movement systems as dynamical systems

Sports Med 2003; 33 (4): 245-260LEADING ARTICLE 0112-1642/03/0004-0245/$30.00/0

© Adis Data Information BV 2003. All rights reserved.

Movement Systems asDynamical SystemsThe Functional Role of Variability and its Implications forSports Medicine

Keith Davids,1 Paul Glazier,2 Duarte Araujo3 and Roger Bartlett4

1 School of Physical Education, University of Otago, Dunedin, New Zealand2 School of Sport, P.E. and Recreation, University of Wales Institute, Cardiff, Wales3 Faculty of Human Kinetics, Technical University of Lisbon, Lisbon, Portugal4 Centre for Sport and Exercise Science, Sheffield Hallam University, Sheffield, England

In recent years, concepts and tools from dynamical systems theory have beenAbstractsuccessfully applied to the study of movement systems, contradicting traditionalviews of variability as noise or error. From this perspective, it is apparent thatvariability in movement systems is omnipresent and unavoidable due to thedistinct constraints that shape each individual’s behaviour. In this position paper,it is argued that trial-to-trial movement variations within individuals and perform-ance differences observed between individuals may be best interpreted as attemptsto exploit the variability that is inherent within and between biological systems.That is, variability in movement systems helps individuals adapt to the uniqueconstraints (personal, task and environmental) impinging on them across differenttimescales. We examine the implications of these ideas for sports medicine, by: (i)focusing on intra-individual variability in postural control to exemplifywithin-individual real-time adaptations to changing informational constraints inthe performance environment; and (ii) interpreting recent evidence on the role ofthe angiotensin-converting enzyme gene as a genetic (developmental) constrainton individual differences in physical performance.

The implementation of a dynamical systems theoretical interpretation ofvariability in movement systems signals a need to re-evaluate the ubiquitousinfluence of the traditional ‘medical model’ in interpreting motor behaviour andperformance constrained by disease or injury to the movement system. Accord-ingly, there is a need to develop new tools for providing individualised plots ofmotor behaviour and performance as a function of key constraints. Coordinationprofiling is proposed as one such alternative approach for interpreting the variabil-ity and stability demonstrated by individuals as they attempt to construct function-al, goal-directed patterns of motor behaviour during each unique performance.Finally, the relative contribution of genes and training to between-individualperformance variation is highlighted, with the conclusion that dynamical systemstheory provides an appropriate multidisciplinary theoretical framework to explaintheir interaction in supporting physical performance.

246 Davids et al.

In the last 20 years, ideas from scientific para- letes during practice or rehabilitation.[12] Further-more, there is growing evidence that the nature ofdigms such as chaos theory and the sciences ofmovement variability is driven by the interaction ofcomplexity have been integrated with concepts andthe various sources of constraints on action, and thistools from dynamical systems theory to re-shape ourleads to the uniqueness of system dynamics for aunderstanding of movement behaviour.[1,2] Dynami-particular performer under a specific set of taskcal systems theory has been successfully applied toconstraints. This task-specific view may provide athe study of coordination in nervous systems andbetter framework for understanding the role of inter-movement control,[3-5] movement development[6-9]

and intra-individual variability in the provision ofand skill acquisition.[2,10] In particular, the dynami-diagnoses and treatment interventions in humancal systems framework has influenced the way thatmovement by sports medicine specialists.movement scientists view inter- and intra-individual

This paper provides a brief overview of the appli-variability in motor performance, as a function ofcation of dynamical systems theory to the study oflearning and development across a lifespan.motor behaviour before making inferences about theTraditional approaches to the study of motorneed to revise understanding of inter- and in-behaviour have operationalised variability withtra-individual levels of variability in sport and exer-measures of variance in motor output (e.g. standardcise, and health behaviours for sports medicine. Wedeviation around the distribution mean of a depend-draw parallels between the biological determinisment variable measured over repeated trials). In theimplicit in traditional conceptions of variability ascognitive sciences, for example, the search fornoise superimposed on the production of idealisedmotor invariance led to a narrow interpretation ofmotor representations, and the biological determin-variability in motor behaviour as evidence of noiseism prevailing in some medical science interpreta-or random fluctuations at different levels of thetions of perceptual-motor disorders and the role ofmovement system (e.g. anatomical, mechanical,genes in constraining health and physical perform-physiological). From a dynamical systems theoreti-ance.cal perspective, variability in movement systems is

It will become clear that a broader understandingconsidered to be a central theoretical issue worthy ofof individual variability, including performancestudy in its own right, and has been related to thevariability, requires careful interpretation of data bysensorimotor equivalence that arises from the abun-sports clinicians, since it can be functional in facili-dance of motor system degrees of freedom charac-tating adaptations of individual biological systemsterising the human body.[11]

to the changing constraints of dynamic environ-From this new perspective, variability of per- ments. To contextualise discussion, we selectively

formance has been viewed as more functional, since focus on two important areas for sports medicine: (i)a consistent outcome can be achieved by different postural control; and (ii), the relative role of genespatterns of joint relations owing to the dynamics of and the environment in constraining inter-individualthe joint biomechanical degrees of freedom (DOF). variations in health and performance. We evaluateFor example, a defining feature of a chaotic system the implications of this new view of variability foris that deterministic processes can drive fluctuations the ubiquitous ‘medical model’ in clinical assess-in system output that apparently seem random. In ments of health and movement behaviours.principle, a range of deterministic and stochasticprocesses could contribute to the observed fluctua- 1. Dynamical Systems Theory and thetions in movement and its outcome. With such a Study of Movement Behaviourview, noise may have a positive role in preventing asystem from becoming too stable in complex envi- Dynamical systems theory is a multidisciplinary,ronments so that functional movement solutions systems-led approach, encompassing mathematics,may be found during exploratory behaviour of ath- physics, biology, psychology and chemistry, to des-

© Adis Data Information BV 2003. All rights reserved. Sports Med 2003; 33 (4)

Movement Systems as Dynamical Systems 247

cribe systems that are constantly changing and biological systems searching for optimal states ofevolving over different timescales.[13] A central ten- organisation.[19,20] Constraints reduce the number ofet of dynamical systems theory is that natural phe- configurations available to a dynamical system bynomena can be explained, at multiple scales of ana- structuring the state space of all possible configura-lysis, with the same underlying abstract principles tions available to it. There are many classes ofregardless of system structure and composition. constraints that can shape the behaviour of a dynam-

ical system and Newell[21] categorised them as orga-nismic (pertaining to the individual), the task and the1.1 Movement Systems asenvironment (see figure 1; note that this paper advo-Dynamical Systemscates an adaptation of the model to the study of

The study of movement systems as dynamical human behaviour including movement coordina-systems contains two strands of modelling work. tion).The first approach investigates information-forcetransactions between a performer and the environ- 2. Bernstein’s (1967) Problemment using the analytical tools of nonequilibrium

