. forces and torques that move the club, is also a bottom-up action.[55] There is segmental summation from the legs, hips, lower trunk, upper trunk, shoulders, arms and wrists. When the joints stop doing work, they become static or move slightly in the opposite direction.[55] The leg muscles were not found to generate much work on the club (around 4%), but gave support to the body and aided the hip motion. The trunk and hip muscles generate most of the work on the club (approximately 70% of the total body work). Furthermore, a smooth, constant torque was recommended in order to achieve high clubhead speed.[55] EMG studies further indicate that professional players activate their muscles in a bottom-up sequence.[61,74,75] Individual muscle activation is also more consistent in better players than in poorer players, although there are large variations between expert players.[77-79] Thus, it seems important to achieve peak maximal angular velocity and to generate work (forces and torques that generate displacement) in a sequence from the segments closest to the ground and up towards the club during full swings.5 The rate of increase or decrease in segmental velocity (acceleration-deceleration) work and power is also an important factor in determining maximal clubhead speed at impact.[45,55] In summary, the initial ball speed and direction, and spin rate and spin axis orientation are

4 Defined as the time between the start of the backswing (0%) and ball impact (100%). 5 See Kreighbaum and Barthels[80] and Putnam[81] for further discussion.

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (9)

Competitive Elite Golf

decided at impact. The D-plane model may be used to explain the relation between impact and ball displacement. The ability to repeat the swing is important to control the ball displacement and to hit many GIR. Players may constrain the right knee and hip movement during the take away to increase control and to create a large X-factor separation in the backswing. They should rapidly increase this separation further in the start of the downswing (X-factor stretch). A stable base with the right foot toward the ground is then important to create necessary forces and torques needed for the weight transfer and pelvic movement. The downswing should have a bottom-up sequential increase in angular velocities and a smooth force and torque production. The upper torso should reach maximum velocity after the hips. The wrists un-cock and the right elbow extends after the shoulders reach maximum angular velocity to maximize the clubhead speed at impact leading to long striking distances. 3. Physique 3.1 Elite Players’ Physique

Golfers may swing better if they are strong and flexible, and have good balance. Elite players (n = 45, handicap <0) showed significantly larger hip, torso and shoulder strength and flexibility than less skilled golfers (n = 92, handicap 10–20).[82] However, the average peak torque was normalized to the subject’s bodyweight before comparing strength between the groups. Such normalization may not be valid in golf, since there are no playing categories based on body mass. Elite players also had better balance with their eyes open using their right legs, but not with left leg or when having their eyes closed. Such a difference indicates that the golf swing requires better balance on the backswing than in the finish. The physical differences between the groups may be due to different volumes of fitness and golf training. Some physical differences between elite and less skilled players may be explained by the demands of the game (i.e. natural selection). Twenty-three Ladies PGA Tour players with rankings of 1–135 on the 1981 money list were examined ª 2009 Adis Data Information BV. All rights reserved.

735

regarding height, mass, percentage body fat, resting heart rate, blood pressure, absolute and relative grip strength. and maximum oxygen uptake/consumption (VO2max).[83] Compared with studies of other, larger female populations, the authors found that mean height and mass were a little higher than average (mean 168.9 cm and 61.8 kg), heart rate and blood pressure were low (67.1 beats/min and 113/73 mmHg, respectively), the right and left grip strengths were good. to excellent (36.5/36.0 kg, respectively), and VO2max was fair (34.2 mL/kg/min). A group of professional male Japanese golfers (n = 11) had significantly greater limb girth than other Japanese golfers (n = 52) and non-golfers (n = 45). Their body types tended to be more mesomorphic (an athletic and muscular appearance).[84] When 15 Caucasian male professionals were compared with 18 sedentary subjects, the only significant difference (p < 0.05) between the groups was a 9% increase in muscle mass in the dominant arm of the professional golfers.[85] Professional players are perhaps taller and heavier than the normal population, but there are also short world-class golfers. 3.2 Physique, Clubhead Speed and Ball Flight

There is an association between physique and driving distance.[86] Muscle power is important to consider when training for clubhead speed. The relationship between muscle power, maximum grip strength, anthropometrics and swing speed was investigated in 40 golfers with handicaps <3. Power was tested by using a modified Sergeant Jump reach test, by a rotary abdominal machine and a mass stack potentiometer, and by a multihip flexor machine and mass stack potentiometer. The four strongest correlations with swing speed were trunk power (r = 0.80), leg and hip power (r = 0.60), combined arm and trunk power (r = 0.58) and grip strength (r = 0.54).[87] Height (r = 0.51) and arm length (r = 0.45) showed weaker but significant (p < 0.05) correlations, whereas body mass (r = 0.22) and shoulder width (r = 0.20) were not significantly related to swing speed. Thus, there are weak to strong correlations between physical attributes and clubhead speed, Sports Med 2009; 39 (9)

736

although experimental studies are needed to establish causality (table III). Long arms, a large usable muscle mass that creates high maximal strength and power, and flexibility are likely to affect the players’ technique and their ability to create clubhead speed. Players can probably utilize longer levers (arm/shaft complex) until the levers become so long that the increase in moment of inertia becomes too great. A quadratic (‘inverted-u’) correlation between arm and shaft length and clubhead speed might be stronger than a linear correlation. Furthermore, there is probably an optimal range of movement in each body segment giving maximal clubhead speed, but golf-related research is currently lacking. Several studies of fitness training have reported increased striking distance or clubhead speed after a period of weight training and/or flexibility training,[88-94] and degree of accuracy does not seem to decrease as a result of weight training.[88,94] However, few studies have trained and tested skilled golfers and used a control group. Pinter[95] examined the effects of weight training, flexibility training and a combination of both on clubhead speed and accuracy in the drive of 25 golfers with a handicap of 4 or better. There were seven subjects in the control group who were allowed to play and practice in their normal fashion, but trained in neither strength nor flexibility. All three experimental groups participated in 50 minutes of workout three times a week for a period of 8 weeks. There were no significant differences between pre-test and post-test clubhead speed scores among the four groups. There was, however, a significant increase in driving accuracy in the groups engaged in weight or flexibility training. Unfortunately, the possible increase in strength was not tested. The subjects were instructed to swing ‘as in a competition’, which might have caused them not to swing at maximum clubhead speed, thus making it more difficult to find significant correlations between physical training and striking distance. The effect of strength training probably depends on the player’s skill level, training history and weight-lifting methods.[94] The change in an object’s momentum depends on the magnitude ª 2009 Adis Data Information BV. All rights reserved.

Hellstro¨m

of the force and the time that the force acts on the object. The player’s initial rate of force development (RFD) and maximum RFD are more important in increasing the object’s speed than the maximal strength factor, when the object is light (<25% of 1 repetition maximum [RM]), and the time to affect it is short (<250 ms).[96] The effect of weight training on the RFD depends on how the weight training is performed and the subject’s fitness status.[97-99] Less powerful players might increase their clubhead speed regardless of how a weight-training programme is designed. However, the downswing takes about 0.2–0.3 seconds for the average tour player,[38,70] and the club is light in relation to the mass that is possible to swing. Well trained top-class players with high clubhead speed may therefore develop a slower clubhead speed after a period of slow velocity strength training, due to the difference in neural activity and RFD compared with power training.[96] Future studies should therefore consider the individual player’s needs and investigate the effects of different power and maximal weight training on clubhead speed. In summary, a long range of motion and a high production of forces and torques are needed to create a large work on the golf club. Golfers therefore need to be flexible and have high maximal strength and power. A large spinal axial rotation, adduction of the left arm and protraction of the left shoulder girdle can lead to a long backswing. Strong legs and hip muscles can create high forces and torques toward the ground. The torso should be able to create high rotary power. A fast adduction of the right arm requires high power generated from the pectoralis major muscle, and the right triceps should have high power in order to extend fast through impact. Large grip strength is needed to control the club during high clubhead speed. In order to achieve a good kinematic sequence, players should probably have their proximal segments in positions so that the distal segments can create high torques without increasing the variability during the swing. Further research is needed to understand what effects physical training can have on the kinematic sequence during the downswing. Sports Med 2009; 39 (9)

Competitive Elite Golf

737

Table III. Investigations into elite players’ physical characteristics, physique and associations with clubhead speed Study

Subjects

Design

Findings

Crews et al.[83]

N = 23; females; professional golfers; age 22–49 years

Group characteristics

Mean height and mass were 168.9 cm and 61.8 kg, resting heart rate and blood pressure were 67.1 beats/min and 113/73 mmHg, respectively, the right and left grip strengths were . 36.5/36.0 kg, and VO2max was 34.2 mL/kg/min

Dorado et al.[85]

N = 33; 15 male professional golfers and 18 sedentary individuals, matched for sex, race, mean age (29 and 25 years, respectively), mean body mass (79 and 74 kg), mean height (1.78 and 1.77 m) and mean percentage body fat (20 and 21%)

Group comparisons

The professional golfers had significantly* larger (9%) muscle mass in the dominant arm. No significant differences in bone mass or density were found between the groups

Kawashima et al.[84]

N = 128; males; professional golfers (n = 11), collegiate golfers (n = 24), general amateur golfers (n = 13), collegiate recreational golfers (n = 15), non-golfing college student (n = 45) and a senior population of non-golfers as a control group (n = 20)

Group comparisons

The professional golfers had significantly* larger limb girth than the other groups. The better players tended to have more athletic somatotypes then the less skilled players

Sell et al.[82]

N = 257; males; three groups: hcp <0 (n = 45, mean age 39.2 years), hcp 1–9 (n = 120, mean age 43.7 years), and hcp 10–20 (n = 92, mean age 50.9 years)

Group comparisons

The best group of golfers had significantly* greater hip strength, torso strength, shoulder strength, shoulder flexibility, hip flexibility, torso flexibility and right leg balance with eyes open, than the players in the highest hcp group

Yoon[87]

N = 41; males; hcp <3

Associative

Clubhead speed correlated significantly* with trunk power (r = 0.80), leg and hip power (r = 0.60), combined arm and trunk power (r = 0.58), grip strength (r = 0.54), height (r = 0.51), and arm length (r = 0.45). Body mass (r = 0.22) and shoulder width (r = 0.20) were not significantly related to swing speed

Pinter[95]

N = 25; males; hcp £4, age 19–23 years. Four groups: strength only (n = 6), flexibility only (n = 6), strength + flexibility (n = 7) and control (n = 6)

Causality. Fitness training 50 minutes, 3 times a week, for 8 weeks

No significant* difference between pretest and post-test clubhead speed scores among the four groups who were instructed to swing as they would in competition. There was a significant increase in accuracy of the golf drive for the groups engaged in strength training only and in flexibility training only

Doan et al.[94]

N = 10; males, hcp » 0; mean age 19.8 years (another group with 6 women is excluded due to hcp >4)

Causality. Fitness training 3 times per week for 11 weeks

Significant* increases were noted for all strength, power and flexibility tests from pre- to post-training. Clubhead speed did not increase significantly. Putting distance control improved significantly

hcp = handicap; * p < 0.05.

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (9)

Hellstro¨m

738

4. Conclusions The literature shows that the best players differ from less skilled players in playing results, technique and physique. These variables are interrelated. Anthropometrics and flexibility can affect the setup and range of motion during the swing. Strength and power can affect the forces and torques that the player is able to exert on the shaft. Creating high total work on the shaft causes high clubhead speed. Higher clubhead speed can lead to higher initial ball speed, longer striking distance and lower scores. Lower scores are also associated with driving accuracy, GIR, accurate short game and number of putts per GIR. Experts around the players should assess and discuss these relations before making training recommendations. Less is known of what physical and technical capacities are needed to control the clubface angle, the direction of the clubhead and submaximal clubhead speed, in order to achieve higher control of ball displacements. Future studies should investigate how successful players’ physique and technique are related, and what physique and technique are needed to control clubhead velocity and clubface angle in a way that repeatedly creates desired impacts and ball flights. Such information should be studied together with more detailed game statistics, where the intended target, the exact distance the ball ends up from the target and the environmental factors are known, in order to obtain a deeper understanding of processes leading to lower scores. Acknowledgements The Swedish National Centre for Research in Sports funded this research. This is a national organization with the task of initiating, coordinating, supporting and informing about sport-related research. The author would also like to thank Leif Isberg, Johnny Nilsson and Robert Neal for their help in the preparation of this paper. The author has no conflicts of interest that are directly relevant to the content of this review.

References 1. Morrow Jr JR, Jackson AW, Disch JG, et al. Measurement and evaluation in human performance. 2nd ed. Champaign (IL): Human Kinetics, 2000

ª 2009 Adis Data Information BV. All rights reserved.

2. Davidson J, Templin T. Determinants of success among professional golfers. Res Q Exerc Sport 1986; 57 (1): 60-7 3. Hale T, Hale G. Lies, damned lies and statistics in golf. In: Cochran AJ, editor. Science and golf: proceedings of the first World Scientific Congress of Golf. London: E & FN Spon, 1990: 165-7 4. Jones RE. A correlation analysis of the Professional Golf Association (USA) statistical rankings for 1988. In: Cochran AJ, editor. Science and golf: proceedings of the first World Scientific Congress of Golf. London: E & FN Spon, 1990: 165-7 5. Dorsel TN, Rotunda RJ. Low scores, top 10 finishes, and big money: an analysis of professional golf association tour statistics and how these relate to overall performance. Percept Mot Skills 2001; 92 (2): 575-85 6. Engelhardt GM. It’s not how you drive, it’s how you arrive: the myth. Percept Mot Skills 1995; 80 (3 Pt 2): 1135-8 7. Engelhardt GM. Differences in shot-making skills among high and low money winners on the PGA tour [letter]. Percept Mot Skills 1997; 84 (3 Pt 2): 1314 8. Jimenez JA, Fierro-Hernandez C. Are European and American golf players different? Reply to Engelhardt (1997). Percept Mot Skills 1999; 89 (2): 417-8 9. Finley PS, Halsey JJ. Determinants of PGA tour success: an examination of relationships among performance, scoring, and earnings. Percept Mot Skills 2004; 98 (3 Pt 1): 1100-6 10. Belkin DS, Gansneder B, Pickens M, et al. Predictability and stability of Professional Golf Association tour statistics. Percept Mot Skills 1994; 78 (3 Pt 2): 1275-80 11. Nix CL, Koslow R. Physical skill factors contributing to success on the professional golf tour. Percept Mot Skills 1991; 73 (3 Pt 2): 1272-4 12. Wiseman F, Chatterjee S. Comprehensive analysis of golf performance on the PGA Tour: 1990-2004. Percept Mot Skills 2006; 102 (1): 109-17 13. Quinn RJ. Exploring correlation coefficients with golf statistics. Teach Stat 2006; 28 (1): 10-3 14. Larkey PD. Comparing players in professional golf. In: Farally MJ, Cochran AJ, editors. Science and golf II: proceedings of the World Scientific Congress of Golf. London: E & FN Spon, 1994: 193-8 15. Fried HO, Lambrinos J, Tyner J. Evaluating the performance of professional golfers on the PGA, LPGA and SPGA tours. Eur J Oper Res 2004; 154 (2): 548-61 16. Thomas FW. The state of the game, equipment and science. In: Farally MJ, Cochran AJ, editors. Science and golf II: proceedings of the World Scientific Congress of Golf. London: E & FN Spon, 1994: 237-46 17. Karlsen J, Smith G, Nilsson J. The stroke has only a minor influence on direction consistency in golf putting among elite players. J Sports Sci 2008; 26 (3): 243-50 18. Pelz D. The short game bible. New York: Broadway Books, 1999 19. Pelz D. Dave Pelz’s putting bible. New York: Random House, 2000 20. Larkey PD, Smith AA. All around improvements. In: Science and golf III: proceedings of the World Scientific Congress of Golf. Champaign (IL): Human Kinetics, 1999: 377-84

Sports Med 2009; 39 (9)

Competitive Elite Golf

21. Wiseman F, Habibullah M, Yilmaz M. A new method for ranking total driving performance on the PGA tour [online]. Sport J 2007; 10 (1). Available from URL: http:// www.thesportjournal.org/article/new-method-ranking-totaldriving-performance-pga [Accessed 2009 Jul 1] 22. Cochran AJ. Club face flexibility and coefficient of restitution. In: Cochran AJ, Farally MR, editors. Science and golf III: proceedings of the World Scientific Congress of Golf. London: E & FN Spon, 1999: 486-92 23. Johnson SH, Hubbel JE. Golf ball rebound enhancement. In: Cochran AJ, Farally MR, editors. Science and golf III: proceedings of the World Scientific Congress of Golf. London: E & FN Spon, 1999: 493-9 24. Werner FD, Greig RC. How golf clubs really work and how to optimize their design. Jackson (WY): Origin Inc., 2000 25. Wiren G. Laws, principles and preferences: a teaching model. In: Cochran AJ, editor. Science and golf: proceedings of the first World Scientific Congress of Golf. London: E & FN Spon, 1990: 3-13 26. PGA of America. PGA teaching manual: the art and science of golf instruction. Palm Beach Gardens (FL): PGA of America, 1990 27. Jorgensen T. The physics of golf. 2nd ed. New York: AIP Press, 1999 28. Miura K, Sato F. The initial trajectory plane after ball impact. In: Farally MR, Cochran AJ, editors. Science and golf III: proceedings of the World Scientific Congress of Golf. Champaign (IL): Human Kinetics, 1999: 535-43 29. Miura K. Mapping clubhead to ball impact and estimating trajectory. In: Thain E, editor. Science and golf IV: proceedings of the World Scientific Congress of Golf. London: Routledge, 2002: 490-500 30. Chou PH, Liang D, Yang J. Contact forces, coefficient of restitution, and spin rate of golf ball impact. Science and golf II: proceedings of the World Scientific Congress of Golf. London: E & FN Spon, 1994: 296-301 31. McLeod WM. Biomechanical variables as predictors of distance and accuracy in the golf drive [dissertation]. Halifax (NS): University of Dalhousie, 1976 32. Shaw M. Aerodynamics of golf balls. In: Cochran A, editor. Golf: the scientific way. Hemel Hempstead: Aston Publishing Group, 1995: 152-5 33. Scheie CE. The golf club-ball collision: 50 000 g’s. In: Cochran AJ, editor. Science and golf: proceedings of the first World Scientific Congress of Golf. London: E & FN Spon, 1990: 237-40 34. Williams KR, Sih BL. Changes in golf clubface orientation following impact with the ball. Sports Eng 2002; 5 (2): 65-80 35. Kelley H. The golfing machine. Seattle (WA): Star System Press, 1982 36. Dalton J. Compression by any other name. Science and golf IV: proceedings of the World Scientific Congress of Golf. London: Routledge, 2002: 319-27 37. Chou A, Gilbert P. Clubhead designs: how they affect ball flight. In: Cochran A, editor. Golf: the scientific way. Hemel Hempstead: Aston Publishing Group, 1995 38. Cochran A, Stobbs J. Search for the perfect swing. Grass Valley (CA): The Booklegger, 1968

ª 2009 Adis Data Information BV. All rights reserved.

739

39. Hunter C. The sequence of hip and selected upper-extremity joint movements during the golf drive [dissertation]. Springfield (MA): Springfield College, 1971 40. Marshall RN, Elliott BC. Long-axis rotation: the missing link in proximal-to-distal segmental sequencing. J Sports Sci 2000; 18 (4): 247-54 41. Soriano BC. A biomechanical investigation of human movement during the golf swing [dissertation]. Irvine (CA): University of California, 1996 42. Fradkin AJ, Sherman CA, Finch CF. How well does club head speed correlate with golf handicaps? J Sci Med Sport 2004; 7 (4): 465-72 43. Watanabe K, Kuroki S, Hokari M, et al. Golf swing and skill. In: Farally MJ, Cochran AJ, editors. Science and golf III: proceedings of the World Scientific Congress of Golf. London: E & FN Spon, 1994: 29-39 44. Barrentine SW, Fleisig GS, Johnson H. Ground reaction forces and torques of professional and amateur golfers. In: Farally MJ, Cochran AJ, editors. Science and golf II: proceedings of the World Scientific Congress of Golf. London: E & FN Spon, 1994: 33-9 45. Nesbit SM. A three dimensional kinematic and kinetic study of the golf swing. J Sports Sci Med 2005; 4 (4): 499-519 46. Fujimoto-Kanatani K. Determining the essential elements of golf swings used by elite golfers [dissertation]. Corvallis (OR): Oregon State University, 1995 47. Zheng N, Barrentine SW, Fleisig GS, et al. Kinematic analysis of swing in pro and amateur golfers. Int J Sports Med 2008; 29 (6): 487-93 48. Milburn PD. Summation of segmental velocities in the golf swing. Med Sci Sports Exerc 1982; 14 (1): 60-4 49. Miura K. Accelerating and decelerating phases of the wrist motion of the golf swing. The engineering of sport: design and development. 2nd International Conference on the Engineering of Sport 1998. Oxford: Blackwell Publishing, 455-63 50. Butler JH, Winfield DC. The dynamic performance of the golf shaft during the downswing. In: Farally MJ, Cochran AJ, editors. Science and golf II: proceedings of the World Scientific Congress of Golf. London: E & FN Spon, 1994: 259-64 51. Horwood GP. Golf shaft: a technical perspective. In: Farally MJ, Cochran AJ, editors. Science and golf II: proceedings of the World Scientific Congress of Golf. London: E & FN Spon, 1994: 247-58 52. Milne RD, Davis JP. The role of the shaft in the golf swing. J Biomech 1992; 25 (9): 975-83 53. Brylawski AM. An investigation of three dimensional deformation of a golf club during downswing. In: Farally MJ, Cochran AJ, editors. Science and golf II: proceedings of the World Scientific Congress of Golf. London: E & FN Spon, 1994: 265-70 54. Miao T, Watari I, Kawaguchi M, et al. A study of clubhead speed as a function of grip speed for a variety of shaft flexibility. In: Cochran AJ, Farally MR, editors. Science and golf III: proceedings of the World Scientific Congress of Golf. Champaign (IL): Human Kinetics, 1999: 554-61 55. Nesbit SM, Serrano M. Work and power analysis of the golf swing. J Sports Sci Med 2005; 4 (4): 520-33

Sports Med 2009; 39 (9)

Hellstro¨m

740

56. Hay J. Golf. In: Hay J, editor. Biomechanics of sports techniques. 4th ed. Englewood Cliffs (NJ): Prentice-Hall, 1993: 276-95 57. Sprigings E, Neal RJ. An insight into the importance of wrist torque in driving the golf ball: a simulation study. J Appl Biomech 2000; 16 (4): 356-67 58. Coleman SG, Rankin AJ. A three-dimensional examination of the planar nature of the golf swing. J Sports Sci 2005; 23 (3): 227-34 59. Tang W-T, Abraham L. Analysis of golf swing optimization models. 19th Congress of the International Society of Biomechanics: the human body in motion. Congress handbook and book of abstracts. Dunedin: University of Otago, 2003: 383 60. Pink M, Jobe FW, Perry J. Electromyographic analysis of the shoulder during the golf swing. Am J Sports Med 1990; 18 (2): 137-40 61. Okuda I, Armstrong CW, Tsunezumi H, et al. Biomechanical analysis of professional golfer’s swing: Hidemichi Tanaka. In: Thain E, editor. Science and golf IV: proceedings of the World Scientific Congress of Golf. London: Routledge, 2002: 19-27 62. Miura K. Parametric acceleration: the effect of inward pull of the golf club at impact stage. Sports Eng 2001; 4 (2): 75-86 63. Kawashima K, Meshizuka T, Takeshita S. A kinematic analysis of foot force exerted on the soles during the golf swing among skilled and unskilled golfers. In: Cochran AJ, Farally MR, editors. Science and golf III: proceedings of the World Scientific Congress of Golf. Champaign (IL): Human Kinetics, 1999: 40-5 64. Wallace ES, Graham D, Bleakley EW. Foot-to-ground pressure patterns during the golf drive: a case study involving a low handicap player and a high handicap player. In: Cochran AJ, editor. Science and golf: proceedings of the first World Scientific Congress of Golf. London: E & FN Spon, 1990: 25-9 65. Koenig G, Tamres M, Mann RW. The biomechanics of the shoe-ground interaction in golf. In: Farally MJ, Cochran AJ, editors. Science and golf II: proceedings of the World Scientific Congress of Golf. London: E & FN Spon 1994: 25-9 66. Ball KA, Best RJ. Different centre of pressure patterns within the golf stroke I: cluster analysis. J Sports Sci 2007; 25 (7): 757-70 67. Ball KA, Best RJ. Different centre of pressure patterns within the golf stroke II: group-based analysis. J Sports Sci 2007; 25 (7): 771-9 68. Grimshaw PN, Burden AM. Case report: reduction of low back pain in a professional golfer. Med Sci Sports Exerc 2000; 32 (10): 1667-73 69. Gulgin HR. Hip rotation range of motion asymmetry in elite female golfers [dissertation]. Toledo (OH): University of Toledo, 2005 70. McTeigue M, Lamb SR, Mottram R, et al. Spine and hip motion analysis during the golf swing. In: Farally MJ, Cochran AJ, editors. Science and golf II: proceedings of the World Scientific Congress of Golf. London: E & FN Spon, 1994: 50-8 71. Cheetham P, Martin P, Mottram R, et al. The importance of stretching the ‘‘X-factor’’ in the downswing of golf: ‘‘The X-factor stretch’’. In: Thomas PR, editor. Optimising per-

ª 2009 Adis Data Information BV. All rights reserved.

72.

73.

74. 75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86. 87.

88.

formance in golf. Brisbane (QLD): Australian Academic Press Ltd, 2001: 192-9 Hume PA, Keogh J, Reid D. The role of biomechanics in maximising distance and accuracy of golf shots. Sports Med 2005; 35 (5): 429-49 Myers J, Lephart S, Tsai YS, et al. The role of upper torso and pelvis rotation in driving performance during the golf swing. J Sports Sci 2008; 26 (2): 181-8 Carlso¨o¨ S. A kinetic analysis of the golf swing. J Sports Med Phys Fitness 1967; 7 (2): 76-82 Bechler JR, Jobe FW, Pink M, et al. Electromyographic analysis of the hip and knee during the golf swing. Clin J Sport Med 1995; 5 (3): 162-6 Newton RU, Murphy AJ, Humphries BJ, et al. Influence of load and stretch shortening cycle on the kinematics, kinetics and muscle activation that occurs during explosive upper-body movements. Eur J Appl Physiol Occup Physiol 1997; 75 (4): 333-42 Hosea TM, Gatt CJ, Galli KM, et al. Biomechanical analysis of the golfer’s back. In: Cochran AJ, editor. Science and golf: proceedings of the first World Scientific Congress of Golf. London: E & FN Spon, 1990: 43-8 Abernethy B, Neal RJ, Moran MJ. Expert-novice differences in muscle activity during the golf swing. In: Cochran AJ, editor. Science and golf: proceedings of the first World Scientific Congress of Golf. London: E & FN Spon, 1990: 55-60 Barclay JK, McIlroy WE. Effect of skill level on muscle activity in neck and forearm muscles during the golf swing. In: Cochran AJ, editor. Science and golf: proceedings of the first World Scientific Congress of Golf. London: E & FN Spon, 1990: 49-53 Kreighbaum E, Barthels KM. Biomechanics: a qualitative approach for studying human movement. 4th ed. Old Tappan (NJ): Allyn & Bacon, 1996 Putnam CA. Sequential motions of body segments in striking and throwing skills: descriptions and explanations. J Biomech 1993; 26 (1): 125-35 Sell TC, Tsai YS, Smoliga JM, et al. Strength, flexibility, and balance characteristics of highly proficient golfers. J Strength Cond Res 2007; 21 (4): 1166-71 Crews DJ, Shirreffs JH, Thomas G, et al. Psychological and physiological attributes associated with performance of selected players of the Ladies Professional Golf Association Tour. Percept Mot Skills 1986; 63 (1): 235-8 Kawashima K, Kat K, Miyazaki M. Body size and somatotype characteristics of male golfers in Japan. J Sports Med Phys Fitness 2003; 43 (3): 334-41 Dorado C, Sanchis Moysi J, Vicente G, et al. Bone mass, bone mineral density and muscle mass in professional golfers. In: Thain E, editor. Science and golf IV: proceedings of the World Scientific Congress of Golf. New York: Routledge, 2002: 54-63 Wiren G. Human factors influencing the golf drive for distance [dissertation]. Corvallis (OR): University of Oregon, 1968 Yoon S. The relationship between muscle power and swing speed in low-handicapped golfers [MS thesis]. Provo (UT): Brigham Young University, 1998 Lanford EE. Effect of strength training on distance and accuracy in golf [dissertation]. Corvallis (OR): University of Oregon, 1976

Sports Med 2009; 39 (9)

Competitive Elite Golf

89. Fletcher IM, Hartwell M. Effect of an 8-week combined weights and plyometrics training program on golf drive performance. J Strength Cond Res 2004; 18 (1): 59-62 90. Hetu FE, Christie CA, Faigenbaum AD. Effects of conditioning on physical fitness and club head speed in mature golfers. Percept Mot Skills 1998; 86 (3 Pt 1): 811-5 91. Thompson CJ, Osness WH. Effects of an 8-week multimodal exercise program on strength, flexibility, and golf performance in 55- to 79-year-old men. J Aging Phys Act 2004; 12 (2): 144-56 92. Jones D. The effects of proprioceptive neuromuscular facilitation flexibility training on the clubhead speed of recreational golfers. In: Farally MR, Cochran AJ, editors. Science and golf III: proceedings of the World Scientific Congress of Golf. Champaign (IL): Human Kinetics, 1999: 46-50 93. Westcott WL, Dolan F, Cavicchi T. Golf and strength training are compatible activities. Strength Cond 1996; 18 (4): 54-6 94. Doan BK, Newton RU, Kwon YH, et al. Effects of physical conditioning on intercollegiate golfer performance. J Strength Cond Res 2006; 20 (1): 62-72

ª 2009 Adis Data Information BV. All rights reserved.

