- Email: [email protected]
The assessment of motor recovery: A new look at an old problem
Kinesiol. Vol. 6, No. 2, pp. 137-145, 1996 Copyright 0 1996 Elsevier Science Ltd. All rights reserved Printed in Great Britain 1050-641l/96 $15.00 + 0.00
J. Electromyogr.
1050-6411(95)00029-1
The Assessment of Motor Recovery: A New Look at an Old Problem* Theo Mulder’,2, Bart Nienhuis’ and John Pauwels’ ‘Department
of Research and Development, Sint Maartenskliniek, PO Box 9011, 6500 GM Nijmegen and 2Neurological Institute, University of Nijrnegen, The Netherlands
Summary: After nervous system damage, functional recovery usually occurs. It is of great clinical importance to follow the course of recovery and, when possible, predict the extent. This measurement and prediction of recovery is one of the main challenges facing clinicians today. The majority of assessment procedures currently employed, however, are impairment-oriented; that is, these procedures are oriented primarily at the disease or organ level. Until now, few procedures have been available that focus at the disability or behavioural level. This lack of disability-oriented assessmentprocedures hinders the development and evaluation of rehabilitation programmes, as impairment-oriented assessment scores have little relevance for the prediction of daily functioning. In this paper, a preliminary task-set is presented that may function as an impetus for the development of novel disability-oriented assessmentprocedures in neurological rehabilitation. Motor behaviour is not assessedin vacua, rather as the end result of a continuous interaction between motor, sensory and cognitive processes. With this task set, it is possible to evaluate the changing influence of sensory and cognitive factors on the quality of motor performance over time. These changes, it is argued, are a crucial component of functional recovery. Copyright 0 1996 Elsevier Science Ltd Key Words: Motor recovery-Assessment-Rehabilitation. J. Electromyogr. Kinesiol., Vol. 6, 137-145, June
interests of the clinicians. Biomedical engineering therefore emphasized the technical improvement of these devices. In the present paper this emphasis is questioned, as it was found that the lack of clinical relevance results primarily from the absence of a general theoretical framework guiding how such devices can be used and not from their poor design. In the present text, a theoretical framework is sketched in general terms (for more details see Refs. 21-24). Clinical movement analysis is not uniform, and consists instead of a set of techniques directed at a diverse field of problems and disciplines. It can, therefore, be argued that the requirements for clinical
INTRODUCTION
In 1987 Messenger and Bowke?’ showed that the majority of clinical movement analysis facilities available in clinics in the UK were not routinely used. This led to the conclusion that the available devices were not advanced enough to serve the
Addresscorrespondence and reprint requeststo Theo Mulder. Dept. of Researchand Developm&t, Sini Maattenskliniek,P.O. Box 9011, 6500 GM Nijmegen,The Netherlands. *This paperwas presentedas a keynotelecturein the 1994ISEK Congressin Charleston.
137
138
T. MULDER ET AL.
movement analysis may differ across different disciplines (e.g. orthopaedics, neurology, rheumatology, and rehabilitation). For example, a detailed electromyographical and kinematic analysis of gait may be of value for the planning of orthopaedic surgery, but it is probably of less value for the physiotherapist who wants to predict daily motor behaviour. This variability underscores the need for differentiated routes in the development of assessment,This paper focuses on these varying needs of rehabilitation medicine; i.e. it focuses on the need for the development of a disability-oriented assessment. Such a disability-oriented approach, it should be noted, is a global approach and is not directed at the level of (isolated) movements, but at the level of skills. IMPAIRMENT
VS. DISABILITY
The terms ‘impairment’ and ‘disability’ play a prominent role in rehabilitation medicine. Impairment refers to the loss of a body structure and represents a disturbance at the organ level (e.g. lack of muscle strength, decreased range of motion). Disability refers to the consequencesof impairments for functional performance and activities and thus represents a disturbance at the level of the acting person. It is important to note that no direct relationship exists between the level of impairment and the level of disability34. Thus, a disability-oriented assessmentneeds to use terms different from those found in an impairment-oriented framework. ESSENTIALS OF MOTOR CONTROL Movement is the output of a hybrid functional system intimately linked to its environment in which sensory, cognitive and motor processes continuously interact. In the organization of a movement no clear separation exists between these processes;their relative weights may change, however, according to the complexity of the task, the skill of the performer and the integrity of the system. Recent models of motor control stress this interactive nature of Con~o~3,18,21-24,33
This construction process, however, is not the outcome of a strictly hierarchical and sequentially organized system, rather it is the result of a heterarchical, multilevel system where ‘higher levels’ control the invariant global aspects of the action and subordinate structures specify details. These ‘high level’ control structures activate functional
groups of muscles, which are constrained to act as a single coordinative structure and can be temporarily organized to accomplish a particular behavioural goal; the dimensionality - or number of degrees of freedom which apply to the control of a given task is thereby reduced, providing the basic organization of movement16. Muscle-specific details, such as force, velocity and spatial accuracy, are considered ‘lower order’ parametric problems which are dependent on the actual biomechanical and environmental context38. Such an architecture enables the system to be functionally specific, that is, to produce movements ideally suited to the specific requirements of attaining the movement goal. Movement is therefore not the result of a fixed sequence of signals transmitted over efferent pathways that determine all muscle-specific details of the required act, but it is the result of a flexible ‘negotiation’ between the available input and the required output. As there is an almost infinite number of ways of or_ramzing to attain movement goals, many solutions may lead to the same output. The ‘selection’ of certain solutions depends on the complexity and novelty of the task, the skill of the performer, the integrity of the system and the requirements of the environment. For example, many overlearned, routine-based tasks such as balance, gait, reaching, eating, etc., require none or only minimal conscious control. On the other hand, a novel or complex task, such as playing a violin, can only be performed under a conscious regime. The ability to shift between levels of control is an important characteristic of normal motor control. It is not a conscious process, rather the result of exploitation of an emergent characteristic of the central nervous system (CNS), namely the capacity to perform cost-benefit analyses leading to the most efficient strategy. For example, when normal automatic performance has been disturbed, the organism is still able to move by recruiting compensatory strategies. Each task situation, combined with the actual state of the organism, leads to its own optimal strategy as the organism continuously shifts among the available means of control. These strategies are to some extent flexible; i.e. small adjustments can be made without changing the strategy. But, if the adjustments are too large, a shift to a novel strategy is necessary. For example, some loss of proprioceptive input can be tolerated by the system, but when loss of proprioceptive input exceeds a certain level the system is forced to shift to another strategy (visual control).