Key ideas from dynamical systems theory havethermodynamics.[14,15] Kinematic patterns in move-been allied to the theoretical insights of the pioneer-ment systems are viewed as the product of forceing Russian physiologist and biomechanist, Nikolai(kinetic) fields, which lawfully give rise to flowBernstein who formulated the fundamental problem(informational) fields. This programme of work exa-for movement systems as “the process of masteringmines coordination as an emergent process in dy-the redundant degrees of freedom” or more succinct-namic environments, based on the intentions of thely “the organisation of the control of the motorperformer.[16,17] A second, related research line isapparatus”.[22] Bernstein used the term redundantfounded on ‘pattern dynamics’ which initiallyDOF to refer to the (bio)mechanical DOF that ex-sought applications of synergetics and nonlinearceed the minimum number required to successfullydynamics to human movement.[18] The goal of thisaccomplish any given motor task.[23] Newell andprogramme is to construct dynamical equations ofVaillancourt[24] proposed that DOF are the “numbermotion of relevant coordination phenomena thatof independent coordinates required to uniquelycapture the stability and loss of stability associateddescribe the configuration of a system”. For in-with persistence and change in movement systems,stance, there are seven DOF in the human arm, threefor example occurring during performance transi-at the shoulder, one at the elbow and three at thetions as well as longer-term changes due to learning,wrist, far more than needed for most typical inter-development or rehabilitation in individuals.ceptive arm movements, providing flexibility andIn dynamical systems, spontaneous pattern for-security in adapting to changing environmental con-mation between component parts has been found toditions.[25,26]emerge through processes of self-organisation. Such

systems are typically ‘open’ thermodynamic sys-tems engaged in constant energy transactions withthe environment.[17] Self-organisation is manifestedas transitions between different organisational statesemerging due to internal and external constraintspressurising system components into change.

The concept of constraints from dynamical sys-tems theory perhaps has the greatest implications fora new view of individual variability for sportsmedicine. Constraints have been defined as bounda-ries or features, which interact to limit the form of

Task

Environment Organism

Physicalperformance

Action(movement)

Perception(information)

Fig. 1. Newell’s model of interacting constraints adapted to illus-trate the resulting effects on variability of physical performance.


248 Davids et al.

At different scales of analysis, the human move- vided the foundations for substantial theorising andexperimentation in motor control. Kelso andment system is replete with redundancy. For exam-Schoner[33] argued that, contrary to traditional theo-ple, molecular biologists have found that part of theretical emphases, a greater understanding of theDNA code specifies which proteins (long strings ofbehaviour of human movement systems qua com-amino acids) are made during development. Twentyplex biological systems would be revealed in or-different kinds of amino acids have been found andganisational principles rather than hard-wired mech-it appears that DNA code is read in groups ofanisms or psychological constructs such as ‘repre-consecutive nucleotides, arranged into triplets. Forsentations’ encoded within the central nervouseach triplet there is one amino acid. With 64 possi-system. Variability of movement has a functionalble triplets and 20 amino acids, there is redundancyrole in selecting for the coordinative structures thateven at this molecular level of the human body (i.e.emerge under constraint as less functional states ofmore than one nucleotide triplet matches the sameorganisation in the motor system are explored andamino acid).[27]

abandoned.[2,17,19,28,34] This retreat from biologicalAt the behavioural level, Bernstein’s initial solu-

determinism recognises that biological systems aretion to the redundant DOF problem was that humans

required to generate both stable (persistent) andeliminated a portion of the DOF by “rigidly and

flexible (variable) behavioural output in response tospastically fixing”[22] joint articulations to restrict changing intentions and dynamic environmentalsegmental rotation and translation. Movement sys- conditions, and signals a new way for understandingtems cope with the redundant DOF by introducing variability observed in movement systems.temporarily strong, rigid couplings between multi-ple DOF resulting in more controllable single ‘virtu-al’ DOF complexes termed ‘coordinative struc- 3. Motor System Variabilitytures’.[28] Coordinative structures constrain the inter-connections made by parts of the movement system As we noted in the introductory section, an im-during functional movements. They have been de- portant theme in dynamical systems accounts offined by Kay[29] as “…an assemblage of many human movement behaviour is that careful infer-microcomponents …assembled temporarily and ences need to be drawn about the variability ob-flexibly, so that a single microcomponent may par- served in movement behaviour by clinicians andticipate in many different coordinative structures on sport scientists.[12,24,35] In some parts of the system,different occasions”. Coordinative structures reduce higher levels of variability can actually reflect sub-the dimensionality (i.e. complexity) of the dynami- conscious compensatory measures by individuals.cal movement system by allowing humans to exploit In sport, compensatory variability can be observedthe inherent interconnectedness of the anatomical in performers who are skilful at exploiting the highsystem. An essential feature of a coordinative struc- dimensionality offered by the many motor systemture is that if one of the component parts introduces DOF. Evidence shows that skilled performers canan error into the common output, the other compo- freeze or unfreeze the DOF in the chain of move-nents automatically vary their contribution to movement as the prevailing task constraints demand.[24]

ment organisation and minimise the original er- Less skilled performers, in comparison, tend to rig-ror.[30] Flexibility in adapting to local conditions is idly fix DOF and may show as much or moreenhanced by the capacity of parameters of coordina- variability that is not functional, owing to weaktive structures to be tuned by information so that adaptation to task constraints. Furthermore, throughmovement goals can be achieved.[28]

ageing and illness, human movement systems tendDespite recent claims that the original problem of to show a loss of complexity, defined as inadequate

motor redundancy was itself inadequately formulat- adaptation to environmental changes and hence re-ed,[30-32] Bernstein’s preliminary insights have pro- duced functionality.[24]



An example of functional variability was report- to reproduce invariant movement patterns despiteed in data on the shooting performance of skilled years of practice.and unskilled marksmen by Arutyunyan and col- Schoellhorn and Bauer[39,40] developed a patternleagues.[36] They identified different levels of vari- recognition algorithm for identifying individualability in each joint of the upper arm in skilled movement patterns for various types of sport andperformers. This finding has since been replicated everyday life tasks based on the production of non-and interpreted as contributing to success in the task,

linear, categorising, self-organised neural networksince the gun and performer’s motor apparatus were

maps. This method was used to qualitatively analyseviewed as forming a complex system.[37] Higher

intra- and inter-individual levels of performancelevels of variability in the shoulder and elbow jointsvariability in athletes during complex sport actions.complemented each other to allow the wrist (andThe algorithm effectively measures the ‘space’ be-therefore gun) to maintain a stable position. Thetween separate performances within and betweeninterface of the gun with the hand provided theeach individual. A smaller amount of space betweenlowest magnitude of intra-trial variability, fitting theperformances could be interpreted as signifyingneed to be very stable at this location. The same highlower levels of inter- or intra-individual variability,levels of variability were not seen in the shoulderdepending on whether the analysis was betweenand elbow joints of unskilled shooters, with thedifferent athletes or over repeated trials for the sameconsequence that the pistol position remained unsta-performer. In this way, clusters of performancesble and more variable during shooting.could be picked up, providing a qualitative measureThe findings of the studies on pistol shoot-of variability in movement patterns. Performanceing[36,37] provide a powerful rationale for cliniciansclusters of higher density signified lower levels ofand coaches to consider functional variability ofvariability, with lower density of clusters implyingmotor behaviour as a key criterion of successfulgreater levels of variability.performance, rather than the ability to replicate an

ideal movement template or common optimal motor In one study,[40] two discus throwers were filmedpattern.[38] That is, motor patterns emerge under with high-speed cameras during several training anddifferent task constraints to achieve stable task out- competition periods. The result was 45 trials forcomes, and do not appear as pre-determined, invari- analysis by a decathlete and eight trials for analysisant, anatomical assemblages. Kugler and Turvey[17]