741

95. Pinter MD. Effect of strength training and flexibility on club head speed and accuracy in the golf drive [dissertation]. Starkville (MI): Mississippi State University, 1992 96. Schmidtbleicher D, Buehrle M. Neuronal adaptation and increase of cross-sectional area studying different strength training methods. In: Johnson B, editor. Biomechanics X-B; vol. 6-B; 615-20 97. Hakkinen K, Alen M, Komi PV. Neuromuscular, anaerobic, and aerobic performance characteristics of elite power athletes. Eur J Appl Physiol Occup Physiol 1984; 53 (2): 97-105 98. Wilson GJ, Newton RU, Murphy AJ, et al. The optimal training load for the development of dynamic athletic performance. Med Sci Sports Exerc 1993; 25 (11): 1279-86 99. Cronin J, McNair PJ, Marshall RN. Developing explosive power: a comparison of technique and training. J Sci Med Sport 2001; 4 (1): 59-70

Correspondence: John Hellstro¨m, Bjo¨rnva¨gen 8a, 18133 Lidingo¨, Sweden. Email: [email protected]

Sports Med 2009; 39 (9)

REVIEW ARTICLE

Sports Med 2009; 39 (9): 743-764 0112-1642/09/0009-0743/$49.95/0

ª 2009 Adis Data Information BV. All rights reserved.

Do ‘Mind over Muscle’ Strategies Work? Examining the Effects of Attentional Association and Dissociation on Exertional, Affective and Physiological Responses to Exercise Erik Lind,1 Amy S. Welch2 and Panteleimon Ekkekakis2 1 Department of Physical Education, State University of New York College at Oneonta, New York, New York, USA 2 Department of Kinesiology, Iowa State University, Ames, Iowa, USA

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Purpose of the Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Literature Search . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Defining and Conceptualizing Attentional Focus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Defining ‘Association’ and ‘Dissociation’ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Conceptual Frameworks for the Study of Attentional Focus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Psychometric Assessment of Attentional Focus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Theoretical Mechanisms of Regulating Focal Awareness. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 ‘Competition of Cues’ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Parallel Processing of Information Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Social Psychophysiological Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4 Social-Cognitive Perspective of Perceived Exertion and Exertion Tolerance . . . . . . . . . . . . . . . . . 4.5 Dual Mode Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Review of Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 Descriptive Reports. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Perceived Exertion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 Affective and other Psychological Responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.1 Affective, Emotional and Mood-Related Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.2 Other Psychological Responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4 Preferred Attentional Focus Style . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5 Personality Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6 Exercise Economy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.1 Heart Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.2 Oxygen Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.3 Ventilatory Responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.4 Respiratory Exchange Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.5 Hormonal Responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.7 Exercise Tolerance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1 Future Directions: What are the Possible Sources of Inconsistencies? . . . . . . . . . . . . . . . . . . . . . . . 6.1.1 Participant Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1.2 Exercise Stimulus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1.3 Experimental Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 A Comment on A/D Guidelines and Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

744 745 745 746 746 746 747 747 748 748 748 749 749 749 749 750 751 751 753 753 754 754 754 755 755 755 756 756 757 757 757 758 759 759

Lind et al.

744

Abstract

Despite the well established physical and psychological benefits derived from leading a physically active life, rates of sedentary behaviour remain high. Dropout and non-compliance are major contributors to the problem of physical inactivity. Perceptions of exertion, affective responses (e.g. displeasure or discomfort), and physiological stress could make the exercise experience aversive, particularly for beginners. Shifting one’s attentional focus towards environmental stimuli (dissociation) instead of one’s body (association) has been theorized to enhance psychological responses and attenuate physiological stress. Research evidence on the effectiveness of attentional focus strategies, however, has been perplexing, covering the entire gamut of possible outcomes (association and dissociation having been shown to be both effective and ineffective). This article examines the effects of manipulations of attentional focus on exertional and affective responses, as well as on exercise economy and tolerance. The possible roles of the characteristics of the exercise stimulus (intensity, duration) and the exercise participants, methodological issues, and limitations of experimental designs are discussed. In particular, the critical role of exercise intensity is emphasized. Dissociative strategies may be more effective in reducing perceptions of exertion and enhancing affective responses at low to moderate exercise intensities, but their effectiveness may be diminished at higher and nearmaximal levels, at which physiological cues dominate. Conversely, associative strategies could enable the exerciser to regulate intensity to avoid injury or overexertion. Thus, depending on intensity, both strategies have a place in the ‘toolbox’ of the public health or exercise practitioner as methods of enhancing the exercise experience and promoting long-term compliance.

Rates of physical inactivity for adults in the US remain high despite the well publicized health benefits derived from leading a physically active life.[1] Other industrialized nations face similar public health challenges.[2-4] Physical inactivity in conjunction with a poor diet accounts for a substantial number of preventable deaths.[5] Various local, national and international health agencies (e.g. WHO) are beginning to closely examine the reasons behind sedentary behaviour and to develop interventions that address physical inactivity. However, approximately 50% of individuals initiating an exercise programme drop out within the first 6 months and psychological interventions aimed at preventing noncompliance have met with modest success.[6] Thus, dropout is a major contributor to the physical inactivity epidemic. It is reasonable to assume that, if dropout rates could be drastically reduced, overall public participation rates could be substantially increased. ª 2009 Adis Data Information BV. All rights reserved.

One possible explanation for the high dropout rate is a causal chain linking exercise intensity, exertional and affective responses (e.g. pleasure vs displeasure, enjoyment vs aversion), and exercise adherence.[7,8] Studies show that, as exercise intensity increases, affective responses become less positive or more negative.[9-12] Other studies show that higher exercise intensity levels are associated with reduced adherence and increased dropout, and this effect cannot be fully accounted for by injuries.[13-17] Providing a possible explanation for these findings, recent evidence demonstrates that affective responses are significant predictors of subsequent physical activity participation.[18] In short, if exercise intensity is too high, and exercise is not perceived as enjoyable, comfortable or tolerable, it is reasonable to assume that people will be less likely to repeat the activity in the future.[19] The problem is that most adults who initiate an exercise programme do so after long periods of Sports Med 2009; 39 (9)

Associative and Dissociative Strategies

sedentary living and, consequently, face the challenges of a low level of cardiorespiratory fitness and potentially high bodyweight, resulting in strenuous levels of intensity. Thus, during the critical early stages of participation, their experiences may be characterized by high levels of perceived exertion and non-positive affective responses. A commonly employed solution has been to use a cognitive strategy designed to alter how the exercise stimulus is experienced (i.e. reduce the perceptions of physical discomfort or improve the affective response to exercise). The armamentarium of such cognitive strategies includes manipulations of attentional focus (i.e. association or dissociation), use of audiovisual stimuli (e.g. music or television), self-talk and bolstering the participant’s sense of self-efficacy. One indicator of how popular attentional focus strategies, in particular dissociation, are in practice can be found in books developed with the fitness professional in mind. Dissociative techniques are routinely recommended for diverting attention away from uncomfortable or displeasurable bodily sensations during exercise. For example, Rejeski and Kenney[20] endorsed dissociation as a means of ‘countering the discomforts of exercise’ (p. 85). The authors suggest focusing on distracting stimuli that are enjoyable, engaging, positive and safe. Similarly, Leith[21] highlighted the risk of associating during exercise by stating, ‘‘Focusing on the physical activity serves to remind us of feelings of fatigue and makes the effort more of a chore’’ (p. 88). 1. Purpose of the Review The aim of this review is 4-fold: (i) to examine the effectiveness of manipulating an exercise participant’s attentional focus by using association or dissociation in controlling perceptions of exertion, enhancing affective responses, and attenuating physiological strain; (ii) consider what role various study characteristics may have played in some equivocal results that have been reported; (iii) highlight the critical importance of the element of exercise intensity in modulating the effectiveness of association and dissociation; and (iv) provide recommendations for future research. ª 2009 Adis Data Information BV. All rights reserved.

745

Although the subject matter for this review is limited to studies of attentional associative and dissociative strategies, it should be noted that there are a number of related techniques that fall under the umbrella of ‘cognitive strategies’. These will be dealt with in subsequent instalments in this series of reviews. Specifically, the use of music and other audiovisual stimuli will be examined in a second review and other, less extensively studied, cognitive strategies (i.e. manipulation of selfefficacy, self-talk, hypnotic suggestion and deception) will be covered in a third. 2. Literature Search To locate studies on the use of association/ dissociation (A/D) during exercise, computer searches were conducted in scientific databases (PsycLit, PubMed, Google Scholar) using the terms (and combinations thereof): ‘exercise’, ‘attentional focus’, ‘association’, ‘dissociation’, ‘focal awareness’, ‘internal focus’, ‘external focus’. Furthermore, the reference lists of the obtained articles were searched for additional pertinent studies. A total of 88 studies related specifically to endurance or aerobic exercise and A/D were retrieved through these methods. Some of the published papers included more than one study relevant to this review. Articles examining the relationship between A/D strategies during resistance and/or strength training were excluded from this review because of the substantially different physiological demand characteristics of such activities. In certain cases, only unpublished manuscripts or abstracts from papers presented at scientific conferences were available. These are not included in the summary table (table I; available online as supplementary material [Supplemental Digital Content 1, http://links.adisonline.com/SMZ/ A2]), but are discussed in the relevant sections of the review, as needed. The studies consist of both descriptive reports and experimental investigations on the influence of A/D on perceived exertion, affective responses, and exercise performance variables including exercise economy and exercise tolerance. Exercise economy is considered here as a physiological index of the efficiency of movement, expressed as physiological Sports Med 2009; 39 (9)

Lind et al.

746

responses required to perform a given exercise workload (e.g. heart rate, oxygen uptake and ventilatory responses). Exercise tolerance is considered here as an index of endurance capacity expressed as maximal exercise duration and/or maximal attained workload during an exercise test (e.g. time to exhaustion, peak power output). Table I (supplementary material) provides detailed information on the characteristics of the sample, the exercise stimuli, study conditions, relevant dependent variables, and highlighted results. The table is divided into subsections, with the studies grouped according to exercise intensity level. 3. Defining and Conceptualizing Attentional Focus 3.1 Defining ‘Association’ and ‘Dissociation’

The way association and dissociation have been defined over the years has varied. Originally, the attentional focus strategy of association was characterized as the focus on bodily sensations necessary for optimal performance[22] and, more specifically, on physical sensations emanating from changes in respiration, temperature and muscular fatigue.[23] At the other end of the attentional focus spectrum, dissociation was characterized as a cognitive process of actively ‘blocking out’ sensations of pain or discomfort related to physical effort. As described by Morgan,[23] an individual who dissociates ‘‘purposefully cuts himself off from the sensory feedback he normally receives from his body’’ (p. 39). 3.2 Conceptual Frameworks for the Study of Attentional Focus

Schomer[24] characterized the attentional strategies along the single dimension of taskrelatedness in his work with marathoners. Integrating Nideffer’s[25] attentional styles concept, association and dissociation were further defined along the dimensions of attentional width (i.e. broad or narrow attentional focus) and attentional direction (i.e. attending to internal or external cues). Schomer[24] organized tape-recorded verbalizations of marathon runners into subª 2009 Adis Data Information BV. All rights reserved.

categories that could be more broadly characterized as associative or task-related on the one hand and dissociative or task-unrelated on the other. Specifically, verbalizations that pertained to feelings and affect, body monitoring, and command and instruction were characterized as reflecting an internal/narrow attentional focus. Verbalizations that pertained to pace monitoring were characterized as reflecting an external/ narrow attentional focus. Verbalizations that pertained to reflective thoughts, personal problem solving, and career-related thoughts comprised an internal/broad attentional focus category. Finally, verbalizations that pertained to running course information and talk and chatter comprised an external/broad attentional focus category. More recently, a two-dimensional classification system of attentional focus was proposed by Stevinson and Biddle.[26,27] According to this system, attentional focus can be characterized along the dimension of relevancy (task-relevant vs task-irrelevant), which relates to factors associated with optimally performing a task, and the dimension of direction (internal vs external), which relates to the subject of focal awareness. In this two-dimensional approach, an individual’s attentional focus strategy can be located within one of four quadrants: (a) internal/task-relevant (e.g. fatigue, breathing, perspiration); (b) internal/ task-irrelevant (e.g. daydreams, imagining music, solving math problems); (c) external/task-relevant (e.g. strategy, split times, conditions); and (d) external/task-irrelevant (e.g. scenery, environment, other competitors). These mutually exclusive categories were employed in a series of studies aimed at understanding the phenomenon of ‘hitting the wall’, in a sample of marathon runners.[26,27] Employing an internal/task-irrelevant strategy was found to be related to ‘hitting the wall’ earlier.[26,27] Similarly, an earlier study had suggested that an internal/task-irrelevant strategy amplified the related sensations of pain and exhaustion.[28] However, other investigators have questioned this relationship. For example, although 73% of participants running in the Melbourne Marathon reported ‘hitting the wall’ after mile 19 (30th km), the use of association or dissociation was unrelated to the phenomenon.[29] Likewise, Sports Med 2009; 39 (9)

Associative and Dissociative Strategies

Buman et al.[30] observed that 43% of a sample of marathon runners reported ‘hitting the wall’, and this phenomenon was related to factors that were both arguably associative (e.g. generalized fatigue, 66%; intentionally slowing pace, 46%; muscle cramping/pain, 50%) and dissociative (e.g. loss of concentration, 17%; deliberate direction of attention away from race, 16%). 3.3 Psychometric Assessment of Attentional Focus

Several attempts have been made to develop general and activity-specific standardized selfreport measures of A/D in addition to the numerous survey instruments used in early studies. Additionally, certain questionnaires with a broader scope (e.g. commitment, discomfort) have also included scales designed to tap A/D-related constructs. This diversity highlights a major challenge to developing a cohesive conceptual framework in this area. It is also important to note that the lack of consistency in operational definitions has been one of the major obstacles in consolidating the research on A/D strategies. Another major challenge has been the absence of extensively supported and elaborated conceptual models or a unifying theoretical framework. Early self-report measures included items pertaining to perceived physical symptoms and affective responses,[31-33] estimated effort and performance time,[34] strategy prevalence,[35,36] specific thought content,[37] and circumstances surrounding strategy usage.[38] Goode and Roth[39] developed the Thoughts During Running Scale (TDRS) to assess participants’ thought content during a run. The TDRS was found to correlate with negative and positive mood states. Specifically, the negative mood state Fatigue was positively correlated with the TDRS scale Association (r = 0.25) and negatively correlated with the scale Dissociation (specifically, with the subscales Interpersonal Relationships [r = -0.21] and Daily Events [r = -0.19]). The positive mood state Vigor was positively correlated with the scale Dissociation (specifically, with the subscales Interpersonal Relationships [r = 0.25], Daily Events [r = 0.23] and External Surrounding [r = 0.23]). The TDRS has ª 2009 Adis Data Information BV. All rights reserved.

747

been used to assess attentional thought content in a number of investigations.[40-42] Brewer et al.[43] developed the Attentional Focus Questionnaire (AFQ), which assesses the broad dimensions of Association and Dissociation but differs from the TDRS in that the AFQ takes into account the influence of exercise intensity. Various items assess the role of specific physical sensations, feelings of fatigue, monitoring technique or pace as well as neutral and valenced (positive or negative) psychological content. The AFQ has also been used extensively in the A/D literature.[44,45] For example, Masters et al.[46] observed significant correlations between the Dissociation scale of the AFQ and the following scales of the Motivations of Marathoners Scale (MOMS): Psychological Coping (r = 0.54), Self-Esteem (r = 0.31) and Life Meaning (r = 0.36). Finally, some instruments have been developed to gauge the overall exercise experience, including attentional focus factors. For example, the Running Discomfort Scale (RDS)[47] includes the scales Disorientation and Task Completion Thoughts, which reflect the thoughts and feelings experienced by runners. The authors reasoned that distance runners tend to focus either on thoughts related to completing the run or on sensations of discomfort emanating from muscular and respiratory strain. Furthermore, they argued that ‘‘under such conditions of perceived discomfort, the mechanisms for the regulation of pain are more likely to stem from psychological than physical bases.’’[47] Similarly, Carmack and Martens[48] developed the Commitment to Running Scale, which included the dissociationlike factor ‘Spin-out’, characterized as ‘a detached or dreamy state of mind’ (p. 35). 4. Theoretical Mechanisms of Regulating Focal Awareness Early interest in A/D strategies can be attributed to the seminal work of Morgan and colleagues[23,49-51] on the physiological and psychological characteristics of long-distance runners. Since then, the role of A/D strategies has been examined in several literature reviews.[52-57] However, although a considerable amount of evidence Sports Med 2009; 39 (9)

Lind et al.

748

has accumulated, studies were rarely designed to examine specific theoretical propositions, and a guiding theoretical framework has yet to emerge. Nevertheless, the literature does include several potentially relevant conceptual models, which are summarized below. 4.1 ‘Competition of Cues’

Pennebaker and Lightner[31] proposed a ‘competition of cues’ mechanism as an explanation for the effects of A/D strategies during exercise. In their study, participants performed 10 minutes of self-paced treadmill exercise while (a) listening to their own breathing (to induce an internal focus), (b) listening to ambient street sounds (to induce an external focus), or (c) wearing headphones without sound (control). The authors noted that after the exercise bout, participants reported perceiving more negative physical symptoms, fatigue and tension during the internal focus condition compared with the other two conditions. In contrast, greater perceptions of enjoyment and satisfaction as well as improved exercise performance were reported under the external focus condition. Cardiovascular measures, however, showed no difference across experimental conditions. The authors concluded that ‘‘To the extent that one source of potential stimuli (e.g. the external environment) is tapped extensively, other sources (e.g. internal sensations) go unused’’ (p. 172). Thus, the concept of ‘competition of cues’ implies that the subjective exercise experience will depend on whether individuals maintain either an internal or an external attentional focus. 4.2 Parallel Processing of Information Model

A model advanced by Leventhal and coworkers,[58,59] which focused on the experience of pain, posited that there are separate but parallel pathways in which informational (i.e. noxious attributes of the stimulus) and emotional (i.e. generation of distress) qualities of pain are processed. The process was thought to consist of both preconscious and conscious phases. During the preconscious phase, a vast amount of sensory data relating to both the stimulus attributes (including ª 2009 Adis Data Information BV. All rights reserved.

intensity, duration and location) and the emotional qualities are gathered by sensory receptors. Whether or not this information is attended to, and thus comes into focal awareness, depends on the status of theorized attentional channels. If the attentional channels carrying pain-related information are selected, as in the case of association, then this information reaches conscious awareness, resulting in observable behavioural responses. Conversely, if attentional channels carrying nonpain-related information are selected, as in the case of dissociation, then pain information is likely to remain outside focal awareness. 4.3 Social Psychophysiological Model

Rejeski[60,61] applied the model of Leventhal and Everhart[59] to the sensations of fatigue and perceived exertion associated with exercise. Rejeski[61] noted that the original model failed to account for ‘‘how environmental and task variables contribute to the perceptual salience of specific physiological variables during exercise’’ (p. 376). He further argued that perception is not a passive process but rather an active process that is amenable to cognitive manipulations. According to Rejeski,[60] perceived exertion is determined by both psychological (e.g. cognitive strategies, individual differences, motivation) and physiological (e.g. lactate, hydrogen ions) factors. During exercise of low- and moderatedemand characteristics (intensity and duration), perceived exertion was theorized to reflect a greater contribution from psychological factors. As the intensity and duration progress, however, the contribution of physiological factors increases (as a still-undetermined mathematical function). The more perceived exertion reflects the influence of physiological cues, the smaller the contribution of psychological factors and, therefore, the smaller the potential impact of cognitive strategies. By acknowledging the contribution of psychological factors, the model also incorporates the notion of individual differences in exercise tolerance. The level of exercise intensity at which there is a critical shift in the balance between psychological and physiological determinants of perceived exertion was not identified. Nevertheless, Sports Med 2009; 39 (9)

Associative and Dissociative Strategies

individuals were thought to exhibit different levels of tolerance around this hypothetical threshold. That is, for some individuals the complete domination of perceived exertion by physiological cues might be at a lower, and for others at a higher, percentage of their actual maximal exercise capacity. The purpose of a psychological intervention, therefore, could be to extend the range of exercise intensity during which psychological factors remain influential. When applied to A/D strategies, Rejeski’s[60] social psychophysiological model implies that such interventions are more likely to be effective when the demand characteristics of the exercise stimulus are submaximal rather than maximal. 4.4 Social-Cognitive Perspective of Perceived Exertion and Exertion Tolerance

Tenenbaum[62] proposed a social-cognitive model of perceived exertion that makes very similar predictions to that proposed by Rejeski.[60,61] This model specifically addresses the relationship between exercise intensity and attentional focus strategies. In particular, at low levels of effort sense, the exerciser maintains the capacity to freely shift between an associative and a dissociative focus. Thus, during exercise of low intensity, the exerciser may shift between attending to and diverting away from discrete physiological sensations such as sweating and heavy breathing. As the exercise intensity and effort sense increase, however, focal awareness becomes predominantly internal and cognitive strategies become gradually less effective at influencing perceived exertion. During high exercise intensities, perceived exertion reflects symptoms of fatigue and exhaustion, and the ability to voluntarily shift attention away from them is diminished. 4.5 Dual Mode Model

The previously described models represent an evolution of conceptual ideas surrounding the factors that shape the exercise experience. The notions of a continuous and dynamic competition between internal and external cues, the active channelling of sensory input into and away from focal awareness, and the intensity-dependent shift ª 2009 Adis Data Information BV. All rights reserved.

749

in the contributions of psychological factors (including differences in attentional focus) and physiological cues constitute important theoretical advances. These ideas have served as the foundation of the Dual Mode Model (DMM).[63] This model further postulates that the ventilatory threshold (VT), because it is associated with an accentuation of several physiological parameters (blood lactate accumulation, frequency and depth of ventilation, sympathetic shift in autonomic regulation), might represent the level of exercise intensity demarcating the transition from a cognition-dominant to an interoception-dominant mode of eliciting both affective[63] and exertional[64] responses. According to the DMM, affective and exertional responses during exercise performed at intensities below and proximally to the VT involve primarily cortical pathways. On the other hand, at intensities that exceed the VT and preclude the maintenance of a physiological steady state, interoceptive afferent cues reach areas of the brain responsible for the elicitation of affective and exertional responses following direct, faster routes, bypassing the cortex.[65] 5. Review of Findings Studies on the effects of association and dissociation in the context of exercise reflect a great diversity of conceptual and methodological approaches. To facilitate the synopsis of this body of evidence, this section is organized into summaries of descriptive reports and experimental studies. This latter section is further divided into studies focusing on: (i) perceptual responses, including perceived exertion and self-reported physical symptoms; (ii) psychological responses, including affect, cognitive performance and exercise adherence; and (iii) physiological variables, including exercise economy and tolerance. 5.1 Descriptive Reports

Descriptive investigations include examinations of the prevalence of and circumstances surrounding the use of attentional focus strategies. However, no consistent trends have emerged. Factors that have been examined as possibly Sports Med 2009; 39 (9)

Lind et al.

750

associated with the use of attentional focus strategies have included ability level, nature of the task, age and sex, temporal patterning (i.e. when each strategy is used) and risk of injury. Specifically, some studies have reported that an associative focus is more frequently employed by elite than non-elite[23,66] and more experienced than inexperienced competitors,[67] among whom it is predictive of success.[68] However, other studies contradict these findings, indicating that ability level does not influence use of attentional focus strategies.[24,69-71] More rigorous control for exercise intensity might be one way to clarify this relationship.[70] Likewise, in some studies, the nature of the task was shown to be related to the selection of the attentional focus strategy, with association being used primarily in competition and dissociation being used primarily during training.[23,40,44,49,51,72] Alternatively, other studies have reported either greater reliance on dissociative focus[35,73] or a mixed attentional focus strategy[29,36] during competitive endurance activities. Once again, however, unmasking the relationship between the nature of the task and A/D usage would require careful control for exercise intensity, something that is not usually feasible in field studies. The relationship between age and use of attentional focus strategies seems consistent with association being reported during a larger percentage of time by younger participants than older participants during endurance events.[73-76] Conversely, the effects of sex are less clear. Although some evidence exists for gender-specific use of associative[44] and dissociative[68,70] focus, other studies have demonstrated no such difference.[35,69] Some researchers have observed that female exercisers gauge progress using a different set of performance information indicators, including associative factors,[77] and others have shown sex differences in the specific thought content of attentional focus. For example, solving personal problems was found to be more common among female runners whereas male runners were more likely to engage in social conversations.[78] These conclusions, however, are again limited by the lack of control for exercise intensity in most studies (see Tenenbaum and Connolly[70] for an exception). ª 2009 Adis Data Information BV. All rights reserved.