ASSESSMENT OF MOTOR RECOVERY
These shifts enable the organism to reach the required goals even under non-optimal conditions. IMPLICATIONS
FOR MOTOR ASSESSMENT
From the above statements it can be inferred that solely analysing the output characteristics of motor behaviour (kinematic and biomechanical aspects of movement) reveals important information about the end result, but not about the strategies leading to that result, as these strategies remain hidden. Insight into the control mechanisms, however, is crucial to understanding the patient’s level of functional recovery. Assessment procedures focusing on functional recovery, therefore, should be designed in such a way that the adaptive processes can be followed over time. THREE PRINCIPLES
OF RECOVERY
Because recovery depends - to a large extent on the ability of the CNS to adapt to (peripheral) changes, the study of recovery is for a large part the study of adaptation. How can this adaptive process be studied over time, and what observable mechanisms can be distinguished are commonly raised questions. From the work performed in Nijmegen during the past 5 yr23-26,three principles of recovery can be distilled: (a) a decrease of cognitive regulation; (b) a decrease of visual dependency; and (c) an improvement of sensorimotor adaptability. Decrease in Cognitive Regulation The performance of a difficult (non-automatized) task interferes with other simultaneously performed tasks (see also Ref. 39). Therefore, if one uses an attention-demanding task, it is possible to use the degree of interference produced by this task on the primary task as a measure of the attention demands of the primary task. This idea has relevance for clinical (movement) analysis. For instance, overlearned motor tasks require very little information processing capacity. If the proficiency in two tasks remains unchanged, regardless of whether they occur simultaneously or separately, then at least one task would seem to be automatic. On the other hand, if a task is performed less well when it is combined with a secondary task, then both tasks are thought to require at least some attentional capacity (cognitive regulation). Thus, a decrease in dual-task interference reflects a decrease in cognitive regulation
139
over time, and may be seen as a sensitive indicator of an improved skill (re)acquisition and automaticity. Many motor control theories assume such a transition from attention-invested (or controlled) performance to attention-free (or automatic) performance during the learning of a new ~kill’*~*~~.These ideas have recently been applied in studies focusing on (disorders of) motor control in the elderly35-37, showing that with increasing age the attentional involvement in tasks such as gait increases substantially. The use of this dual task methodology in recovery research, however, is still extremely rare. In a series of experiments performed by Geurts et a1.*-13,the validity of this idea was tested. Their experiments focused on the recovery of balance after lower limb amputation. Amputation leads to a massive disruption of the afferent-efferent equilibrium and forces the CNS to develop alternative control strategies. It can be argued that as long as these novel strategies are not available, the system is vulnerable. Subjects with a leg amputation who were wearing their first prosthesis were instructed to stand as still as possible with their hands folded at the back. It was shown that, even at the very beginning of the rehabilitation process, patients were able to perform this task; i.e. their balance performance did not differ significantly from the performance of healthy age-matched controls. Patients stood on a dual-plate force platform which registered centre of pressure (cp) displacements. After 15 s, a slide was projected in front of the subject. The slide contained the words ‘red’, ‘blue’ and ‘green’. The words, however, were printed in a colour different from the meaning of the words, e.g. the word ‘red’ was printed in blue, the word ‘blue’ was printed in green. The patients were instructed to name as quickly as possible the colour of the words while suppressing the strong tendency to recite the word that was written. This is an attentiondemanding task, and the question was whether this task would interfere with the ability of the patient to maintain his/her upright balance. The results indicate that at the beginning of the rehabilitation process, the balance performance of patients was significantly hindered by the performance of a concurrent cognitive task; i.e. the sway increased under the influence of the dual-task performance. Such a dual-task effect could not be shown in healthy age-matched subjects. Most interesting, however, was that this effect slowly diminished over
140
T. MULDER ET AL.
time, indicating a re-automatization of balance contrOP.
With amputation, the system is suddenly deprived of its ‘normal’ input. An interesting question, therefore, is whether this dual-task effect would also be seen in patients with a chronic, slowly progressing disease leading to balance disturbances. Consequently, the experiment described above was repeated with patients suffering from hereditary motor and sensory neuropathy (HMSN) type I or II. These patients generally have distal muscle weakness as well as atrophy and loss of somesthesis, in particular, deep sensibility, Also, foot-ankle deformities develop, including pes cavus, transversoplanus, equinovarus and hammer toes9. Problems in postural control are one of the most frequent consequences of HMSN. The results of this study indicated that although these patients suffered from balance disorders, they were not hindered by the dual-task performance. One can conclude that as a result of the chronic character of the disease, the brain had ample time to reorganize its control procedures and to create a novel efferent-afferent equilibrium. Now, let us take the next step: what would happen if these patients were fitted with corrective orthopaedic footwear? The results were surprising. Geurts et al.9 showed that dual-task performance had no effect on balance control in barefoot conditions, but that it significantly hindered balance control (increased sway) in conditions where patients wore new rehabilitation footwear. Although at first sight this result seems remarkable, it is in accordance with the above described ‘fluid’ character of the CNS and underscores the permanent adaptive character of (motor) control. As soon as movement constraints are changed, the system is forced to adapt and shift to another control strategy. As such, a movement constraint can be defined as any variable that limits the organization and control of a movement. Decrease in Visual Dependency Normally a moving individual receives, interprets and acts appropriately on somesthetic input from the periphery. This interplay between motor and sensory information tunes movement to the specific context of the situation1**27*28.It can be argued, however, that when the somatosensoryinput is either distorted or absent, this interplay is (partly) lost, with vision becoming the most important input
source as a result of a strategic shift. In such a case, postural tasks are impossible to perform without vision. The system can no longer ‘build’ on the incoming somesthetic information, because this information has lost its redundant character. One can therefore predict that in early rehabilitation phases patients will show a marked visual dependency. Visual dependency, in this context, implies a disproportional dependency. Visual information (online) information is always used, and vision provides crucial information - particularly for the prospective control of gait 2*17,30-32, but the information is not used all the time to guide and control the action. We are aware, at an anecdotal level if not a scientific one, that while walking it is possible to shift one’s eyes from the trajectory without falling down or becoming disoriented. Thus, it is the degree of dependency that matters. We argue that this dependency is disproportionally large in patients with sensorimotor disorders. The question remains as to whether this dependency decreases over (recovery) time. To answer this question, postural control was assessedin patients with a unilateral leg amputation before and after a traditional rehabilitation training programme. The experimental set-up was basically the same as that employed in the above-mentioned studies: the centre-of-pressure fluctuations during quiet, upright standing on a force platform were registered with and without visual information. Although there was a small improvement in balance control with full visual information, the major improvement was a decrease in visua1 dependency over time’, with the visual-dependency scores approaching normal values at the end of the rehabilitation. This decrease in visual dependency within subjects clearly indicates a central re-integration of sensory input from the amputated limb into the multisensory control of posture. Although the general effect was significant, it is important to note that substantial inter-individual differences existed. Some subjects showed a relatively rapid decrease of visual dependency, whereas visual dependency for others either changed slowly or showed little change. It should be borne in mind that healthy adults also vary substantially in the degree of visual dependency. Nonetheless, the results are important because they again indicate that relevant aspects of the recovery process remain hidden from the observer as long as specific process-oriented manipulations are not used.
