Sports Med 2005; 35 (11): 935-950

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 935 1. Defining and Modelling the Postural Control System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 937 1.1 Terminology and the Task of Postural Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 937 1.2 Modelling Postural Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 937 1.3 Postural Steadiness and the Mechanical Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 938 1.4 Limitations of the Biomechanical Approach to Measuring Postural Steadiness . . . . . . . . . . . . . . 939 2. An Alternative Theoretical Proposition: Measuring Complexity in a Nonlinear Postural Control System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 940 2.1 Evidence for Postural Steadiness Emerging from a Nonlinear System . . . . . . . . . . . . . . . . . . . . . . 940 2.2 Change in Complexity and Postural Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 942 2.3 Approximate Entropy as a Measure of Postural Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 942 3. Recovery of Postural Control After Cerebral Concussion: New Insights Using the Nonlinear Dynamic Theoretical Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 943 3.1 Basic Concepts of Sport-Related Cerebral Concussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 943 3.2 Recovery of Postural Stability After Cerebral Concussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 944 3.3 Detecting Altered Postural Control in Athletes Without Postural Instability . . . . . . . . . . . . . . . . . . 945 3.4 Comparing the Recovery of Postural Stability with the Recovery of Altered Centre of Pressure Regularity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 946 4. Conclusions and Recommendations for Postural Control Assessment Approaches in Sports Medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 947 Recent research suggests that traditional biomechanical models of postural Abstract stability do not fully characterise the nonlinear properties of postural control. In 936 Cavanaugh et al. sports medicine, this limitation is manifest in the postural steadiness assessment approach, which may not be sufficient for detecting the presence of subtle physiological change after injury. The limitation is especially relevant given that return-to-play decisions are being made based on assessment results. This update first reviews the theoretical foundation and limitations of the traditional postural stability paradigm. It then offers, using the clinical example of athletes recovering from cerebral concussion, an alternative theoretical proposition for measuring changes in postural control by applying a nonlinear dynamic measure known as ‘approximate entropy’. Approximate entropy shows promise as a valuable means of detecting previously unrecognised, subtle physiological changes after concussion. It is recommended as an important supplemental assessment tool for determining an athlete’s readiness to resume competitive activity. Sports injuries are not commonly associated with defined using an amplitude metric (e.g. magnitude ‘balance problems’. Reports of disequilibrium or of sway) that denotes precision; i.e. optimal control falls after injury are rare, because athletes generally is evidenced by less movement, or error, about a are in excellent health or because most injuries are target position. Athletes who demonstrate normal isolated to relatively few musculoskeletal structures. postural stability (relative to their age-matched Sport activities, however, often demand exquisite peers) are generally assumed to have a healthy posbody control, such that even subtle impairments tural control system. may interfere with optimal performance without This assumption may be fundamentally flawed. producing obvious unsteadiness. Perhaps more so During the last decade, a variety of studies have than most individuals, athletes utilise a wide array of revealed that the variability of centre of pressure complex strategies involving arms, legs, torso, neck (COP) location during quiet standing is not the and head for control of body position to gain advanresult of random error.[7-12] Instead, COP oscillatage over their competitors. Injuries producing tions, despite appearing erratic and irregular, conweakness, sensory impairment, diminished joint tain a ‘hidden’ structure, or orderliness, that emerges range of motion, or alterations in neural processing in time, presumably as a result of interactions among potentially can affect the ability to control the orienunderlying postural control system components. Retation of the body in space.[1,2] cent evidence supports the idea that with advanced age and disease, the complexity of temporal strucThe medical assessment of postural control after ture breaks down, resulting in more regular COP injury often includes the determination of postural oscillations.[13,14] Importantly, changes in the regustability in quiet standing. In the tradition of Romlarity of COP oscillations can occur in the absence berg,[3] postural stability assessment typically reof changes in postural stability.