Related Topic(s):

Electroencephalography and Sport; Review and Future Directions

Vietta E. Wilson, Ph.D.
York University

A chapter from Textbook of Neurofeedback, EEG Biofeedback, qEEG and Brain Self Regulation by Rob Kall and Joe Kamiya

Purpose of the paper

I. PSYCHOLOGICAL SKILLS NECESSARY FOR SPORTS PERFORMANCE

A. Traits of Optimum Performers
B. States of Optimum Performance

II. RESEARCH ON EEG IN SPORT

A. Genetics vs Learning

B. EEG and Sport Performance
Pre-performance
background EEG
evoked response potentials-ERP
During the event
Pre event to post event
Imagery of the event
Biofeedback
Cautions and Summary of Section

III. SPORT CONSIDERATIONS

FUTURE DIRECTIONS

REFERENCES
The purpose of this paper is to briefly review the psychological traits and states that are believed necessary for performance in competitive sport. This information may help guide areas of brain research that could have a large practical impact upon enhancing performance. The paper will then focus on reviewing studies which used electroencephalography (EEG) to assess brain processing in athletes. The final section will present considerations necessary for conducting sport research followed by some future directions.



PSYCHOLOGICAL SKILLS NECESSARY FOR SPORT PERFORMANCE

The need for psychological skills in learning, maintaining or improving performance in sport, especially under the stress of competition, has been anecdotal documented in the popular press and sport literature for decades . More recently research. has identified some of the psychological predispositions or traits as well as the mental states necessary for elite performance in sport.


Traits of Optimum Performers


Coaches who have worked with athletes as well as the "normal' population will generally concur that athletes are different from non-athletes in more than motor skills. A controversy surrounded the use of personality testing of athletes in North America in the 1960's. With the integration of an Interaction Model (traits X states) and sport specific inventories many of the previous objections about trait research in sport psychology have been lessened.

An example of the extensive work in North America in the 1960's and "70's was Ogilvie and Tutko's ( Ogilvie 1968 ) sport specific personality test that was found to differentiate athletes from non-athletes and further differentiated among professional, university and high school athletes in several team sports. Common traits listed by Ogilvie as being important in athletes of high calibre were trust, extroversion, tough mindedness, self-controlled and intelligence. Eysenck (1984) reviewed studies tests using the Eysenck Personality Inventory for trait assessment in sport and concluded that there is ample evidence to support extroversion as being a trait of athletes.


States of Optimal Performance

Most of the work from the mid-1970's to today has focussed upon theoretical testing of generally one personality concept or trait such as anxiety. Williams and Krane's (1993) review of interviews and studies with athletes, coaches, sport psychologists and professional scouts suggests that there are common similarities across sports of the mental states or feelings that are present when athletes report experiencing peak performances. The common states can be generalized into arousal control, high self-confidence, attentional control, and determination. The skills taught in mental training programmes for athletes in attempts to achieve these successful states usually include but are not restricted to goal setting, relaxation/energization, imagery, self-talk for performance and coping, and attentional strategies. Research reviews (Greenspan & Feltz,1989,Vealey,1994)) have shown that these skills can be successfully taught and are related to improved sport performance. For our purposes all the above skills are placed into the categories of either arousal control or attention control

Raglin's (1992) review of the arousal/anxiety research indicates that the Inverted U hypothesis of there being an optimal range of arousal for performance, not too high nor too low, is not supported. Rather he concludes Hanin's (1978,1986) theory of the Individual Zones of Optimal Functioning (IZOF) can more accurately explain the relationship between arousal/anxiety and competitive sport performance. The IZOF states that an individual performs best when his/her pre-competitive anxiety is within a relatively narrow range and that many successful performances are produced under both high and low levels of anxiety. Rundle and Weinberg (1997) found no support for the ZOF when team athletes are used. The suggestion that arousal/anxiety levels for successful performance is specific to each athlete has a great deal of face validity. Gould and Udry ( 1994) summarize the literature which indicates that arousal/anxiety control for sport can successfully be taught and indicate the strengths and weakness of studies in the area.