Research on the temporal patterning of A/D use clearly illustrates that a mix of attentional focus strategies is used over the course of endurance events.[26,37,74,75,79-81] Although association appears to be the primary strategy throughout an event, the ratio of associative to dissociative focus approaches an equal distribution during the middle stages before becoming more associative towards the finish. The individual ability to shift between association and dissociation during exercise, termed attentional flexibility, has been identified as a topic that deserves future research consideration.[56,69,82] Moreover, although Schomer[24,83-85] has consistently emphasized that injury prevention might be one of the most important possible benefits of using association, other studies have failed to find a relationship between using a particular attentional focus strategy and the occurrence of injuries.[44,86] However, there is evidence that having sustained a previous injury contributes to using a more associative focus during subsequent exercise.[76] 5.2 Perceived Exertion

Ratings of perceived exertion (RPE) have been perhaps the most widely studied outcome in investigations examining the effectiveness of A/D strategies (35 out of 88 studies). Perceived exertion represents a ‘gestalt’ of all sensory inputs pertaining to the intensity of exercise. Theoretically, an attentional focus strategy that amplifies physical sensations, as in the case of association, should result in consistently higher perceived exertion ratings. Conversely, any attentional focus strategy that attenuates physical sensations, as in the case of dissociation, should result in consistently lower ratings. A review of the studies investigating the relationship between A/D strategies and perceived exertion, however, reveals that findings have been inconclusive. In some cases, these results may be due to the confounding influence of sex or uncontrolled individual-difference variables. For example, Wrisberg et al.[87] reported that, under a self-focused (i.e. associative), low-intensity exercise condition, male participants displayed higher heart rates and lower Sports Med 2009; 39 (9)

Associative and Dissociative Strategies

perceived exertion ratings, whereas female participants exhibited lower heart rates and higher perceived exertion ratings. On the other hand, female participants classified as ‘externals’ on a locus-of-control scale (i.e. tended to attribute outcomes to external causes) reliably reported higher perceived exertion ratings across cycle ergometer and treadmill exercise conditions compared with a group of ‘internal’ female participants.[88,89] Some evidence suggests that both associative and dissociative strategies can result in higher perceived exertion ratings. For example, some studies have demonstrated higher perceived exertion ratings during short-[90,91] and long-[24,78,83-85] distance running and rowing[45] related to associative strategies or thinking. On the other hand, other studies have shown that dissociative thinking can also result in higher RPE.[43,92,93] Delignie´res and Brisswalter[94] noted higher perceived exertion scores when participants performed a dissociative task (i.e. reaction time) while cycling at 20%, 40%, 60% and 80% maximal oxygen uptake . (VO2max). Conversely, other investigations have noted that dissociation results in lower perceived exertion ratings during running,[42,95] cycle ergometry,[96,97] moderate-intensity exercise[94,95] and various self-paced physical activities.[34] Several researchers, using self-reported physical symptoms as a complement to perceived exertion, have observed fewer physical symptoms when focusing externally or dissociating compared with associating.[31,32] Finally, a number of studies have found no difference in RPE between association and dissociation strategies during swimming,[98,99] outdoor versus indoor running,[100] cycling at low, moderate and high exercise intensities,[33,101,102] self-paced running[22] and military marching.[103] Evidence suggests that lower perceived exertion may be related to dissociation at lower exercise intensities and to association at higher exercise intensities.[70,104] In fact, some researchers have suggested that a shift from dissociation to association appears to be initiated around a rating of 13 (‘somewhat hard’)[105] or when relative exercise intensities exceed 50% of maximal workload.[70] ª 2009 Adis Data Information BV. All rights reserved.

751

5.3 Affective and other Psychological Responses

Studies examining the effect of A/D strategies on psychological responses have focused on a wide range of variables, including affective responses, cognitive performance and programme adherence. Some researchers have proposed that exercise-induced affective responses represent a type of associative experience. For example, in a series of studies on marathon running and selfregulatory processes, Schomer[24,83-85] argued that an internal/task-related associative strategy consisted, in part, of ‘‘feelings and affect’’ (p. 45).[85] These thoughts were composed of general wholebody sensations, feelings of vitality or fatigue, and nonspecific overall body tiredness and/or stiffness. Researchers have examined the range of psychological responses, from basic affect to specific emotional feeling states to broad mood states. 5.3.1 Affective, Emotional and Mood-Related Measures

Affective, emotional and mood-related responses have received less attention than RPE within the A/D literature (29 of 88 studies). Investigations of the basic affective dimension of pleasure-displeasure in A/D research have been based on the Feeling Scale, an 11-point rating scale ranging from ‘I feel very good’ (during exercise) to ‘I feel very bad’.[106] Based on the results of studies using this measure, both association and dissociation have been found to be related to declines in pleasure. Researchers have observed declining pleasure ratings with a dissociative strategy during . treadmill exercise at 90% VO2max,[92] as well as greater post-exercise distress reports in untrained participants performing stair-climbing exercise.[43] Baden et al.[95] observed a relationship between more negatively valenced affective responses and greater associative thinking during 20 minutes of treadmill running at 75% peak treadmill running speed. Participants exercised under conditions in which they were: (i) informed of how long they would be running (‘20 minutes’); (ii) told they would run for 10 minutes and then unexpectedly were told to run for 10 additional minutes Sports Med 2009; 39 (9)

752

(‘10 minutes’); or (iii) not informed of the duration (‘UN’). In each 20-minute condition, there was a significant linear increase in associative thinking over time. There was a significant decline in pleasure ratings between minutes 10 and 11 during the ‘10 minutes’ condition compared with either the ‘20 minutes’ or ‘UN’ conditions. Other authors have also commented on the phenomenon of parallel increases in associative thought content and decreases in pleasure and enjoyment. For example, Brewer et al.[43] noted that ‘‘focusing on distress cues while performing an endurance task is counterproductive in terms of both performance and quality of experience (i.e. pain, affect)’’ (p. 12). At the other end of the attentional focus continuum, association and negatively valenced affective ratings have also been found to be related. Welch and colleagues[105] noted declining pleasure ratings reported by young physically inactive women during a cycle ergometer test to volitional exhaustion. This decline in affective valence was paired with more associative thinking, particularly beyond the VT. The authors noted that, ‘‘on average, participants held a greater awareness of the physical sensations of the physiological changes around the VT and beyond, which is likely to manifest itself in both the type of attentional focus reported and the affect experienced’’ (p. 416). Besides A/D, other concurrent cognitive appraisals may also be influential. Cioffi[107] had participants perform 10 minutes of cycle ergo. metry at 60% VO2max either with (association) or without (dissociation) instructions to closely monitor physical sensations. Half of the participants within each condition were then informed that they could be randomly shocked during the trial. Post-experimental examination of the physical sensations experienced revealed that, regardless of receiving or not receiving instructions to monitor physical sensations, individuals who had received the threat rated their physical sensations as more unpleasant compared with the no-threat group. Other investigations of A/D strategies have focused on distinct feeling states. The most commonly used instrument to measure these specific states has been the Exercise-induced ª 2009 Adis Data Information BV. All rights reserved.

Lind et al.

Feeling Inventory (EFI).[108] It should be noted that, unlike the Feeling Scale, the EFI has typically been administered before and after exercise, not during exercise. Contrary to the previously discussed findings, dissociation has been consistently linked to improvements in the feeling states of revitalization, positive engagement and tranquillity, and reductions in physical exhaustion during submaximal aerobic exercise in young, healthy participants.[42,109] Studies examining other affective states have reported varying relationships with A/D strategies. For example, Durtschi and Weiss[66] found that ‘non-elite’ Olympic-trial marathon runners were more anxious in the days prior to and immediately before the event compared with their ‘elite’ (invited) counterparts. Subsequent analyses of thought-content reports provided by the nonelite competitors showed greater dissociative thinking than elite competitors during the event. Using a similar sample but investigating a rather different phenomenon, Masters[110] reported a significant positive correlation between dissociation and the euphoric ‘runner’s high’ among marathon competitors. More recently, Couture et al.[103] found that only the control group reported lower perceived fatigue scores during a military march, whereas the experimental groups of association (i.e. biofeedback), dissociation (i.e. meditation) and combined associationdissociation (i.e. biofeedback and meditation) did not. Finally, some researchers have focused on broad mood states. With respect to ultraendurance events, association has been found to be related to worsening mood states,[81] and the variance in negative mood states can be almost entirely accounted for by pain sensations.[79] The effects of dissociation, on the other hand, appear less consistent. Reports of no effect[33] or fewer physical symptoms and more positive mood with dissociative strategies[32] have been published, even from the same laboratory. However, exercise intensity was not precisely controlled in these studies. Pennebaker and Skelton[111] provided a helpful theoretical basis for understanding the link between psychological responses and A/D Sports Med 2009; 39 (9)

Associative and Dissociative Strategies

strategies. They argued that simply attending to physical symptoms intensifies the sensations and that these sensations are interpreted based on contextual cues. Mood states can serve as contextual cues. Results from their investigations highlight low to modest correlations between negative mood states and physical-symptom reporting. Given that physical symptoms are influenced by both attentional focus and interpretive cues, these investigators recommended that future research should examine ‘‘which situational variables force attention to the body and bring into play various [interpretive] sets’’ (p. 529). 5.3.2 Other Psychological Responses

Some investigations have demonstrated that association is related to longer reaction times[112] and more response errors, specifically at high heart rates. This has been interpreted as suggestive of ‘‘an internalizing of attention as individuals focus on internal signals of pain and fatigue rather than upon the external stimuli.’’[113] Other studies of cognitive tasks, however, have shown either no decrement[81] or improved performance.[103] According to some researchers, performance outcomes depend on dissociative complexity.[101] However, the lack of control for relative exercise intensity also cannot be discounted as a possible reason for the inconsistent results. Studies of exercise compliance are similarly inconclusive. On the one hand, thematic analysis of case vignettes in a qualitative study showed that both attentional focus strategies would contribute to improved compliance.[114] On the other hand, while a dissociative compared with an associative strategy was found to improve both immediate and long-term exercise programme adherence,[115] other evidence suggests that use of internal (i.e. associative) or external (i.e. dissociative) self-statements was unrelated to run distance or adherence at 6 months.[116] 5.4 Preferred Attentional Focus Style

An area that warrants future research consideration involves individual differences in the ª 2009 Adis Data Information BV. All rights reserved.

753

preference for a particular attentional focus strategy. Although some studies have found a strong preference for one strategy over the other (e.g. association over dissociation),[86] others suggest a more equal division during an endurance activity.[103] Some researchers have proposed examining attentional flexibility during exercise,[56,82] and there appears to be support for the tendency to shift between strategies[24,80] regardless of age or running distance.[74,75] For example, Saintsing et al.[117] noted that individuals assigned to association and dissociation groups used the designated strategy 62% and 43% of the time, respectively, for the duration of a 1.5-mile run. In other words, participants opted to use a strategy other than their assigned one between one-third to over half of the time it took to complete the run. Surprisingly, despite the indications that individuals tend to use both A and D strategies within a single exercise bout, there is presently no further evidence on the prevalence or effectiveness of the combined use of the two strategies. Some investigators[80,98,118,119] have noted difficulty in getting participants to adhere to the assigned attentional focus strategy. In some cases, there has been outright refusal by the participants to adopt a specific strategy.[84,85] In other cases, participants were grouped depending on the strategy they actually used.[118] To further complicate matters, there may be discrepancies between which strategy is preferred and which improves performance,[120] although there is some support for improved performance and easier strategy adherence when using the preferred technique.[121] Tailoring strategies to an individual’s preference may improve compliance[98] as well as improve work output.[121] The preference for an attentional focus strategy is as unique as the individual, and what is attended to depends, in part, on past experience and the importance assigned to stimuli.[122] As noted by Pargman,[122] ‘‘in regard to certain contextual demands, some styles are more supportive of efficient, accurate, or desirable outcomes’’ (p. 396). It should be kept in mind, however, that during exercise of high intensity, association might be unavoidable.[70,83,90,105,123] Sports Med 2009; 39 (9)

Lind et al.

754

5.5 Personality Factors

The preference for a predominantly associative or dissociative strategy may be accounted for, in part, by individual dispositional tendencies. For example, some investigators[88,89] have observed that locus of control influences perceived exertion during exercise. Locus of control refers to how people process the information about and reinforcement from a behaviour. Individuals with an external locus of control attribute reinforcement to factors beyond (i.e. external to) their control and attend to non-relevant information. In contrast, individuals with an internal locus of control attribute reinforcement to factors emanating from within (i.e. internal) and seek out information relevant to the activity. The authors argued that differences in information-processing capacities between ‘internals’ and ‘externals’ would be evident in how somatic cues were rated.[89] When applied to exercise, individuals with an internal locus of control are theorized to actively seek out information regarding the activity (e.g. effort sense), whereas those with an external locus of control divert their attention outward. Results from exercise studies have shown that individuals with an external locus of control report higher perceived exertion ratings compared with individuals with an internal locus of control, particularly at higher workloads.[88,89] It should be noted, however, that the hypothesized link between an internal locus of control and association, and an external locus of control and dissociation, has not yet been established. Additional traits may serve as a proxy to either associative or dissociative cognitive styles. For example, of Jung’s four basic personality dimensions of extraversion-introversion, sensing-intuition, thinking-feeling and judging-perceiving, sensing appears most relevant to A/D research. Specifically, sensing is related to how an individual perceives and understands his or her environment by relying on the five senses (i.e. attend to bodily sensations). In a sample of 50 competitive marathoners, Gontang et al.[124] reported that the most common personality profile was introvert-sensingthinking-judging (ISTJ). Some support for this has also been provided by other researchers.[125] ª 2009 Adis Data Information BV. All rights reserved.

Other, more specific, dispositions have also been examined, including competitiveness, commitment and motivation. It is difficult to draw any definitive conclusions from these relatively few studies. There appears to be some support for association to be positively correlated with the trait competitiveness.[40,46] Dissociation, on the other hand, appears more related to the individual’s commitment to[48] and motivation for running,[72] as well as his or her thought content.[36] Furthermore, sex might play a role in some situations, as female runners have been shown to be more likely to engage in ‘personal problem solving’ during marathon training.[78] 5.6 Exercise Economy

In addition to the self-reported and other psychological outcomes summarized in the previous sections, the effects of A/D on a range of physiological outcomes have also been investigated. In the following sections, we examine studies investigating effects on heart rate, oxygen consumption, ventilatory responses, the respiratory exchange ratio and hormonal responses. 5.6.1 Heart Rate

Measuring absolute (HRpeak, HRmax) or relative (%HRpeak, %HRmax) heart rate as well as blood pressure while associating or dissociating has been common practice within the attentional focus literature. Twenty-one of the 88 studies have included heart rate data. Findings from studies in which such measures were taken have shown equivocal results. Several investigators have reported no changes in absolute heart rate or blood pressure[22,31,90,97,101,126] under either association or dissociation conditions. Alternatively, other researchers have observed that association results in lower[103] as well as higher[45] heart rate. For example, Rushall et al.[127] noted significantly higher heart rates while using task-relevant statements (i.e. association) compared with a control condition in a sample of competitive cross-country skiers. Similarly, dissociation has been found to decrease[103,104] as well as increase[128] heart rate. For example, Morgan and colleagues[50] observed lower heart Sports Med 2009; 39 (9)

Associative and Dissociative Strategies

rates during the initial phase (minute 5) of an incremental treadmill test under a dissociative condition compared with both placebo and control conditions. These differences, however, were eliminated by the final minute of the test. 5.6.2 Oxygen Consumption

. . Measures. of absolute . (VO2peak, VO2max) and relative (%VO2peak, %VO2max) oxygen consumption are another studied index of exercise intensity level and economy within the A/D literature (six of 88 studies). Unlike the conflicting findings on heart rate, results on oxygen consumption have typically shown no effect. Smith and colleagues[128] found no difference in oxygen consumption (mL/kg) per kilometre between a control condition and both passive and [50] active association. . Morgan et .al. failed to find differences in VO2max or %VO2max under dissociation at any stage of an incremental treadmill test. Finally,. Hatfield et al.[126] observed no differences in VO2 between a feedback (i.e. association), a distraction and a control condition during a submaximal treadmill run. Only Martin et al.[129] noted that competitive runners who scored high on a self-attention questionnaire, and therefore could be classified as having a more associative orientation, demonstrated better running economy – defined as lower oxygen uptake relative to bodyweight (e.g. mL/kg/min). 5.6.3 Ventilatory Responses

Ventilatory measures, including minute . venti. . lation . . (VE) and ventilatory equivalents (VE/VO2, VE/VCO2), also provide information as to the exercise intensity level or economy during an exercise bout. In general, it appears that association has a beneficial effect on ventilatory responses (two of 88 studies). For example, Hatfield et al.[126] had participants complete a 36-minute submaximal (sub-ventilatory threshold) treadmill run under the conditions of biofeedback, distraction (reaction time task) and control. The researchers observed significant differences in numerous ventilatory variables between the feedback and other conditions. Specifically, the . . . feedback condition elicited lower VE/VO2, VE (L/min), respiration rate, tidal volume and ª 2009 Adis Data Information BV. All rights reserved.

755

pressure of end-tidal oxygen and carbon dioxide compared with the other conditions. These results confirmed an earlier study by Hatfield et al.,[130] in which an associative strategy (visual feedback of ventilatory . . responses) resulted in significantly lower VE/VO2 compared with both a control and a dissociative condition. Collectively, these results prompted the researchers to suggest a link between associative coping strategies and ventilatory efficiency and to conclude that ‘‘psychological processes may alter metabolic efficiency during intense activity’’ (p. 441). Attempts have been made to design interventions aimed at helping runners tune into their ventilatory responses. Simes[131] developed a cognitive coping strategy that incorporated both associative and dissociative elements (PaceAssisted Dissociation/Association; PADA) in addition to running mechanics. This strategy involved the ‘‘coordination of respiration with stride frequency with continuous attention to it maintained by counting respirations’’ (p. 2). This strategy was thought to be most beneficial during uphill running to avoid the transition into anaerobic supplementation. Simes[131] stated, ‘‘By keeping the respiration and stride frequency in synchrony on the uphill grade, the runner naturally shortens the stride length and thus stays closer to anaerobic threshold or the optimum metabolic workload’’ (p. 2). 5.6.4 Respiratory Exchange Ratio

Another index of exercise economy that has received little attention in the A/D literature (two of 88 studies) is the respiratory exchange ratio (RER). This measure provides another index of exercise economy by highlighting the relative contribution of either carbohydrate or fat oxidation towards energy expenditure. Despite the relatively few studies that have included this measure, there appears to be support for an associative strategy resulting in a lower ratio (i.e. higher percentage of fat oxidation). For example, Hatfield and colleagues[126] reported significantly lower RER values in participants using biofeedback (i.e. association) compared with either a distraction or a control condition during a run just below the ventilatory threshold. Sports Med 2009; 39 (9)

Lind et al.

756

In a similar study, Smith et al.[128] observed that the most economical runners (those showing . lower HR, VE and RER) reported significantly less use of dissociation compared with the least economical runners. However, the most and least economical runners did not differ in the use of association. As in many other studies in this literature, intensity was not precisely controlled. 5.6.5 Hormonal Responses

To date, only one known study has examined the influence on A/D strategies on stress hormone markers. To examine the effects of differences in attentional focus on the stress hormone response, Harte and Eifert[100] had participants run outdoors (dissociation) or indoors on a treadmill for 45 minutes with either an internal (association) or external (dissociation) focus. The researchers noted that adrenaline (epinephrine) did not appreciably differ between conditions, but that both cortisol and noradrenaline were higher under the indoor-internal focus condition. Moreover, participants rated the indoor-internal focus as least pleasing compared with the other conditions. However, the results are confounded by a notable limitation of the study: specifically, exercise intensity was not controlled and therefore the effects of the physical stress of exercise and the A/D intervention could not be teased apart. 5.7 Exercise Tolerance

The effects of A/D strategies on exercise tolerance and other performance measures have been examined in 28 of 88 studies in the A/D literature. A variety of measures of performance have been used, such as distance covered (e.g. metres rowed, running distance), time elapsed (e.g. running time, time to exhaustion) or work reproduction. The majority of investigations support association as a more beneficial strategy compared with either dissociation or no strategy for enhancing performance times during swimming,[119,120] cross-country skiing,[127] running,[42,117,132,133] rowing,[45,134] race walking,[135] triathlon[38] and submaximal cycle ergometry.[136] Conversely, dissociative thought content has been shown to be related to slower marathon ª 2009 Adis Data Information BV. All rights reserved.

times[44,66] as well as experiencing the phenomenon of ‘hitting the wall’ earlier and for a longer duration.[27] It should be noted, however, that an alternative explanation for these results is that the performance demands may dictate the attentional focus strategy utilized. This implies that it might not be that the use of association led to improved performance but rather that, under the conditions of maximal effort required to achieve a great performance, attentional focus might be forced to shift toward association.[60-63] However, there have also been some inconsistent findings of the effectiveness of A/D strategies on performance. According to Lorentzen and Sime (personal communication), in some cases, an equal number of respondents have reported perceived running performance improvements with association and dissociation. Nietfield[137] failed to find a significant correlation between the strategy of monitoring performancerelated factors (i.e. association) and performance on a 1-mile run. Moreover, using a dissociative strategy during a graded exercise test has been shown to result in both performance improvements[50] and decrements.[104] Furthermore, an attentional strategy commonly used under one set of circumstances may not be beneficial under another.[80] It should also be pointed out that not all authors agree that altering one’s perception of pain or exertion can be seen as a beneficial strategy. For example, Guyot[138] noted that runners who pushed themselves to the point of feeling pain did not have better running statistics compared with those who did not push themselves to pain during a run. Of particular interest is the author’s conclusion that ‘‘it makes little sense to take risks associated with medical symptoms and injury when the main goals of running are improved health and fitness’’ (p. 460). In studies of work reproduction or recall, some results suggest a greater ability to reproduce running times[139] with an associative strategy, whereas other investigations report no difference between association and dissociation during submaximal cycling[140] or self-paced running.[32] These conflicting results could be due to differences in the demand characteristics of the exercise bouts or Sports Med 2009; 39 (9)

Associative and Dissociative Strategies

differences in the participants’ experiences with the task in different studies. Exercise intensity was controlled in only one study.[140] Finally, it has also been suggested that, compared with the importance of training variables and other psychological factors (e.g. self-efficacy), A/D strategies might be less critical for performance.[118] 6. Discussion As can be seen from this review, there continues to be strong research interest in attentional focus strategies within the context of physical activity. Yet, despite some refinements in the conceptualization of association and dissociation, there remain many inconsistencies in the findings and, accordingly, important inconsistencies in the recommendations issued to exercisers and practitioners. For example, a common theme that has emerged in this review is the lack of experimental control for exercise intensity. Given that the focus of one’s attention depends, in part, on what cues one is most aware of, it is surprising that greater attention to exercise intensity was not observed in more investigations. This review examined the effects of attentional focus strategies on specific perceptual, affective and physiological variables associated with exercise. In general, studies of exercise economy, tolerance, affective responses and perceptions of exertion have not yielded consistent findings. Some studies suggest that improvements in one factor may be at the expense of another. For example, during 5 km runs, LaCaille et al.[42] noted improved running performance times with an associative strategy but significant improvements in feeling states using a dissociative strategy. Attempting to find the point where physiological risk and psychological benefit are balanced seems critical. Indeed, numerous investigators have commented on the reciprocal relationship between increased physiological stress and deteriorating psychological responses.[43] For example, Harte and Eifert,[100] in their study of hormonal marker changes under associative versus dissociative conditions, concluded that ‘‘patterns of urinary adrenaline, noradrenaline, and cortisol excretion and ª 2009 Adis Data Information BV. All rights reserved.

757

concomitant emotional change differ when environmental setting and the focus of attention are altered and a normally pleasant task becomes tedious and negatively evaluated’’ (p. 54). Furthermore, an important effect of attentional focus strategies might be during the post-exercise experience.[141] Influencing how the exercise stimulus is perceived and registered in memory might be critical for an exerciser’s subsequent motivation to continue engaging in exercise. Researchers have noted differences in psychological responses to experimentally induced pain depending on whether the individual monitored was distracted from or suppressed the painful sensation. Cioffi and Holloway[141] found that use of a dissociative coping strategy (i.e. distraction and suppression) during the exposure to a painful stimulus was associated with higher pain ratings during the postexposure period. In contrast, monitoring specific sensory components of pain allowed individuals to assume control over the stimulus, resulting in nonvalenced descriptions of pain. In other words, an individual was more likely to describe the experience in negative affective terms after being distracted from or suppressing thoughts of pain during exposure. While actively monitoring the sensations, however, the individual was more likely to describe the pain in neutral terms (e.g. ‘‘The more I paid attention to it, the more the pain started to feel more like an itch’’).[141] 6.1 Future Directions: What are the Possible Sources of Inconsistencies?

The studies highlighted in this review have contributed considerably to the current understanding of associative and dissociative cognitive strategies across the realm of exercise. There remains, however, a need to address and clarify some key issues that might have contributed to the inconsistent results. 6.1.1 Participant Characteristics

Arguably, because of the pioneering work of Morgan,[23,49] much of the focus on the associative versus dissociative styles debate has centered on elite athletes. Of the 88 studies included in this review, 57 investigations (»65%) used participants Sports Med 2009; 39 (9)

Lind et al.

758

who could be classified as moderately to well trained athletes, competitive at the local, regional or national level in their sport. For example, Beaudoin. et al.[92] investigated highly trained men (average VO2: 70.5 mL/kg/min), Takai[139] sampled endurance runners averaging approximately 145 km/wk, Netfield[137] tested competitive mile runners (average times: men 4 : 29 minutes; women 5 : 27 minutes), Morgan et al.[51] used runners with average 10 km times of 27 : 49 minutes, and Durtschi and Weiss[66] researched competitive male (average time 2 : 20 hours) and female (average time 2 : 50 hours) marathon runners. Conversely, only nine studies (»10%) used participants who could be classified as ‘sedentary’, with the remaining 22 studies (25%) sampling participants characterized as ‘healthy’. Consequently, the samples represented in this review reflect a bias in favour of competitive athletes and, therefore, also a possible bias in favour of using or preferring associative strategies. Given Morgan’s[23] assertion that dissociation might be more advisable for helping the untrained individual tolerate an exercise bout, it seems difficult to arrive at any definitive conclusions given the paucity of research using physically inactive individuals. Drawing generalizable conclusions is further complicated because participants were primarily <30 years of age in most studies (55%). Consequently, it appears premature (and potentially hazardous) to recommend either association or dissociation to middle-aged, overweight, sedentary individuals who may be at risk for chronic disease (probably the typical profile of beginning exercisers in the US and other industrialized countries with high rates of physical inactivity). 6.1.2 Exercise Stimulus

The second issue deals with the nature of the exercise stimulus. Understandably, because of the early work with marathoners, running has continued to be the most studied exercise mode in attentional focus research (»47%). Following running, the most common modalities are, in order: cycle ergometry (»22%), treadmill exercise (e.g. exercise testing, »14%), rowing ergometry (»5%), swimming (»5%), walking (3%) and ª 2009 Adis Data Information BV. All rights reserved.

stairclimbing, cross-country skiing and military marching (1% each). The most commonly studied intensity level has been submaximal (63 studies; »71%). Within these studies, the range has varied from 4 minutes of submaximal cycle ergometry[89] to the sustained effort needed to complete an ultraendurance run.[35] In 13 studies (»15%), the intensity level was characterized as ‘self-selected’ or ‘self-paced’. Of these investigations, the exercise stimulus ranged from running 1 mile[32] to 30 minutes of cycle ergometry.[34] Intensity levels that approached or reached maximal were investigated in 12 (»14%) of the studies. Examples range from . performing a 30-minute treadmill run at 90% VO2max[92] to cycling to volitional exhaustion.[105] As can be seen in table I (supplemental material), the wide range of intensity levels and durations of the exercise stimuli make it difficult to conclude when each strategy is most effective. It seems particularly noteworthy that exercise intensity has been identified as a critical moderator of the effectiveness of A/D strategies in most of the theoretical models discussed earlier.[60-63] Yet, with rare exceptions,[70,123] A/D studies have not been specifically designed to test the proposed moderating role of exercise intensity. This seems surprising, as several authors have found that attentional focus tends to become more associative as exercise intensity increases.[70,83,90,96,105,123] Others have commented that a level of exercise intensity that is high enough could render any attempt at manipulating one’s attentional focus ineffective.[23,62,63,140] For example, Siegal et al.[140] speculated that an increased workload might ‘‘eventually negate the attentional effect’’ (p. 152). Given the central role that has been attributed to exercise intensity through the years, the absence of more experimental investigations that have directly compared the effectiveness of A/D strategies across different levels of exercise intensity is striking. In anticipation of future experimental investigations on the role of exercise intensity, it might be worth highlighting the possibility that the intensitydependent changes in the effectiveness of A/D strategies might be linked to important physiological landmarks such as the ventilatory threshold and respiratory compensation point.[63,65] Intuitively, it Sports Med 2009; 39 (9)

Associative and Dissociative Strategies

seems reasonable to suggest that a dissociative strategy would be mostly effective when peripheral physiological changes first start to generate a bombardment of the brain’s interoceptive mechanisms with salient cues. This is likely initiated when an exerciser reaches his or her ventilatory threshold. It is at that level of intensity that the notion of ‘competition of cues’, first advanced by Pennebaker and Lightner[31] as the mechanism by which dissociation exerts its beneficial effects, appears most relevant. Conversely, when exercise intensity is raised to a level that does not permit the maintenance of a physiological steady state and homeostasis is in jeopardy, the need for survival and successful adaptation dictates a direct, veridical link between bodily cues and conscious awareness. One could speculate that this switch to an associationdominant mode is probably unavoidable (i.e. not amenable to deliberate manipulation)[62,63] and takes place proximally to the respiratory compensation point or the maximal lactate steady state. 6.1.3 Experimental Designs

The final issue deals with the level of methodological quality that characterizes many investigations in this area. There seems to have been a preponderance of studies lacking a control group or condition and an over-reliance on correlational designs and observational methods of data collection. Twenty-one of the 88 studies (»24%) reviewed used self-reports of attentional focus during the exercise bout. Although 42 studies (»48%) compared different experimental conditions (e.g. association group vs dissociation group), only 18 investigations (»20%) had a control group or condition. It is clear that, although quasi-experimental, correlational and observational studies have been instrumental in raising awareness for the possible importance of A/D strategies in influencing both performance and the experience that exercisers derive, definitive results can only be obtained through tightly controlled, hypothesis-driven experimental investigations. In that sense, it could be argued that the level of evidence that is currently available, despite its volume, continues to be preliminary in nature and, as such, it forms a relatively weak foundation for deriving meaningful practical recommendations. ª 2009 Adis Data Information BV. All rights reserved.