by a specialist discus thrower. Data were obtainedargued that descriptions of goal-directed movement on positions and angular velocities of all body andbehaviour should focus on the functions of an ac-

limb joints from the point of the last step of thetion, not on the specific anatomical units involved

athlete to the moment the discus left the hand. Thebecause “… an act is functionally specific, its vari-

data revealed clustering across different training andability is in reference to preserving the function itcompetitive sessions for each individual. The move-fulfils rather than preserving any particular aggrega-ments were more similar during each individualtion of body parts that it happens to involve”.session and varied between different sessions. Thesefindings belie the traditional view of expert perform-4. Coordination Profilingance being characterised by invariant features.Clearly, even elite athletes come to training andThe new view of variability signals that the motorcompetition sessions unable to reproduce identicalsystem’s inherent noisiness results in variability be-movement patterns despite years of practice. In an-ing omnipresent, unavoidable, and yet functional inother study by Bauer and Schoellhorn[40] higherhelping to produce stable movement outcomes. De-levels of inter-individual variation were found with-spite the search for invariance in cognitive sciencein the clusters of the international athletes, com-studies of motor control over the past decades, stud-pared with the national athletes, rejecting the idea ofies have revealed that even elite athletes are unable


250 Davids et al.

the existence of common optimal movement pat- on sport performance suggest that practitionersshould engage in ‘coordination profiling’, and resistterns in highly skilled performers.interpreting performance data in terms of proximityFurther evidence of both inherent and functionalof individuals to a perceived common optimal motorvariability has emerged from studies of interceptivepattern.[38] Coordination profiling refers to the use ofactions in sport such as long jumping, triple jump-individualised, in-depth analyses to examine howing, and table tennis reviewed by Davids and col-each individual performer uniquely satisfies specificleagues.[41] In long jumping, for example, the litera-task constraints during goal-directed behaviour.[43] Itture suggests that both expert and novice performersrecognises that individuals approach performance,exhibit a two-phase approach strategy that includes:training, practice, and rehabilitation with distinct(i) an initial acceleration phase during which ath-intrinsic movement system dynamics shaped byletes attempt to maintain a stereotypical stride pat-many important constraints. With the increasing ca-tern while progressively increasing their stridepacity for in-depth analysis of movement coordina-length as they accelerate down the track; followedtion brought about by technological advances, theby (ii), a zeroing-in phase during which athletesexistence of subtle individualities or ‘signature’ pat-attempt to modify stride length parameters over theterns is becoming apparent, even in highly con-final strides. Analyses of inter-trial footfall variabil-strained tasks.[3] The powerful constraints on indi-

ity in relation to the take-off board have revealed anvidual performance variation underlines why the

ascending-descending trend that corresponds withstatistical technique of pooling individual data in the

the two phases of the run-up. During the accelera-study of motor coordination (i.e. reporting group

tion phase, small inconsistencies in stride length,means and standard deviations) may have limited

representative of inherent variability effects invalue since, according to Kelso,[3] “one might as

motor systems, are accumulated until approximatelywell average apples and oranges”.

4 strides before the take-off board. After optimalhorizontal velocity has been reached, visual control

5. Coordination Profiling and thetakes over during the zeroing-in phase and strideMedical Modellength is regulated to remove the initial variability

created during the acceleration phase, indicating that Parallels can be drawn between the sports sciencethe variability observed in the last four strides is and medical literature. The misconception of ‘com-functional. This approach has also been observed in mon optimal movement patterns’ also exists in theother sports such as golf. In a study examining the study of perceptual-motor disorders and has impli-variability of ground reaction forces during the cations for sports medicine. For example, in the areabackswing, Koenig and coworkers[42] reported that of disability studies, the implicit ‘medical model’ orall golfers (n = 14) regardless of handicap (range ‘disability as tragedy model’, used by many clini-from 1–18) showed increasing variability of force cians provides a unitary, biologically determinedproduction from the initiation of the back swing to perspective of health and movement behaviour inthe mid-point of the downswing, followed by a which variability, viewed as deviation from an ‘ac-decrease during the impact phase, and a further cepted’ norm, is seen as dysfunctional and an indexincrease in the follow-through. of abnormality.[44] In this model, health and per-

In summary, it seems that increasing expertise formance behaviours are identified as problems fordoes not lead to movement invariance and the con- the individual since they may deviate from what isstruction of a single, pre-determined motor pattern, perceived as population norms. Rather, the alterna-as argued in cognitive science theories of motor tive view prompted by dynamical systems theory iscontrol. This misconception is an example of the that variability in behaviour may be viewed as anbiological determinism that influences much theory adaptation to unique organismic constraints such asand experimentation in human behaviour.[27] Data genes, motor system structure or personality. Vari-



ability has a functional role in helping individuals In the concluding sections of this paper, we ex-amine the implications of these ideas for sportsadapt to ever-changing constraints imposed on themmedicine by examining the study of postural controlby environmental, anatomical and physiologicaland for assessing the genetic basis of physical per-changes due to disease, illness, injury and ageing.formance.An implication of this view for clinicians is that

behaviours exist on a spectrum characterising theboundaries of naturally occurring variability. 6. Interacting Constraints on

Postural ControlThe terms ‘impaired’ and ‘elite’ performance insport and exercise need to be understood in relation

During the past decade, an increasing number ofto the constraints on each individual.[45] This isinvestigations into postural control have been con-

because the precise location that a movement systemducted within a dynamical systems framework,[46-48]

inhabits on the performance spectrum emerges fromwith an important line of enquiry attempting to

the multitude of constraints pressurising it at specif-determine how different sensory systems (i.e. visual,

ic points during the lifespan. Since the constraints onacoustic, haptic, proprioceptive) interact and con-

each individual are many and unique, it follows thattribute to postural behaviour.[49,50] In many studies,

coordination solutions will differ within and be-movement scientists and clinicians have typically

tween individuals in order to maintain functionality.used a ‘constraints-led’ approach whereby different

As Latash and Anson[45] note, the “phenomena ofinformational and task constraints have been experi-

variability of voluntary movements by themselvesmentally manipulated to examine their effects on

indicate that ‘correct’ peripheral motor patterns may postural control in individuals with and withoutform a rather wide spectrum”. This fundamental movement disorders.[51] Controlling posture andprinciple applies to healthy individuals across the maintaining balance whilst performing rudimentarylifespan as well as those with injuries, diseases and motor tasks such as standing or walking is a majorperceptual-motor disorders. Latash and Anson[45]

challenge for many elderly people and for patientsproposed that adaptations to constraints should not with movement disorders. The potential for severenecessarily be perceived as pathological since motor injury or even fatality caused by accidental fallingpatterns may be optimal for the conditions affecting due to a loss of postural stability has promptedthe individual’s motor system at any given point in extensive analysis of the task and individual con-time. Typically, although motor patterns in individ- straints on the regulation of posture.[52-54]

uals with cognitive, perceptual and motor deficits One example of a ‘constraints-led’ approach wasmay be defined by some clinicians as ‘abnormal’ proposed in a study by Davids et al.[47] which exa-compared with ‘common optimal motor patterns’ mined the interaction of task and informational con-idealised in the ‘medical model’, they may be better straints on postural behaviour in healthy individualsviewed as functional and emergent under the conflu- and patients with anterior cruciate ligament (ACL)ence of constraints that each individual needs to deficiency. The ACLs represent an important part ofsatisfy. the knee joint complex as they provide mechanical