Nideffer (1976)extended theoretical work on attentional processing by proposing that different types of attention could be assessed, and applied in sport. He identified the need for controlling width (broad vs narrow), direction ( internal vs external), and the flexibility to shift attention. While the research using his Test of Attentional Style has shown equivocal results, the importance of attention as paramount to sport performance has been amply demonstrated. Training in controlling one's attentional state has also been found to enhance performance in sport (Nideffer, 1993)..


The use of imagery in mental training programmes for both arousal control and attentional control is widely advocated in sport and will be included as part of the attention control. The effects of imagery upon various psychological and physiological indices have been documented by the National Institute of Health among others (NIH,1995). Meta-analysis reviews have noted the beneficial effects of imagery in sport (Feltz,1987, Landers, et al 1983).



II. RESEARCH ON EEG and SPORT


An underlying problem with the research on the mental skills through the above methods is that the intervening process of what is happening within the athlete's head has to be inferred. This means that true understanding and control can not be attained until measures are taken of both the covert internal processing of the brain and the overt motor and behavioural processes resulting from the internal processing.

Measures of peripheral physiology (eg. heart rate, temperature) or behavioural outcomes (eg anxiety, performance errors) have been the basis of scientific investigations in sport for the past 70 years but only recently has the technology and expertise become available to assess central nervous system physiology in sport. With better EEG and computer capabilities we can more easily and directly measure the electrical activity of the brain implicated in specific attention, arousal, affect and cognitive processes that are either invoked or evoked by the performance demands in sport..

The task demands are different between those who participate in sport for recreational or fitness purposes and those who compete to win. Thus, for our purposes, an athlete is defined as one who is both skilled in and competes in a particular sport


A. Genetics vs Learning

Since there are suggestions that EEG has a genetic base (Allen,Reiner,Katsanis,Iacono,1997, Christian, Morzorati, Norton, Williams, O'Connor & Li,1996) it would seem logical that one would ask are there similarities in EEG patterns that distinguish whether a person is predisposed to perform well in sports? Additionally, are there are similarities in EEG patterns after individuals have "learned' a motor task?


The differences in personality that predispose individuals to participate and succeed in different sports has been reported for decades but the assessment has generally been limited to paper and pencil tests. Based upon Strelau's (1977) work that a resting baseline EEG percent time alpha distinguished pilots with good performance skills from pilots with poorer skills under the stress of flying, Cummings & Wilson (1978) had a track coach rate those athletes who performed well under the stress of competition. The baseline per cent time alpha from O1- T3 was significantly higher for those designated as good copers under stress. Wilson, Ainsworth & Bird, (1984) then used a nationally ranked mens volleyball team rated by coaches as either good or poor copers/concentrators under the stress of competition and again found higher baseline alpha for the good performers. The possibility of using EEG for assessing predisposing mental processes is currently being investigated in other areas such as exercise dependency (Beh, Mathers & Holden,1996).


The second consideration is whether or not the learning of a motor skill affects the subsequent neural processing as assessed by EEG .Gliner et al (1983) reported changes in EEG as individuals learned a motor task but the design did not allow for determining whether they were measuring performance changes or whether learning had occurred. Etnier's and others (1996) study of changes in EEG after a person practices the motor tasks included retention tests in their design to ensure that "learning' of the motor task had occurred. After no initial EEG differences between the experimental and control group in the baseline, the experimental group had significantly greater alpha power at the end of the training and which remained after two retention tests. This confirms that learning a motor task changes the EEG of the participants in future performances of that task.

Further support for an inclusion of both traits (genetics) and states (learning) has been noted by.
Rosenfeld, Reinhart & Srivastava (1997). They have recently shown that alpha and beta entrainment are dependent upon baseline EEG .