759

6.2 A Comment on A/D Guidelines and Applications

The theme that emerges from the present review appears to be one of inconsistency in the findings and lack of systematicity in the investigational approaches. Nevertheless, efforts have been made to issue A/D-related recommendations to exercisers and sport competitors. For the reasons noted in the previous sections, these should be viewed with caution. Some authors have outlined comprehensive psychological skills packages designed to improve performance in a variety of endurance sports.[142,143] These packages include cognitive strategies such as self-talk, imagery, arousal regulation and attentional focus manipulation. Similarly, some researchers have endorsed the benefits of one attentional focus strategy over the other.[23,71] For example, some have argued that an outward distraction ‘‘allows the enjoyment derived from the atmosphere to be fully appreciated, and this degree of distraction minimizes the effect of any discomfort.’’[27] Stanley et al.[96] recommended dissociation for recreational athletes ‘‘as long as the physical effort is tolerable’’ (p. 361), while Berger[144] concluded that dissociation could ‘‘be achieved most easily at a gentle, slow pace’’ (p. 45). Conversely, other authors have recommended an associative strategy as most beneficial for self-regulation[24,56,82-85,98] and prevention of dropout.[98] Schomer[24] observed that his sample of marathon runners ‘‘preferred to deal with pain or discomfort associatively by talking about their origin, and adjusting pace and stride to alleviate the symptoms’’ (p. 55). As noted earlier, the intensity level of the exercise bout appears to be a constant theme in each of these recommendations and it deserves greater attention in future research.[123] Increasing levels of exercise intensity are associated with unpleasant sensations of pain and fatigue,[145] prompting some to question whether cognitive strategies can compete with salient physical sensations.[62] Studies have shown that thought content becomes more associative as intensity increases.[70,83,90,96,105,123] Other researchers contend that the decision to use either attentional focus strategy depends on whether there is a specific desired outcome or on Sports Med 2009; 39 (9)

Lind et al.

760

the experience level of the individual. A number of researchers have noted the benefit of using association to improve performance and dissociation to enhance the physical activity experience.[45,80] In contrast, other investigators have advanced the idea of association as the optimal strategy for individuals experienced with sensations generated internally and their meaning, whereas dissociation as the preferred strategy for inexperienced individuals for whom processing of internal sensations might be unfamiliar.[53,56] However, at present, evidence for the validity of these interpretations remains scant. There appears to be a paradox within the attentional focus literature. More specifically, it seems that using attentional associative strategies for fit individuals allows for better regulation of effort and pace. Alternatively, attentional dissociative strategies have been recommended for individuals initiating an exercise programme as a means of better tolerating the physiological strain associated with exercise. However, there are several scenarios (e.g. previously sedentary and inexperienced individuals initiating an unsupervised programme of physical activity, and overweight or obese individuals) in which intensity may rise to a level where focal awareness becomes centered exclusively on somatic cues. As suggested by the theoretical models outlined earlier, this may happen despite an individual’s effort to divert attention away from internal cues and towards external cues.[60-63] In addition, in some exercise environments, such as rehabilitation programmes, individuals are often encouraged to focus attention on bodily cues in order to monitor intensity and effort level to prevent potentially dangerous levels of exertion. This paradox has prompted other investigators to call for studies using improved conceptual models[146] and identifying the link between attentional focus and bodily responses.[31,53,119,147] Possibly the most persuasive argument was outlined by Masters and Ogles.[55] In concluding their review of 20 years of A/D research, the authors stated: ‘‘The theoretical foundations of A/D need further development. The field presently operates on somewhat implicit and unexamined theoretical underpinnings. Since Morgan and Pollock’s (1977) ª 2009 Adis Data Information BV. All rights reserved.

initial study, little has been done to directly advance our theoretical understanding of why, when, how, and in what context, and for whom A/D operates. Studies that offer theoretical proposals and then test them empirically are encouraged’’ (p. 267). The lack of consensus and the current state of attentional focus research is still well captured by Sachs’[80] earlier conclusion: ‘‘It is clear that we cannot yet provide definitive recommendations on the use of associative and dissociative strategies’’ (p. 300). Thus, it appears that, based on the findings of this review of the A/D literature, little substantive progress has been made in the past decade since the last major review.[55] Given the obesity and physical inactivity epidemic currently impacting most industrialized nations, it seems imperative that future A/D research should consider the limitations that have characterized past investigations and give particular attention to the critical role of exercise intensity before any definitive recommendations can be advanced to the general public. Acknowledgements The authors wish to thank Dr Spiridoula Vazou for critical insights and helpful comments and suggestions on an earlier version of the manuscript. No sources of funding were used to assist in the preparation of this review. The authors have no conflicts of interest that are directly relevant to the content of this review.

References 1. United States Department of Health and Human Services. Physical activity and health: a report from the Surgeon General. Atlanta (GA): Department of Health and Human Services, Centers for Disease Control and Prevention, Office of Chronic Disease Prevention and Health Promotion, 1996 2. Armstrong T, Bauman A, Davies J. Physical activity patterns of Australian adults: results of the 1999 National Physical Activity Survey. Canberra (ACT): Australian Institute of Health and Welfare, 2000 3. Canadian Fitness and Lifestyle Research Institute. 2003 physical activity monitor. Ontario: Health Canada, 2003 4. Sproston K, Primatesta P. Health survey for England. London: The Stationary Office, 2004 5. Mokdad AH, Marks JS, Stroup DF, et al. Actual causes of death in the United States, 2000. JAMA 2005; 293: 293-4

Sports Med 2009; 39 (9)

Associative and Dissociative Strategies

6. Dishman RK, Buckworth J. Increasing physical activity: a quantitative synthesis. Med Sci Sports Exerc 1996; 28: 706-19 7. Ekkekakis P. The study of affective responses to acute exercise: the dual-mode model. In: Stelter R, Roessler KK. editors. New approaches to exercise and sport psychology. Oxford: Meyer & Meyer Sport, 2005: 119-49 8. Ekkekakis P, Lind E. Exercise does not feel the same when you are overweight: the impact of self-selected and imposed exercise intensity on affect and exertion. Int J Obesity (Lond) 2006; 30: 652-60 9. Acevedo EO, Kraemer RR, Haltom RW, et al. Perceptual responses proximal to the onset of blood lactate accumulation. J Sports Med Phys Fitness 2003; 6: 267-73 10. Bixby WR, Spalding TW, Hatfield BD. Temporal dynamics and dimensional specificity of the affective response to exercise of varying intensity: differing pathways to a common outcome. J Sport Exerc Psychol 2001; 23: 171-90 11. Ekkekakis P, Hall EE, Petruzzello SJ. Practical markers of the transition from aerobic to anaerobic metabolism during exercise: rationale and a case for affect-base exercise prescription. Prev Med 2004; 38: 149-59 12. Hall EE, Ekkekakis P, Petruzzello SJ. The affective beneficence of vigorous exercise revisited. Br J Health Psychol 2002; 7: 47-66 13. Cox KL, Burke V, Gorely TJ, et al. Controlled comparison of retention and adherence in home- vs center-initiated exercise interventions in women ages 40–65 years: the S.W.E.A.T. study (Sedentary Women Exercise Adherence Trial). Prev Med 2003; 36: 17-29 14. Epstein LH, Koeske R, Wing RR. Adherence to exercise in obese children. J Card Rehabil 1984; 4: 185-95 15. Lee JY, Jensen BE, Oberman A, et al. Adherence in the training levels comparison trial. Med Sci Sports Exerc 1996; 28: 47-52 16. Perri MG, Anton SD, Durning PE, et al. Adherence to exercise prescriptions: effects of prescribing moderate versus higher levels of intensity and frequency. Health Psychol 2002; 21: 452-8 17. Sallis JF, Haskell WL, Fortmann SP, et al. Predictors of adoption and maintenance of physical activity in a community sample. Prev Med 1986; 15: 331-41 18. Williams DM, Dunsiger S, Ciccolo JT, et al. Acute affective response to a moderate-intensity exercise stimulus predicts physical activity participation 6 and 12 months later. Psych Sport Exerc 2008; 9: 231-45 19. Emmons RA, Diener E. A goal-affect analysis of everyday situational choices. J Res Pers 1986; 20: 309-26 20. Rejeski WJ, Kenney EA. Fitness motivation: preventing participant dropout. Champaign (IL): Human Kinetics, 1988 21. Leith LM. Exercising your way to better mental health. Morgantown (WV): Fitness Information Technology, Inc., 1998 22. Weinberg RS, Smith J, Jackson A, et al. Effect of association, dissociation and positive self-talk strategies on endurance performance. Can J Appl Sport Sci 1984; 9: 25-32 23. Morgan WP. Mind of the marathoner. Psychol Today 1978; 11: 38-49

ª 2009 Adis Data Information BV. All rights reserved.

761

24. Schomer HH. Mental strategies and the perception of effort of marathon runners. Int J Sport Psychol 1986; 17: 41-59 25. Nideffer RM. The ethics and practice of applied sport psychology. New York (NY): Mouvement Publications, 1981 26. Stevinson CD, Biddle SJH. Cognitive orientations in marathon running and ‘‘hitting the wall.’’ Br J Sports Med 1998; 32: 229-35 27. Stevinson CD, Biddle SJH. Cognitive strategies in running: a response to Masters and Ogles (1998). Sport Psych 1999; 13: 235-6 28. Summers JJ, Machin VJ, Sargent GI. Psychosocial factors related to marathon running. J Sport Psychol 1983; 5: 314-31 29. Summers JJ, Sargent GI, Levey AJ, et al. Middle-aged, non-elite marathon runners: a profile. Percept Mot Skills 1982; 54: 963-9 30. Buman MP, Brewer BW, Cornelius AE, et al. Hitting the wall in the marathon: phenomenological characteristics and associations with expectancy, gender, and running history. Psych Sport Exerc 2008; 9: 177-90 31. Pennebaker JW, Lightner JM. Competition of internal and external information in an exercise setting. J Pers Soc Psychol 1980; 39: 165-74 32. Fillingim RB, Fine MA. The effects of internal versus external information processing on symptom perception in an exercise setting. Health Psychol 1986; 5: 115-23 33. Fillingim RB, Roth DL, Haley WE. The effects of distraction on the perception of exercise-induced symptoms. J Psychosom Res 1989; 33: 241-8 34. Padgett VR, Hill AK. Maximizing athletic performance in endurance events: a comparison of cognitive strategies. J Appl Soc Psychol 1989; 19: 331-40 35. Acevedo EO, Dzewaltowski DA, Gill DL, et al. Cognitive orientations of ultramarathoners. Sport Psychol 1992; 6: 242-52 36. Freischlag J. Selected psycho-social characteristics of marathoners. Int J Sport Psychol 1981; 12: 282-8 37. Hollander DB, Acevedo EO. Successful English Channel swimming: the peak experience. Sport Psych 2000; 14: 1-16 38. Baker J, Coˆte´ J, Deakin J. Cognitive characteristics of expert, middle of the pack, and back of the pack ultra-endurance triathletes. Psych Sport Exer 2005; 6: 551-8 39. Goode KT, Roth DL. Factor analysis of cognitions during running: association with mood change. J Sport Exerc Psychol 1993; 15: 375-89 40. Bachman AD, Brewer BW, Petitpas AJ. Situation specificity of cognitions during running: replication and extension. J Appl Sport Psychol 1997; 9: 204-11 41. Butryn TM, Furst DM. The effects of park and urban setting on the moods and cognitive strategies of female runners. J Sport Beh 2003; 26 (4): 335-55 42. LaCaille RA, Masters KS, Heath EM. Effects of cognitive strategy and exercise setting on running performance, perceived exertion, affect, and satisfaction. Psych Sport Exer 2004; 5: 461-76

Sports Med 2009; 39 (9)

762

43. Brewer BW, Van Raalte JL, Linder DE. Attentional focus and endurance performance. Appl Res Coaching Athl Ann 1996; 11: 1-14 44. Masters KS, Ogles BM. The relations of cognitive strategies with injury, motivation, and performance among marathon runners: results from two studies. J Appl Sport Psychol 1998; 10: 281-96 45. Connolly CT, Janelle CM. Attentional strategies in rowing: performance, perceived exertion, and gender considerations. J Appl Sport Psychol 2003; 15: 195-212 46. Masters KS, Ogles BM, Jolton JA. The development of an instrument to measure motivation for marathon running: the Motivations of Marathoners Scales (MOMS). Res Q Exerc Sport 1993; 64: 134-43 47. Tenenbaum G, Fogarty G, Stewart E, et al. Perceived discomfort in running: scale development and theoretical considerations. J Sports Sci 1999; 17: 183-96 48. Carmack MA, Martens R. Measuring commitment to running: a survey of runners’ attitudes and mental states. J Sport Psychol 1979; 1: 25-42 49. Morgan WP, Pollock ML. Psychologic characterization of the elite distance runner. Ann N Y Acad Sci 1977; 301: 382-403 50. Morgan WP, Horstman DH, Cymerman A, et al. Facilitation of physical performance by means of cognitive strategy. Cognit Ther Res 1983; 7: 251-64 51. Morgan WP, O’Connor PJ, Ellickson KA, et al. Personality structure, mood states, and performance in elite male distance runners. Int J Sport Psychol 1988; 19: 247-63 52. Crews DJ. Psychological state and running economy. Med Sci Sports Exerc 1992; 24: 475-82 53. Hardy L, Nelson D. Self-regulation training in sport and work. Ergonomics 1988; 31: 1573-83 54. Laasch C. Cognitive strategies and long-distance running. Imag Cognit Pers 1994-1995; 14: 317-32 55. Masters KS, Ogles BM. Associative and dissociative cognitive strategies in exercise and running: 20 years later, what do we know? Sport Psych 1998; 12: 253-70 56. Moran AP. Cognitive style constructs in sport: explanatory and attentional processes in athletes. Int J Educ Res 1998; 29: 277-86 57. Rose JMC. Association and dissociation cognitive strategies in endurance performance: a brief review. In: Watkins J, Reilly T, Burwitz L, editors. Sports science: proceedings of the VIII Commonwealth and International Conference on Sport, Physical Education, Dance, Recreation and Health. New York (NY): E. & F.N. Spon, 1986: 229-36 58. Leventhal H. A perceptual-motor processing model of emotion. In: Pliner P, Blankstein RK, Spigel IM, editors. Perception of emotion in self and others: advances in the study of communication and affect. New York (NY): Plenum, 1979: 1-46 59. Leventhal H, Everhart D. Emotion pain, and physical illness. In: Izard CE, editor. Emotion and psychopathology. New York (NY): Plenum, 1979: 263-99 60. Rejeski WJ. The perception of exertion: a social psychophysiological integration. J Sport Psychol 1981; 4: 305-20 61. Rejeski WJ. Perceived exertion: an active or passive process? J Sport Psychol 1985; 7: 371-8

ª 2009 Adis Data Information BV. All rights reserved.

Lind et al.

62. Tenenbaum G. A social-cognitive perspective of perceived exertion and exertion tolerance. In: Singer RN, Hausenblas HA, Janelle C, editors. Handbook of sport psychology. New York (NY): Wiley, 2001: 810-22 63. Ekkekakis P. Pleasure and displeasure from the body: perspectives from exercise. Cogn Emot 2003; 17: 213-39 64. Hall EE, Ekkekakis P, Petruzzello SJ. Is the relationship of RPE to psychological factors intensity-dependent? Med Sci Sports Exerc 2005; 37: 1365-73 65. Ekkekakis P, Acevedo EO. Affective responses to acute exercise: toward a psychobiological dose-response model. In: Acevedo EO, Ekkekakis P, editors. Psychobiology of physical activity. Champaign (IL): Human Kinetics, 2006: 91-109 66. Durtschi SK, Weiss MR. Psychological characteristics of elite and nonelite marathon runners. In: Landers DM, editor. The 1984 Olympic Scientific Congress proceedings: sport and elite performers. Champaign (IL): Human Kinetics, 1984: 73-80 67. McDonald D, Kirkby RJ. Use of dissociation strategies when running becomes difficult: levels of ability and gender differences. Eur J High Ability 1995; 6: 73-81 68. Silva III JM, Appelbaum MI. Association-dissociation patterns of United States Olympic marathon trial contestants. Cognit Ther Res 1989; 13: 185-92 69. Antonini-Philippe R, Reynes E, Bruant G. Cognitive strategy and ability in endurance activities. Percept Mot Skills 2003; 96: 510-16 70. Tenenbaum G, Connolly CT. Attention allocation under varied workload and effort perception in rowers. Psych Sport Exerc 2008; 9: 704-17 71. Wrisberg CA, Pein RL. Past running experience as a mediator of the attentional focus of male and female recreational runners. Percept Mot Skills 1990; 70: 427-32 72. Ogles BM, Lynn SJ, Masters KS, et al. Runners’ cognitive strategies and motivations: absorption, fantasy style, and dissociative experiences. Imag Cognit Pers 1993/1994; 13: 163-74 73. Callen KE. Mental and emotional aspects of long-distance running. Psychosomatics 1983; 24: 133-41 74. Okwumabua TM. Psychological and physical contributions to marathon performance: an exploratory investigation. J Sport Beh 1985; 8: 163-71 75. Okwumabua TM, Meyers AW, Santille L. A demographic and cognitive profile of master runners. J Sport Beh 1987; 11: 212-20 76. Ungerleider S, Golding JM, Porter K, et al. An exploratory examination of cognitive strategies used by master’s track and field athletes. Sport Psychol 1989; 3: 245-53 77. Ebbeck V. Sources of performance information in the exercise setting. J Sport Exerc Psychol 1990; 12: 56-65 78. Schomer HH, Connolly MJ. Cognitive strategies used by marathoners in each quartile of a training run. S Afr J Res Sport Phys Educ Rec 2002; 24: 87-99 79. Kirkby RJ. Ultraendurance running: a case study. Int J Sport Psychol 1996; 27: 109-16 80. Sachs ML. The mind of the runner: cognitive strategies used during running. In: Sachs ML, Buffone GW, editors. Running as therapy: an integrated approach. Lincoln (NE): University of Nebraska Press, 1984: 288-303

Sports Med 2009; 39 (9)

Associative and Dissociative Strategies

81. Sacks MH, Milvy P, Perry SW, et al. Mental status and psychological coping during a 100-mile race. In: Sacks MH, Sachs ML, editors. Psychology of running. Champaign (IL): Human Kinetics, 1981: 166-75 82. Moran AP. The psychology of concentration in sport performers: a cognitive analysis. East Sussex (UK): Psychology Press, 1996 83. Schomer HH. The relationship between cognitive strategies and perceived effort of marathon runners. S Afr J Res Sport Phys Educ Rec 1987; 10: 37-64 84. Schomer HH. Mental strategy training programme for marathon runners. Int J Sport Psychol 1987; 18: 133-51 85. Schomer HH. A cognitive strategy training programme for marathon runners: ten case studies. S Afr J Res Sport Phys Educ Rec 1990; 134: 47-78 86. Masters KS, Lambert MJ. The relations between cognitive coping strategies, reasons for running, injury, and performance of marathon runners. J Sport Exerc Psychol 1989; 11: 161-70 87. Wrisberg CA, Franks BD, Birdwell MW, et al. Physiological and psychological responses to exercise with an induced attentional focus. Percept Mot Skills 1988; 66: 603-16 88. Hassme`n P, Koivula N. Ratings of perceived exertion by women with internal or external locus of control. J Gen Psychol 1996; 123: 297-307 89. Koivula N, Hassmen P. Central, local, and overall ratings of perceived exertion during cycling and running by women with an external or internal locus of control. J Gen Psychol 1998; 125: 17-29 90. Baden DA, McLean TL, Tucker R, et al. Effect of anticipation during unknown or unexpected exercise duration on rating of perceived exertion, affect, and physiological function. Br J Sports Med 2005; 39: 742-6 91. Tammen VV. Elite middle and long distance runner’s associative/dissociative coping. J Appl Sport Psychol 1996; 8: 1-8 92. Beaudoin CM, Crews DJ, Morgan DW. Influence of psychogenic factors during a prolonged maximal run. J Sport Beh 1998; 21: 377-81 93. Russell WD, Weeks DL. Attentional style in ratings of perceived exertion during physical exercise. Percept Mot Skills 1994; 78: 779-83 94. Delignie´res D, Brisswalter J. Influence of an added perceptual motor task on perceived exertion: a test of the dissociation effect. Percept Mot Skills 1994; 78: 855-58 95. Baden DA, Warwick-Evans L, Lakomy J. Am I nearly there? The effect of anticipated running distance on perceived exertion and attentional focus. J Sport Exerc Psychol 2004; 26: 215-31 96. Stanley CT, Pargman D, Tenenbaum G. The effect of attentional coping strategies on perceived exertion in a cycling task. J Appl Sport Psychol 2007; 19: 352-63 97. Johnson JH, Siegel DS. Active vs passive attentional manipulation and multidimensional perceptions of exercise intensity. Can J Sport Sci 1987; 12: 41-5 98. Couture RT, Jerome W, Tihanyi J. Can associative and dissociative strategies affect the swimming performance of recreational swimmers? Sport Psych 1999; 13: 334-43

ª 2009 Adis Data Information BV. All rights reserved.

763

99. Couture RT, Tihanyi J, St Aubin M. Can performance in a distance swim be improved by increasing a preferred cognitive thinking strategy? Sport J 2003; 6 (2): 1-6 100. Harte JL, Eifert GH. The effects of running, environment, and attentional focus on athletes’ catecholamine and cortisol levels and mood. Psychophysiology 1995; 32: 49-54 101. Siegal D, Johnson J, Davis C. Attention and perception of intensity of work. Percept Mot Skills 1981; 53: 331-7 102. Stamford BA, Weltman A, Foulke E. Information processing and perceived effort during bicycle ergometer work [abstract]. Med Sci Sport 1979; 11: 78 103. Couture RT, Singh M, Lee W, et al. The effect of mental training on the performance of military endurance tasks in the Canadian infantry. Int J Sports Psychol 1994; 25: 144-57 104. Franks BD, Myers BC. Effects of talking on exercise tolerance. Res Q Exerc Sport 1984; 55: 237-41 105. Welch AS, Hulley A, Ferguson C, et al. Affective responses of inactive women to a maximal incremental exercise test: a test of the dual-mode model. Psych Sport Exer 2007; 8: 401-23 106. Hardy CJ, Rejeski WL. Not what, but how one feels: the measurement of affect during exercise. J Sport Exerc Psychol 1989; 11: 304-17 107. Cioffi D. Sensory awareness versus sensory impression: affect and attention interact to produce somatic meaning. Cogn Emot 1991; 5: 275-94 108. Gauvin L, Rejeski WL. The exercise-induced feeling inventory: development and initial validation. J Sport Exerc Psychol 1993; 15: 403-23 109. Blanchard CM, Rodgers WM, Gauvin L. The influence of exercise duration and cognitions during running on feeling states in an indoor running track environment. Psych Sport Exer 2004; 5: 119-33 110. Masters KS. Hypnotic susceptibility, cognitive dissociation, and runner’s high in a sample of marathon runners. Am J Clin Hypn 1992; 34: 193-201 111. Pennebaker JW, Skelton JA. Psychological parameters of physical symptoms. Pers Soc Psychol Bull 1978; 4: 524-30 112. Coˆte´ J, Salmela J, Papathanasopoulu KP. Effects of progressive exercise on attentional focus. Percept Mot Skills 1992; 75: 351-4 113. Salmela JH, Ndoye OD. Cognitive distortions during progressive exercise. Percept Mot Skills 1986; 63: 1067-72 114. Stetson BA, Frommelt SJ, Boutelle KN, et al. Exerciserelated thoughts in cardiac exercise programs: a study of exercise-adherent cardiac rehabilitation patients. Int J Rehabil Health 1995; 1: 125-36 115. Martin JE, Dubbert PM, Katell AD, et al. Behavioral control of exercise in sedentary adults: studies 1 through 6. J Con Clin Psychol 1984; 52: 795-811 116. Welsh MC, Labbe´ EE, Delaney D. Cognitive strategies and personality variables in adherence to exercise. Psychol Rep 1991; 68: 1327-35 117. Saintsing DE, Richman CL, Bergey DB. Effects of three cognitive strategies on long-distance running. Bull Psychon Soc 1988; 26: 34-6

Sports Med 2009; 39 (9)

764

118. Okwumabua TM, Meyers AW, Schleser R, et al. Cognitive strategies and running performance: an exploratory study. Cognit Ther Res 1983; 7: 363-70 119. Spink KS, Longhurst K. Cognitive strategies and swimming performances: an exploratory study. Aust J Sci Med Sport 1986; 18: 9-13 120. Rushall BS, Shewchuk ML. Effects of thought content instructions on swimming performance. J Sports Med Phys Fitness 1989; 29: 326-34 121. Baghurst T, Thierry G, Holder T. Evidence for a relationship between attentional styles and effective cognitive strategies during performance. Athl Insight 2004; 6 (1): 36-51 122. Pargman D. Individual differences: cognitive and perceptual styles. In: Singer RN, Murphey M, Tennant LK, editors. Handbook of sport psychology. New York (NY): Macmillan, 1993: 379-401 123. Hutchinson JC, Tenenbaum G. Attention focus during physical effort: the mediating role of task intensity. Psych Sport Exerc 2007; 8: 233-45 124. Gontang A, Clitsome T, Kostrubala T. A psychological study of 50 sub-3-hour marathoners. Ann N Y Acad Sci 1977; 301: 1020-8 125. Hartung GH, Farge EJ. Personality and physiological traits in middle-aged runners and joggers. J Gerontol 1977; 32: 541-8 126. Hatfield BD, Spalding TW, Mahon AD, et al. The effect of psychological strategies upon cardiorespiratory and muscular activity during treadmill running. Med Sci Sports Exerc 1992; 24: 218-25 127. Rushall BS, Hall M, Roux L, et al. Effects of three types of thought content instructions on skiing performance. Sport Psych 1988; 2: 283-97 128. Smith AL, Gill DL, Crews DJ, et al. Attentional strategy use by experienced long distance runners: physiological and psychological effects. Res Q Exerc Sport 1995; 66: 142-50 129. Martin JJ, Craib M, Mitchell V. The relationships of anxiety and self-attention to running economy in competitive male distance runners. J Sports Sci 1995; 13: 371-6 130. Hatfield BD, Spalding TW, Mahon AD, et al. Ventilatory economy as a function of attentional self-focus during treadmill running [abstract]. Psychophysiology 1986; 23: 440-1 131. Simes WE. A new look at association/dissociation in long distance runners. Running Psychol 1982; 4: 1-3 132. Donohue B, Barnhart R, Covassin T, et al. The development and initial evaluation of two promising mental preparatory methods in a sample of female cross country runners. J Sport Beh 2001; 24: 19-30

ª 2009 Adis Data Information BV. All rights reserved.