ASSESSMENTOF MOTOR RECOVERY
Improvement of Sensorimotor Adaptability A crucial prerequisite for successful locomotion is the capacity to adapt movement to the environmental demands and the organism’s goals. Two aspects of adaptation are distinguished here: anticipation and reaction. Although in most actions both aspects are served at the same time, for the sake of clarity they will be discussed separately here. Anticipation refers to the capacity of the organism to adjust the present locomotor dynamics to foreseeable environmental changes. It should, therefore, be distinguished from anticipatory postural adjustments occurring in association with voluntary movements, which are aimed at minimizing the equilibrium disturbances produced by the movement performance. Anticipatory movements require the integration of various information sources such as somatosensory, visual and auditory. In humans even cognitive and emotional information may be used. Because these adjustments reflect a feedforward postural control, which occurs before posture and equilibrium are disturbed, past locomotor experience must be continually combined with the actual perceptual input concerning both the body and the environment (see also Ref. 5). These necessary adaptive modifications are effected by feedback from the peripheral motor apparatus, from changes in tactual motion and from the perceptual pick-up of information. As mentioned earlier, visual input plays an important role in the detection of obstacle characteristics, of surface characteristics and of terrain layout3’. Examples of surfaces that require anticipatory control are slopes, stairs and surfaces with different frictional characteristics such as carpets or an ice-covered path. Visually detected information plays a fundamental role in permitting anticipatory changes in the basic locomotor pattern through feedforward contro130. Although the way in which the visual system extracts relevant information from the environment is not yet understood, it is probably fair to argue that identification of the safeness and traversability of the path constitutes the minimum information needed for the control of locomotion. In this context there is a real symbiotic relationship between perception and action, as the organism’s relationship with the environment is mediated primarily through the visual system. The integration of visual information in high level (cortical) loops seems to be necessary to perform the intentional regulations of locomotion mentioned above4. An important com-
plicating factor, however, is that (visual) information does not always have the same ‘weight’. What may be useful for one task may be useless for another task. An intention-coupled selection mechanism which can tune the motor system for the reception of the most relevant information, therefore, seems necessa+. Therefore, adequate anticipatory responses are largely based on central set effects allowing descending commands to prepare sensory and motor systems for anticipated stimulus and task conditions. Indeed, setting aspects of the response in advance can decrease the time it takes the CNS to transform an eliciting stimulus into an appropriate response. This is an important mechanism because postural responses are often initiated before the availability of relevant peripheral information. Central set allows individuals to modify their automatic responses to a postural stimulus by taking into account prior experience with perturbation characteristics and the effectiveness of prior responses. Central set, however, leads to errors in responses when the stimulus or external conditions unexpectedly change. In this respect, anticipation can be distinguished from reaction. Reaction refers to the ability of the organism to deal adequately with unexpected and sudden hindrances, obstacles, or changes. Even at a relatively low level of control, this ability becomes visible. Phase-dependent reflex modulation is one example. If during walking the leg flexion responses elicited by stimuli from different phases of the step cycle were always the same, problems would arise in the different phases. Duysens et a1.6showed that reactions to unexpected perturbations were modulated according to the phase of the step-cycle. Patla et a1.32also suggested a complex, task-specific modulation of the locomotor synergy to alter step length. This seemingly simple ability reflects a process of extreme complexity. The moving organism has to detect dangerous situations and use that information for an on-line modification of the activated locomotory programmes. The difference between anticipation and reaction is clear; with the former the subject is able to plan the movement beforehand, whereas with the latter this is not possible. It can be argued, therefore, that the ability to react requires a higher level of sensorimotor (re-)integration than anticipation does. Clinical studies on sensorimotor adaptability are rare. The majority have been performed with young, healthy adults performing obstacle-avoidance tasks
142
T. MULDER ET AL.
in a laboratory context. McFadyen and Winterr9, for example, showed that when a fixed, visible obstacle is placed at mid-swing a few strides in front of the subject, distinct changes in lower limb motor coordination are observed just prior to foot clearance. Large obstacles (up to 10% of leg length) induced a complete reorganization of the motor coordination of the lower limb. MacFadyen and Winter” termed these changes anticipatory locomotor adjustments. An interesting point is that these adaptive changes are very task specific, with the strategic changes fully tuned to the specific characteristics of the environment-performer relationship. The size of the obstacle, the moment of its appearance and the amount of visual information determines the nature of this adaptation. At present a series of experiments focusing on both aspects of sensorimotor adaptability is being performed in our laboratory. In these experiments, patients with lower limb amputation or total hip replacement have to negotiate a clearly visible hindrance with a height of approximately 6 inches. They must step on the hindrance (resembling a sidewalk) without changing walking speed. While walking towards the hindrance, a concurrent task must be performed (e.g. solving a cognitive, attention-demanding task). The results show that in the dual-task condition (compared to the normal, undisturbed condition) the walking speed decreased significantly, a difference that diminished over time. The same was true for the cognitive task: when compared with the end of rehabilitation ones sees that more errors were made during the first rehabilitation phases. In these experiments the cognitive task consisted of a woman’s voice pronouncing the words ‘high’ and ‘low’ with either a high or low pitch. There were congruent stimuli (saying ‘low’ with a low pitch and ‘high’ with a high pitch) and incongruent stimuli (saying ‘low’ with a high pitch and ‘high’ with a low pitch). The subjects had to indicate, while walking, whether the tone presented was low or high, regardless of voice information (the word spoken) PRELIMINARIES FOR A DISABILITY ORIENTED TASK SET FOR BALANCE AND GAIT We designed a task set that differs substantially from the conventional clinical movement-analysis procedures. Table 1 shows the ingredients for a bal-
ance assessment task, whereas Table 2 shows the ingredients for a gait assessmenttask set. MOTOR TASK SPECIFICATIONS Balance Task 1: Basic measurements: The patient is asked to stand as still as possible without any constraint so researcherscan gather individual reference values. Task 2: Perceptual manipulations: The patient is asked to maintain his/her balance for a period of 30 s under conditions of impaired vision so researchers can gather individual information concerning the degree of visual dependency. The patient wears special glasses with controlled transparency. Task 3: Cognitive manipulations (standing while per$orming a non-motor concurrent task): The patient is asked to stand as still as possible for a period of 30 s. The concurrent task is presented aurally and regulated by a computer. The task is either a mental calculation task or an auditory Stroop task. With
TABLE 1.
Design of disability-oriented motor tasks for the assessment of balance Description
Aim
Basic measurements:
Quiet upright standing under optimal conditions
Determination of reference values
Perceptual manipulations:
Quiet upright standing under impoverished illuminated conditions (dark glasses, milky white glasses) Quiet upright standing while performing a concurrent cognitive task (mental calculation) Quiet upright standing while waiting for a disturbance. Type and moment of perturbation is known (anticipation); type and moment of perturbation is not known (reaction)
Determination of visual dependency
Cognitive manipulations:
Motor manipulations:
Determination of cognitive dependency
Determination of the capability of the system to implement future disturbances into a motor programme under construction and to determine the capability of the system to react adequately to sudden environmental changes
143
ASSESSMENTOF MOTOR RECOVERY l
TABLE 2. Design of disability-oriented motor tasks for the assessment of locomotion Description
Aim
Basic measurements:
Flat-level walking, optimal conditions
Determination of reference values
Perceptual manipulations:
Walking under impoverished illumination conditions. Walking with ‘impaired peripheral vision. Walking with ‘impaired’ central vision
Determination of visual dependency
Cognitive manipulations:
Walking under conditions of loud noise (traffic); walking while performing a concurrent task (auditory Stroop task)
Determination cognitive dependency
Motor manipulations:
Walking at a certain (firm) speed while negotiating a clearly visible hindrance resembling a sidewalk (anticipation). Walking at a certain (firm) speed while approaching an unpredictable hindrance: a lightspot on the floor (reaction)
Determination of the system’s capacity to anticipate future expected disturbances and to react to sudden unexpected changes in the environment
l
Errors on the concurrent task. Increase/decrease of the velocity of centre of pressure (Vcp) as a result of the selected manipulations.
The sway parameters are measured by means of a dual-plate force-platform (for more details about the intrasubject variability see Ref. 13). Gait
of
the Stroop task, a loudspeaker voice speaks the words ‘high’ and ‘low’ in either a high or low pitch. The patient is instructed to name the pitch and to suppress the tendency to repeat the word using the same pitch as was spoken. This is an attention-demanding task, and the aim is to determine the cognitive dependency of task performance to gain more insight about the level of (re-)automatization. Task 4: Motor manipulation: The patient is asked to stand as still as possible while waiting for a (known/unknown) type of perturbation. The aim of this task is to gain more insight about the system’s adaptability. Dependent variables in disability-oriented balance assessment The following dependent variables have been recorded during the laboratory measurements:
Task 1: Basic measurement: Patients walk along a straight line without any constraint so reference values can be gathered. Task 2: Perceptual manipulations (walking under diferent visual conditions): Patients have to walk under conditions of impaired visual input in order to gather information concerning the degree of visual dependency. The patient walks while wearing special glasses with controlled transparency. The switching command of the glasses is regulated by the computer. While the patient is walking, the glasses change from the initial configuration to the final one. Possible configurations are: both lenses open; one lens open and the other closed; and both lenses 50% overshadowed. The timing of the configuration change can be controlled. Task 3: Cognitive manipulations (walking while petiorming a concurrent task): To evaluate the degree of cognitive dependency, patients have to walk while performing a non-motor concurrent task. The task is presented aurally and consists of a mental calculation task or an auditory Stroop task (see Task 3 of the balance assessmentset-up). Task 4: Motor manipulations (anticipation and reaction): In order to determine to what degree patients are capable of anticpating future (but expected) events, they walk while approaching a sidewalk. The patient has to step upon this sidewalk without decreasing velocity. This task is performed with and without a concurrent task. To assess the patient’s capability of reacting to sudden unexpected changes in the environment the patient is instructed to avoid a ‘hindrance’ (a spot of light) that is projected on the ground just before the subject reaches that area. The moment of projection as well as the place of projection can be manipulated. Dependent variables in disability-oriented analysis
gait
The following dependent variables have been measured during laboratory measurements:
144
T. MULDER ET AL.
0 Walking velocity. l Step frequency (measurement of intra-subject variability across the different task manipulations and over (recovery) time. l Step-length (measurement of intra-subject variability across the different task manipulations and over (recovery) time. 0 Pelvic rotation parameters reflecting the efficiency of locomotion. l Movement and velocity of the trunk in anteriorposterior and medial-lateral directions. The gait parameters are measured by means of a PRIMAS 3-D motion analysis system. CONCLUSION AND FUTURE PERSPECTIVES The suggested task set reflects all theoretical aspects discussed in this paper. This task set should enable clinicians to gain insight, not only into the visible end result of motor performance, but also into the processes leading to this end result. It is argued here that shifts in these (control) processes over time may give an objective indication of recovery. It is further argued that these aspects of recovery have more relevance for the prediction of daily activities than the conventional biomechanical impairment-related parameters. If a patient at the end of a treatment programme still shows a strong cognitive and/or visual dependency, as well as an impaired ability to negotiate obstacles, it will be clear that this patient is at risk under normal environmental circumstances. Assessmentprocedures currently used do not deliver this information because they focus primarily on output parameters acquired under optimal conditions. At this moment only a prototype version of the task set exists, and the generalization value of the collected data has to be established before it is clear whether this approach has real additional value for rehabilitation research. Therefore, patients who perform poorly at the end of rehabilitation should be followed over time to test whether they are at risk in terms of safety, dependency, skill etc., as compared to patients who performed well at the end of rehabilitation. Such a comparison requires the ambulatory monitoring of motor activities under daily conditions. Such a system (AMMA) has been developed in Rotterdam in the context of a European research project named CAMARC-II (Computer Aided Movement Analysis in a Rehabilitation
Context14*15).The system is able to store relevant information from several sensors placed on the patient’s body over a period of several hours. Because the system is also sensitive for sounds and ambient lighting levels, it is possible to study the relationship between environmental factors and the quality of gait. More research is needed to determine the dependent variables most sensitive for recovery. Therefore, the journey towards a disability-oriented motor assessment task has just begun, with only the first few steps described in this paper. REFERENCES 1. Adams JA: A closed-loop theory of motor learning. J Mot Behav 3:111-149, 1971. 2. Bardy BG, Baumberger B, Fluckiger M, Laurent M: On the role of global and local visual information in goal-directed walking. Acta Psychologica 81: 199-210, 1992. 3. Baron EU. The cerebral computer: introduction to the computational structure of the human brain. Erlbaum. Hillsdale, 1987. 4. Drew T: Visuomotor coordination in locomotion. Curr 0pin Neurobiol 1:652-657, 1991. 5. Dufosse M, Massion J: Posturo-kinetic interactions: modeling and modes of control. In: Tutorials in Motor Behaviour II, ed by Stelmach GE, Requin J. North Holland Publishing, Amsterdam, pp 125-155, 1992. 6. Duysens J, Trippel M, Horstmann GA Dietz V: Gating and reversal of reflexes in ankle muscles during human walking. Exp Brain Res 82:351-358, 1990. 7. Fitts PM, Posner MI: In Human Performance, Brooks/Cole, Monterey; 1967. 8. Geurts ACH, Mulder Th, Nienhuis B, Rijken RAJ: Dualtask assessmentof reorganization of postural control in persons with lower limb amputation. Arch Phys Med Rehabif. 72:1059-1064, 1991. 9. Geurts ACH, Mulder Th, Nienhuis B, Rijken RAJ: Postural reorganization following lower limb amputation: possible motor and sensory determinants of recovery. Stand J Rehabil Med 24:83-90, 1992a. 10. Geurts ACH, Mulder Th, Nienhuis B, Rijken RAJ: Postural organization in patients with heriditary motor and sensory neuropathy. Arch Phys Med Rehabil 73:569-572, 1992b. 11. Geurts ACH, Mulder Th, Nienhuis B, Rijken RAJ: Influence of orthopaedic footwear on postural control in patients with hereditary motor and sensory neuropathy. J Rehabil Sci 5: 3-9, 1992~. 12. Geurts ACH, Mulder Th: Attentional demands in balance recovery following lower limb amputation. J Mot Behav 26:162-170, 1993. 13. Geurts ACH, Nienhuis B, Mulder Th: Intrasubject variability of selected force-platform parameters in the quantification of postural control. Arch Phys Med Rehabil 74:1144-l 150, 1993. 14. Groeneveld WH, Waterlander KJ, Kil A, Van Riel MPJM, Konijnendijk HJ: A solid-state recording system for ambulatory monitoring of postural signals. Proc I lth Symp Biotelemerry, Yokohama, Japan, pp 334-338, 1990. 15. Groeneveld WH: Instrumentation for ambulatory monitoring. In: Deliverable, 26, CAMA RC (A-2002)IAIMID G XIII, part B, 1994. 16. Kelso JA, Tuller B: Toward a theory of apractic syndromes. Bruin Lang 12:224-245, 1981.
ASSESSMENT OF MOTOR RECOVERY 17. Laurent M, Thomson JA: The role of visual information in control of a constrained locomotor task. J Mot Behav 20: 17-37, 1988. 18. Massion J: Postural control system. Cut-r Opin Neurobiol 4: 877-887, 1994. 19. McFadyen BJ, Winter DA: Anticipatory locomotor adjustments during obstructed human walking. Neurosci Res Commm 9:3744, 1991. 20. Messenger N, Bowker P: The role of gait analysis in clinical medicine: a survey of UK centres. Engng Med 16:221227, 1987. 21. Mulder Th: A process-oriented model of motor behavior: towards a theory-based rehabilitation auproach. Phvs Ther __ 71:157-164, 1991. 22. Mulder Th: Current ideas on motor control and learning: imulications for theraov. In: Suinal Cord Dvshmctions: Vol. 1I:‘lmplications for thk;apy, ed by Illis L. Oxford University Press, Oxford, pp 187-209, 1992. 23. Mulder Th: Current topics in motor control: implications for rehabilitation. In: Neurological Rehabilitation, ch 11, ed by Greenwood RJ, Barnes MP, McMillan TM, Ward CD. Churchill Livingstone, Edinburgh; pp 125-133, 1993a. 24. Mulder Th: The learning machine: ideas about adaptation and learning following nervous system damage. In: L&rung
29.
30. 31. 32.
33. 34. 35.
Motor Skills: Proceedings of the Third Svmoosium of the Graduate Institute of Human “Movement. ch 2, kd by Do&en-
25. 26. 27.
28.
bosch CAM, Out L, Commissaris DACM, Oudejans, RRD, Wimmers RH, Roszek B, Stins JF. Free University Press, Amsterdam, pp 29-47, 1993b. Mulder Th, Geurts ACH: The assessmentof motor dysfunctions: preliminaries to a disability-oriented approach. Hum Movem Sci 10565-574, 1991. Mulder Th, Geurts ACH: Recovery of motor skill following nervous system disorder: a behavioral approach. Bailliere’s Clin New-01 2:1-13, 1993. Nashner LM, Shumway-Cook A, Marin 0: Stance posture control in select groups in select groups of children with cerebral palsy: deficits in sensory integration and muscular coordination. Exp Brain Res 49:393409, 1983. Nashner LM, Shupert CL, Horak FB, Black FO: Organization
36. 37.