Sport Performance

Both types of EEG investigations, background EEG and event related potentials ERP, have been used in sport. Background EEG is a measurement of the variety of electrical signals or potentials that spontaneously occur within the person and is used to investigate mental processes across time. Sport research has used the traditionally classifications of a specified range of frequencies or speed cycles per second or hertz (Hz). Data have traditionally been reported in specified ranges, such as beta 12-20Hz, alpha 8-12 Hz, theta 4-8Hz and delta 0-4Hz. Few studies used the recommended smaller 1 Hz bandwidth .

Quantitative electroencephalography, QEEG , simultaneously measures a large number of electrode sites then the waves are digitized and mathematically analysed. Coloured three dimensional maps of the brain are usually produced to illustrate the specific brain frequencies in the various brain regions. Only one study was located which used QEEG in sport related setting.

Since background EEG contains so many signals which could be related to innumerable events, researches devised a method to determine what part of the EEG signal is a response to a particular event or an evoked response potential (ERP). They do this by repeatedly presenting the same stimulus, such as squeezing a trigger in archery / shooting or initiating a putt in golf, and observing the background EEG for a brief period of time before and after the presentation. With enough repetitions, most of the background EEG potentials will cancel each other out. What remains is a signal that is specifically related to only the stimulus that was used to elicit the response. This signal is generally measured immediately prior to and following the stimulus and the resulting electrical spikes, called positive and negative, are generally named according to the milliseconds that occur before or after the stimulus: for example, P300 is a positive spike 300 milliseconds following the stimulus. This technique can help clarify the athlete's response to a particular display and help identify the various components of the ERP such as readiness to respond, different types of attentional and cognitive focussing. Evoked potentials have the disadvantage of not being able to determine the complexity of the total brain/body involvement in the sport process.



Pre-performance

Background EEG

Early work by Lander's group (Hatfield, Landers & Ray 1984) established that rifle shooters had an increase in temporal alpha activity in the left hemisphere immediately prior to trigger pulls resulting in good scores. They hypothesised that this was indicative of more efficient mental processing and perhaps represented mental quieting through less self talk. This reduction in activation, particularly of the left hemisphere was also noted in later studies of rifle shooters(Bird,1987, Hatfield et al, 1987) archers (Salazar et al, 1990) and golfers (Crews and Landers,1993) and was attributed to the attentional demands being processed in the right hemisphere. The last study also found increased activity in the right hemisphere which they believe is due to both hands being required to execute the motor skill unlike the previous skills.

Contrary to the Lander's group, Collins, Powell and Davies (1990, 1991a, 1991b) demonstrated increased alpha power in the temporal and central locations in both hemispheres prior to successful performance in karate, soccer and cricket tasks performed in the laboratory. Trials deemed failures were associated with decreased alpha activity in both hemispheres. The differences may be due to the demands of the tasks since Landers used predominately skills that were self initiated actions while Collins used skills that required responses to others. The use of the right hemisphere for processing visual-perceptual information has been documented in pilots by Sterman et al (1994) and would explain the increased processing in Collin's work compared to Landers..


Using computer generated tasks designed to elicit spread attention (broad), selective attention (narrow), readiness period and reaction period believed to represent open skill sport attentional situations, Fontani, Voglino and Girolami (1996) report using central EEG electrodes(sites not specified) to assess the differences among females from volleyball, basketball, swimming and a sedentary control group. All groups showed an increase in alpha activity from spread to selective attention with the swimmers showing the lowest alpha throughout all tests and the highest proportion of beta. All groups showed an increase in alpha in the readiness and reaction periods. Volleyball players had the largest frequency band variability across tests and differed from the other groups by having high levels of low frequency during the first second of the readiness period but prevailing beta bands in the last second of the readiness period. The authors (Fontani, et al, 1989) claim that this is a replication of reduction in frequency during the readiness period as measured "on field' for fencers and volleyball players and may represent anticipation of a response. While the authors fail to explain the inconsistencies to be expected for "open' skills versus "closed skills', the study does suggest that one must be aware of the different attentional skills necessary for different sports.