Therefore, treatment interventions in sports stability to prevent anterior displacement of the tibiamedicine should not be directed towards the on the femur and proprioceptive information aboutachievement and maintenance of an ‘ideal’ motor postural sway.[55-58] Postural control was assessedpattern. This is not the point of therapy or rehabilita- using a static postural sway meter and a dynamiction programmes. The overarching aim of an inter- stabilometer under six different conditions: standingvention should be to help individuals satisfy the on both legs with eyes open and closed; standing onunique conflagration of constraints that impinge up- the injured leg with eyes open and closed; andon them in order to improve their functional capaci- standing on the non-injured leg with eyes open andty in the performance contexts they face. closed.[47] Under static task conditions, results show-


252 Davids et al.

ed that proprioceptive information was not a majorfactor in postural control (figure 2 and figure 3).Under more dynamic task constraints, significantdifferences in postural control between the ACL-deficient group and the control group were observedonly when vision was unavailable suggesting thatvision is the most important source of expropriocep-tive information. Clearly, these findings have sub-stantial practical implications for the design ofclinical tests to assess functional postural behav-iours.

Some interesting findings on centre of pressure(COP) dynamics from the postural sway meter werealso reported.[47] In postural control research, COPprofiles have been routinely used to indicate postu-ral stability as they are related to postural sway.[59]

Davids et al.[47] found that the mean and maximumCOP velocity of the control group was greater thanthat of the ACL-deficient group (see figure 4 andfigure 5). According to previous research, these

ACLDCON

02468

10121416182022242628

Conditions

Mea

n nu

mbe

r of

con

tact

s

BO BC IO IC NIO NIC

Fig. 3. Data on the number of contacts with the ground on thestabilometer under the dynamic task constraints as a function oftype of stance and informational constraints (reproduced from Da-vids et al.,[47] with permission). ACLD = anterior cruciate ligamentdeficient group; BC = standing on both feet with eyes closed; BO =standing on both feet with eyes open; CON = control group; IC =standing on the injured leg with eyes closed; IO = standing on theinjured leg with eyes open; NIC = standing on the non-injured legwith eyes closed; NIO = standing on the non-injured leg with eyesopen.

contradictory findings provide an indication ofgreater postural stability in the ACL-deficient groupthan in the controls since COP velocity is consideredto have an inverse relationship with the ability tocontrol posture.[60,61] Alternative explanations ofparticipants reducing postural sway in order to re-duce pain or injury were discounted because therewere no reports of discomfort or self-concern duringthe performance of this static balancing task. Davidset al.[47] suggested that the higher COP velocity inthe control group should not be interpreted asgreater instability but rather as indicative of normalexploratory behaviour (i.e. generation of posturalsway) in discovering stable solutions to the posturalcontrol problem.[49,50,62,63]

This line of reasoning has also been followed in anumber of other studies investigating the relation-ship between COP variability and stability in indi-viduals with movement disorders such as Parkin-son’s disease.[51] A feature of individuals withParkinson’s disease is the inherent lack of posturalstability and the proneness to accidental falling. Animplicit assumption of much previous research is

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

BO BC IO IC NIO NIC

Conditions

Pos

tura

l sw

ay (

degr

ees)

ACLDCON

Fig. 2. Whole body sway data under the static task constraints ofthe postural sway meter as a function of type of stance and infor-mational constraints (reproduced from Davids et al.,[47] with permis-sion). ACLD = anterior cruciate ligament deficient group; BC =standing on both feet with eyes closed; BO = standing on both feetwith eyes open; CON = control group; IC = standing on the injuredleg with eyes closed; IO = standing on the injured leg with eyesopen; NIC = standing on the non-injured leg with eyes closed; NIO= standing on the non-injured leg with eyes open.



0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

Conditions

Vel

ocity

(m

/sec

)

ACLDCON

BO BC IO IC NIO NIC

Fig. 4. Mean centre of pressure velocity data under the static taskconstraints of the postural sway meter as a function of type ofstance and informational constraints (reproduced from Davids etal.,[47] with permission). ACLD = anterior cruciate ligament deficientgroup; BC = standing on both feet with eyes closed; BO = standingon both feet with eyes open; CON = control group; IC = standing onthe injured leg with eyes closed; IO = standing on the injured legwith eyes open; NIC = standing on the non-injured leg with eyesclosed; NIO = standing on the non-injured leg with eyes open.

ACLDCON

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

ConditionsV

eloc

ity (

m/s

ec)

BO BC IO IC NIO NIC

Fig. 5. Maximum centre of pressure velocity data under the statictask constraints of the postural sway meter as a function of type ofstance and informational constraints (reproduced from Davids etal.,[47] with permission). ACLD = anterior cruciate ligament deficientgroup; BC = standing on both feet with eyes closed; BO = standingon both feet with eyes open; CON = control group; IC = standing onthe injured leg with eyes closed; IO = standing on the injured legwith eyes open; NIC = standing on the non-injured leg with eyesclosed; NIO = standing on the non-injured leg with eyes open.

that postural instability is associated with large and 7. Genes and the Environment asConstraints on Human Behaviourhighly variable COP oscillations.[64] However, in-

vestigations conducted within a dynamical systemsIn science, perhaps the most controversial ana-framework suggest that under certain task con-

lysis about variability in human behaviour concernsstraints healthy individuals exhibit greater COP os-the relative influence of genetic and environmental

cillations than individuals with Parkinson’s dis- influences. How these constraints shape variationsease.[65,66] These findings suggest that movement in human performance is a question of increasing

interest in sports medicine. This problem is a mani-variability plays a functional role in the detectionfestation of the longstanding nature-nurture debateand exploration of stability boundaries during bipe-that has dogged biological science, education, soci-dal stance and locomotion.[67,68] Additionally, higherology, philosophy and psychology for many years,

levels of postural sway variability, found in partici- meaning the identification of the precise proportionpants with eyes closed (compared with eyes open), of performance variation in a population accountedindicated greater use of perceptual information in for by genetic characteristics or environmental in-

fluences. Much has been written about this particu-order to control posture and reduce effects of ran-lar dualism in science, and resolution of the debatedom fluctuations in the musculature of the posturalhas proved difficult. For excellent analyses see

control system.[67,68] Riley and Turvey’s[12] com-Lewontin,[27] and Johnston and Edwards.[69]

ments on these findings succinctly capture the impli-From a dynamical systems perspective, it is clear

cation for sports medicine: “More variable does not that both nature and nurture can act as constraints onmean more random, and more controllable does not behaviour, although there have been few attempts toalways mean more deterministic”. determine how organismic constraints of a genetic