38.
39.
145
of posture controls: an analysis of sensory and mechanical constraints. In: Progress in Brain Research, ed by Allum JHJ, Hulliger M. Vol. 80, Elsevier Science, Amsterdam, pp 411423, 1989. Neumann 0: Automatic processing: a review of recent findings and a plea for an old theory. In: Cognition and Motor Processes, ed by Prim W, Sanders AE, Springer, Berlin, pp 255-293, 1984. Patla AE: In search of laws for the visual control of locomotion: some observations. .I exp Psychol: Hum Percept Pet-f 15:624-628, 1989. Patla AE: Adaptability of Human Gait: Implications for the Control of Locomotion. North Holland, Amsterdam, 1991. Patla AE, Robinson C, Samways M, Armstrong CJ: Visual control of step length during overground locomotion: task specific modulation of the locomotor synergy. J exp Psychol: Hum Percent Perf 15:603-617. 1989. Rosenbaum‘ DA: “Human Motor Control. Academic Press, New York, 1991. Rozendal RH: Gait analysis and ICIDH. J Rehabil Sci 2: 89-93, 1989. Teasdale N, Bard C, Dadouchi F, Fleury M, Lame J, Stelmath GE: Posture and elderly persons: evidence for deficits in the central integrative mechanisms. In: Tutorials in Motor Behavior II, ed by Stelmach GE, Requin, J. North Holland Publishing, Amsterdam, pp 917-931, 1992. Teasdale N, Bard C, Lame J, Fleury M: On the cognitive penetrability of posture control. Exp Aging Res 19:1-13, 1993. Teasdale N, Lajoie J, Bard C, Fleury M, Courtemanche R: Cognitive processes involved for maintaining postural stability while standing and walking. In: Sensorimotor Impairment in the Elderly, ed by Stelmach GE, Homberg V. Kluwer Publishing, Dordrecht, pp 157-168, 1993. Van Galen GP, Van Doom RRA, Schomaker LR: Effects of motor programming on power spectral density function of finger and wrist movements. J exp Psychol: Hum Percept Perf 16:755-765, 1990. Wickens CD: Engineering Psychology and Human Perhormante. Harper Collins Publishing, New York, 1992.
J. Electromyogr.
1050-6411(95)00029-1
The Assessment of Motor Recovery: A New Look at an Old Problem* Theo Mulder’,2, Bart Nienhuis’ and John Pauwels’ ‘Department
of Research and Development, Sint Maartenskliniek, PO Box 9011, 6500 GM Nijmegen and 2Neurological Institute, University of Nijrnegen, The Netherlands
Summary: After nervous system damage, functional recovery usually occurs. It is of great clinical importance to follow the course of recovery and, when possible, predict the extent. This measurement and prediction of recovery is one of the main challenges facing clinicians today. The majority of assessment procedures currently employed, however, are impairment-oriented; that is, these procedures are oriented primarily at the disease or organ level. Until now, few procedures have been available that focus at the disability or behavioural level. This lack of disability-oriented assessmentprocedures hinders the development and evaluation of rehabilitation programmes, as impairment-oriented assessment scores have little relevance for the prediction of daily functioning. In this paper, a preliminary task-set is presented that may function as an impetus for the development of novel disability-oriented assessmentprocedures in neurological rehabilitation. Motor behaviour is not assessedin vacua, rather as the end result of a continuous interaction between motor, sensory and cognitive processes. With this task set, it is possible to evaluate the changing influence of sensory and cognitive factors on the quality of motor performance over time. These changes, it is argued, are a crucial component of functional recovery. Copyright 0 1996 Elsevier Science Ltd Key Words: Motor recovery-Assessment-Rehabilitation. J. Electromyogr. Kinesiol., Vol. 6, 137-145, June
interests of the clinicians. Biomedical engineering therefore emphasized the technical improvement of these devices. In the present paper this emphasis is questioned, as it was found that the lack of clinical relevance results primarily from the absence of a general theoretical framework guiding how such devices can be used and not from their poor design. In the present text, a theoretical framework is sketched in general terms (for more details see Refs. 21-24). Clinical movement analysis is not uniform, and consists instead of a set of techniques directed at a diverse field of problems and disciplines. It can, therefore, be argued that the requirements for clinical
INTRODUCTION
In 1987 Messenger and Bowke?’ showed that the majority of clinical movement analysis facilities available in clinics in the UK were not routinely used. This led to the conclusion that the available devices were not advanced enough to serve the
Addresscorrespondence and reprint requeststo Theo Mulder. Dept. of Researchand Developm&t, Sini Maattenskliniek,P.O. Box 9011, 6500 GM Nijmegen,The Netherlands. *This paperwas presentedas a keynotelecturein the 1994ISEK Congressin Charleston.
137
138
T. MULDER ET AL.
movement analysis may differ across different disciplines (e.g. orthopaedics, neurology, rheumatology, and rehabilitation). For example, a detailed electromyographical and kinematic analysis of gait may be of value for the planning of orthopaedic surgery, but it is probably of less value for the physiotherapist who wants to predict daily motor behaviour. This variability underscores the need for differentiated routes in the development of assessment,This paper focuses on these varying needs of rehabilitation medicine; i.e. it focuses on the need for the development of a disability-oriented assessment. Such a disability-oriented approach, it should be noted, is a global approach and is not directed at the level of (isolated) movements, but at the level of skills. IMPAIRMENT
VS. DISABILITY
The terms ‘impairment’ and ‘disability’ play a prominent role in rehabilitation medicine. Impairment refers to the loss of a body structure and represents a disturbance at the organ level (e.g. lack of muscle strength, decreased range of motion). Disability refers to the consequencesof impairments for functional performance and activities and thus represents a disturbance at the level of the acting person. It is important to note that no direct relationship exists between the level of impairment and the level of disability34. Thus, a disability-oriented assessmentneeds to use terms different from those found in an impairment-oriented framework. ESSENTIALS OF MOTOR CONTROL Movement is the output of a hybrid functional system intimately linked to its environment in which sensory, cognitive and motor processes continuously interact. In the organization of a movement no clear separation exists between these processes;their relative weights may change, however, according to the complexity of the task, the skill of the performer and the integrity of the system. Recent models of motor control stress this interactive nature of Con~o~3,18,21-24,33
This construction process, however, is not the outcome of a strictly hierarchical and sequentially organized system, rather it is the result of a heterarchical, multilevel system where ‘higher levels’ control the invariant global aspects of the action and subordinate structures specify details. These ‘high level’ control structures activate functional
groups of muscles, which are constrained to act as a single coordinative structure and can be temporarily organized to accomplish a particular behavioural goal; the dimensionality - or number of degrees of freedom which apply to the control of a given task is thereby reduced, providing the basic organization of movement16. Muscle-specific details, such as force, velocity and spatial accuracy, are considered ‘lower order’ parametric problems which are dependent on the actual biomechanical and environmental context38. Such an architecture enables the system to be functionally specific, that is, to produce movements ideally suited to the specific requirements of attaining the movement goal. Movement is therefore not the result of a fixed sequence of signals transmitted over efferent pathways that determine all muscle-specific details of the required act, but it is the result of a flexible ‘negotiation’ between the available input and the required output. As there is an almost infinite number of ways of or_ramzing to attain movement goals, many solutions may lead to the same output. The ‘selection’ of certain solutions depends on the complexity and novelty of the task, the skill of the performer, the integrity of the system and the requirements of the environment. For example, many overlearned, routine-based tasks such as balance, gait, reaching, eating, etc., require none or only minimal conscious control. On the other hand, a novel or complex task, such as playing a violin, can only be performed under a conscious regime. The ability to shift between levels of control is an important characteristic of normal motor control. It is not a conscious process, rather the result of exploitation of an emergent characteristic of the central nervous system (CNS), namely the capacity to perform cost-benefit analyses leading to the most efficient strategy. For example, when normal automatic performance has been disturbed, the organism is still able to move by recruiting compensatory strategies. Each task situation, combined with the actual state of the organism, leads to its own optimal strategy as the organism continuously shifts among the available means of control. These strategies are to some extent flexible; i.e. small adjustments can be made without changing the strategy. But, if the adjustments are too large, a shift to a novel strategy is necessary. For example, some loss of proprioceptive input can be tolerated by the system, but when loss of proprioceptive input exceeds a certain level the system is forced to shift to another strategy (visual control).