Lind et al.

133. Miller A, Donohue B. The development and controlled evaluation of athletic mental preparation strategies in high school distance runners. J Appl Sport Psychol 2003; 15: 321-34 134. Scott LM, Scott D, Bedic SP, et al. The effect of associative and dissociative strategies on rowing ergometer performance. Sport Psych 1999; 13: 57-68 135. Clingman JM, Hilliard DV. Race walkers quicken their pace by tuning in, not stepping out. Sport Psych 1990; 4: 25-32 136. Goudas M, Theodorakis Y, Laparidis K. The effect of external versus internal types of feedback and goal setting on endurance performance. Athl Insight 2007; 9 (3): 57-66 137. Nietfield JL. An examination of metacognitive strategy use and monitoring skills by competitive middle-distance runners. J Appl Sport Psychol 2003; 15: 307-20 138. Guyot WG. Psychological and medical factors associated with pain running. J Sports Med Phys Fitness 1991; 31: 452-60 139. Takai K. Cognitive strategies and recall of pace by longdistance runners. Percept Mot Skills 1998; 86: 763-70 140. Siegal D, Johnson J, Kline G. Attentional load and the reproduction of physical work. Res Q Exerc Sport 1984; 55: 146-52 141. Cioffi D, Holloway J. Delayed costs of suppressed pain. J Pers Soc Psychol 1993; 64: 274-82 142. Orlick T. Psyching for marathons: in pursuit of excellence. Ottawa (ON): Canadian Coaching Association, 1980 143. Ungerleider S. Mental training for peak performance. Emmaus (PA): Rodale Press, Inc., 1996 144. Berger B. Running strategies for women and men. In: Sachs ML, Buffone GW, editors. Running as therapy: an integrated approach. Lincoln (NE): University of Nebraska Press, 1984: 23-62 145. Roth WT. Some motivational aspects of exercise. J Sports Med 1974; 14: 40-7 146. Rejeski WJ, Kenney EA. Distracting attention from fatigue: does task complexity make a difference? J Sport Exerc Psychol 1987; 9 (1): 66-73 147. Johnson JH, Siegel DS. Effects of association and dissociation on effort perception. J Sport Beh 1992; 15: 119-29

Correspondence: Dr Panteleimon Ekkekakis, 235 Barbara E. Forker Building, Department of Kinesiology, Iowa State University, Ames, IA 50011, USA. E-mail: [email protected]

Sports Med 2009; 39 (9)

Sports Med 2009; 39 (9): 765-777 0112-1642/09/0009-0765/$49.95/0

REVIEW ARTICLE

ª 2009 Adis Data Information BV. All rights reserved.

Rest Interval between Sets in Strength Training Belmiro Freitas de Salles,1,2 Roberto Sima˜o,2 Fabrı´cio Miranda,2 Jefferson da Silva Novaes,2 Adriana Lemos2 and Jeffrey M. Willardson3 1 Laboratory for Clinical and Experimental Research in Vascular Biology (BioVasc), Biomedical Center, State University of Rio de Janeiro, Rio de Janeiro, Brazil 2 School of Physical Education and Sports, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil 3 Kinesiology and Sports Studies Department, Eastern Illinois University, Charleston, Illinois, USA

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Literature Search . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Acute Responses and the Rest Interval between Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Influence of Rest Interval Length on Repetition Performance Over Multiple Sets . . . . . . . . . . . . . 2.2 Influence of Rest Interval Length on Repeated Maximal Strength Assessments . . . . . . . . . . . . . . 2.3 Influence of Rest Interval Length on Acute Expression of Muscular Power . . . . . . . . . . . . . . . . . . . 2.4 Influence of Rest Interval Length on Acute Hormonal Responses and their Influence on Muscular Hypertrophy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Chronic Adaptations and the Rest Interval between Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Influence of Rest Interval Length on Muscular Strength and Power Adaptations . . . . . . . . . . . . . 3.2 Influence of Rest Interval Length on Muscular Endurance Adaptations. . . . . . . . . . . . . . . . . . . . . 4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

765 766 767 767 769 770 770 772 772 774 775 776

Strength training has become one of the most popular physical activities for increasing characteristics such as absolute muscular strength, endurance, hypertrophy and muscular power. For efficient, safe and effective training, it is of utmost importance to understand the interaction among training variables, which might include the intensity, number of sets, rest interval between sets, exercise modality and velocity of muscle action. Research has indicated that the rest interval between sets is an important variable that affects both acute responses and chronic adaptations to resistance exercise programmes. The purpose of this review is to analyse and discuss the rest interval between sets for targeting specific training outcomes (e.g. absolute muscular strength, endurance, hypertrophy and muscular power). The Scielo, Science Citation Index, National Library of Medicine, MEDLINE, Scopus, Sport Discus and CINAHL databases were used to locate previous original scientific investigations. The 35 studies reviewed examined both acute responses and chronic adaptations, with rest interval length as the experimental variable. In terms of acute responses, a key finding was that when training with loads between 50% and 90% of one repetition maximum, 3–5 minutes’ rest between sets allowed

De Salles et al.

766

for greater repetitions over multiple sets. Furthermore, in terms of chronic adaptations, resting 3–5 minutes between sets produced greater increases in absolute strength, due to higher intensities and volumes of training. Similarly, higher levels of muscular power were demonstrated over multiple sets with 3 or 5 minutes versus 1 minute of rest between sets. Conversely, some experiments have demonstrated that when testing maximal strength, 1-minute rest intervals might be sufficient between repeated attempts; however, from a psychological and physiological standpoint, the inclusion of 3- to 5-minute rest intervals might be safer and more reliable. When the training goal is muscular hypertrophy, the combination of moderate-intensity sets with short rest intervals of 30–60 seconds might be most effective due to greater acute levels of growth hormone during such workouts. Finally, the research on rest interval length in relation to chronic muscular endurance adaptations is less clear. Training with short rest intervals (e.g. 20 seconds to 1 minute) resulted in higher repetition velocities during repeated submaximal muscle actions and also greater total torque during a high-intensity cycle test. Both of these findings indirectly demonstrated the benefits of utilizing short rest intervals for gains in muscular endurance. In summary, the rest interval between sets is an important variable that should receive more attention in resistance exercise prescription. When prescribed appropriately with other important prescriptive variables (i.e. volume and intensity), the amount of rest between sets can influence the efficiency, safety and ultimate effectiveness of a strength training programme.

Strength training has been heavily studied during the last 50 years, and is now an integral component of a well-rounded exercise programme. Strength training has been shown to contribute to improvements in sports performance, as well as treatment and prophylaxis of some illnesses.[1,2] Studies have focused on manipulation of the different variables involved in resistance exercise prescription to gain a better understanding of how to best achieve different muscular characteristics. Strength training has been proven to stimulate chronic adaptations within the muscles that lead to increases in strength, hypertrophy, muscular endurance and power. According to the American College of Sports Medicine,[1,2] the main methodological variables of prescription are the intensity, number of sets and repetitions (i.e. volume), rest interval between sets, order of exercises, movement velocity and training frequency. Among such variables, the rest interval between sets has received little attention, relative to other prescriptive variables ª 2009 Adis Data Information BV. All rights reserved.

such as intensity and volume. The existing research has demonstrated that different rest intervals between sets can produce differing acute responses and chronic adaptations in the neuromuscular and endocrine systems.[3-15] The rest interval between sets is commonly prescribed based on the training goal (e.g. strength, power, muscular hypertrophy and endurance).[1,16] However, there are still conflicting findings in the literature, which often makes the identification of the appropriate rest interval difficult. Therefore, the purpose of this review is to analyse and discuss the rest interval between sets for targeting specific training outcomes (i.e. absolute strength, muscular endurance, hypertrophy and power). 1. Literature Search The Scielo, Science Citation Index, National Library of Medicine, MEDLINE, Scopus, Sport Discus and CINAHL databases were used to locate previous literature original scientific Sports Med 2009; 39 (9)

Rest between Sets

investigations. The studies reviewed examined both acute responses and chronic adaptations, with rest interval length as the experimental variable. The search utilized the following keywords: ‘rest interval’, ‘rest period’, ‘recovery’, ‘recovery time’ and ‘fatigue’, combined with the keywords ‘training volume’, ‘repetitions’, ‘sets’, ‘resistance training’, ‘resistance exercise’ and ‘strength training’. The names of authors cited in some studies were also utilized. Hand searches of relevant journals and reference lists obtained from articles were also conducted in the Federal University of Rio de Janeiro and State University of Rio de Janeiro libraries. Such combinations resulted in the inclusion of 35 original research articles addressing the rest interval between sets in strength training. The last search was performed on 17 July 2009. Criteria for inclusion were as follows: 1. Only studies from journals with an impact factor ‡1.0. We included this impact factor because the median impact factor for journals in sport science is 1.0, which is similar to that of many other biomedical disciplines. Those with an impact factor <1.0 are considered low-impact journals by the Sports Science Organization.[17] 2. Studies must have examined the effects of the rest interval as the experimental variable on performance, strength, power, hypertrophy and/or muscular endurance. 2. Acute Responses and the Rest Interval between Sets 2.1 Influence of Rest Interval Length on Repetition Performance Over Multiple Sets

Several studies have demonstrated that the execution of a single set becomes less effective for trained people versus performing multiple sets;[18,19] however, the rest interval between sets may determine the effectiveness of performing a multiple set programme. Although recommendations concerning rest intervals between multiple sets are based on training goals such as strength, power, hypertrophy and muscular endurance, the achievement of these goals may depend on the ability to maintain the number of ª 2009 Adis Data Information BV. All rights reserved.

767

repetitions within a prescribed zone over consecutive sets.[16] With regard to the maintenance of repetitions over consecutive sets, some authors demonstrated that resting less than 3 minutes can result in a significant decrease in repetitions (see table I). Kraemer[20] tested the effects of 1- and 3-minute rest intervals on the total number of repetitions completed in three consecutive sets with a 10 repetition maximum (RM) load on the bench press and leg press exercises. Twenty American football players who had participated in resistance exercise for two consecutive years participated in the study. The findings indicated that resting 3 minutes between sets was sufficient to allow for completion of 10 repetitions on each set. However, resting 1 minute between sets resulted in a significant decrease in the total repetitions completed. Conversely, Richmond and Godard[21] found that for 12RM loads, 3- and 5-minute rest intervals were not sufficient to maintain repetitions over two consecutive sets. In this experiment, 3and 5-minute rest intervals allowed for completion of approximately 8 and 10 repetitions on the second set, respectively. Such results were in conflict with the results reported by Kraemer,[20] in which 3 minutes allowed for consistent repetitions. This discrepancy between studies might be due to the samples utilized; for example, highly trained athletes in the study by Kraemer[20] versus healthy recreationally trained men in the study by Richmond and Godard.[21] Therefore, practitioners must consider the training status of individuals when prescribing the rest interval between sets. For untrained individuals, 5 minutes’ rest between sets might be necessary if the goal is consistency in repetitions over high-intensity sets. Corroborating with Richmond and Godard,[21] Willardson and Burkett[22,23] demonstrated that 3- and 5-minute rest intervals were not sufficient to maintain consistent repetitions in recreationally trained men who had performed approximately three strength workouts per week during the previous 3 years. For example, Willardson and Burkett[22] compared three different rest intervals on the number of repetitions completed Sports Med 2009; 39 (9)

De Salles et al.

768

Table I. Acute effect of different rest intervals between sets over the number of repetitions maximum (RM). Values are expressed as RM Study

Load

Exercises and intervals

Kraemer[20]

10RM

Bench press + leg press

Richmond and Godard[21]

Willardson and Burkett[22]

75% 1RM

8RM

Set 1

Set 2

1 min

10

8

3 min

10

10

Set 3

Set 4

Set 5

7.1 10

Bench press 1 min

11.9

3 min

11.5

6 8

5 min

11.5

10

Bench press 1 min

7.4

4.4

2.8

2.4

2 min

7.7

5.7

4.2

3.9

5 min

7.6

6.5

6.0

5.6

1 min

7.8

5.9

4.4

4.2

3 min

8.0

6.6

6.0

4.8

5 min

8.0

7.8

7.0

6.0

Squat

Willardson and Burkett[23]

80% 1RM 50% 1RM

Willardson and Burkett[24]

15RM

Bench press 1 min 80%

9.3

3.3

2.0

1.6

1.6

1 min 50%

29.8

10.0

7.0

6.1

6.0

2 min 80%

9.1

5.1

3.3

2.8

2.5

2 min 50%

29.9

14.8

11.1

9.7

9.1

3 min 80%

9.1

5.9

4.6

3.8

3.5

3 min 50%

30.4

18.2

14.0

12.6

12.2

Bench press 30 sec

14.9

4.9

2.4

1.8

1.5

1 min

14.6

5.9

3.6

3.3

2.8

2 min

14.6

8.6

5.6

5.3

4.9

Squat 30 sec

15.6

10.1

6.8

5.9

5.4

1 min

15.4

10.6

8.4

6.2

6.3

2 min

15.4

12.5

10.6

9.4

8.6

for the squat and bench press exercises. Three test sessions were conducted, during which four sets of the squat and bench press were performed with a constant 8RM load and 1, 2 or 5 minutes’ rest between sets. For each exercise, a significant decline occurred in the number of repetitions completed between the first and the fourth sets; however, for a given rest interval there were greater total repetitions performed for the squat versus the bench press. This finding suggests that the muscles of the lower body possessed greater endurance characteristics versus the ª 2009 Adis Data Information BV. All rights reserved.

muscles of the upper body. These results indicated that the specific combination of muscles involved affects the prescription of the rest interval; therefore, practitioners may prescribe longer rest intervals for compound upper body exercises (e.g. bench press) and shorter rest intervals for compound lower body exercises (e.g. barbell back squat). Willardson and Burkett[24] also compared 30-second and 1- and 2-minute rest intervals on the number of repetitions completed for the squat and bench press over five sets with a constant Sports Med 2009; 39 (9)

Rest between Sets

15RM load. Significant declines in the number of repetitions occurred between the first and the fifth sets, irrespective of the rest interval. For both exercises, the 2-minute rest interval resulted in significantly greater repetitions versus the 30-second rest interval. These results suggest that when short rest intervals are utilized to develop muscular endurance, the intensity (i.e. load) may need to be progressively lowered over subsequent sets to sustain repetitions within the range conducive to this training goal. Future research should establish the extent to which the resistance should be lowered between sets in order to sustain repetitions for muscular endurance training. A limitation of the aforementioned studies was the evaluation of only one or two exercises; only one study to date has examined the influence of different rest intervals in the context of a typical training session that involves multiple exercises. Miranda et al.[25] compared the effects of 1- versus 3-minute rest intervals between sets on the number of repetitions completed for each exercise during an upper body workout. Fourteen recreationally trained men with a minimum of 2 years and a mean of 6.34 years of experience performed two training sessions; both sessions consisted of three sets with 8RM loads for six upper extremity exercises (e.g. wide grip lat pull-down, close-up grip pull-down, machine seated row, barbell row lying onto bench, dumbbell seated arm curl and machine seated arm curl). Miranda et al.[25] demonstrated that for all exercises significantly fewer repetitions were completed when resting 1 minute between sets. However, significant reductions were noted between the first and third sets in four of the six exercises, irrespective of the rest interval. These results suggest that when training for muscular strength, resting for ‡3 minutes might be advantageous to accumulate a higher training volume while also maintaining the intensity of the load lifted. It is important to emphasize that for any intensity or objective in strength training, the rest interval length may vary when sets are not performed to concentric failure, according to the exercises used and/or practitioner’s level of conditioning. Additionally, recent studies demonstrated that pracª 2009 Adis Data Information BV. All rights reserved.

769

titioners should also consider age and sex when prescribing rest intervals between sets.[26,27] 2.2 Influence of Rest Interval Length on Repeated Maximal Strength Assessments

The rest interval between sets when testing maximal strength is crucial because the reliability of testing may depend on the ability to recover. Contrary to what has been practiced in most experiments (i.e. utilizing longer rest intervals between sets when testing for maximal strength), some studies have demonstrated that 1-minute rest intervals were sufficient between repeated 1RM attempts.[28,29] Weir et al.[28] examined the effect of rest interval length on repeated 1RM bench press performance. On the first day of testing, the 1RM load was determined for 16 recreationally trained men with a minimum of 2 years of bench press experience who could bench press at least 125% of their bodyweight. Four subsequent test sessions involved the performance of two 1RM attempts, separated by 1, 3, 5 or 10 minutes’ rest. The results demonstrated that all of the rest intervals tested allowed for approximately the same number of subjects to successfully complete the second attempt. Similarly, Matuszak et al.[29] compared the effects of different rest intervals in 17 recreationally trained men with a minimum of 2 years of squat experience who could back squat at least 1.5 times their bodyweight. Three subsequent test sessions involved the performance of two 1RM attempts, separated by 1, 3 or 5 minutes’ rest. The results demonstrated that for the 1-minute rest condition, 75% of the subjects completed the second attempt successfully, with 94.1% and 88.2% successful completions for the 3- and 5-minute rest conditions, respectively. The results of these studies indicate that in some cases 1-minute rest intervals might be sufficient; however, from a psychological and physiological standpoint, the inclusion of 3- to 5-minute rest intervals might be safer and more reliable. In short, testing for maximal strength is a process that should not be rushed, especially when testing compound exercises like the back squat and bench press, which require high levels of neuromuscular coordination. Sports Med 2009; 39 (9)

De Salles et al.

770

2.3 Influence of Rest Interval Length on Acute Expression of Muscular Power

Power performance is highly dependent on anaerobic energy metabolism (primarily the phosphagen system). The rest interval between sets when training for muscular power should closely match the time required for replenishment of phosphocreatine (PCr), which requires a minimum of 4 minutes’ recovery.[30] If the rest interval is not sufficient to allow for replenishment of PCr, energy production shifts to emphasize the glycolytic system. This results in the accumulation of H+ ions and disturbances in the concentration gradients of other ions (i.e. Na+, K+, Ca2+ Mg2+, Cl-), resulting in a lowered intracellular pH. At low pH values, both the peak isometric force and the maximal velocity of shortening are substantially depressed.[7] Abdessemed et al.[31] examined the effect of recovery duration on muscular power and blood lactate concentration during the performance of ten sets of six maximal effort bench press repetitions performed at 70% of 1RM and with 1, 3 or 5 minutes’ rest between sets in ten untrained men. Measurements of force and displacement of the bar and mean power during each repetition were calculated. Blood lactate was evaluated before and immediately after each interval. No significant variation in mean power occurred between the first and the tenth sets when resting 3 or 5 minutes betweens sets; lactate did not increase significantly from baseline with either of these rest conditions. Conversely, the 1-minute rest condition resulted in a significant decrease in mean power and a significant elevation in blood lactate. These results suggest that the 1-minute rest condition was not sufficient to replenish PCr, which placed greater emphasis on glycolytic energy production, as demonstrated by the significant elevation in blood lactate.[31] Therefore, resting 3–5 minutes between sets may allow for maintenance of force and power production over multiple sets and repetitions. However, longitudinal research is necessary to determine whether greater acute power during individual workouts would translate into performance gains ª 2009 Adis Data Information BV. All rights reserved.

in activities that require high power output, such as the vertical jump. 2.4 Influence of Rest Interval Length on Acute Hormonal Responses and their Influence on Muscular Hypertrophy

The maintenance of training intensity is not the main focus in strength training directed toward muscular hypertrophy, and research suggests that successive sets should be performed prior to full recovery. Several cross-sectional studies suggest that short periods (£1 minute) might provide a superior stimulus for hypertrophy due to the acute elevations in growth hormone.[3-5,10,31-32] Kraemer et al.[4] compared the acute hormonal responses to different resistance exercise protocols. Nine recreational trained young men (not competitive lifters) performed a protocol that involved three sets of eight exercises with a 10RM load and a 1-minute interval between sets, and another protocol that involved five sets of five exercises with a 5RM load and a 3-minute rest interval between sets. Blood hormonal concentrations of total testosterone, free testosterone, cortisol, growth hormone and blood lactate were collected prior to the exercise session and at 0, 5, 15, 30, 60, 90 and 120 minutes following the session. The results indicated that acute elevations in growth hormone were significantly greater for the protocol that involved 1-minute rest intervals and 10RM loads. However, a limitation of this study was that changes in muscular hypertrophy were not examined over time. The studies by Kraemer et al.[4] in young men and women[5] indicated that shorter rest intervals (i.e. 1 vs 3 minutes) were associated with greater acute elevations in growth hormone. However, they also observed higher values for corticotropin and cortisol, which have antagonistic effects to growth hormone in terms of the catabolic effects on skeletal muscle. Therefore, the acute elevations in growth hormone may not reflect the long-term potential for muscular hypertrophy. Goto et al.[10] conducted a study that examined both acute responses and chronic adaptations to hypertrophy- and strength-oriented programmes. All the subjects were recreationally Sports Med 2009; 39 (9)

Rest between Sets

771

trained, but they had not participated in a regular training programme for at least 6 months prior to commencement of the study. Acute elevations in growth hormone were measured in response to three leg extension workouts, which included: (i) moderate intensity (nine sets of approximately 10RM, with 30-second rest intervals and intensity reduction as the sets progressed); (ii) high intensity (five sets at 90% of 1RM and 3-minute rest intervals); or (iii) combined (high-intensity plus one low-intensity set after 30 seconds). Acute elevations in growth hormone post-exercise were significantly different between protocols, in the following order: moderate intensity > combined > high intensity. In addition, after 4 weeks of training, the combined programme demonstrated significantly larger increases versus the high-intensity programme in quadriceps cross-sectional area, 1RM leg press, maximal isokinetic strength and muscular endurance for the leg extension. Conversely, Ahtiainen et al.[11] indicated that hormonal responses and hypertrophic adaptations did not vary with 2- or 5-minute rest intervals in 13 recreationally trained men (with an experience of 6.6 – 2.8 years of continuous strength training). This experiment involved a crossover design so that two groups trained for 3 months with each rest condition. The maximal strength of the leg

extensors and the quadriceps cross-sectional area were assessed before and after completion of each condition. Other variables that were assessed included electromyographic activity of leg extensor muscles, concentrations of total testosterone, free testosterone, cortisol, growth hormone and blood lactate. The results demonstrated that, for both conditions, acute responses and chronic adaptations were similar in terms of the hormonal concentrations, strength development and increases in quadriceps cross-sectional area. A key finding by Ahtiainen et al.[11] was that the 5-minute rest interval allowed for the maintenance of a higher training intensity (approximately 15% higher); however, the volume of training was equalized so that the 2-minute condition required more sets at a lower intensity, while the 5-minute condition required less sets at a higher intensity. Thus, the strength and hormonal responses appeared to be somewhat independent of training intensity as long as an equal volume was performed. Buresh et al.[33] also compared the chronic effects of different interset rest intervals after 10 weeks of strength training. Twelve untrained males were assigned to strength training programmes using either 1- or 2.5-minute rest between sets, with a load that elicited failure only on the third set of each exercise. Measures of body composition,

Table II. Acute effects of different rest intervals on serum levels of growth hormone (GH) Study

Training

Intervals

Measurements

Results

Kraemer et al.[4]

Session 1: three sets of 10RM with 1-min interval in eight exercises; session 2: five sets of 5RM with 3-min interval in five exercises

1 and 3 min

Serum GH before the session and at 0, 5, 15, 30, 60, 90 and 120 min after

GH concentrations were significantly higher for the protocol using 1-min intervals for all the measurements used

Goto et al.[10]

Session 1: nine sets of 10RM with 30-sec interval; session 2: five sets at 90% of 1RM and 3-min intervals; session 3: five sets at 90% of 1RM and 3-min intervals + one set of 10RM after 30 sec

30 sec and 3 min

Serum GH before the session and at 5, 15, 30 and 60 min after

The increased GH levels post-exercise were shown to be significantly dependent on the protocols in the following order: session 1 > session 3 > session 2

Ahtiainen et al.[11]

Five sets in leg press at approximately 10RM

2 and 5 min

Serum GH before and immediately after

Both protocols resulted in acute increases in serum GH concentrations, without any difference between intervals

Bottaro et al.[12]

Three sets of 10RM in four exercises for lower limbs

30, 60 and 120 sec

Serum GH before the session and at 0, 5, 15 and 30 min after

All protocols led to acute increases in GH concentrations after each training session, while GH concentration was higher for 30 sec compared with other intervals

RM = repetition maximum.

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (9)

772

hormone response, thigh and arm indirectly cross-sectional area, and 5RM loads on squat and bench press were assessed before and after a 10-week programme. The results showed that 10 weeks of both strength training programmes resulted in similar significant increases in 5RM squat and bench press strength, thigh and arm cross-sectional area, and lean mass. However, 1-minute rest elicits a greater hormonal response than 2.5-minute rest intervals in the first training weeks, but these differences disappear after 10 weeks of training. These results also suggest that hormonal response may not necessarily be predictive of hypertrophic gains after a 10-week training programme performed by untrained healthy males. Recently, Bottaro et al.[12] examined the acute hormonal responses to three different rest intervals between sets for strength training sessions that involved lower body exercises. Twelve recreationally trained women completed three sessions of strength training with either 30, 60 or 120 seconds between sets. The sessions consisted of three sets of four exercises (e.g. knee extension, hack squat, knee flexion and leg press), performed to concentric failure with 10RM loads. Growth hormone and cortisol concentrations were measured prior to exercise, immediately following each session and 5, 15 and 30 minutes following the session. Following the sessions, significantly greater elevations of growth hormone were demonstrated for the 30-second rest condition.[12] Contrary to previous studies, cortisol was not significantly different between rest conditions.[4,5] Thus, the combination of moderate-intensity sets with very short rest intervals seemed to be most effective for acute elevations in growth hormone (see table II). However, more research is needed examining the hypertrophic adaptations consequent to such training prescriptions. One must take into consideration that the number of motor units increases with the increasing load. Although a 10RM load appears to be ideal, this represents a relatively low intensity, and in several studies the load was lowered progressively over consecutive sets. In such cases, there may not be adequate stimulation to higher ª 2009 Adis Data Information BV. All rights reserved.