254 Davids et al.

basis interact with environmental or task constraints, receiving considerable attention in the sportsmedicine literature and in section 8 we evaluate theas predicted by Newell’s (1986) model. Neverthe-evidence for its role as a genetic constraint on varia-less, it is possible to interpret existing data of rele-tion in physical endurance performance.vance to sports medicine, based on the interaction of

genetic and environmental constraints. For example,8. The Angiotensin-Converting Enzymethere is growing consensus in the study of human(ACE) Geneobesity that the contribution of genetic factors is

exacerbated in environments that differ in cultural The ACE gene is one of a number of candidateconstraints including caloric availability.[70] Genetic genes associated by research with inter-individualpropensity towards adiposity has less of a con- variability in physical endurance performance.[76-85]

straining influence on individuals in environments In muscle, the angiotensin I-converting enzyme haswhere caloric availability was lower, whereas these the role of degrading vasodilators, bradykinin andsame individuals would be at greater risk in other tachykinin, and stimulating production of the vaso-environments. Such environments can be cat- constrictor angiotensin II during physical perform-egorised as high or low risk, depending on the ance.[86] Sequence variation in DNA has led to theprevalence of other significant cultural constraints existence of two alleles of the ACE gene, I and D (Iincluding work patterns imposed on traditional meal for insertion and D for deletion), which combine totimes, and popularity of non-physically active pas- produce polymorphisms in genotype (II, ID, DD).times such as computer games and TV watching. The D variant of the gene has been linked withThus, the effects of the interaction of genes and relatively higher ACE activity. For example, earlyenvironment on phenotypic expression of behaviour work with army recruits found that the II polymor-can be best understood at the level of individual risk, phism of the gene has been associated with lowerrather than as defective behaviour. activity levels of ACE in muscle than the DD allele,

and an increased response to physical training.[87]The past few years have also seen increasingRecruits with ACE genotype II differed by as muchwork on the role of genes in defining the level ofas 1100% in response to repetitive upper-arm exer-athletic performance attainable by individuals. Incises compared with DD genotype peers. Individu-the main, research has focused on genetic and envi-als with a heterogeneous genotype (DI) were associ-ronmental contributions to physical, typically en-ated with levels of performance between both homo-durance, performance[71] although there have alsozygous genotypes. In sport, a higher prevalence ofbeen some attempts to evaluate relative contribu-the II genotype has been found in elite endurancetions to acquisition of motor skill.[72] The significantathletes including mountaineers able to climb toamount of inter-individual variation observed in re-7000m without the aid of oxygen, Olympic-standardsponse to training of the cardiovascular system hasendurance runners and elite rowers.[78,87,88]led to many investigators questioning the extent to

which genetic diversity may be responsible for the8.1 The ACE Gene: Some Critical Issuesdata.[73]

The search for the genetic basis of many human Although some work on endurance performancecapacities, such as physical performance, has engen- in elite athletes has failed to support the more func-dered strong rhetoric in some quarters, with some tional role of the I allele of the ACE gene,[80] themolecular biologists calling it the ‘biological coun- cohort of athletes in that study consisted of 120terpart to the holy grail’,[74] and some sport scientists performers chosen from a variety of sports withasserting that genes are responsible for up to half of different task constraints which possibly maskedthe variation in physical performance between indi- any observed effects (including 26 hockey players,viduals within a population.[75] Currently, the role of 25 cyclists, 21 skiers, 15 track and field athletes, 13the angiotensin-converting enzyme (ACE) gene is swimmers, 7 rowers and 5 gymnasts). Such a mixed



group of athletes may not have had the requisite ly). Nevertheless, future research needs to ascertainwhether: (i) the effect of the I allele of the ACE genelevels of phenotypic homogeneity to lead to validon endurance performance is mediated via periph-estimates of the genetic basis of performance. More-eral muscle effects and changes in efficiency; andover, it has become clear that carriers of the D allele(ii) the effect of the D allele on performance underappear to have an advantage in training and per-power task constraints is mediated via increasedformance when task constraints emphasise powerangiotensin II acting as a local hypertrophic factor inover a shorter duration.[78,79] In fact, the D allele hasmuscle.been related to increased quadriceps muscle strength

gains following 9 weeks of isometric training.[89] It In interpreting the data from studies of the genet-is possible that the D allele may confer some per- ic constraints on physical performance and motorformance and training benefit in task constraints skill acquisition, there is enormous potential forrequiring power (perhaps through its effect on confusion amidst the rhetoric. On the face of it, thegreater angiotensin II genotype and muscle hyper- data in favour of a strong genetic constraint ontrophy), and similarly the I allele an effect under physical performance seem compelling. Whilstendurance task constraints. The implication is that most researchers working in the area of geneticvariability at the level of individual genes provides variations in human performance agree with Hop-functionality and adaptability in movement systems kins’[75] opinion that athletes are born and made, aneeding to perform a variety of activities in a com- clear interpretation of the data on the ACE gene isplex environment. This suggestion also emphasises needed to understand how athletic performancethat, in experiments on the ACE gene variants and emerges under interacting constraints.sport performance, a clear understanding of differ- An example of the need for care in drawingences in task constraints is needed to ensure homog- appropriate conclusions from data, in the face of theenous cohorts of athletes are carefully examined to rhetoric that human physical performance is strong-avoid the loss of genetic association. ly influenced by genetic factors,[78] was provided in

a study by Bouchard and colleagues.[91] They at-Finally, research on the ACE gene is progressingtempted to estimate the proportion of influence at-rapidly and there are some indications that its role intributable to genetic and environmental constraintsconstraining physical performance may be some-on familial resemblance for VO2max during exercisewhat different than originally perceived. For exam-on a cycle ergometer in sedentary individuals.[91] Forple, there has been some doubt cast on the relation-this purpose, exercise performance of fathers,ship between the presence of the I allele of the ACEmothers, sons and daughters was measured in 86gene and the responsiveness to endurance trainingnuclear families. Maximum heritability includingoriginally proposed in some studies.[87,88,90] The lo-genetic and non-genetic causes for physical per-cus of the ACE gene has been identified as chromo-formance accounted for 51% of the total adjustedsome 17q23 and genomic scanning for presence ofphenotype variance. Several models of interactingcandidate genes for baseline maximum oxygen up-constraints were tested and results showed that theretake (VO2max) performance or responsiveness towas 2.6–2.9 times more variance between familiestraining failed to confirm evidence of linkage.[91]

than within families.These findings on a sedentary population were sup-ported by a frequency analysis, which failed to find Unfortunately, the approach taken in this studya relationship between the accumulation of alleles I by Bouchard et al.[91] meant that genetic and familialand II and endurance performance in 192 elite ath- environmental influences could not be fully quanti-letes (skiers, runners and cyclists) and 189 con- fied separately although “inferences about their re-trols.[71] Interestingly, the highest frequencies respective contributions to the phenotype varianceported for both the elite athlete group and controls could be made by inspection of the pattern of famili-were for the ID genotype (0.46 and 0.47, respective- al correlations”. It is important to note here that


256 Davids et al.

correlations do not imply causation, and that the the most ardent geneticists agree that transmissionfocus on the constraints imposed by the shared fa- of genetic information between generations is lessmilial environment only, precludes the influence of than perfect.[92] DNA is simply a copy of informa-wider environmental constraints, such as socio-cul- tion transmitted between generations, which is readtural changes in society, including impact of media by cellular machinery in the production of proteinsimages, government education programmes, and that create the individual, part by part. This is impor-peer group pressure, being factored into the equa- tant to note because the view of DNA as informationtion. bearer has been replaced with the fallacy of DNA as

blueprint, plan, or master molecule.[27]Moreover, conclusions on the contribution of thegenetic component to performance were somewhatspeculative. Despite the fact that maximal heritabili- 8.3 Genes and Variability inties reported in this study were inflated by familial Movement Systemsnon-genetic contributions, the effects of maternal