ASSESSMENT OF MOTOR RECOVERY
These shifts enable the organism to reach the required goals even under non-optimal conditions. IMPLICATIONS
FOR MOTOR ASSESSMENT
From the above statements it can be inferred that solely analysing the output characteristics of motor behaviour (kinematic and biomechanical aspects of movement) reveals important information about the end result, but not about the strategies leading to that result, as these strategies remain hidden. Insight into the control mechanisms, however, is crucial to understanding the patient’s level of functional recovery. Assessment procedures focusing on functional recovery, therefore, should be designed in such a way that the adaptive processes can be followed over time. THREE PRINCIPLES
OF RECOVERY
Because recovery depends - to a large extent on the ability of the CNS to adapt to (peripheral) changes, the study of recovery is for a large part the study of adaptation. How can this adaptive process be studied over time, and what observable mechanisms can be distinguished are commonly raised questions. From the work performed in Nijmegen during the past 5 yr23-26,three principles of recovery can be distilled: (a) a decrease of cognitive regulation; (b) a decrease of visual dependency; and (c) an improvement of sensorimotor adaptability. Decrease in Cognitive Regulation The performance of a difficult (non-automatized) task interferes with other simultaneously performed tasks (see also Ref. 39). Therefore, if one uses an attention-demanding task, it is possible to use the degree of interference produced by this task on the primary task as a measure of the attention demands of the primary task. This idea has relevance for clinical (movement) analysis. For instance, overlearned motor tasks require very little information processing capacity. If the proficiency in two tasks remains unchanged, regardless of whether they occur simultaneously or separately, then at least one task would seem to be automatic. On the other hand, if a task is performed less well when it is combined with a secondary task, then both tasks are thought to require at least some attentional capacity (cognitive regulation). Thus, a decrease in dual-task interference reflects a decrease in cognitive regulation
139
over time, and may be seen as a sensitive indicator of an improved skill (re)acquisition and automaticity. Many motor control theories assume such a transition from attention-invested (or controlled) performance to attention-free (or automatic) performance during the learning of a new ~kill’*~*~~.These ideas have recently been applied in studies focusing on (disorders of) motor control in the elderly35-37, showing that with increasing age the attentional involvement in tasks such as gait increases substantially. The use of this dual task methodology in recovery research, however, is still extremely rare. In a series of experiments performed by Geurts et a1.*-13,the validity of this idea was tested. Their experiments focused on the recovery of balance after lower limb amputation. Amputation leads to a massive disruption of the afferent-efferent equilibrium and forces the CNS to develop alternative control strategies. It can be argued that as long as these novel strategies are not available, the system is vulnerable. Subjects with a leg amputation who were wearing their first prosthesis were instructed to stand as still as possible with their hands folded at the back. It was shown that, even at the very beginning of the rehabilitation process, patients were able to perform this task; i.e. their balance performance did not differ significantly from the performance of healthy age-matched controls. Patients stood on a dual-plate force platform which registered centre of pressure (cp) displacements. After 15 s, a slide was projected in front of the subject. The slide contained the words ‘red’, ‘blue’ and ‘green’. The words, however, were printed in a colour different from the meaning of the words, e.g. the word ‘red’ was printed in blue, the word ‘blue’ was printed in green. The patients were instructed to name as quickly as possible the colour of the words while suppressing the strong tendency to recite the word that was written. This is an attentiondemanding task, and the question was whether this task would interfere with the ability of the patient to maintain his/her upright balance. The results indicate that at the beginning of the rehabilitation process, the balance performance of patients was significantly hindered by the performance of a concurrent cognitive task; i.e. the sway increased under the influence of the dual-task performance. Such a dual-task effect could not be shown in healthy age-matched subjects. Most interesting, however, was that this effect slowly diminished over
140
T. MULDER ET AL.
time, indicating a re-automatization of balance contrOP.
With amputation, the system is suddenly deprived of its ‘normal’ input. An interesting question, therefore, is whether this dual-task effect would also be seen in patients with a chronic, slowly progressing disease leading to balance disturbances. Consequently, the experiment described above was repeated with patients suffering from hereditary motor and sensory neuropathy (HMSN) type I or II. These patients generally have distal muscle weakness as well as atrophy and loss of somesthesis, in particular, deep sensibility, Also, foot-ankle deformities develop, including pes cavus, transversoplanus, equinovarus and hammer toes9. Problems in postural control are one of the most frequent consequences of HMSN. The results of this study indicated that although these patients suffered from balance disorders, they were not hindered by the dual-task performance. One can conclude that as a result of the chronic character of the disease, the brain had ample time to reorganize its control procedures and to create a novel efferent-afferent equilibrium. Now, let us take the next step: what would happen if these patients were fitted with corrective orthopaedic footwear? The results were surprising. Geurts et al.9 showed that dual-task performance had no effect on balance control in barefoot conditions, but that it significantly hindered balance control (increased sway) in conditions where patients wore new rehabilitation footwear. Although at first sight this result seems remarkable, it is in accordance with the above described ‘fluid’ character of the CNS and underscores the permanent adaptive character of (motor) control. As soon as movement constraints are changed, the system is forced to adapt and shift to another control strategy. As such, a movement constraint can be defined as any variable that limits the organization and control of a movement. Decrease in Visual Dependency Normally a moving individual receives, interprets and acts appropriately on somesthetic input from the periphery. This interplay between motor and sensory information tunes movement to the specific context of the situation1**27*28.It can be argued, however, that when the somatosensoryinput is either distorted or absent, this interplay is (partly) lost, with vision becoming the most important input
source as a result of a strategic shift. In such a case, postural tasks are impossible to perform without vision. The system can no longer ‘build’ on the incoming somesthetic information, because this information has lost its redundant character. One can therefore predict that in early rehabilitation phases patients will show a marked visual dependency. Visual dependency, in this context, implies a disproportional dependency. Visual information (online) information is always used, and vision provides crucial information - particularly for the prospective control of gait 2*17,30-32, but the information is not used all the time to guide and control the action. We are aware, at an anecdotal level if not a scientific one, that while walking it is possible to shift one’s eyes from the trajectory without falling down or becoming disoriented. Thus, it is the degree of dependency that matters. We argue that this dependency is disproportionally large in patients with sensorimotor disorders. The question remains as to whether this dependency decreases over (recovery) time. To answer this question, postural control was assessedin patients with a unilateral leg amputation before and after a traditional rehabilitation training programme. The experimental set-up was basically the same as that employed in the above-mentioned studies: the centre-of-pressure fluctuations during quiet, upright standing on a force platform were registered with and without visual information. Although there was a small improvement in balance control with full visual information, the major improvement was a decrease in visua1 dependency over time’, with the visual-dependency scores approaching normal values at the end of the rehabilitation. This decrease in visual dependency within subjects clearly indicates a central re-integration of sensory input from the amputated limb into the multisensory control of posture. Although the general effect was significant, it is important to note that substantial inter-individual differences existed. Some subjects showed a relatively rapid decrease of visual dependency, whereas visual dependency for others either changed slowly or showed little change. It should be borne in mind that healthy adults also vary substantially in the degree of visual dependency. Nonetheless, the results are important because they again indicate that relevant aspects of the recovery process remain hidden from the observer as long as specific process-oriented manipulations are not used.