De Salles et al.

threshold muscle fibres capable of the greatest increases in hypertrophy. There might be higher acute elevations in growth hormone with 30 seconds’ to 1 minute’s rest between sets, but this represents one variant that may or may not be associated with long-term increases in muscular hypertrophy. Other anabolic hormones such as testosterone and insulin-like growth factor-I are not elevated with short rest intervals between sets, possibly due to the influence of acidity and heat shock proteins. Therefore, there is still much research to be done examining how the rest interval should be structured to promote muscular hypertrophy on a long-term basis. 3. Chronic Adaptations and the Rest Interval between Sets 3.1 Influence of Rest Interval Length on Muscular Strength and Power Adaptations

For programmes targeting absolute strength and power development, the American College of Sports Medicine[1,2] recommended 2- to 3-minute rest intervals between sets when performing multi-joint exercises and 1- to 2-minute rest intervals between sets for single-joint exercises. Longer rest intervals may allow for maximal voluntary activation of motor units and maintenance of training intensity. These recommendations were validated by Pincivero et al.[8] for isokinetic-type training. Fifteen untrained men were divided into group 1 (40 seconds) and group 2 (160 seconds). One leg of each subject was assigned to a 4-week, 3-days-per-week isokinetic protocol that involved concentric knee extension and flexion muscle actions performed at 90/sec. Changes in quadricep and hamstring function were evaluated with five repetitions performed at 60/sec and 30 repetitions performed at 180/sec. The 160-second rest group demonstrated significantly greater increases in peak torque, average power and total work at 180/sec. Robinson et al.[6] demonstrated findings that were consistent with Pincivero et al.[8] for free weight training. In this study, the effects of three different intervals (3 minutes, 90 seconds and Sports Med 2009; 39 (9)

Rest between Sets

773

Table III. Chronic effects of different rest intervals on strength gains and muscular power Study

Duration (wk)

Training

Intervals (sec)

Measurements

Results

Pincivero et al.[8]

4

Isokinetic training of the knee at 90/sec, 3· wk

Group 1: 40; group 2: 160

Isokinetic strength, peak torque, power and total work at 60/sec with 5 rep and also at 180/sec with 30 rep

Group 2 showed greater peak torque, maximum power and total work than their contralateral limb and group 1

Pincivero and Campy[9]

6

Isokinetic training of the knee at 180/sec, 2· wk

Group 1: 40; group 2: 160

Isokinetic strength, torque peak, power and total work of the quadriceps at 180/sec in 30 contractions

Group 2 showed greater strength and peak torque; no changes were observed in group 1 and control

Robinson et al.[6]

5

Isotonic training of lower limbs, 4· wk

Group 1: 180; group 2: 90; group 3: 30

Vertical jump power and 1RM in squat

Group 1 showed greater 1RM values, while no influence was observed in power

rep = repetitions; RM = repetition maximum.

30 seconds) were compared on vertical jump power and maximum strength. Thirty-three moderately trained college-age men performed a free weight training programme 4 days per week for 5 weeks. The group that rested 3 minutes between sets demonstrated significantly greater increases in maximal squat strength versus the 90-second and 30-second rest groups; however, none of the groups demonstrated significant improvements in vertical jump power. Willardson and Burkett[15] compared squat strength gains and volume components with 2 minutes’ versus 4 minutes’ rest between sets over 13 weeks. After the first squat 1RM assessment, 15 recreationally trained men were divided into group 1 (2 minutes) and group 2 (4 minutes). Each group performed the same training programme, with the only difference being the length of the rest interval between sets. Subjects performed two squat workouts per week. The squat workouts varied in the intensity, number of sets and repetitions performed per set in a nonlinear periodized manner. Differences in strength gains and volume components (load utilized per set, repetitions performed per set, intensity per set and volume performed per workout) were compared between groups. The key finding was that during the entire training period, group 2 (4 minutes) demonstrated significantly higher total volumes during the high-intensity workouts.[15] However, the ª 2009 Adis Data Information BV. All rights reserved.

groups were not significantly different in squat strength gains. These findings suggest that there was a threshold in terms of the volume necessary to gain a certain amount of strength. Resting 2 minutes between sets resulted in sufficient volume to achieve the same strength gains as resting 4 minutes between sets. Therefore, athletes attempting to achieve specific volume goals may need longer rest intervals initially but may later adapt, so that shorter rest intervals can be utilized without excessive fatigue, leaving additional time to focus on other conditioning priorities. The findings of these studies suggest that longer rest intervals (i.e. 2–3 minutes) result in significantly greater increases in strength compared with shorter rest intervals (i.e. 30–90 seconds) [table III]. Longer rest intervals allow for higher intensities and volumes of training. Furthermore, the evidence also indicates that excessively long rest intervals (i.e. 4 minutes) are not necessary, and may detract from other conditioning priorities.[6,8,15] However, the few longitudinal studies conducted on different rest interval lengths have focused solely on lower body strength and did not examine the full spectrum of rest intervals (i.e. from short to long to very long) within the same study. Therefore, more research is necessary to ascertain the effects of different rest interval lengths on upper body exercises, and the influence of very dissimilar rest interval lengths (i.e. 1- vs 5-minute rest) on strength development. Sports Med 2009; 39 (9)

De Salles et al.

774

Table IV. Chronic effects of different rest intervals on muscular endurance Study

Duration (wk)

Training

Intervals

Measurements

Results

Garcı´a-Lo´pez et al.[41]

5

Three sets to concentric failure at 60–75% of MVC in the machine seated arm curl exercise, 2· wk

Group 1: 1 min; group 2: 4 min

Number of RM and average velocity of performance in moderate intensity (60% of MVC) before and after the training period

The average velocity presented by group 1 in post-training was higher than the corresponding average velocity in pre-training conditions, while no significant difference was observed in group 2 and in the control group

Hill-Haas et al.[14]

5

Two to five sets of 15–20RM in 11 whole body exercises, 2· wk

Group 1: 20 sec; group 2: 80 sec

Strength for 3RM in leg press, total torque in cycle ergometer during five maximum sprints, with 6 sec duration and 30 sec intervals, before and after each training period

Strength for 3RM was greater in group 2 (45.9%) compared with group 1 (19.6%); total torque during the test was higher in group 1 (12.5%) compared with group 2 (5.4%)

EMG = electromyography activity; MVC = maximal voluntary contraction; RM = repetition maximum.

3.2 Influence of Rest Interval Length on Muscular Endurance Adaptations

Local muscular endurance can be defined as the capacity to sustain submaximal muscle actions for an extended period of time. In strength training, where the goal is the development of muscular endurance, the typical recommendation has been to utilize relatively low-intensity resistance combined with high repetitions and very short rest intervals between sets.[34-36] As a result of the use of low intensities, the prescription of short rest intervals has been theorized to allow for sufficient recovery betweens sets; however, the ability to recover may depend on whether sets are being performed to the point of voluntary exhaustion.[24,37-40] Muscular endurance training has been thought to stimulate increases in mitochondria and capillary density, allowing for submaximal muscle actions to continue because of greater reliance on oxidative metabolism.[3,34,36] Ratamess et al.[13] examined the effects of different rest intervals on the intensity, volume and metabolic responses to the bench press exercise. Eight trained men (minimum 3 years of experience with the bench press exercise) performed ten randomized protocols (five bench press sets at 75% or 85% of 1RM for ten repetitions and five repetitions, respectively, using different intervals between sets [30 seconds, 1, 2, 3, 5 minutes]). The oxygen consumption was measured during exercise ª 2009 Adis Data Information BV. All rights reserved.

and for 30 minutes thereafter. For the 30-second and 1-minute rest intervals, 15–55% reductions in intensity and volume were observed (sets 5 < 4 < 3 < 2 < 1). For the 2-minute rest interval, the performance was maintained during the first two sets, but declined 8–29% during the third, fourth and fifth sets. For the 3-minute rest interval, a volume reduction was noted for the fourth and fifth sets (approximately 21% lower than the first, second and third sets). At 5 minutes, a reduction was observed only for the fifth set. Overall, the greatest reductions in performance occurred with very short rest intervals (<1 minute) and performance was maintained during the first 3–4 sets when 3- to 5-minute rest intervals were utilized. The mean oxygen consumption and ventilation progressively increased as the rest interval decreased. As could be expected, the mean oxygen consumption was higher when ten repetitions versus five repetitions were performed, irrespective of the intensity level. Following each bench press workout, oxygen consumption, ventilation and respiratory exchange ratio were still elevated at 30 minutes. These data demonstrate that rest interval length of £1 minute leads to a more continuous elevation in oxygen consumption, which could have potential ramifications for training programmes targeting muscular endurance or aerobic fitness. Garcı´ a-Lo´pez et al.[41] examined the effects of different rest intervals over 5 weeks on muscular endurance performance and mean repetition Sports Med 2009; 39 (9)

Rest between Sets

velocity during a moderate-intensity set of elbow flexion contractions (60% maximal voluntary contraction [MVC]). Twenty-one untrained subjects were divided into three groups: group 1 (1 minute), group 2 (4 minutes) and a control group. Groups 1 and 2 performed three sets to concentric failure in the seated arm-curl machine, 2 days a week, for 5 weeks, with moderate loads (60–75% MVC). Group 2 demonstrated a significantly higher total training volume than group 1 following the intervention. However, both training groups demonstrated similar gains in muscular endurance performance, with an increased repetitions performance with the pre-training 60% MVC. The mean repetition velocity demonstrated by group 1 following training at 40%, 50%, 60%, 70%, 80% and 90% of the total number of repetitions completed was significantly higher than the corresponding mean repetition velocity prior to training, while no significant difference was observed for group 2 or in the control group. Coaches can apply this finding when training athletes for sports such as soccer and basketball, which require repeated submaximal muscle actions while maintaining high velocities of movement. Hill-Haas et al.[14] also examined muscular endurance-related performance adaptations resulting from different rest intervals during strength training. Eighteen active women, who had not undertaken strength training for a least 1 year, were divided into group 1 (20 seconds) and group 2 (80 seconds). The intensity and volume were equated between groups; each group trained 3 days per week for 5 weeks, performing 2–5 sets of 15–20RM for 11 exercises encompassing the entire body. Subjects in each group were evaluated for 3RM leg press strength, total torque in cycle ergometer during the performance of five maximum sprints of 6 seconds duration and 30-second rest intervals between sprints, and anthropometric measures. The results demonstrated that the percentage increase in 3RM leg press strength was significantly higher in group 2 (80 seconds; 45.9%) versus group 1 (20 seconds; 19.6%). Conversely, the total torque during the cycle ergometer test was significantly higher in group 1 (20 seconds; 12.5%) versus group 2 (80 seconds; 5.4%); no change was observed in skinfold (% fat) and circumference measures. ª 2009 Adis Data Information BV. All rights reserved.

775

These results show that despite a smaller increase in strength, extremely short intervals may allow for greater maintenance of relative force levels. This study demonstrates that high repetition can improve repeated-sprint ability in untrained recreationally active team-sport players, and that this improvement is greater when there is a shorter rest interval between sets, with training load and volume matched. In summary, shorter intervals between sets (i.e. <1 minute) benefit performance variables related to localized muscular endurance (table IV) and aerobic endurance development due to greater oxygen consumption.[13] However, the utilization of extremely short rest intervals may not allow for the maintenance of repetitions per set, even when utilizing relatively low-intensity loads. Therefore, when utilizing extremely short rest intervals (i.e. <1 minute), the load should be reduced as needed over consecutive sets to maintain repetitions within the range conducive to this training goal.[24] 4. Conclusions When training for muscular strength with loads smaller than 90% of 1RM (up to 50%) for multiple sets, 3- to 5-minute rest intervals are necessary to maintain the number of repetitions performed per set within the prescribed zone without great reductions in training intensity. On the other hand, contrary to what was observed in most of the experiments concerning muscular strength, some evidence suggests that 1-minute intervals allowed for sufficient recovery during repeated 1RM attempts; however, from a psychological and physiological standpoint, the inclusion of 3- to 5-minute rest intervals might be safer and more reliable. The acute expression of muscular power was best maintained when including 3- or 5-minute rest intervals versus 1-minute rest intervals between sets. When the training goal is muscular hypertrophy, the combination of moderate-intensity sets with short intervals of 30–60 seconds might be the best alternative, due to higher acute increases of growth hormone, which can contribute to the hypertrophic effect. Finally, similar to hypertrophy training, extremely short intervals (e.g. 20 seconds to a minute) between sets allowed for greater Sports Med 2009; 39 (9)

De Salles et al.

776

muscular endurance development. In the case of muscular endurance training, a progressive reduction of training intensity during the performance of consecutive sets may be a way to maintain the number of repetitions within the prescribed zone. It is worth noting that, for any intensity or goal in strength training, the rest interval between sets may vary between practitioners of different ages, when the sets are not performed to concentric failure, according to the exercise and/or according to the athlete’s level of conditioning.

5. Recommendations There is a scarcity of studies concerning chronic adaptations and rest interval length. This review should provide some direction for future studies investigating aspects related to strength, power, hypertrophy and muscular endurance development. Additional investigations concerning acute responses are also necessary, involving women or individuals with different levels of physical conditioning, and potential interactions between the order of exercises and rest interval length. Furthermore, comparison between different rest intervals for exercises that involve relatively large versus small muscle groups would be useful from a practical standpoint. Overall, there is still much research to be done on this topic. Acknowledgements Dr Roberto Sima˜o would like to thank the Brazilian National Board for Scientific and Technological Development (CNPq) and Research and Development Foundation of Rio de Janeiro State (FAPERJ) for the research grant support. The authors have no conflicts of interest that are directly relevant to the content of this review.

References 1. American College of Sports Medicine. Position stand: progression models in resistance training for healthy adults. Med Sci Sports Exerc 2002; 34: 364-80 2. American College of Sports Medicine. American College of Sports Medicine Position Stand. Appropriate physical activity intervention strategies for weight loss and prevention of weight regain for adults. Med Sci Sports Exerc 2009; 41: 459-71

ª 2009 Adis Data Information BV. All rights reserved.

3. Kraemer WJ, Noble BJ, Clark MJ, et al. Physiologic responses to heavy-resistance exercise with very short rest periods. Int J Sports Med 1987; 8: 247-52 4. Kraemer WJ, Marchitelli L, Gordon SE, et al. Hormonal and growth factor responses to heavy resistance exercise protocols. J Appl Physiol 1990; 69: 1442-50 5. Kraemer WJ, Fleck SJ, Dziados JE, et al. Changes in hormonal concentrations after different heavy resistance exercise protocols in women. J Appl Physiol 1993; 75: 594-604 6. Robinson JM, Stone MH, Johnson RL, et al. Effects of different weight training exercise/rest intervals on strength, power, and high intensity exercise endurance. J Strength Cond Res 1995; 9: 216-21 7. Larson GD, Potteiger JA. A comparison of three different rest intervals between multiple squat bouts. J Strength Cond Res 1997; 11: 115-8 8. Pincivero DM, Lephart SM, Karunakara RG. Effects of rest interval on isokinetic strength and functional performance after short-term high intensity training. Br J Sports Med 1997; 31: 229-34 9. Pincivero DM, Campy RM. The effects of rest interval length and training on quadriceps femoris muscle, part I: knee extensor torque and muscle fatigue. J Sports Med Phys Fitness 2004; 44: 111-8 10. Goto K, Nagasawa M, Yanagisawa O, et al. Muscular adaptations to combinations of high and low intensity resistance exercises. J Strength Cond Res 2004; 18: 730-7 11. Ahtiainen JP, Pakarinen A, Alen M, et al. Short vs long rest period between the sets in hypertrophic resistance training: influence on muscle strength, size, and hormonal adaptations in trained men. J Strength Cond Res 2005; 19: 572-82 12. Bottaro M, Martins B, Gentil P, et al. Effects of rest duration between sets of resistance training on acute hormonal responses in trained women. J Sci Med Sport 2009; 12: 73-8 13. Ratamess NA, Falvo MJ, Mangine GT, et al. The effect rest interval length on metabolic responses to the bench press exercise. Eur J Appl Physiol 2007; 100: 1-17 14. Hill-Haas S, Bishop D, Dawson B, et al. Effects of rest interval during high-repetition resistance training on strength, aerobic fitness, and repeated-sprint ability. J Sports Sci 2007; 25: 619-28 15. Willardson JM, Burkett LN. The effect of different rest intervals between sets on volume components and strength gains. J Strength Cond Res 2008; 22: 146-52 16. Willardson JM. A brief review: factors affecting the length of the rest interval between resistance exercise sets. J Strength Cond Res 2006; 20: 978-84 17. Hopkins WG. Olympian impact factors: top journals in exercise and sports science and medicine for 2008. Eur Sportsci 2008; 12: 22-4 18. Rhea MR, Alvar BA, Burkett LN. Single versus multiple sets for strength: a meta-analysis to address the controversy. Res Q Exerc Sport 2002; 73: 485-8 19. Peterson MD, Rhea MR, Alvar BA. Applications of the dose-response for muscular strength development: a review of meta-analytic efficacy and reliability for designing training prescription. J Strength Cond Res 2005; 19: 950-8

Sports Med 2009; 39 (9)

Rest between Sets

20. Kraemer WJ. A series of studies: the physiological basis for strength training in American football: fact over philosophy. J Strength Cond Res 1997; 11: 131-42 21. Richmond SR, Godard MP. The effects of varied rest periods between sets of failure using bench press in recreationally trained men. J Strength Cond Res 2004; 18: 846-9 22. Willardson JM, Burkett LN. A comparison of 3 different rest intervals on the exercise volume completed during a workout. J Strength Cond Res 2005; 19: 23-6 23. Willardson JM, Burkett LN. The effect of rest interval length on bench press performance with heavy vs light load. J Strength Cond Res 2006; 20: 396-9 24. Willardson JM, Burkett LN. The effect of rest interval length on the sustainability of squat and bench press repetitions. J Strength Cond Res 2006; 20: 400-3 25. Miranda H, Fleck SJ, Sima˜o R, et al. Effect of two different rest period lengths on the number of repetitions performed during resistance training. J Strength Cond Res 2007; 21: 1032-6 26. Theou O, Gareth JR, Brown LE. Effect of rest interval on strength recovery in young and old women. J Strength Cond Res 2008; 22: 1876-81 27. Faigenbaum AD, Ratamess NA, McFarland J, et al. Effect of rest interval length on bench press performance in boys, teens, and men. Pediatr Exerc Sci 2008; 20: 457-69 28. Weir JP, Wagner LL, Housh TJ. The effect of rest interval length on repeated maximal bench presses. J Strength Cond Res 1994; 8: 58-60 29. Matuszak ME, Fry AC, Weiss LW, et al. Effect of rest interval length on repeated 1 repetition maximum back squats. J Strength Cond Res 2003; 17: 634-7 30. Harris RC, Edwards RH, Hultman E, et al. The time course of phosphorylcreatine resynthesis during recovery of the quadriceps muscle in man. Pflugers Arch 1976; 28: 137-42 31. Abdessemed D, Duche P, Hautier C, et al. Effect of recovery duration on muscular power and blood lactate during the bench press exercise. Int J Sports Med 1999; 20: 368-73 32. Mccall GE, Byrnes WC, Fleck SJ, et al. Acute and chronic hormonal responses to resistance training designed to

ª 2009 Adis Data Information BV. All rights reserved.

777

33.

34.

35.

36.

37.

38.

39.

40.

41.

promote muscle hypertrophy. Can J Appl Physiol 1999; 24: 96-107 Buresh R, Berg K, French J. The effect of resistive exercise rest interval on hormonal response, strength, and hypertrophy with training. J Strength Cond Res 2009; 23: 62-71 Anderson T, Kearney JT. Effects of three resistance training programs on muscular strength and absolute and relative endurance. Res Q Exerc Sport 1982; 53: 1-7 Campos GER, Luecke TJ, Wendeln HK, et al. Muscular adaptations in response to three different resistancetraining regimens: specificity of repetition maximum training zones. Eur J Appl Physiol 2002; 88: 50-60 Stone WJ, Coulter SP. Strength/endurance effects from three resistance training protocols with women. J Strength Cond Res 1994; 8: 231-4 Stull GA, Clarke DH. Patterns of recovery following isometric and isotonic strength decrement. Med Sci Sports Exerc 1971; 3: 135-9 Yates JW, Kearney JT, Noland MP, et al. Recovery of dynamic muscular endurance. Eur J Appl Physiol 1987; 56: 662-7 Sahlin K, Ren JM. Relationship of contraction capacity to metabolic changes during recovery from a fatiguing contraction. J Appl Physiol 1989; 67: 648-54 Bilcheck HM, Kraemer WJ, Maresh CM, et al. The effects of isokinetic fatigue on recovery of maximal isokinetic concentric and eccentric strength in women. J Strength Cond Res 1993; 7: 43-50 Garcı´ a-Lo´pez D, De Paz JA, Moneo E, et al. Effects of short vs long rest period between sets on elbow-flexor muscular endurance during resistance training to failure. J Strength Cond Res 2007; 21: 1320-4

Correspondence: Roberto Sima˜o, PhD, School of Physical Education and Sports, Rio de Janeiro Federal University, Av. Carlos Chagas Filho, Cidade Universita´ria, Rio de Janeiro 21941-590, Brazil. E-mail: [email protected]

Sports Med 2009; 39 (9)

Sports Med 2009; 39 (9): 779-795 0112-1642/09/0009-0779/$49.95/0

REVIEW ARTICLE

ª 2009 Adis Data Information BV. All rights reserved.

The Quantification of Training Load, the Training Response and the Effect on Performance Jill Borresen and Michael Ian Lambert MRC/UCT Research Unit for Exercise Science and Sports Medicine, Department of Human Biology, University of Cape Town, Cape Town, South Africa

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Quantifying Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Questionnaires and Diaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Physiological Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.1 Heart Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.2 Oxygen Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.3 Lactate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.4 Rating of Perceived Exertion (RPE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.5 Critical Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Direct Observation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4 Indices of Training Stress. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4.1 Training Impulse (TRIMP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4.2 Session RPE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4.3 Summated Heart Rate Zone Score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4.4 Lucia’s TRIMP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Modelling the Relationship between Training and Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Fitness and Fatigue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Physiological Correlates of Fitness and Fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Influence Curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Recursive Least Squares Algorithm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 Threshold Saturation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6 Limitations to Modelling the Training-Performance Relationship . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Variability in the Physiological Response to Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Variability in the Relationship between Training Adaptations and Performance . . . . . . . . . . . . . . . . . 5. Summary and Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

779 781 781 781 781 782 782 783 783 784 784 784 785 786 786 787 787 788 789 789 790 790 791 791 792

Historically, the ability of coaches to prescribe training to achieve optimal athletic performance can be attributed to many years of personal experience. A more modern approach is to adopt scientific methods in the development of optimal training programmes. However, there is not much research in this area, particularly into the quantification of training programmes and their effects on physiological adaptation and subsequent performance. Several methods have been used to quantify training load, including questionnaires, diaries, physiological monitoring and direct observation. More recently,

Borresen & Lambert

780

indices of training stress have been proposed, including the training impulse, which uses heart rate measurements and training load, and session rating of perceived exertion measurements, which utilizes subjective perception of effort scores and duration of exercise. Although physiological adaptations to training are well documented, their influence on performance has not been accurately quantified. To date, no single physiological marker has been identified that can measure the fitness and fatigue responses to exercise or accurately predict performance. Models attempting to quantify the relationship between training and performance have been proposed, many of which consider the athlete as a system in which the training load is the input and performance the system output. Although attractive in concept, the accuracy of these theoretical models has proven poor. A possible reason may be the absence of a measure of individuality in each athlete’s response to training. Thus, in the future more attention should be directed towards measurements that reflect individual capacity to respond or adapt to exercise training rather than an absolute measure of changes in physiological variables that occur with training.

The ultimate goal of any sports coach and athlete is to produce a winning or personal best performance at a specific time, preferably in competition. The prescription of the training required to achieve this goal has been largely instinctive, resulting from years of personal experience. As such, the ability to achieve peak fitness and performance coinciding with dates of competition is met with varying degrees of success. It is generally believed that increasing training will result in improvements in sporting performance and physical well-being. However, although widely accepted, this vague approach to prescribing training may be tenuous, especially because random increases in training volume, intensity or frequency may also increase the likelihood of injury and symptoms of overtraining.[1-5] The role of scientific research in this process is becoming more important in order to prescribe optimal training programmes that prevent both underand overtraining and increase the chance of achieving desired performances. The frequency, duration and intensity of exercise all contribute to the nature and magnitude of the training effect. However, relatively little research has been conducted into the quantification of training programmes and their effects on physiological adaptation and subsequent performance. This is surprising because peaking for sporting performance requires an understanding ª 2009 Adis Data Information BV. All rights reserved.

of the quantifiable effects of training on performance so that optimal training and rest regimens may be planned in preparation for the event. Optimizing training first involves quantifying what the athlete is currently doing. Several methods have been suggested to quantify exercise bouts, some of which will be reviewed below. Secondly, it needs to be established whether the athlete is adapting favourably to certain levels of exertion. Thereafter, training can be adjusted to optimize the athlete’s improvement to meet a specific goal within a specified time. This review therefore presents methods currently being used to quantify training load, and assesses literature investigating the relationship between training load and the physiological response to training and performance. Finally, the importance of considering the variability in individual responses to training will be highlighted in the assessment of the training and performance relationship. A PubMed search of the academic literature was performed using the following terms: ‘quantification of training load’, ‘quantify exercise intensity’, ‘modelling training and performance’, ‘endurance training adaptations’, ‘training impulse’, ‘session RPE’ and ‘physiological response to training’, limited to English papers and human subjects. Literature was also sourced from links to related articles, hand searches and the bibliographies of academic papers. The searches retrieved Sports Med 2009; 39 (9)

Training Quantification and Performance Models

2553 papers, of which 237 abstracts were reviewed and 114 full articles were evaluated. 1. Quantifying Training 1.1 Questionnaires and Diaries

Questionnaires and diaries obtain data recalled from the athlete (with diaries being completed more frequently [daily] than questionnaires), and are used to examine physical activity during the past week, month or even years.[6,7] The use of questionnaires to assess habitual physical activity and exercise, especially in large populations, is popular because their administration is easy, cost effective and does not impede training. However, their weakness is the fact that the athletes’ responses are subjective.[6] Borresen and Lambert[8] studied the relationship between what athletes say they do in training and what they actually do. Twenty-four percent of the participants overestimated the duration of training they were doing, and 17% underestimated their training duration. Because this margin of error in selfreported data may significantly affect the prescription of training, it was recommended that the error be accounted for or, where possible, physiological measurements be used to corroborate self-reported data.[8] The use of data collected by questionnaires to quantify exercise load is also limited by inadequate reliability and validity compared with laboratory measures.[9] For example, reliability decreases as the time between the activity and recall increases, because this is dependent on human memory.[6,9] A sports score derived from the Baecke questionnaire assesses the intensity of physical activity and has been used to estimate weekly training load.[10] The sports score is calculated using the duration (h/wk), frequency (mo/y) and intensity (dimensionless codes based on energy costs) of the activity.[10] There are, however, problems with questionnaires that assess the type, intensity, frequency and duration of the exercise and the environmental conditions in which the exercise was performed. For example, duration and frequency may be over-reported, especially if the person is influenced by the response he or she ª 2009 Adis Data Information BV. All rights reserved.