Evoked Response Potentials (ERP)

Rossi and Zani (1990) have reviewed an extensive series of their ERP studies which were used to demonstrate the interface between cognitive and sport psychology. They reported the exogenous, responses that are automatically evoked without conscious processing and believed related to sustained attention, ERP's of fencers, pentathletes and clay pigeon shooters showed a predominance of the right hemisphere. They also noted that the time of day, noise and menstrual cycles of the females affected the speed of processing of information.


In studies of endogenous, responses that represent mental processing, the development of strategies or styles, ERP's of athletes Rossi and Zani (1990) found differences in information processing between specialist in various sport disciplines. They suggested that this demonstrates that the attention styles of athletes are a function of their experience in the sport. For example, younger athletes have less ability to make use of information and have more problems with programming and inhibiting motor responses than older athletes in the same sport. Based upon the amplitude and latency of N2 and P300 they suggest that skeet shooter's EEG excelled at the prediction of information while trap shooter's EEG patterns suggest constant vigilance. Both these patterns of responses are congruent with the requirements and training of their respective sports.

Konttinen and Lyytenen(1992) used the event-related potential method to assess the EEG differences (Fz, C3,C4 & Oz) in experienced rifle shooters prior to the trigger pull. They found a decrease in negativity in successful shots which they explained as being a state of lowered but optimal arousal. In a follow-up study to determine whether the results were due to the actual movement needed to control the rifle or due to the aiming task, they used inexperienced shooters in a variety of holding and aiming tasks. In summary, they found that motor activity necessary for gun stabilization was associated with slow-wave positivity while the aiming task had frontal-central negativity. These results replicate Deecke"s,et al (1984) findings for a laboratory tracking task. They suggest that the balance between movement control and aiming is dependent upon the strategy, skill and experience of the shooter .



During the Event

Only one abstract was located which reported using EEG during competition. (Fontani, Tarricone,Vigni & Zalaffi, 1989) in which they report using EEG telemetry on the prefrontal cortical areas of three female fencers during competition. They report finding that the higher frequency bands were found in the higher skill level athletes.


Sterman's work with pilots in a simulator may provide direction as to the development of valid and reliable tests which nearly duplicate the conditions of the task . One must be cautious about simulations as one can not duplicate the emotions of competition. The perception of the athlete of the importance of the competition is a prime factor in their cognitive and motor responses. Only when the EEG can be easily and reliably used in high level competition will the true nature of brain functioning under the stress of competition be determined.

Prior to Post Event EEG

Weng (1987) recorded EEG 2-3 days prior to a marathon and one-half to three and one-half hours following the marathon. He reported a reduction in the power spectrum following the race especially in the total alpha band indicating that distance running affects the CNS as well as muscular and cardiopulmonary systems.

Twelve elite wrestlers had occipital and precentral mean alpha frequency (MAF) recorded prior to and following two training competitions (Weiss,Beyer & Hanson,1991). MAF at the precentral site was increased in 20 of 24 sessions while at the occipital region MAF was increased in 22 of 24 sessions. This was viewed as a sign of higher CNS activation and corresponded with previous reported results of increased MAF for the "imagined' wrestling moves. The authors conclude that this concordance would suggest that motor imagery may well be a good model for studying activation processes in sport.


Imagery of the Event

There is a recent review of the ample evidence of the changes in EEG during the task of imaging motor skills but athletes were not used as subjects. Jennerod (1994) reviews the neurophysiological studies of movement imagery and suggest that when an individual is imaging self-performed movements both the motor and visual-spatial systems of the brain are involved.


DeBease (1989) used softball players in an attempt to determine whether doing visual versus kinesthetic imagery would result in different areas of the brain being utilized. She found more alpha in the occipital region, the visual area, as compared with the central or motor regions regardless of whether the athlete was using the visual or kinesthetic perspective. .