ASSESSMENTOF MOTOR RECOVERY
Improvement of Sensorimotor Adaptability A crucial prerequisite for successful locomotion is the capacity to adapt movement to the environmental demands and the organism’s goals. Two aspects of adaptation are distinguished here: anticipation and reaction. Although in most actions both aspects are served at the same time, for the sake of clarity they will be discussed separately here. Anticipation refers to the capacity of the organism to adjust the present locomotor dynamics to foreseeable environmental changes. It should, therefore, be distinguished from anticipatory postural adjustments occurring in association with voluntary movements, which are aimed at minimizing the equilibrium disturbances produced by the movement performance. Anticipatory movements require the integration of various information sources such as somatosensory, visual and auditory. In humans even cognitive and emotional information may be used. Because these adjustments reflect a feedforward postural control, which occurs before posture and equilibrium are disturbed, past locomotor experience must be continually combined with the actual perceptual input concerning both the body and the environment (see also Ref. 5). These necessary adaptive modifications are effected by feedback from the peripheral motor apparatus, from changes in tactual motion and from the perceptual pick-up of information. As mentioned earlier, visual input plays an important role in the detection of obstacle characteristics, of surface characteristics and of terrain layout3’. Examples of surfaces that require anticipatory control are slopes, stairs and surfaces with different frictional characteristics such as carpets or an ice-covered path. Visually detected information plays a fundamental role in permitting anticipatory changes in the basic locomotor pattern through feedforward contro130. Although the way in which the visual system extracts relevant information from the environment is not yet understood, it is probably fair to argue that identification of the safeness and traversability of the path constitutes the minimum information needed for the control of locomotion. In this context there is a real symbiotic relationship between perception and action, as the organism’s relationship with the environment is mediated primarily through the visual system. The integration of visual information in high level (cortical) loops seems to be necessary to perform the intentional regulations of locomotion mentioned above4. An important com-
plicating factor, however, is that (visual) information does not always have the same ‘weight’. What may be useful for one task may be useless for another task. An intention-coupled selection mechanism which can tune the motor system for the reception of the most relevant information, therefore, seems necessa+. Therefore, adequate anticipatory responses are largely based on central set effects allowing descending commands to prepare sensory and motor systems for anticipated stimulus and task conditions. Indeed, setting aspects of the response in advance can decrease the time it takes the CNS to transform an eliciting stimulus into an appropriate response. This is an important mechanism because postural responses are often initiated before the availability of relevant peripheral information. Central set allows individuals to modify their automatic responses to a postural stimulus by taking into account prior experience with perturbation characteristics and the effectiveness of prior responses. Central set, however, leads to errors in responses when the stimulus or external conditions unexpectedly change. In this respect, anticipation can be distinguished from reaction. Reaction refers to the ability of the organism to deal adequately with unexpected and sudden hindrances, obstacles, or changes. Even at a relatively low level of control, this ability becomes visible. Phase-dependent reflex modulation is one example. If during walking the leg flexion responses elicited by stimuli from different phases of the step cycle were always the same, problems would arise in the different phases. Duysens et a1.6showed that reactions to unexpected perturbations were modulated according to the phase of the step-cycle. Patla et a1.32also suggested a complex, task-specific modulation of the locomotor synergy to alter step length. This seemingly simple ability reflects a process of extreme complexity. The moving organism has to detect dangerous situations and use that information for an on-line modification of the activated locomotory programmes. The difference between anticipation and reaction is clear; with the former the subject is able to plan the movement beforehand, whereas with the latter this is not possible. It can be argued, therefore, that the ability to react requires a higher level of sensorimotor (re-)integration than anticipation does. Clinical studies on sensorimotor adaptability are rare. The majority have been performed with young, healthy adults performing obstacle-avoidance tasks
142
T. MULDER ET AL.
in a laboratory context. McFadyen and Winterr9, for example, showed that when a fixed, visible obstacle is placed at mid-swing a few strides in front of the subject, distinct changes in lower limb motor coordination are observed just prior to foot clearance. Large obstacles (up to 10% of leg length) induced a complete reorganization of the motor coordination of the lower limb. MacFadyen and Winter” termed these changes anticipatory locomotor adjustments. An interesting point is that these adaptive changes are very task specific, with the strategic changes fully tuned to the specific characteristics of the environment-performer relationship. The size of the obstacle, the moment of its appearance and the amount of visual information determines the nature of this adaptation. At present a series of experiments focusing on both aspects of sensorimotor adaptability is being performed in our laboratory. In these experiments, patients with lower limb amputation or total hip replacement have to negotiate a clearly visible hindrance with a height of approximately 6 inches. They must step on the hindrance (resembling a sidewalk) without changing walking speed. While walking towards the hindrance, a concurrent task must be performed (e.g. solving a cognitive, attention-demanding task). The results show that in the dual-task condition (compared to the normal, undisturbed condition) the walking speed decreased significantly, a difference that diminished over time. The same was true for the cognitive task: when compared with the end of rehabilitation ones sees that more errors were made during the first rehabilitation phases. In these experiments the cognitive task consisted of a woman’s voice pronouncing the words ‘high’ and ‘low’ with either a high or low pitch. There were congruent stimuli (saying ‘low’ with a low pitch and ‘high’ with a high pitch) and incongruent stimuli (saying ‘low’ with a high pitch and ‘high’ with a low pitch). The subjects had to indicate, while walking, whether the tone presented was low or high, regardless of voice information (the word spoken) PRELIMINARIES FOR A DISABILITY ORIENTED TASK SET FOR BALANCE AND GAIT We designed a task set that differs substantially from the conventional clinical movement-analysis procedures. Table 1 shows the ingredients for a bal-
ance assessment task, whereas Table 2 shows the ingredients for a gait assessmenttask set. MOTOR TASK SPECIFICATIONS Balance Task 1: Basic measurements: The patient is asked to stand as still as possible without any constraint so researcherscan gather individual reference values. Task 2: Perceptual manipulations: The patient is asked to maintain his/her balance for a period of 30 s under conditions of impaired vision so researchers can gather individual information concerning the degree of visual dependency. The patient wears special glasses with controlled transparency. Task 3: Cognitive manipulations (standing while per$orming a non-motor concurrent task): The patient is asked to stand as still as possible for a period of 30 s. The concurrent task is presented aurally and regulated by a computer. The task is either a mental calculation task or an auditory Stroop task. With
TABLE 1.
Design of disability-oriented motor tasks for the assessment of balance Description
Aim
Basic measurements:
Quiet upright standing under optimal conditions
Determination of reference values
Perceptual manipulations:
Quiet upright standing under impoverished illuminated conditions (dark glasses, milky white glasses) Quiet upright standing while performing a concurrent cognitive task (mental calculation) Quiet upright standing while waiting for a disturbance. Type and moment of perturbation is known (anticipation); type and moment of perturbation is not known (reaction)
Determination of visual dependency
Cognitive manipulations:
Motor manipulations:
Determination of cognitive dependency
Determination of the capability of the system to implement future disturbances into a motor programme under construction and to determine the capability of the system to react adequately to sudden environmental changes
143
ASSESSMENTOF MOTOR RECOVERY l
TABLE 2. Design of disability-oriented motor tasks for the assessment of locomotion Description
Aim
Basic measurements:
Flat-level walking, optimal conditions
Determination of reference values
Perceptual manipulations:
Walking under impoverished illumination conditions. Walking with ‘impaired peripheral vision. Walking with ‘impaired’ central vision
Determination of visual dependency
Cognitive manipulations:
Walking under conditions of loud noise (traffic); walking while performing a concurrent task (auditory Stroop task)
Determination cognitive dependency
Motor manipulations:
Walking at a certain (firm) speed while negotiating a clearly visible hindrance resembling a sidewalk (anticipation). Walking at a certain (firm) speed while approaching an unpredictable hindrance: a lightspot on the floor (reaction)
Determination of the system’s capacity to anticipate future expected disturbances and to react to sudden unexpected changes in the environment
l
Errors on the concurrent task. Increase/decrease of the velocity of centre of pressure (Vcp) as a result of the selected manipulations.