The presence of genetic material should not betransmission of mitochondrial DNA to the fertilisedviewed as a blueprint for success in sport. As John-zygote were proposed as being optimally allied toston and Edwards[69] have pointed out, it is "a verythe father’s environmental contribution. Bouchardlong step from polypeptide sequences to behaviour –et al.[91] argued that the data ‘revealed’ that “mater-a step …that covers much incompletely understoodnal influence, perhaps by mitochondrial inheritance,territory". Attempts to see genes as building plans oraccounts for as much as 30% of the familial trans-coded blueprints are one of the great artificialisms ofmission”. The authors’ conclusion that, “Based onhuman conceptualisations of nature.[93] It has be-the present results, we estimate that ‘mitochondrial’come a ‘central dogma’ of how people think aboutheritability is in the range of 30–35%”, seems athe process of evolution.[94] In relation to this argu-speculative interpretation of the data based on corre-ment, it is interesting to note that a major aspect oflational statistics, a limited range of environmentalDarwin’s legacy was to change the emphasis inconstraints impacting on model construction and noevolutionary theory from a transformational modelevidence from DNA analysis.(focus on group modal properties) to a variationalmodel (focus on variability amongst individual8.2 Problems of Interpreting Data frommembers of a species). Moreover, a major argumentGenetic Studiesagainst the blueprint conceptualisation of the role ofgenes is found in evidence that identical twins areIn order to avoid confusion over the relativenot identical! For example, evidence from the studyinfluences of genetic and environmental constraintsof phenotypically identical twins showed that theiron behaviour it is worth reiterating what is alreadyfingerprints differ and the shape of their brains canknown in this area of work. It is important to notediffer by as much as 40%.[95] This change in empha-that heritability of a trait is constrained by geneticsis was geared to the idea that the internal con-and environmental factors to some extent and thatstraints on variability (e.g. genes) are causallyresearch in behavioural genetics is concerned withdependent on the environmental constraints that se-explanations of hereditary influences at the level oflect for them.[27]populations, not individuals. As most geneticists

working on physical performance understand, genes Parallels can be drawn with the revision of under-work in combination to influence biological func- standing of the concept of movement variability intion, refuting the idea of successful athletes being dynamical systems accounts of behaviour. Lewon-differentiated on the presence of a single gene (for a tin’s criticisms of the stance of biological determin-similar argument in developmental theory see John- ism[27] is reflected in the medical model’s rejectionston and Edwards[69]). It has also become clear that of polymorphism and the implicit notion of variabil-genes are not biologically determinate, since even ity as deviation from a ‘perfect ideal’. Genetic diver-



sity is the norm and biological systems are not tion of the role of variability implies that it should beDNA-determined. There is no single, standard, nor- studied as an inherent part of motor system output,mal DNA sequence that we all share and estimates and a concomitant focus of sports medicine shouldare that we differ in DNA sequence by 0.1% (~3 be on developing techniques, such as coordinationmillion nucleotides) including inherited sequences profiling, in order to study the relationship betweenfrom parents. It takes more than DNA to produce a variability of performance and the nature of theliving organism, which cannot be computed from interacting constraints impinging on each individu-DNA sequences. According to Lewontin,[27] “A liv- al. To support our arguments, we examined some ofing organism at any moment of its life is the unique the current literature on postural control and theconsequence of a developmental history that results genetic basis of endurance performance as vehiclesfrom the interaction of and determination by internal for understanding the pervasiveness of biologicaland external forces”. determinism in traditional theories of human beha-

viour, exemplified by the search for motor invari-In summary, it seems that the evidence stronglyance or ‘common optimal motor patterns’ in sportsupports the view that genetic and environmentaland exercise science and the pervasiveness of thefactors do interact to constrain performance vari-‘medical model’ in sports medicine.ability observed between sport performers. Genetic

diversity may be responsible for a small part of The main implications for sports medicine in-training or performance response in individuals. It clude: (i) an integrated explanatory framework mayappears that the influence of genetic factors is often be more beneficial in understanding physical per-very small, and only when there is a favourable formance in sport, rather than unitary scales of ana-interaction with important environmental con- lysis, (e.g. social, psychological, physiological, mo-straints are performance benefits observed. The im- lecular); (ii) variability in movement behaviour mayplication is that elite endurance athletes of a less be functional due to the interacting constraints onfavourable genotypic disposition can succeed with behaviour; and (iii), a better comprehension ofthe appropriate training environment, but it can be human functioning (and malfunctioning) may beconcluded that performers with a more favourable obtained by describing the dynamics of differentgenotype, who interact with their training environ- sub-systems of the body and by understandingment in an appropriate manner, are more likely to macro- (performer-environment) and micro- (at thereceive a greater training response. level of muscles and joint complexes) level interac-

tions.9. Conclusions

AcknowledgementsIn this article we have highlighted how dynami-

The authors would like to acknowledge the help of Mau-cal systems theory has paved the way for a newreen Hazelwood in the preparation of this manuscript, and the

interpretation of movement variability, avoiding the comments of Major David Woods, RAMC, on a previousbiological determinism implicit in traditional theo- version of this article. No sources of funding were used tories of motor control. It is becoming clear that the assist in the preparation of this manuscript. The authors have

no conflicts of interest that are directly relevant to the contenthuman motor system is intrinsically noisy, and theof this manuscript.new perspective on movement variability is seen as

functional in allowing individuals to adapt to theReferencesmultitude of unique constraints on performance. We

1. Beek PJ, Meijer OG. On the nature of ‘the’ motor-action contro-have argued that sports clinicians need to understandversy. In: Meijer OG, Roth K, editors. Complex movement

the nature of the interacting constraints, which shape behaviour: the motor-action controversy. Amsterdam: ElsevierScience, 1988: 157-85emergent movement behaviour, warranting recon-

2. Davids K, Button C, Bennett SJ. Coordination and control ofsideration of some practices and approaches in movement in sport: an ecological approach. Champaign (IL):sports medicine. The new, more positive, interpreta- Human Kinetics, 2003


258 Davids et al.

3. Kelso JAS. Dynamic patterns: the self-organisation of brain and 25. Sparrow WA. Energetics of human activity. Champaign (IL):behaviour. Cambridge (MA): MIT Press, 1995 Human Kinetics, 2000

4. Davids K, Williams AM, Button C, et al. An integrative model- 26. Rushton SK, Wann JP. Weighted combination of size anding approach to the study of intentional movement behaviour. dispartity: a computational model for timing a ball catch. NatIn: Singer RN, Hausenblas H, Jannelle C, editors. Handbook of Neurosci 1999; 2: 186-90sport psychology. 2nd ed. New York: John Wiley & Sons, 27. Lewontin R. It ain’t necessarily so: the dream of the human2001: 144-73 genome and other illusions. London: Granta Books, 2000

5. van Emmerik REA, van Wegen EEH. On the functional aspects 28. Turvey MT. Coordination. Am Psychol 1990; 45: 938-53of variability in postural control. Exerc Sport Sci Rev 2002;