Beyer,Weiss,Hansen, Wolf and Seidel (1990) claim that the mean alpha frequency increased over the left occipital and pre-central areas during imagery of a swimming task. They used three imagery trials and found the second trial to produce the larges increase in mean alpha frequency. They also recommend the use of imagery to assess the mental processes in sport as it seems to represent actual sport and has a minimum of artefacts.

Wilson, Bird, Schwartz & Williams (1994) used quantitative or Q EEG to assess visual and kinesthetic imagery of a 100 metre race with elite swimmers. There were no differences in alpha between the two perspectives but females had significantly more left temporal beta while males had more right frontal beta during kinesthetic imagery. This was interpreted as the utilizing of more thinking with language by the females and more thinking with images by the males.

BIOFEEDBACK

The use of EEG biofeedback to enhance motor skills of atypical patients has been demonstrated in several studies (Birbaumer, 1997). Hemispheric changes due to biofeedback of slow cortical potentials has been demonstrated (Rockstroh, Elbert, Birbaumer & Lutzenberger, 1990).

From the previous pre-performance EEG studies which showed hemispheric asymmetries prior to the execution of a skill, Landers et al (1991) used ERP biofeedback to determine if athletes could learn hemisphere differentiation and whether this would affect sport performance. Pre-elite but experienced archers were assigned to either a correct (decrease left hemisphere activity), incorrect (decrease right hemisphere) ERP's or to a control group. Slow potential shifts were presented from the data of the few seconds prior to arrow release with a visual bar display which also had computer controls for movement artifacts. The archers had warm up trials followed by 27 data collection trials with right and left temporal electrodes. There were no differences between pre and post test performance scores for the control group. The incorrect feedback group had poorer post-treatment scores while the correct feedback group significantly improved their post treatment archery scores.

A number of authors have reported clinically using EEG biofeedback for enhancing sport performance but none were located in the research literature. Additionally, these studies reported at conferences included other mental training skills that are known to have an impact upon sport performance, eg imagery, relaxation, etc, so the contribution of EEG biofeedback to the reported athlete improvement can not be assessed.


Cautions and Summary of Section

Much of the research reported here should be considered preliminary as the methodology was not always reported and the quality of the research could not be determined. Additionally most studies had very few athletes and their selection was not always described. The diversity of countries reporting sport EEG research suggests that there is an interest and need for research into the assessment and training of athletes.

Following are summary statements of research previewed in this paper:

The baseline EEG of athletes who perform well in the stress of competition shows a higher percentage of time in alpha
The learning of motor skills changes EEG patterns.
EEG alpha increased in left frontal prior to successful performance when there is no visual-perceptual processing required

EEG alpha increases in both right and left temporal/central regions when the athlete is responding to others actions or both hands are required
EEG alpha increased when athletes moved from spread to selective attention
Low frequencies dominated the first second of the readiness period while beta dominated the last second of the readiness period
ERP's of athletes showed a predominance of right hemisphere activity
Different ERP measures distinguished the type of skill necessary for different rifle shooting events
Frontal-central ERP negativity is related to aiming in shooting tasks while slow wave positivity is related to gun stabilization.
In recordings during competition, higher frequency EEGs were recorded for the better athletes
Pre to post sport performance shows enhanced alpha activity
Increases in alpha are found in athletes who do sport skill imagery but only beta was different when comparing visual vs kinesthetic imagery.
ERP biofeedback was successfully learned by athletes and resulted in improved sport performance


IVSPORT CONSIDERATIONS


Athletes


It is important to properly identify the population of athletes for important characteristics that can influence EEG assessment and training. Factors to consider are motivation, attention, arousal and cognition. It is imperative that skill level be identified and verified, amount and quality of experience, as well as physical fitness level.

Sport Task Specificity

Identification of tasks must be specific to what is needed in that sport and/or specific skill. For example one does not merely assess attention, but what type of attention. The attention of a goalie in hockey has to shift from broad open focus to narrow closed focus depending on where the puck is at the moment. This is quite different from the fixed attention required in archery.