[15] Optimal postural quires an athlete to stand as still as possible under control (in quiet standing), therefore, appears to be any one of a number of base of support (double-limb characterised by COP oscillations that not only are stance, tandem stance, or single limb stance) or small in amplitude but also are relatively unconsensory (eyes open vs eyes closed) conditions.[4-6] strained and irregular. More stable athletes are assumed to be able to stand with less postural sway about a central equilibrium For injured athletes, the return of optimal posturpoint. In this sense, postural control is operationally al control is an important rehabilitation goal. Cur 2005 Adis Data Information BV. All rights reserved. Sports Med 2005; 35 (11) A Nonlinear Dynamic Approach for Evaluating Postural Control 937 rent measures of postural stability, however, may be or in constant motion (dynamic equilibrium), i.e. not inadequate for helping athletes, coaches and medical acted upon by unbalanced external forces.[17] Balprofessionals judge when complete recovery of posance, often used synonymously with equilibrium, tural control has occurred. The aims of this article refers to equilibrium about a specified axis, such as are: (i) to review the theoretical basis and limitations the vertical axis in upright standing.[19] Postural of current approaches to postural control assesscontrol and balance control, therefore, can be used ment; (ii) to offer an alternative measurement apinterchangeably to refer to the act of maintaining or proach based on a nonlinear dynamics theoretical returning the body close to a state of static or dyframework of motor control; and (iii) using sport-renamic equilibrium.[19] lated cerebral concussion as an example, to review recent evidence that highlights the limitations of 1.2 Modelling Postural Control current approaches and supports the application of a Current models of the postural control system nonlinear dynamic theoretical framework for regenerally have evolved from the seminal work of vealing unique changes in postural control after Nikolai Bernstein (1896–1966), for whom the injury. model of nervous system function considered the whole body as a mechanical system subject to gravi1. Defining and Modelling the Postural tational and inertial forces.[20] For Bernstein, coordiControl System nated movement, including postural control, was a problem that involved mastering the many redun1.1 Terminology and the Task of dant degrees of freedom defined over several levels Postural Control of biological analysis. The large number of potential degrees of freedom in the control system precluded The task of postural control involves controlling the possibility that each is controlled individually at the body’s position in space for the dual purposes of every point in time; thus, Bernstein proposed that stability and orientation.[16] Orientation is defined as control of integrated movement was probably disthe ability to maintain an appropriate relationship tributed throughout many interacting systems workbetween body segments and between the body and ing cooperatively. Consistent with this idea, current the environment. Stability, in the broadest sense, models of postural control can be broadly organised refers to the ability of a system to resist perturbainto two categories: models that ‘describe’ a system tions.[17] Postural stability defines the ability to of interacting components and models that seek to maintain a desired postural orientation, either at rest ‘predict’ how components interact to achieve posor during movement, in response to perturbations tural control. Descriptive models are useful for congenerated from either internal or external sources. ceptualising the multidimensional nature of the posFor human functional activities performed standing tural control system but do not guide the measureor sitting, postural stability specifically refers to the ment of postural control system changes after injury ability to resist perturbations such that the whole or in response to rehabilitation.[1,16] body centre of mass is maintained within the limits of the base of support. Postural steadiness, a special Predictive models of postural control system outcase of postural stability, defines the ability to stand put have evolved from both linear and nonlinear as motionless as possible.[18] Equilibrium, a term dynamics frameworks (table I). Linear dynamic derived from Newtonian mechanics, refers to condimodelling is based on a stimulus-response parations in which an object is at rest (static equilibrium) digm, in which system output can be predicted from  2005 Adis Data Information BV. All rights reserved. Sports Med 2005; 35 (11) 938 Cavanaugh et al. Table I. General characteristics of linear and nonlinear postural control system models Characteristic Linear model Nonlinear model Paradigm Stimulus response Self-organisation Degree of postural control reflected Predicted based on input (perturbation) Dependent on initial conditions and nature in the output signal parameters and known linear relationships of multi-linked interactions among system among system components components. Predicted over various time