781

believes is sought by the investigators. Seasonal variations in duration and frequency of training may also not be taken into consideration. Perceptions of intensity may differ depending on experience or tolerance of the person, particularly if asked to report intensity as simply light, moderate, hard or very hard. The environmental conditions under which the activity is performed may have important motivational, psychological and physical effects on the person, but these are often overlooked. Responses can also be influenced by differences in human understanding, which may be the result of cultural factors or of the translation of the questionnaire. The length of the questionnaire and the detail required from the participant may also affect results, as the person may become bored or confused with exhaustive questioning.[6,9] Therefore, although questionnaires may assist with monitoring general changes in population activity, attempts to quantify exercise dosage from data collected with questionnaires remain inadequate.[9] 1.2 Physiological Measures 1.2.1 Heart Rate

Heart rate monitoring has become a popular method for measuring exercise intensity.[11] This method is based on the principle that there is a linear relationship between heart rate and steadystate work rate.[6,12,13] Although absolute measures of intensity are commonly used, the relative equivalent may be more informative because considerable inter- and intraindividual differences may exist in the way people respond to various modes of exercise. Percent maximum/ competition heart rate has been used to prescribe exercise intensity,[6] but Karvonen and Vuorimaa[14] suggest the use of percent heart rate reserve (equation 1) as a more accurate means of quantifying and prescribing intensity, because this method considers the fact that resting heart rate varies with age and fitness level, and maximal heart rate decreases with age. % heart rate reserve ¼

ðHRex  HRrest Þ  100 ðEq: 1Þ HRmax  HRrest

Sports Med 2009; 39 (9)

Borresen & Lambert

782

where HRex is average heart rate of the exercise session, HRrest is resting heart rate and HRmax is maximal heart rate. Although heart rate monitors have been found to measure heart rate accurately during physical activity, many factors may influence the relationship between work load and heart rate. The day-to-day variation in heart rate is approximately 6 beats/min,[15] or <6.5%.[16] However, if the factors affecting heart rate, such as state of training, environmental conditions, diurnal changes,[13] exercise duration, hydration status, altitude[11,13] and medication, are controlled, the accuracy with which heart rate can be used as a marker of exercise intensity improves.[15] 1.2.2 Oxygen Consumption

Because it is generally accepted that the. relationship between oxygen consumption. (VO2) and steady-state work rate is linear,[13] VO2 has been promoted as a valid measure of exercise intensity during steady-state exercise, but not interval, supramaximal exercise bouts. In a review of the oxygen kinetics during exercise, Xu and Rhodes[17] point out that when exercising . at a work rate below the lactate threshold, VO2 increases exponentially to a steady-state level, .but when exercising above the lactate threshold, VO2 kinetics become more complex. Relative . . (%VO2max) rather than absolute values of VO2 have been used to compare the exercise intensities executed by athletes of differing physiological [6] However, it and performance characteristics. . is exercise mode has been found that VO 2max . specific, so VO2max needs to be determined for each mode before exercise. can be prescribed or quantified using relative VO2 values.[18] Oxygen consumption reserve (VO2R, equation 2) has been suggested as a more accurate means with which . to prescribe exercise intensity than %VO2max: ðVO2 ex  VO2 rest Þ  100 %VO2 R ¼ ðEq: 2Þ VO2 max  VO2 rest It has been shown in cycling and running that whereas calculating exercise intensity using heart rate reserve and VO2R give similar results, . the exercise intensities calculated using %VO2max differ.[19,20] Baldwin et al.[21] found that heart ª 2009 Adis Data Information BV. All rights reserved.

rate and plasma markers of exercise stress such as lactate, ammonia and hypoxanthine at . 70%VO2peak were different between trained and untrained individuals. . This supports the suggestion that using %VO2peak does not necessarily produce the same physiological response in .different people. It has also been found that the VO2 kinetics at the onset of exercise may differ with level of physical . training, age and disease. As such, the use of VO2 may be inappropriate as a means with which to prescribe relative exercise intensity.[21,22] 1.2.3 Lactate

The measurement of blood lactate concentrations has become easier with the development of portable measurement instruments and requiring the collection of only one drop of blood from a finger prick. Nevertheless, it remains impractical to measure lactate frequently during every training session in order to prescribe or quantify intensity. Particular attention has been paid to determining the lactate threshold, which is defined as the exercise intensity at a fixed or maximal steady-state blood lactate level.[23] It has been proposed as a measure of endurance fitness, but also a means with which to standardize training intensity. The steady-state exercise intensity that elicits a lactate concentration of approximately 4 mmol/L has been suggested as the most favourable for inducing optimal physiological adaptations for endurance events.[24,25] However, Stegmann et al.[26] warn that this ‘optimal’ lactate level may range from 2 to 7.5 mmol/L among athletes. The inherent inter- and intraindividual differences in the extent to which lactate accumulates during exercise are two limitations of many in the use of lactate to prescribe exercise intensity. Extraneous factors such as ambient temperature and dehydration may influence the interpretation of lactate measurements. Mode of exercise can also be an influence, as it alters the muscle mass used during exercise[27] such that.the same lactate concentration occurs at different VO2 levels during running and cycling. Exercise duration, intensity and the rate of change in exercise intensity may also influence lactate concentration, as may prior Sports Med 2009; 39 (9)

Training Quantification and Performance Models

exercise, diet and muscle glycogen content.[24,27] Exercising with damaged muscles has also been shown to cause an increase in lactate levels.[27] Improvements in training status as well as overtraining have both been associated with decreases in maximal and submaximal blood lactate concentration,[28-30] which may lead to erroneous interpretations of lactate measurements and incorrect exercise prescription.[27] The interpretation of lactate concentration may further be affected by sampling and measurement procedures such as the time and site of blood sampling, measurement techniques and dilution volume.[24,27] The extent to which the abovementioned factors affect the way lactate accumulates, independent of exercise intensity, makes the importance of the lactate threshold less definitive,[26] thus limiting its usefulness in monitoring and prescribing training intensity. 1.2.4 Rating of Perceived Exertion (RPE)

A rating of perceived exertion (RPE) is based on the understanding that athletes can inherently monitor the physiological stress their bodies experience during exercise, and thus be able to adjust their training intensity using their own perceptions of effort.[13] This principle has been demonstrated during steady-state exercise[13] and high-intensity interval cycling[31] where the athletes’ reported RPE correlated well to average heart rate[13] and acute changes in heart rate.[31] However, poor correlations have also been found between heart rate and RPE responses during short-duration, high-intensity soccer drills[32] and during step dance sessions.[33] A meta-analysis of the literature concluded that although the Borg scale has been shown to be a valid measure of exercise intensity, the validity coefficients between the Borg 6–20 RPE scale and physiological criterion variables are not as high as previously thought.[34] For example, the weighted mean validity coefficients were 0.62. for heart rate, 0.57. for blood lactate, 0.64 for %VO2max, 0.63 for VO2, 0.61 for ventilation and 0.72 for respiration rate. Thus, further research is required to ascertain the physiological mechanisms behind our cognitive perception of effort, which may clarify exactly what RPE represents. ª 2009 Adis Data Information BV. All rights reserved.

783

1.2.5 Critical Power

Critical power is a theoretical concept that represents an estimation of the maximal power output that can be maintained at physiological steady state without fatigue.[35,36] The identification of such a marker would be useful for the prescription of training or the assessment of training response following an intervention.[37] However, if critical power has physiological meaning, then a steady-state physiological response during exercise at critical power would be required.[37] This hypothesis has been investigated. Brickley et al.[37] found that the work rate when exercising . at critical power was approximately 80%VO2max and that a physiological steady state was not reached, with oxygen uptake, blood lactate concentration and heart rate increasing over time. There was also considerable interindividual variability in the time that exercise could be sustained.[37] For subjects cycling at 20 W above the maximal lactate steady state (close to their critical [36] found that blood power), Pringle and Jones . . lactate concentration, VO2 and VE increased significantly over time. Although critical power was significantly greater than the maximal lactate steady state, the two variables were strongly correlated.[36] Brickley et al.[37] concluded that critical power does not represent a sustainable steadystate intensity and that a more appropriate definition for critical power would be ‘‘the highest, non-steady-state intensity that can be maintained for a period in excess of 20 min, but generally no longer than 40 min’’.[37] Dekerle et al.[35] found that critical power, calculated for a range of exhaustion times with subjects exercising at self-selected cadence on a cycle ergometer, was significantly higher than maximal lactate steady state, suggesting that these two variables represent different physiological phenomena. The authors concluded that the physiological significance of the intensity at critical power remains unknown and that further research is required to define the accurate physiological meaning of critical power.[35] Vanhatalo et al.[38] investigated whether critical power may be an indicator of the heavy/severe domain of exercise intensity. They found that critical power was not significantly different from power output Sports Med 2009; 39 (9)

Borresen & Lambert

784

towards the end of a 3-minute all-out cycling test. As such, the authors suggest that there may be potential for the use of this test in the determination of critical power in place of the conventional protocol of multiple exhaustive exercise tests.[38] 1.3 Direct Observation

Direct observation is usually carried out by a coach during the training session and may record training components such as exercise mode, duration and absolute/relative intensity.[6] Speed, for example, may be a useful measure of intensity in swimming or on a flat, measured, indoor running track. However, in other modes such as cycling, skiing and cross-country running, the influence of factors such as terrain, environmental conditions and equipment may alter the accuracy with which speed reflects intensity.[39] Direct observation may also include subjective measures such as the coach’s perception of whether or not an athlete is overtraining.[6] Perceptions by coaches and athletes of the same training have been studied by Foster et al.[40] They showed significant differences between the training that the coaches prescribed and the training the athletes actually did. This disassociation may have a significant impact on the effectiveness of the training the coach prescribes. The extent to which training can be quantified based on direct observations may therefore also be limited. Because this method requires the presence of an observer at every training session, which may be impractical or impossible, the amount of data able to be collected in order to monitor training accurately may be inadequate.[6] New technology using a global positioning system (GPS) offers innovative ways to track distance covered and speed during training.[41-43] The accuracy of these techniques has improved so that for distance mean errors of 0.04–0.7% have been found, and for position mean errors of 1.94–2.13 m have been found.[44] Schutz and Herren[45] found that the accuracy of speed prediction had a standard deviation of 0.08 km/h for walking and 0.11 km/h for running, yielding coefficients of variation of 1.38% and 0.82%, reª 2009 Adis Data Information BV. All rights reserved.

spectively.[45] Studies have shown that GPS can be used to quantify training load in horses.[46,47] Although to our knowledge there are no published studies using this technology to quantify training load in humans, there does appear to be potential for its use under certain circumstances. 1.4 Indices of Training Stress 1.4.1 Training Impulse (TRIMP)

Banister et al.[48] proposed a method of quantifying a training session into a unit ‘dose’ of physical effort. They suggested that a person’s heart rate response to exercise, along with the exercise duration, collectively called a training impulse (TRIMP), may be a plausible measure of physical effort, as it is based on the extent to which exercise raises heart rate between resting and maximal levels.[48,49] A TRIMP is calculated using training duration, maximal heart rate, resting heart rate and average heart rate during the exercise session (equation 3). TRIMP ðwðtÞÞ ¼ duration of training ðminÞ  DHR ratio  Y HRex  HRrest where DHR ratio ¼ HRmax  HRrest

ðEq: 3Þ where Y = 0.64e1.92x for males, Y = 0.86e1.67x for females, e = 2.712 and x = DHR ratio. Y is a weighting factor that emphasizes highintensity exercise and is also applied to the equation to avoid giving disproportionate importance to long-duration, low-intensity exercise compared with intense, short-duration activity.[48] The Y factor is based on the lactate profiles of trained men and women relative to increases in exercise intensity. The ability to quantify and reduce training to a single figure/factor, as is possible with this equation, is appealing in terms of its practical application. However, the use of this method of quantification is limited by the necessity to use heart rate monitors throughout training. It is also understood to require steadystate heart rate measurements, thus limiting the accuracy with which exercise of an interval nature can be quantified. Busso et al.[50] simplified the TRIMP equation by multiplying the average fraction of maximum aerobic power output during exercise to the session duration, thereby Sports Med 2009; 39 (9)

Training Quantification and Performance Models

limiting the training stimulus to external loading. A further practical limitation of TRIMP as a measure of training load is the inability to quantify non-aerobic modes of exercise such as resistance training. This is because heart rate increases disproportionately during resistance exercise, and the heart rate responses required for the calculation of TRIMP are not elicited. The equation of Busso et al.[50] was thus adapted to be used in weightlifting by replacing heart rate reserve with %1RM and duration with number of lifts. However, alternative attempts to resolve this limitation resulted in the inclusion of RPE in the quantification of exercise intensity.[51-54] 1.4.2 Session RPE

In an attempt to simplify the quantification of training load, Foster et al.[55] introduced the use of session RPE instead of using heart rate data or having to measure the intensity or type of exercise being performed. The session RPE is a rating of the overall difficulty of the exercise bout obtained 30 minutes after the completion of the exercise. Session load is calculated by multiplying session RPE by session duration of aerobic exercise (in minutes).[55] Foster et al.[56] compared the session RPE method with the summated heart rate zone score[57] (described below) during aerobic exercise and found that the pattern of differences between the two methods was very consistent. They proposed that session RPE was a valid and reliable measure of exercise intensity.[58] However, no correlation coefficients were provided in Foster et al.,[56] and although individual correlations between the two methods ranged between r = 0.75 and r = 0.90 in Foster[59] statistical methods were not explained. Impellizzeri et al.[60] found that individual correlations between the session RPE method and Banister’s TRIMP method ranged between r = 0.50 and r = 0.77; individual correlations between the session RPE method and the summated heart rate zone method ranged from r = 0.54 to r = 0.78, and from r = 0.61 to r = 0.85 between the session RPE method and the TRIMP methods in soccer players during training and matches. They suggest that the session RPE-based score cannot yet reª 2009 Adis Data Information BV. All rights reserved.

785

place the heart rate-based methods as a valid measure of exercise intensity, as only 50% of the variation they measured in heart rate could be explained by the session RPE.[60] A study by Borresen and Lambert[61] found correlations of r = 0.76 between TRIMP and session RPE and r = 0.84 between summated heart rate zone method and the session RPE method. The complex interaction of many factors that contribute to the personal perception of physical effort, including hormone concentrations (e.g. catecholamines), substrate concentrations (e.g. glucose, glycogen and lactate), personality traits, ventilation rate, neurotransmitter levels, environmental conditions or psychological states, may limit the use of RPE in accurately quantifying or prescribing exercise intensity.[1] Although using objective physiological measurements such as heart rate may be a more accurate way of calculating training load, the subjective measure of RPE remains useful. Thus, if heart rate monitors are not available, or an easier means of reporting and calculating training load is required, then the RPE method may still give reasonably accurate assessments of aerobic training load. For resistance exercise, session load is calculated by multiplying session RPE by the number of repetitions performed in resistance exercise.[51-53,55,59,62] The use of session RPE to quantify training load has potential in being a mode- and intensity-independent method that can be used for multiple types of exercise such as high-intensity or non-steady-state exercise such as resistance training, high-intensity interval training or plyometric training.[56] However, there remain limitations to its use in both aerobic and resistance training. RPE is influenced more by resistance load than by volume, so performing more repetitions with a lighter load is perceived as being easier than performing fewer repetitions with a heavier load.[51,62] Sweet et al.[51] and McGuigan et al.[52] found that the RPE varies significantly among different muscle groups used because of differences in muscle mass (and hence metabolic demand), range of motion and the number of joints involved in a movement. The order in which the exercises are performed, the fibre type of the muscle used, the mode of exercise for Sports Med 2009; 39 (9)

Borresen & Lambert

786

which the athlete is trained (i.e. the level of experience the athlete has in resistance training) and the time at which RPE is reported may also affect RPE.[51,62] 1.4.3 Summated Heart Rate Zone Score

The summated heart rate zone score is a modification to the calculation of training impulses that facilitates the quantification of interval training.[57] The accumulated duration (minutes) spent in each of five heart rate zones is calculated (i.e. 50–60%, 60–70%, 70–80%, 80–90% and 90–100% of maximal heart rate) and then multiplied by a multiplier factor for each zone (50–60% = 1, 60–70% = 2, 70–80% = 3, 80–90% = 4 and 90–100% = 5). The results are then summated.[57] Borresen and Lambert[61] attempted to identify characteristics that may explain the variance not accounted for in the relationship between the objective (TRIMP and summated heart rate zone score) and subjective (session RPE) methods of quantifying training load. The results suggested that for athletes who spent a greater percentage of their training time doing high-intensity exercise, the objective (heart rate-based) equations may overestimate training load compared with the subjective (RPE-based) method. Alternatively, the session RPE method may underestimate training load compared with the objective methods for these athletes. Conversely, in athletes who spent proportionally more of their training time doing low-intensity exercise, the heart rate-based methods may underestimate training load when compared with the training load calculated using the session RPE method, or the session-RPE method may overestimate training load.[61] The authors suggest that it may be the weighting system used in this equation that limits its accuracy. Because a weighting factor is applied to each zone comprising a range of heart rates, the lowest heart rate and the highest heart rate in each zone will be weighted the same, despite a difference in the physiological load. Under certain circumstances a change in heart rate of only 1 beat/min will change the weighting factor of the zone, thereby increasing or decreasing the calculated load disproportionately.[61] After an extensive ª 2009 Adis Data Information BV. All rights reserved.

review of the literature, there appears to be no evidence that this method of quantification has been validated. The summated heart rate zone equation may therefore have been derived theoretically and not through experimentation, raising the question of the legitimacy of validating the session RPE method against this heart ratebased method. 1.4.4 Lucia’s TRIMP

Recently, a modified version of the summated heart rate zone equation has been used by Earnest et al.[63] and Lucia et al.[64] and referred to as ‘Lucia’s TRIMP’ by Impellizzeri et al.[60] In this method the duration spent in each of three heart rate zones (zone 1: below the ventilatory threshold; zone 2: between the ventilatory threshold and the respiratory compensation point; and zone 3: above the respiratory compensation point) is multiplied by a coefficient (k) relative to each zone (k = 1 for zone 1, k = 2 for zone 2, and k = 3 for zone 3) and the adjusted scores are then summated. The original source of this equation, however, was not referenced in these studies. This method of quantifying training load shares the same limitation as the summated heart rate zone method, in so far as the weighting of each zone increases in a linear fashion, which does not reflect physiological responses to exercise above the anaerobic threshold.[65] Anaerobic threshold may vary between individuals with equal aerobic power, and therefore the metabolic stress experienced by individuals may be different even when exercising at the same percentage of maximal heart rate.[65] The TRIMP, session RPE and summated heart rate zone methods are becoming popular methods of quantifying training load. The accuracy of these methods in assessing internal training stress is important if training is to be prescribed based on these results in order to produce more predictable performances. However, it is not only the quantification of training load but knowledge of the physiological mechanisms involved in the exercise response and the ability to measure and quantify training-induced adaptations that will allow more accurate prescription of training and prediction of performance. Sports Med 2009; 39 (9)

Training Quantification and Performance Models

2. Modelling the Relationship between Training and Performance At the biological level, exercise training may be interpreted as a stimulus that causes a disturbance in homeostasis, which is restored during recovery after the training session. After several training sessions, the efficiency with which the physiological systems underlying homeostatic control function is altered, so that subsequent exercise at the same intensity may cause less disturbance to homeostasis. To produce optimal adaptations, training load and recovery must be balanced so that the athlete’s physiological systems are sufficiently stimulated to adapt and yet recovery is not impaired.[59,66-68] To this end, the dose-response effect of training impulses on subsequent performances has been investigated. Foster et al.[55] presented quantified observations of the performance response of competitive athletes to changes in training load but found no significant correlations between the improvements in time trial performance and training time, duration of high-intensity training, training intensity (reflected in RPE ratings) or training load (calculated as session RPE · duration). These findings further emphasize the complex relationship between a number of training variables that may contribute to training load, the body’s adaptive response and subsequent performance. Models of the relationship between training and performance have been proposed that consider the athlete as a system in which the training load is the input and performance the system output. The systems models are attractive in their potential to allow more accurate prediction of performance at specific times, or conversely to enable the design of optimal training programmes towards a specific performance goal.[49,69,70] 2.1 Fitness and Fatigue

Banister et al.[48] proposed an equation to assess the training effect (dose) on performance (response) in an attempt to establish a quantifiable relationship. They suggested that, in its simplest form, performance could be defined by two components, a ‘fitness impulse’ and ‘fatigue ª 2009 Adis Data Information BV. All rights reserved.

787

impulse’, and that at any time their difference (fitness – fatigue) can predict an athlete’s performance[48,49] (equation 4): Predicted performance ¼ Fitness  Fatigue aðtÞ ¼ k1 wðtÞet=t1  k2 wðtÞet=t2 ðEq: 4Þ where k1 and k2 are weighting factors (initially k1 = 1 for fitness and k2 = 2 for fatigue) such that the fitness impulse (k1w[t]) and the fatigue impulse (k2w[t]) can be calculated by multiplying the training impulse (w[t]) by the appropriate weighting factor (k1 or k2). The equation thus comprises two functions in which one represents a positive influence on performance and the other represents a negative influence on performance. Between training sessions the fitness and fatigue variables decline exponentially but at different rates. The contribution of a training impulse to fatigue is proportionally larger than to fitness. However, the decay time constant of fitness is longer. Banister et al.[48] suggested that the fitness decay time constant (t1) may be estimated initially as 45 days and the fatigue decay time constant (t2) as 15 days. These values, as with the initial values for the weighting factors (k1 = 1 and k2 = 2), are only estimates that allow for an approximation (prediction) to be made of future performance. Data from real performances are then collected and compared with the approximated (predicted) performance, and the decay time constants and weighting factor constants adjusted if discrepancies occur between the predicted and real performance.[48,49] Busso et al.[70] subsequently tested the accuracy of a simplified form of the above model, comprising only the fitness impulse [a(t) = w(t)k1et=t1 ]. They found that it produced a similar fit of estimated and real performances, accounting for 61–87% of the total variation in estimated and actual performances. However, they pointed out that the fatigue effect may have been underestimated because of the low-intensity endurance training the subjects underwent, and as such the fatigue effect did not contribute substantially to changes in performance. They suggested that Sports Med 2009; 39 (9)

Borresen & Lambert

788

future studies should include more strenuous and varied training programmes. The researchers further acknowledged the low precision of the performance measures, with the standard error of the estimated and real performances ranging from 3.6 to 5.9 performance units (estimated) and 97 to 152 performance units (real). They ascribed this low precision to external factors such as daily stress,[70] which cannot be controlled but needs to be recognized as an integral part of performance. 2.2 Physiological Correlates of Fitness and Fatigue

Wood et al.[71] explored possible physiological and psychological correlates of the positive and negative components of the Banister model in an attempt to validate the parameters with physiological markers. They found that running speed at ventilatory threshold and running economy correlated with the fitness parameter (r = 0.94 and r = -0.61, respectively), whereas a fatigue subset of the Profile of Mood States (POMS) questionnaire correlated moderately (r = 0.75) with the fatigue parameter of the equation. The authors suggested that the fatigue subset of POMS might reflect a more global fatigue (comprising various stressors of occupation, lifestyle and illness, etc.), whereas the fatigue component of the equation may represent only exercise-induced fatigue. The validity of the fatigue component will thus remain unclear until an accurate measure/marker of exercise-induced fatigue is found. Either that or the fatigue component of the performance equation does not accurately represent exerciseinduced fatigue.[71] Lambert and Borresen[72] reviewed several methods being used to monitor fatigue, including the evaluation of recovery and the assessment of muscle soreness after exercise training. In addition, methods of quantifying accumulated (chronic) fatigue were reviewed, including the Daily Analysis of Life Demands for Athletes test, POMS questionnaire, and heart rate measurements of variability, recovery and resting heart rate. Many other physiological adaptations that occur in response to prolonged exercise training have been investigated as possible markers to diª 2009 Adis Data Information BV. All rights reserved.

rectly measure and monitor fitness, fatigue, overtraining and recovery. Jones and Carter,[73] in their review of the effects of endurance training, identify four key parameters of aerobic fitness, . namely VO2max, exercise economy, lactate and ventilatory thresholds, and critical power. They suggest that an improvement in any one or more of these parameters will produce an improvement in performance. However, it must be acknowledged that many other factors, such as environmental conditions, race tactics and psychological factors, may also influence the outcome of a competitive performance.[73] Submaximal blood lactate concentration has also been proposed as a means with which to monitor changes in endurance fitness, because blood lactate concentration decreases at the same absolute and relative intensity after endurance training.[24] The absolute work rate at which the onset of blood lactate accumulation occurs also increases after 6 weeks of training.[74] Pyne et al.[23] found a direct relationship between improvements in lactate parameters and maximal 200 m test time in swimmers after a 20-week training period. However, lactate parameters were unrelated to international competition performance. Because numerous factors besides training have an effect on blood lactate concentration, and as a result of the necessity for standardized testing conditions, the usefulness and accuracy with which lactate profiling can be used to monitor training adaptations remains questionable.[23] The free circulating testosterone : cortisol ratio has been proposed as an indicator of physiological anabolic/catabolic balance.[75] A low free testosterone : cortisol ratio (<30%) has been suggested as a marker of a catabolic (overtrained or over-reached) state.[2,75] The circulating testosterone : cortisol ratio has also been proposed as a predictor of performance. However, consensus has not yet been reached on how testosterone and cortisol concentrations change in response to training and how this relates to performance.[76] It has been proposed that the measurement of serum iron, serum ferritin and transferrin may be used to identify the inflammatory response to muscle damage, as well as the state of acute and chronic recovery.[77] Serum iron and ferritin Sports Med 2009; 39 (9)

Training Quantification and Performance Models

concentrations have been found to be reduced in chronically exercising individuals, particularly those training at higher intensities.[78-80] It has been suggested that such decreases may have a negative effect on performance.[78,79] Hawley and Stepto[81] presented a theoretical model of training-induced adaptations in skeletal muscle that are likely to influence performance in elite cyclists. The model includes changes in muscle morphology such as increases in neural recruitment, capillary density, enzyme concentration and activity, and type I fibre content.[81] The decrease in muscle glycogen utilization and increase in intramuscular triglyceride oxidation[81-84] during prolonged submaximal exercise is also included. The model also considers acid-base status in terms of an increased lactate threshold and transport capacity. However, Hawley and Stepto[81] conclude that knowledge of the effect of physiological adaptations on performance is limited. Despite years of research, consensus has not yet been reached concerning the effects of training and overtraining on each of the many physiological ‘markers’ that have been investigated. As such, no single measure has been identified that can accurately assess how an athlete is responding to training. The correlation between training and the observed changes in these physiological variables is highly personal and dependent on individual tolerance of an exercise load, which may be a culmination of many internal and external factors. 2.3 Influence Curves

Fitz-Clarke et al.[69] proposed the use of influence curves to show conceptually how each consecutive day’s training affects a subsequent performance. The influence curve is a line representing the effect of a training impulse imposed at any time on performance at a specific future time. Performance on a specific day may be considered as a summation of the contributions of each day’s training impulse prior to the day of competition and decayed over the time between the training impulse and competition day. Each training impulse adds a contribution to performance according to its initial magnitude. ª 2009 Adis Data Information BV. All rights reserved.