29. Kay B. The dimensionality of movement trajectories and the30: 177-83degrees of freedom problem: a tutorial. Hum Mov Sci 1988; 7:6. Thelen E, Smith LB. A dynamic systems approach to the343-64development of cognition and action. Cambridge (MA): MIT

30. Latash ML, Scholz JP, Schoner G. Motor control strategiesPress, 1994revealed in the structure of motor variability. Exerc Sport Sci7. Muchisky M, Gershkoff-Cole L, Cole E, et al. The epigeneticRev 2002; 30: 26-31landscape revisited: a dynamical interpretation. In: Rovee-

31. Gelfand IM, Latash ML. On the problem of adequate languageCollier C, Lipsitt LP, editors. Advances in infancy research.in movement science. Motor Control 1998; 2: 306-13Norwood (NJ): Ablex, 1996: 121-59

32. Latash ML. There is no motor redundancy in human move-8. Newell KM, Liu Y-T, Mayer-Kress G. Time scales in motorments. There is motor abundance. Motor Control 2000; 4:learning and development. Psychol Rev 2001; 108 (1): 57-82259-619. Piek JP. The role of variability in early motor development. Inf

Behav Dev 2003; 25: 452-65 33. Kelso JAS, Schoner G. Self-organisation of coordinative move-ment patterns. Hum Mov Sci 1988; 7: 27-4610. Newell KM. Change in movement and skill: learning, retention

and transfer. In: Latash ML, Turvey MT, editors. Dexterity and 34. Goldfield EC. Emergent forms: origins and early developmentits development. Mahwah (NJ): Erlbaum, 1996: 393-430 of human action and perception. Oxford: Oxford University

11. Newell KM, Corcos DM. Issues in variability and motor con- Press, 1995trol. In: Newell KM, Corcos DM, editors. Variability and 35. Hasan Z, Thomas JS. Kinematic redundancy. In: Binder MD,motor control. Champaign (IL): Human Kinetics, 1993: 1-12 editor. Progress in brain research. Amsterdam: Elsevier, 1999:

12. Riley MA, Turvey MT. Variability and determinism in motor 379-87behavior. J Motor Behav 2002; 34: 99-125 36. Arutyunyan GH, Gurfinkwl VS, Mirskii ML. Investigation of

13. Williams AM, Davids K, Williams JG. Visual perception and aiming at a target. Biophysics 1968; 14: 1162-7action in sport. London: Routledge, Taylor & Francis, 1999 37. Scholz JP, Schoner G, Latash ML. Identifying the control struc-

14. Iberall AS. A field and circuit thermodynamics for integrative tures of multijoint coordination during pistol shooting. Expphysiology. I. Introduction to the general notions. Am J Physi- Brain Res 2000; 135: 382-404ol 1977; 233: R171-80

38. Brisson TA, Alain C. Should common optimal movement pat-15. Prigogine I, Stengers I. Order out of chaos. New York: Bantam terns be identified as the criterion to be achieved? J Motor

Books, 1984 Behav 1996; 28: 211-2316. Kugler PN, Kelso JAS, Turvey MT. On the control and coordi-

39. Schoellhorn WI, Bauer HU. Identifying individual movementnation of naturally developing systems. In: Kelso JAS, Clarkstyles in high performance sports by means of self-organizingJE, editors. The development of movement control and coordi-Kohonen maps. In: Riehle HJ, Vieten M, editors. Proceedingsnation. New York: Wiley, 1982: 5-78of the XVI Annual Conference of the International Society for

17. Kugler PN, Turvey MT. Information, natural law, and the self- Biomechanics in Sport; 1998 Jul 7-12: Konstanz, Germany.assembly of rhythmic movement. Hillsdale (NJ): Lawrence Konstanz: International Society for Biomechanics in Sport,Erlbaum Associates, 1987 1998

18. Haken H. Synergetics: an introduction: non-equilibrium phase40. HU, Schoellhorn W. Self-organizing maps for the analysis oftransitions and self-organisation in physics, chemistry and

complex movement patterns. Neural Processing Lett 1997; 5:biology. Berlin: Springer-Verlag, 1983193-9

19. Kugler PN, Kelso JAS, Turvey MT. On the concept of coordina-41. Davids K, Savelsbergh GJP, Bennett SJ, et al. Interceptivetive structures as dissipative structures: I. Theoretical lines of

actions in sport: information and movement. London: Rout-convergence. In: Stelmach GE, Requin J, editors. Tutorials inledge, 2002motor behavior. Amsterdam: North-Holland, 1980: 3-47

42. Koenig G, Tamres M, Mann RW. The biomechanics of the20. Clark JE. On becoming skillful: patterns and constraints. Res Qshoe-ground interaction in golf. In: Cochran AJ, Farrally FR,Exerc Sport 1995; 66: 173-83editors. Science and golf II: proceedings of the 1994 Scientific21. Newell KM. Constraints on the development of coordination.Congress of Golf. London: E & FN Spon, 1994: 40-5In: Wade MG, Whiting HTA, editors. Motor development in

43. Button C, Davids K. Interacting intrinsic dynamics and inten-children: aspects of coordination and control. Dordrecht: Mar-tionality requires coordination profiling. In: Thomson J,tinus Nijhoff, 1986: 341-60Grealey M, editors. Studies in perception and action. Mahway22. Bernstein NA. The coordination and regulation of movements.(NJ): Erlbaum Associates, 1999: 314-8Oxford: Pergamon Press, 1967

44. Barnes C, Mercer G. Disability culture: assimilation or inclu-23. Vereijken B, van Emmerik REA, Whiting HTA, et al.sion? In: Albrecht G, Seelman KD, Bury M, editors. HandbookFree(z)ing degrees of freedom in skill acquisition. J Motorof disability studies. London: Sage Publications, 2001: 515-34Behav 1992; 24: 133-42

24. Newell KM, Vaillancourt D. Dimensional change in motor 45. Latash ML, Anson JG. What are ‘normal movements’ in atypi-learning. Hum Mov Sci 2001; 20: 695-715 cal populations? Behav Brain Sci 1996; 19: 55-106



46. Jeka JJ, Lackner JR. The role of haptic cues from rough and 69. Johnston TD, Edwards L. Genes, interactions and the develop-slippery surfaces in human postural control. Exp Brain Res ment of behaviour. Psychol Rev 2002; 109: 26-341995; 103: 267-76 70. Barsh GS, Farooqi IS, O’Rahilly S. Genetics of body-weight

47. Davids K, Kingsbury D, George K, et al. Interacting constraints regulation. Nature 2000; 404: 644-51and the emergence of postural behavior in ACL-deficient 71. Rankinen T, Perusse L, Gagnon J, et al. Angiotensin-convertingsubjects. J Motor Behav 1999; 31: 358-66 enzyme ID polymorphism and fitness phenotype in the HERI-

48. Riley MA, Balasubramaniam R, Turvey MT. Recurrence quan- TAGE Family Study. J Appl Physiol 2000; 88: 1029-35tification of analysis of postural fluctuations. Gait Posture 72. Fox PW, Hershberger SL, Bouchard TJ. Genetic and environ-1999; 9: 65-78 mental contributions to the acquisition of a motor skill. Nature

49. Riley MA, Mitra S, Stoffregen TA, et al. Influences of body 1996; 384: 356-8lean and vision of unperturbed postural sway. Motor Control 73. Feitosa MF, Gaskill SE, Rice T, et al. Major gene effects on1997; 1: 229-46 exercise ventilatory threshold: the HERITAGE Family Study.