Other considerations if laboratory tasks are to be devised include who controls the signal to begin the task, is it a consistent or inconsistent probability of occurrence, the body position (luge athletes had different physiological responses to imagery depending on whether they were lying on a bed or lying on their sleds), whether internal timing is necessary and whether the response is a pre-determined response or one that depends on another person requiring you to respond to them.

The testing and training paradigm needs to consider the past and current sport training practice of the athlete. For example, when recording psycho physiological measures, including EEG, of wrestlers who were viewing or imaging a novel throw in wrestling, a significant change in parameters consistently occurred after 10 trials even through the study had 20 trials. While I thought I had found the limits of learning, the wrestlers laughed and said they had trained almost every wrestling skill in drills of 10 for their entire career. They habitually shut off after 10 trials even though the experimenter requested they do 20 trials.

Based upon 25 years of mental training elite athletes I suggest that the skill one practices in a practice setting is not the same skill one uses in competition since the intervening variable of the meaning and importance of the outcome seriously affect brain processing (attention, affect, arousal and cognition)and usually affect overt motor and behavioural responses.

EEG Recording

The setting and timing of EEG recording and training is also critical. Allen et.al.(1997) noted changes in baseline EEG by the minute. Our experience with athletes would suggest that pre-session physical activity, sleep deprivation, and current mood states are important factors to be recorded if not controlled.
. Our experience with swimmers and triathletes found that for some athletes their preferences for imaging the event were dominant over our instructions even when they were trying to be compliant with the requests. Thus, manipulation checks of what the athlete actually did do during the trial are essential if there is any cognitive demand placed on the subject.


IV. FUTURE DIRECTIONS

Perhaps pre-performance ERP laboratory studies would be the most efficacious technique for investigating the tasks of closed sport skills, such as shooting, golf, and archery. These sports also lend themselves to being assessed in the practice setting. The use of laboratory tasks designed specifically for sports, such as sport specific attentional games, need to be validated against performance in competition since there may be little correlation between laboratory tasks and field performance. If at all possible, EEG's should be taken during real competitive situations.

Data bases should establish what athletes, skills and conditions were observed. Consistency of terminology of what is a "peak' performance.(pre-determined maximum performance for the competitive season?), or "elite' performance (high consistency across time ?) or maximum performance (one time best ever?) would be helpful.

The following research questions could be readily investigated .Does delta indicate when an athlete is paying attention to internal processing (Harmony, et al, 1996) and is that beneficial or detrimental for skills, such as a gymnastics events? Can we use the R/L frontal asymmetry as an indicator of anxiety and determine who may be predisposed to poorer performance in competition and/or train individuals to change the asymmetry? Can we identify baseline EEG profiles that differentiate the various skill levels within the sport? Is there a different brain signature of elite performers by gender?

I envision the future of EEG in sport to include full psycho physiological assessment of who would make a good athlete (compared to a data base) , EEG training to enhance task and character requirements and on-site telemetry evaluation of performance similar to how videos are now provided players following plays in football. I believe the training of EEG will be done by the athletes who will use portable trainers .Computers will integrate all aspects of their performance into a profile of personal performance including physiological measures such as strength, biomechanical measures such as segment positioning, behavioural measures such as the number and quality of repetitions, as well as psycho physiological measures of how their body is responding in all of its modalities.

Our new Electronic Toll Road (ETR) electronically monitors a car entering and leaving the freeway and calculates the distance by cost and sends you a monthly statement. I am sure the authorities also know It could also be used to charge you with speeding and add a fine! One day we may have ETR's in the gym that telemetrically monitor whether your brain came to the gym with you(fines?). How about the office- no brain, no pay?

A philosophic question then arises. By doing assessment and training with EEG for the expressed purpose of improving performance, are we not providing a direction for the meaning of sport? Is this what we want and what will it cost, in human as well as financial terms?