789

The impulse response encompasses both the negative and positive influences of each day’s training from the start of a programme until competition day. The influence curve thus allows for the identification of the number of days before competition, at which time training load needs to be reduced because training after that day will contribute more to the fatigue impulse than to the fitness impulse. The influence curve can also identify the number of days prior to competition when training is most beneficial to performance on competition day. This model therefore has the potential to design an optimal training programme able to produce a specific performance at a particular time. Influence curves may therefore be useful when giving training advice for one event. However, analysis of training and performance for several competitive events during a season, as is common for elite athletes, is more complex. Influence curves show that the prescription of optimal training for each event becomes a challenge, as the ideal training and rest periods for one performance will impact suboptimally on subsequent performances.[69] 2.4 Recursive Least Squares Algorithm

In the equation of Banister et al.[48] the model parameters t1, t2, k1 and k2, which are initially estimates that are adjusted to suit the individual (after fitting the equation’s predicted performance to a real performance), are subsequently kept constant (i.e. time invariant) for the duration of the study period.[48,49] Busso et al.[85] investigated using a recursive least squares algorithm incorporating parameters that are free to vary over time to more accurately illustrate changes in performance after training. They suggested that each training response may be influenced by previous training bouts and that the day-to-day variation in the model parameters may provide important information on the cumulative effects of training.[85,86] The relationship between predicted and real performances was better using a time-varying model than the time-invariant model, with coefficients of variation for the former being 0.875 and 0.879 for the two subjects tested, compared with 0.682 and 0.666 for the Sports Med 2009; 39 (9)

Borresen & Lambert

790

time-invariant model.[85] However, because the parameters in the time-varying model are estimated at any given time from the previous and present data, this type of model would be limited in its ability to predict performance in response to future training, unless the parameters themselves change in a predictable way.[87] Using their recursive least squares algorithm, Busso et al.[86] later studied the effect of an increase in training frequency on exercise-induced fatigue and found that the time needed to regain a previous performance level increased as training frequency increased. The positive effect of a given training load on performance also decreased when training frequency was increased. Therefore, reducing recovery time between training bouts resulted in an increase in accumulated fatigue. Subsequently, they proposed a model that would account for greater fatigue resulting from increased training frequency.[88] The model comprises a fatigue component that varies over time and with the intensity of past training bouts. It offers the possibility to more accurately describe the dose-response relationship between cumulative training loads and training response, and thus to study training periodization.[88] 2.5 Threshold Saturation

Hellard et al.[89] proposed a model that includes a saturation threshold in which the impact of training on performance is nonlinear and has an upper limit. This method introduces the possibility of identifying an upper limit to the training stimulus of an athlete so that training intensity and duration can be kept below this threshold in order to optimize physiological adaptations. Training maintained above this level may induce excessive chronic fatigue and lead to a decline in performance.[89] Hellard et al.[89] studied Olympic swimmers over 4 years and found that the modified model improved the fit between training and performance compared with the Banister model. However, the training variables still explained only 30% of the variation in performance. They suggested a number of reasons for this discrepancy: (i) an individual’s response to the same exercise load may differ ª 2009 Adis Data Information BV. All rights reserved.

between seasons; (ii) there may be indirect effects of training, for example adaptations to one mode of exercise may influence/change the way in which the body responds to another exercise mode; (iii) variations in technique; or (iv) the fact that swimmers react differently to the same training stimulus.[89] The model parameters were also assumed to remain constant throughout the duration of the study, whereas regular adjustment of these parameters may have improved the fit between predicted and real performances.

2.6 Limitations to Modelling the TrainingPerformance Relationship

The proposed mathematical models attempt to describe the effect of training bouts on performance as a dose-response relationship comprising fitness and fatigue impulses. However, although attractive in concept, the accuracy of these theoretical models has proven to be poor, as is evident in weak correlations between the predicted and measured performances in response to training.[70,85] Considering that the smallest changes in performance could have a significant impact on the outcome of elite competition, it becomes vitally important that models predict performance with only the slightest margin of error. These disparities may be the result of the fact that the values for the fitness impulse, the fatigue impulse and their decay rates are initially estimates and that the adjustment of these values for each individual requires regular criterion performance measures comparable with competition conditions. Not only is this difficult and impractical, but it allows only retrospective adjustments to be made to the equation rather than the ability to prospectively prescribe training to achieve a desired performance. More importantly, the equations lack an accurate measure of individualism in how each athlete responds to training, which may contribute substantially to the inaccuracy of the model. Thus, although much has been reported on the correlations between endurance training and physiological adaptations, specific markers that facilitate the quantification of a dose-response relationship between Sports Med 2009; 39 (9)

Training Quantification and Performance Models

training, adaptation and performance remain to be identified. 3. Variability in the Physiological Response to Training An important limiting factor in the establishment of a quantifiable relationship between training, physiological adaptations and performance is that the variability in the way individuals react to training is not being accounted for. The intersubject variance in training-induced adaptations may be the result of several factors such as: age; sex; training history; psychological factors; initial training status; mode, duration, intensity and frequency of training;[90] recovery potential; exercise capacity; non-training stress factors; stress tolerance; [73,91] and genetics.[92] Jones and Carter,[73] in their review of the effects of endurance training on the parameters of aerobic fitness, note that the . magnitude of change in VO2max may be governed by many of the above factors, and that exercise economy differs significantly between individuals and may be influenced by the velocities/power outputs at which they habitually train. Bell et al.[93] found that concurrent strength and endurance training resulted in training adaptations that differed from those that occurred after either strength or endurance training alone, emphasizing the importance of considering exercise mode when assessing training-induced adaptations. Avalos et al.[94] found that they could separate swimmers into groups that reacted well to either a long-term, mid-term or short-term training period, emphasizing the need to individualize the distribution of training loads throughout a season in order to facilitate optimal adaptation in each athlete. Skinner et al.[22] found . that individual changes in power output and VO2max in response to 20 weeks of cycling training varied significantly among people who trained at heart rates asso. ciated with the same %VO2max. Al-Ani et al.[95] reported an increase in heart rate variability, measured as the difference between maximum and minimum R-R interval in a respiratory cycle, in nine of 11 people after 6 weeks of endurance training. However, with the heart rate variability of the two other subjects having had the opposite ª 2009 Adis Data Information BV. All rights reserved.

791

response after training, the individuality of the training response is again highlighted. Bouchard and Rankinen[96] and Rice et al.[97] found that initial training status and familial factors contributed significantly to the interindividual variation in ‘trainability’ of individuals (measured. as changes in heart rate, blood pressure and VO2max). Wilmore et al.[98] found that sex, race and age contributed to this heterogeneity. Conversely, Skinner et al.[99] found that age, sex, race . and initial fitness had little influence on how VO2max changed after a standardized endurance training programme. The authors showed that there were low, medium and high responders in both sexes, at all ages and at all fitness levels, and suggested that genetics may be responsible for the wide variation in individual responses. This individuality in training response thus requires further investigation. 4. Variability in the Relationship between Training Adaptations and Performance Genetic traits may contribute substantially not only to the way in which athletes adapt to training but also to the observed heterogeneity in performance ability between athletes.[100,101] This variance may depend on different numbers and types of genes being activated in response to different intensities, durations and frequencies of exercise. Further variation may occur during DNA sequencing, gene transcription and protein translation. The potential for improvement in performance or optimal adaptation may also be influenced by a genetic predisposition for performance in a specific mode of exercise.[92] For example, the type of predominant muscle fibre in a muscle may predispose a person to better performance in either an endurance- or resistance-type sport.[92,102] Recently, interest in the identification of specific genes that may be associated with performance has increased.[101] One potential gene is ACTN3, which encodes the protein a-actinin-3 that forms part of the sarcomeres of fast twitch muscle fibres. There is evidence that the 577R allele and 577RR genotype may be associated with sprinting and/or power performance, whereas no association has been found Sports Med 2009; 39 (9)

792

between the R577X polymorphism and endurance performance.[101,103,104] In a study examining the effect of training on swimming performance, Mujika et al.[105] noted significant variance among elite swimmers in a number of systems model variables, including the time constants of decay of the fitness and fatigue impulses, and the fitness and fatigue multiplying factors, amongst others. The model explained between 45% and 85% of the variations in real performance. This emphasizes a limitation in the design of the performance models, which do not take into account the fact that individuals respond and adapt differently to training. However, a limitation in the study was the assumption that the model parameters were constant throughout the study (44 weeks), which ignores the possibility that adaptations to training may alter how an athlete responds to exercise.[105] Indeed, Banister et al.[48] suggested that the period within which the model parameters may be assumed to be constant is 60–90 days, after which the parameters need to be reset by comparing predicted performance to real performance. Hellard et al.[89] reported that training variables explained only 30% of the variation in the performance of Olympic swimmers studied for 4 years. The authors suggest that performance may be affected by the phase of training in which the athlete is, as there were differing short-, intermediate- and long-term effects on performance. In addition, swimmers react differently to the same training load (interindividual differences) and between consecutive seasons (intraindividual differences). Bagger et al.[16] describe the magnitude of individual variation in a number of factors often used to assess training adaptations, because this is necessary to distinguish whether a change in the variable is the result of training or of random biological fluctuation. They found that performance and physiological measurements such .as HRmax, HR10km, RERsubmax, HRsubmax, and VO2max, amongst others, had the lowest total coefficient of variation (13%), whereas metabolic and hormonal variables had the highest coefficient of variation (37%). The variance as a result of within-subject variation in metabolic and hormonal variables was 53%, compared with only ª 2009 Adis Data Information BV. All rights reserved.

Borresen & Lambert

13% in performance and physiological measurements.[16] Thus, training adaptations are a highly individual phenomenon, and the variation within and between athletes needs to be considered when assessing training-induced changes in performance.

5. Summary and Future Research There is currently no accurate quantitative means with which to prescribe the pattern, duration and intensity of exercise required to produce specific physiological adaptations. Added to this is the fact that individuals adapt differently to the same exercise stimulus. Banister et al.[48] found that the constants used in the performance models need to be reset after a period of approximately 60–90 days. Mujika et al.[105] suggest that individual chronic training adaptation profiles could be developed by studying individual fatigue and fitness curves in order to better understand an individual’s response to a training bout. In the development of the performance prediction model, Banister et al.[48] suggest that ‘‘it may be assumed that all the constants of the model that are used to obtain the fit are exactly those that are peculiar to the individual being modeled’’. Thus, studying individual fitness and fatigue curves may allow the quantification of individual response and adaptation to training. The accuracy of the mathematical models in predicting performance may also be improved when the physiological meaning of the modelled responses is better defined.[70,87] As such, the search continues to find easily measurable physiological markers of ‘fitness’ and ‘fatigue’ to improve the accuracy with which performance can be predicted.[87] Despite years of research, no single physiological marker has been identified that can quantify the fitness and fatigue responses to exercise or predict performance with accuracy. The correlation between training and the observed changes in these physiological variables is highly personal and depends on many factors that influence an individual’s tolerance of an exercise load. Thus, more emphasis needs to be directed towards the measurement of markers that reflect an individual’s global capacity to respond or adapt to training, Sports Med 2009; 39 (9)

Training Quantification and Performance Models

rather than an absolute measure of the changes in physiological variables in response to exercise. Acknowledgements The research undertaken in this study is funded in part by the University of Cape Town, Discovery Health, National Research Foundation, Ernst & Ethel Eriksen Foundation and Deutscher Akademischer Austausch Dienst. The authors have no conflicts of interest that are directly related to the content of this review.

References 1. Williams JG, Eston RG. Determination of the intensity dimension in vigorous exercise programmes with particular reference to the use of the rating of perceived exertion. Sports Med 1989; 8 (3): 177-89 2. Budgett R. Fatigue and underperformance in athletes: the overtraining syndrome. Br J Sports Med 1998; 32 (2): 107-10 3. Halson SL, Jeukendrup AE. Does overtraining exist? An analysis of overreaching and overtraining research. Sports Med 2004; 34 (14): 967-81 4. Urhausen A, Kindermann W. Diagnosis of overtraining: what tools do we have? Sports Med 2002; 32 (2): 95-102 5. Budgett R, Newsholme E, Lehmann M, et al. Redefining the overtraining syndrome as the unexplained underperformance syndrome. Br J Sports Med 2000 Feb 1; 34 (1): 67-8 6. Hopkins WG. Quantification of training in competitive sports: methods and applications. Sports Med 1991; 12 (3): 161-83 7. Lambert MI, Bryer L, Hampson DB, et al. Accelerated decline in running performance in a masters runner with a history of a large volume of training and racing. J Aging Phys Activ 2002; 10: 314-21 8. Borresen J, Lambert MI. Validity of self-reported training duration. Int J Sports Sci Coach 2006; 1 (4): 353-9 9. Shephard RJ. Limits to the measurement of habitual physical activity by questionnaires. Br J Sports Med 2003; 37 (3): 197-206 10. Buchheit M, Gindre C. Cardiac parasympathetic regulation: respective associations with cardiorespiratory fitness and training load. Am J Physiol Heart Circ Physiol 2006; 291 (1): H451-8 11. Achten J, Jeukendrup AE. Heart rate monitoring: applications and limitations. Sports Med 2003; 33 (7): 517-38 12. Arts FJ, Kuipers H. The relation between power output, oxygen uptake and heart rate in male athletes. Int J Sports Med 1994; 15 (5): 228-31 13. Robinson DM, Robinson SM, Hume PA, et al. Training intensity of elite male distance runners. Med Sci Sports Exerc 1991; 23 (9): 1078-82 14. Karvonen J, Vuorimaa T. Heart rate and exercise intensity during sports activities: practical application. Sports Med 1988; 5 (5): 303-11

ª 2009 Adis Data Information BV. All rights reserved.

793

15. Lambert MI, Mbamba ZH, St Clair Gibson A. Heart rate during training and competition for long-distance running. J Sports Sci 1998; 16: S85-90 16. Bagger M, Petersen PH, Pedersen PK. Biological variation in variables associated with exercise training. Int J Sports Med 2003; 24 (6): 433-40 17. Xu F, Rhodes EC. Oxygen uptake kinetics during exercise. Sports Med 1999; 27 (5): 313-27 18. Kohrt WM, Morgan DW, Bates B, et al. Physiological responses of triathletes to maximal swimming, cycling, and running. Med Sci Sports Exerc 1987; 19 (1): 51-5 19. Swain DP, Leutholtz BC, King ME,. et al. Relationship between % heart rate reserve and % VO2 reserve in treadmill exercise. Med Sci Sports Exerc 1998; 30 (2): 318-21 20. Swain rate reserve is equivalent to . DP, Leutholtz BC. Heart . %VO2 reserve, not to % VO2max. Med Sci Sports Exerc 1997; 29 (3): 410-4 21. Baldwin J, Snow RJ, Febbraio MA. Effect of training status and relative exercise intensity on physiological responses in men. Med Sci Sports Exerc 2000; 32 (9): 1648-54 22. Skinner JS, Wilmore KM, Krasnoff JB, et al. Adaptation to a standardized training program and changes in fitness in a large, heterogeneous population: the HERITAGE Family Study. Med Sci Sports Exerc 2000; 32 (1): 157-61 23. Pyne DB, Lee H, Swanwick KM. Monitoring the lactate threshold in world-ranked swimmers. Med Sci Sports Exerc 2001; 33 (2): 291-7 24. Jacobs I. Blood lactate: implications for training and sports performance. Sports Med 1986; 3 (1): 10-25 25. Weltman A, Seip RL, Snead D, et al. Exercise training at and above the lactate threshold in previously untrained women. Int J Sports Med 1992; 13 (3): 257-63 26. Stegmann H, Kindermann W, Schnabel A. Lactate kinetics and individual anaerobic threshold. Int J Sports Med 1981; 2 (3): 160-5 27. Swart J, Jennings CL. Use of blood lactate concentration as a marker of training status. S Afr J Sports Med 2004; 16: 3-7 28. Jeukendrup AE, Hesselink MK. Overtraining: what do lactate curves tell us? J Sports Med 1994; 28 (4): 239-40 29. Urhausen A, Gabriel HH, Weiler B, et al. Ergometric and psychological findings during overtraining: a long-term follow-up study in endurance athletes. Int J Sports Med 1998; 19 (2): 114-20 30. Kirwan JP, Costill DL, Flynn MG, et al. Physiological responses to successive days of intense training in competitive swimmers. Med Sci Sports Exerc 1988; 20 (3): 255-9 31. Green JM, McLester JR, Crews TR, et al. RPE association with lactate and heart rate during high-intensity interval cycling. Med Sci Sports Exerc 2006; 38 (1): 167-72 32. Little T, Williams AG. Measures of exercise intensity during soccer training drills with professional soccer players. J Strength Cond Res 2007; 21 (2): 367-71 33. Ozkan A, Kin-Isler A. The reliability and validity of regulating exercise intensity by ratings of perceived exertion in step dance sessions. J Strength Cond Res 2007; 21 (1): 296-300 34. Chen MJ, Fan X, Moe ST. Criterion-related validity of the Borg ratings of perceived exertion scale in healthy

Sports Med 2009; 39 (9)

Borresen & Lambert

794

35.

36.

37.

38.

39.

40.

41. 42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

individuals: a meta-analysis. J Sports Sci 2002; 20 (11): 873-99 Dekerle J, Baron B, Dupont L, et al. Maximal lactate steady state, respiratory compensation threshold and critical power. Eur J Appl Physiol 2003; 89 (3): 281-8 Pringle J, Jones A. Maximal lactate steady state, critical power and EMG during cycling. Eur J Appl Physiol 2002; 88 (3): 214-26 Brickley G, Doust J, Williams CA. Physiological responses during exercise to exhaustion at critical power. Eur J Appl Physiol 2002; 88 (1): 146-51 Vanhatalo A, Doust JH, Burnley M. Determination of critical power using a 3-min all-out cycling test. Med Sci Sports Exerc 2007; 39 (3): 548-55 Jeukendrup A, van Diemen A. Heart rate monitoring during training and competition in cyclists. J Sports Sci 1998; 16: S91-9 Foster C, Heimann KM, Esten PL, et al. Differences in perceptions of training by coaches and athletes. S Afr J Sports Med 2001; 8 (2): 3-7 Larsson P. Global positioning system and sport-specific testing. Sports Med 2003; 33 (15): 1093-101 Townshend AD, Worringham CJ, Stewart IB. Assessment of speed and position during human locomotion using nondifferential GPS. Med Sci Sports Exerc 2008; 40 (1): 124-32 Rodriguez DA, Brown AL, Troped PJ. Portable global positioning units to complement accelerometry-based physical activity monitors. Med Sci Sports Exerc 2005; 37 (11 Suppl.): S572-81 Larsson P, Henriksson-Larsen K. The use of dGPS and simultaneous metabolic measurements during orienteering. Med Sci Sports Exerc 2001; 33 (11): 1919-24 Schutz Y, Herren R. Assessment of speed of human locomotion using a differential satellite global positioning system. Med Sci Sports Exerc 2000; 32 (3): 642-6 Vermeulen AD, Evans DL. Measurements of fitness in thoroughbred racehorses using field studies of heart rate and velocity with a global positioning system. Equine Vet J Suppl 2006; (36): 113-7 Kingston JK, Soppet GM, Rogers CW, et al. Use of a global positioning and heart rate monitoring system to assess training load in a group of thoroughbred racehorses. Equine Vet J Suppl 2006; (36): 106-9 Banister EW, MacDougall JD, Wenger HA, et al. Modeling elite athletic performance: physiological testing of the high-performance athlete. Campaign (IL): Human Kinetics Books; 1991: 403-25 Morton RH, Fitz-Clarke JR, Banister EW. Modeling human performance in running. J Appl Physiol 1990; 69 (3): 1171-7 Busso T, Ha¨kkinen K, Pakarinen A, et al. A systems model of training responses and its relationship to hormonal responses in elite weight-lifters. Eur J Appl Physiol 1990; 61 (1): 48-54 Sweet TW, Foster C, McGuigan MR, et al. Quantitation of resistance training using the session rating of perceived exertion method. J Strength Cond Res 2004; 18 (4): 796-802 McGuigan MR, Egan AD, Foster C. Salivary cortisol responses and perceived exertion during high intensity and

ª 2009 Adis Data Information BV. All rights reserved.

53. 54.

55. 56.

57. 58.

59.

60.

61.

62.

63.

64.

65.

66.

67. 68.

69.

70.

71.

72.

low intensity bouts of resistance exercise. J Sports Sci Med 2004; 3: 8-15 McGuigan MR, Foster C. A new approach to monitoring resistance training. Strength Cond J 2004; 26 (6): 42-7 Egan AD, Winchester JB, Foster C, et al. Using session RPE to monitor different methods of resistance exercise. J Sports Sci Med 2006; 5: 289-95 Foster C, Daines E, Hector L, et al. Athletic performance in relation to training load. Wis Med J 1996; 95 (6): 370-4 Foster C, Florhaug JA, Franklin J, et al. A new approach to monitoring exercise training. J Strength Cond Res 2001; 15 (1): 109-15 Edwards S. The heart rate monitor book. Sacramento (CA): Fleet Feet Press, 1993 Herman L, Foster C, Maher MA, et al. Validity and reliability of the session RPE method for monitoring exercise training intensity. S Afr J Sports Med 2006; 18 (1): 14-7 Foster C. Monitoring training in athletes with reference to overtraining syndrome. Med Sci Sports Exerc 1998; 30 (7): 1164-8 Impellizzeri FM, Rampinini E, Coutts AJ, et al. Use of RPE-based training load in soccer. Med Sci Sports Exerc 2004; 36 (6): 1042-7 Borresen J, Lambert MI. Quantifying training load: a comparison of subjective and objective methods. IJSPP 2008; 3: 16-30 Day ML, McGuigan MR, Brice G, et al. Monitoring exercise intensity during resistance training using the session RPE scale. J Strength Cond Res 2004; 18 (2): 353-8 Earnest CP, Jurca R, Church TS, et al. Relation between physical exertion and heart rate variability characteristics in professional cyclists during the Tour of Spain. Br J Sports Med 2004; 38 (5): 568-75 Lucia A, Hoyos J, Santalla A, et al. Tour de France versus Vuelta a Espana: which is harder? Med Sci Sports Exerc 2003; 35 (5): 872-8 Stagno KM, Thatcher R, van Someren KA. A modified TRIMP to quantify the in-season training load of team sport players. J Sports Sci 2007; 25 (6): 629-34 Kuipers H, Keizer HA. Overtraining in elite athletes: review and directions for the future. Sports Med 1988; 6 (2): 79-92 Kuipers H. Training and overtraining: an introduction. Med Sci Sports Exerc 1998; 30 (7): 1137-9 Uusitalo ALT. Overtraining: making a difficult diagnosis and implementing targeted treatment. Phys Sportsmed 2001; 29 (5): 35-50 Fitz-Clarke JR, Morton RH, Banister EW. Optimizing athletic performance by influence curves. J Appl Physiol 1991; 71 (3): 1151-8 Busso T, Carasso C, Lacour JR. Adequacy of a systems structure in the modeling of training effects on performance. J Appl Physiol 1991; 71 (5): 2044-9 Wood RE, Hayter S, Rowbottom D, et al. Applying a mathematical model to training adaptation in a distance runner. Eur J Appl Physiol 2005; 94 (3): 310-6 Lambert MI, Borresen J. A theoretical basis of monitoring fatigue: a practical approach for coaches. Int J Sports Sci Coach 2006; 1 (4): 371-88

Sports Med 2009; 39 (9)

Training Quantification and Performance Models

73. Jones AM, Carter H. The effect of endurance training on parameters of aerobic fitness. Sports Med 2000; 29 (6): 373-86 74. Mayes R, Hardman AE, Williams C. The influence of training on endurance and blood lactate concentration during submaximal exercise. Br J Sports Med 1987; 21 (3): 119-24 75. Urhausen A, Gabriel H, Kindermann W. Blood hormones as markers of training stress and overtraining. Sports Med 1995; 20 (4): 251-76 76. Filaire E, Bernain X, Sagnol M, et al. Preliminary results on mood state, salivary testosterone: cortisol ratio and team performance in a professional soccer team. Eur J Appl Physiol 2001; 86 (2): 179-84 77. Smith DJ, Roberts D. Effects of high volume and/or intense exercise on selected blood chemistry parameters. Clin Biochem 1994; 27 (6): 435-40 78. Angeli A, Minetto M, Dovio A, et al. The overtraining syndrome in athletes: a stress-related disorder. J Endocrinol Invest 2004; 27 (6): 603-12 79. Beard J, Tobin B. Iron status and exercise. Am J Clin Nutr 2000; 72 (2 Suppl.): 594S-7S 80. Wilkinson JG, Martin DT, Adams AA, et al. Iron status in cyclists during high-intensity interval training and recovery. Int J Sports Med 2002; 23 (8): 544-8 81. Hawley JA, Stepto NK. Adaptations to training in endurance cyclists: implications for performance. Sports Med 2001; 31 (7): 511-20 82. Hawley JA. Adaptations of skeletal muscle to prolonged, intense endurance training. Clin Exp Pharmacol Physiol 2002; 29 (3): 218-22 83. Horowitz JF, Klein S. Lipid metabolism during endurance exercise. Am J Clin Nutr 2000; 72 (2 Suppl.): 558S-63S 84. Schmitt B, Fluck M, Decombaz J, et al. Transcriptional adaptations of lipid metabolism in tibialis anterior muscle of endurance-trained athletes. Physiol Genomics 2003; 15 (2): 148-57 85. Busso T, Denis C, Bonnefoy R, et al. Modeling of adaptations to physical training by using a recursive least squares algorithm. J Appl Physiol 1997; 82 (5): 1685-93 86. Busso T, Benoit H, Bonnefoy R, et al. Effects of training frequency on the dynamics of performance response to a single training bout. J Appl Physiol 2002; 92 (2): 572-80 87. Taha T, Thomas SG. Systems modelling of the relationship between training and performance. Sports Med 2003; 33 (14): 1061-73 88. Busso T. Variable dose-response relationship between exercise training and performance. MedSciSports Exerc 2003; 35 (7): 1188-95 89. Hellard P, Avalos M, Millet G, et al. Modeling the residual effects and threshold saturation of training: a case study of Olympic swimmers. J Strength Cond Res 2005; 19 (1): 67-75 90. Tremblay MS, Copeland JL, Van Helder W. Effect of training status and exercise mode on endogenous steroid hormones in men. J Appl Physiol 2004; 96 (2): 531-9

ª 2009 Adis Data Information BV. All rights reserved.

795

91. Lehmann M, Foster C, Keul J. Overtraining in endurance athletes: a brief review. Med Sci Sports Exerc 1993; 25 (7): 854-62 92. Heck AL, Barroso CS, Callie ME, et al. Gene-nutrition interaction in human performance and exercise response. Nutrition 2004; 20 (7-8): 598-602 93. Bell GJ, Syrotuik D, Martin TP, et al. Effect of concurrent strength and endurance training on skeletal muscle properties and hormone concentrations in humans. Eur J Appl Physiol 2000; 81 (5): 418-27 94. Avalos M, Hellard P, Chatard JC. Modeling the trainingperformance relationship using a mixed model in elite swimmers. Med Sci Sports Exerc 2003; 35 (5): 838-46 95. Al Ani M, Munir SM, White M, et al. Changes in R-R variability before and after endurance training measured by power spectral analysis and by the effect of isometric muscle contraction. Eur J Appl Physiol Occup Physiol 1996; 74 (5): 397-403 96. Bouchard C, Rankinen T. Individual differences in response to regular physical activity. Med Sci Sports Exerc 2001; 33 (6 Suppl.): S446-51 97. Rice T, An P, Gagnon J, et al. Heritability of HR and BP response to exercise training in the HERITAGE Family Study. Med Sci Sports Exerc 2002; 34 (6): 972-9 98. Wilmore JH, Stanforth PR, Gagnon J, et al. Heart rate and blood pressure changes with endurance training: the HERITAGE Family Study. Med Sci Sports Exerc 2001; 33 (1): 107-16 99. Skinner JS, Jaskolski A, Jaskolska A, et al. Age, sex, race, initial fitness, and response to training: the HERITAGE Family Study. J Appl Physiol 2001; 90 (5): 1770-6 100. Macarthur DG, North KN. Genes and human elite athletic performance. Hum Genet 2005; 116 (5): 331-9 101. MacArthur DG, North KN. ACTN3: A genetic influence on muscle function and athletic performance. Exerc Sport Sci Rev 2007; 35 (1): 30-4 102. Brooks GA, Fahey TD, White TP, et al. Exercise physiology: human biogenetics and its applications. Columbus (OH): McGraw-Hill Companies, Inc., 2000 103. Yang N, MacAruthur J. ACTN3 genotype is associated with human elite athletic performance. Am J Hum Genet 2003; 73: 627-31 104. Saunders CJ, September AV, Xenophontos SL, et al. No association of the ACTN3 gene R577X polymorphism with endurance performance in Ironman triathlons. Ann Hum Genet 2007; 71 (6): 777-81 105. Mujika I, Busso T, Lacoste L, et al. Modeled responses to training and taper in competitive swimmers. Med Sci Sports Exerc 1996; 28 (2): 251-8

Correspondence: Prof. Mike Lambert, MRC/UCT Research Unit for Exercise Science and Sports Medicine, Sports Science Institute of South Africa, Boundary Road, Newlands 7700, South Africa. E-mail: [email protected]

Sports Med 2009; 39 (9)