50. Riley MA, Balasubramaniam R, Mitra S, et al. Visual influ- J Appl Physiol 2002; 93: 1000-6ences on centre of pressure dynamics in upright posture. Eco- 74. Kevles DJ, Hood L. The code of codes: scientific and sociallogical Psychology 1998; 10: 65-91 issues in the human genome project. Boston (MA): Harvard

51. van Emmerik REA, van Wegen EEH. On variability and stabili- University Press, 1992ty in human movement. J Appl Biomech 2000; 16: 394-406 75. Hopkins WG. Genes and training for athletic performance [on-

52. Lee DN, Aronson E. Visual proprioceptive control of standing line]. Available from URL: http://www.sportsci.org/jour/in human infants. Percept Psychophys 1974; 15: 529-32 0101/whgene.htm [Accessed 2003 Jan 23]

53. Lee DN, Lishman R. Visual proprioceptive control of stance. J 76. Alvarez R, Terrados N, Ortolano R, et al. Genetic variation inHum Mov Stud 1975; 1: 87-95 the renin-angiotensin system and athletic performance. Eur J

Appl Physiol 2000; 82: 117-2054. Schumway-Cook A, Woollacott MH. The growth of stability:postural control from a developmental perspective. J Motor 77. Montgomery HE, Clarkson P, Barnard M, et al. Angiotensin-Behav 1985; 17: 131-47 converting enzyme gene insertion/deletion polymorphism and

response to physical training. Lancet 1999; 353: 541-555. Crichton KJ, Fricker PA, Purdam CR, et al. Injuries to the pelvisand lower limb. In: Bloomfield J, Fricker PA, Fitch KD, 78. Myerson S, Hemingway H, Budget R, et al. Human angiotensineditors. Textbook of science and medicine in sport. London: I-converting enzyme gene and endurance performance. J ApplBlackwell Scientific Publications, 1992: 381-419 Physiol 1999; 87: 1313-6

56. Carter ND, Jenkinson TR, Wilson D, et al. Joint position sense 79. Nazarov IB, Woods DR, Montgomery HE, et al. The angioten-and rehabilitation in the anterior cruciate ligament deficient sin converting enzyme I/D polymorphism in Russian athletes.knee. Br J Sports Med 1997; 31: 209-12 Eur J Hum Genet 2001; 9: 797-801

57. Barrett DS. Proprioception and function after anterior cruciate 80. Taylor RR, Mamotte CDS, Fallon K, et al. Elite athletes and thereconstruction. J Bone Joint Surg 1991; 73: 833-7 gene for angiotensin-converting enzyme. J Appl Physiol 1999;

87: 1035-758. Corrigan J, Cashman W, Brady M. Proprioception in the cruci-ate deficient knee. J Bone Joint Surg 1992; 74B: 247-50 81. Williams AG, Rayson MP, Jubb M, et al. Physiology: the ACE

gene and muscle performance. Nature 2000; 403: 61459. Palmieri RM, Ingersoll CD, Stone MB, et al. Center-of-pressureparameters used in the assessment of postural control. J Sports 82. Woods DR, Humphries SE, Montgomery HE. The ACE I/DRehabil 2002; 11: 51-66 polymorphism and human physical performance. Trends En-

docrinol Metab 2000; 11: 416-2060. Le Clair K, Riach C. Postural stability measures: what to mea-sure and for how long. Clin Biomech 1996; 11: 176-8 83. Woods DR, Hickman M, Jamshidi Y, et al. Elite swimmers and

the D allele of the ACE I/D polymorphism. Hum Genet 2001;61. Ekdahl C, Jarnlo GB, Andersson SI. Standing balance in healthy108: 230-2subjects. Evaluation of a quantitative test battery on a force

platform. Scand J Rehabil Med 1989; 21: 187-95 84. Woods DR, World M, Rayson MP, et al. Endurance enhance-ment related to the human angiotensin I-converting enzyme I-62. Dijkstra TMH, Schoner G, Gielen CCAM. Temporal stability ofD polymorphism is not due to differences in the cardiorespira-the action-perception cycle for postural control in a movingtory response to training. Eur J Appl Physiol 2002; 86: 240-4visual environment. Exp Brain Res 1994; 97: 477-86

85. Jones A, Montgomery HE, Woods DR. Human performance: a63. Dijkstra TMH, Schoner G, Gielen CCAM. Frequency depend-role for the ACE genotype? Exerc Sport Sci Rev 2002; 30:ence of the action-perception cycle for postural control in a184-90moving visual environment: relative phase dynamics. Biol

Cyberns 1994; 71: 489-501 86. Sonna LA, Sharp MA, Knapik JJ, et al. Angiotensin-convertingenzyme genotype and physical performance during US Army64. Rogers MW. Disorders of posture, balance and gait in Parkin-basic training. J Appl Physiol 2001; 91: 1355-63son’s disease. Clin Geriatr Med 1996; 12: 825-45

65. Horak FB, Nutt JG, Nashner LM. Postural inflexibility in 87. Montgomery HE, Marshall R, Hemingway H, et al. HumanParkinsonian subjects. J Neurol Sci 1992; 111: 46-58 gene for physical performance. Nature 1998; 393: 221-2

66. Schieppati M, Hugon M, Grasso A, et al. The limits of equilibri- 88. Gayagay G, Yu B, Hambly B, et al. Elite endurance athletes andum in young and elderly normal subjects and in Parkinsonians. the ACE I allele: the role of genes in athletic performance.Electroencephalogr Clin Neurophysiol 1994; 93: 286-98 Hum Genet 1998; 103: 48-50

67. Riccio GE. Information in movement variability about the quali- 89. Folland J, Leach B, Little T, et al. Angiotensin-convertingtative dynamics of posture and orientation. In: Newell KM, enzyme genotype affects the response of human skeletal mus-Corcos DM, editors. Variability and motor control. Champaign cle to functional overload. Exp Physiol 2000; 85: 575-9(IL): Human Kinetics, 1993: 317-57 90. Hagberg JM, Ferrell RE, McCole SD, et al. VO2max is associat-

68. Riccio GE, Stoffregen TA. Affordances as constraints on the ed with ACE genotype in postmenopausal women. J Applcontrol of stance. Hum Move Sci 1988; 7: 265-300 Physiol 1998; 85: 1842-6


260 Davids et al.

91. Bouchard C, Daw EW, Rice T, et al. Familial resemblance for 95. Yates FE. Self-organizing systems. In: Boyd C, Noble D, edi-VO2max in the sendentary state: the HERITAGE family study. tors. Logic of life. Oxford: Oxford University Press, 1993:Med Sci Sport Exerc 1998; 30: 252-8

189-21892. Jones RS. Almost like a whale: the origin of the species updated.

London: Anchor Press, 199993. van Geert P. Dynamic systems of development: change between

Correspondence and offprints: Professor Keith Davids,complexity and chaos. New York: Harvester Wheatsheaf,1994 School of Physical Education, University of Otago, PO Box

94. Oyama S. The ontogeny of information: developmental systems56, Dunedin, New Zealand.and evolution. 2nd ed. Durham (NC): Duke University Press,E-mail: [email protected]


movement systems as dynamical systems

Documents