The sway parameters are measured by means of a dual-plate force-platform (for more details about the intrasubject variability see Ref. 13). Gait
of
the Stroop task, a loudspeaker voice speaks the words ‘high’ and ‘low’ in either a high or low pitch. The patient is instructed to name the pitch and to suppress the tendency to repeat the word using the same pitch as was spoken. This is an attention-demanding task, and the aim is to determine the cognitive dependency of task performance to gain more insight about the level of (re-)automatization. Task 4: Motor manipulation: The patient is asked to stand as still as possible while waiting for a (known/unknown) type of perturbation. The aim of this task is to gain more insight about the system’s adaptability. Dependent variables in disability-oriented balance assessment The following dependent variables have been recorded during the laboratory measurements:
Task 1: Basic measurement: Patients walk along a straight line without any constraint so reference values can be gathered. Task 2: Perceptual manipulations (walking under diferent visual conditions): Patients have to walk under conditions of impaired visual input in order to gather information concerning the degree of visual dependency. The patient walks while wearing special glasses with controlled transparency. The switching command of the glasses is regulated by the computer. While the patient is walking, the glasses change from the initial configuration to the final one. Possible configurations are: both lenses open; one lens open and the other closed; and both lenses 50% overshadowed. The timing of the configuration change can be controlled. Task 3: Cognitive manipulations (walking while petiorming a concurrent task): To evaluate the degree of cognitive dependency, patients have to walk while performing a non-motor concurrent task. The task is presented aurally and consists of a mental calculation task or an auditory Stroop task (see Task 3 of the balance assessmentset-up). Task 4: Motor manipulations (anticipation and reaction): In order to determine to what degree patients are capable of anticpating future (but expected) events, they walk while approaching a sidewalk. The patient has to step upon this sidewalk without decreasing velocity. This task is performed with and without a concurrent task. To assess the patient’s capability of reacting to sudden unexpected changes in the environment the patient is instructed to avoid a ‘hindrance’ (a spot of light) that is projected on the ground just before the subject reaches that area. The moment of projection as well as the place of projection can be manipulated. Dependent variables in disability-oriented analysis
gait
The following dependent variables have been measured during laboratory measurements:
144
T. MULDER ET AL.
0 Walking velocity. l Step frequency (measurement of intra-subject variability across the different task manipulations and over (recovery) time. l Step-length (measurement of intra-subject variability across the different task manipulations and over (recovery) time. 0 Pelvic rotation parameters reflecting the efficiency of locomotion. l Movement and velocity of the trunk in anteriorposterior and medial-lateral directions. The gait parameters are measured by means of a PRIMAS 3-D motion analysis system. CONCLUSION AND FUTURE PERSPECTIVES The suggested task set reflects all theoretical aspects discussed in this paper. This task set should enable clinicians to gain insight, not only into the visible end result of motor performance, but also into the processes leading to this end result. It is argued here that shifts in these (control) processes over time may give an objective indication of recovery. It is further argued that these aspects of recovery have more relevance for the prediction of daily activities than the conventional biomechanical impairment-related parameters. If a patient at the end of a treatment programme still shows a strong cognitive and/or visual dependency, as well as an impaired ability to negotiate obstacles, it will be clear that this patient is at risk under normal environmental circumstances. Assessmentprocedures currently used do not deliver this information because they focus primarily on output parameters acquired under optimal conditions. At this moment only a prototype version of the task set exists, and the generalization value of the collected data has to be established before it is clear whether this approach has real additional value for rehabilitation research. Therefore, patients who perform poorly at the end of rehabilitation should be followed over time to test whether they are at risk in terms of safety, dependency, skill etc., as compared to patients who performed well at the end of rehabilitation. Such a comparison requires the ambulatory monitoring of motor activities under daily conditions. Such a system (AMMA) has been developed in Rotterdam in the context of a European research project named CAMARC-II (Computer Aided Movement Analysis in a Rehabilitation
Context14*15).The system is able to store relevant information from several sensors placed on the patient’s body over a period of several hours. Because the system is also sensitive for sounds and ambient lighting levels, it is possible to study the relationship between environmental factors and the quality of gait. More research is needed to determine the dependent variables most sensitive for recovery. Therefore, the journey towards a disability-oriented motor assessment task has just begun, with only the first few steps described in this paper. REFERENCES 1. Adams JA: A closed-loop theory of motor learning. J Mot Behav 3:111-149, 1971. 2. Bardy BG, Baumberger B, Fluckiger M, Laurent M: On the role of global and local visual information in goal-directed walking. Acta Psychologica 81: 199-210, 1992. 3. Baron EU. The cerebral computer: introduction to the computational structure of the human brain. Erlbaum. Hillsdale, 1987. 4. Drew T: Visuomotor coordination in locomotion. Curr 0pin Neurobiol 1:652-657, 1991. 5. Dufosse M, Massion J: Posturo-kinetic interactions: modeling and modes of control. In: Tutorials in Motor Behaviour II, ed by Stelmach GE, Requin J. North Holland Publishing, Amsterdam, pp 125-155, 1992. 6. Duysens J, Trippel M, Horstmann GA Dietz V: Gating and reversal of reflexes in ankle muscles during human walking. Exp Brain Res 82:351-358, 1990. 7. Fitts PM, Posner MI: In Human Performance, Brooks/Cole, Monterey; 1967. 8. Geurts ACH, Mulder Th, Nienhuis B, Rijken RAJ: Dualtask assessmentof reorganization of postural control in persons with lower limb amputation. Arch Phys Med Rehabif. 72:1059-1064, 1991. 9. Geurts ACH, Mulder Th, Nienhuis B, Rijken RAJ: Postural reorganization following lower limb amputation: possible motor and sensory determinants of recovery. Stand J Rehabil Med 24:83-90, 1992a. 10. Geurts ACH, Mulder Th, Nienhuis B, Rijken RAJ: Postural organization in patients with heriditary motor and sensory neuropathy. Arch Phys Med Rehabil 73:569-572, 1992b. 11. Geurts ACH, Mulder Th, Nienhuis B, Rijken RAJ: Influence of orthopaedic footwear on postural control in patients with hereditary motor and sensory neuropathy. J Rehabil Sci 5: 3-9, 1992~. 12. Geurts ACH, Mulder Th: Attentional demands in balance recovery following lower limb amputation. J Mot Behav 26:162-170, 1993. 13. Geurts ACH, Nienhuis B, Mulder Th: Intrasubject variability of selected force-platform parameters in the quantification of postural control. Arch Phys Med Rehabil 74:1144-l 150, 1993. 14. Groeneveld WH, Waterlander KJ, Kil A, Van Riel MPJM, Konijnendijk HJ: A solid-state recording system for ambulatory monitoring of postural signals. Proc I lth Symp Biotelemerry, Yokohama, Japan, pp 334-338, 1990. 15. Groeneveld WH: Instrumentation for ambulatory monitoring. In: Deliverable, 26, CAMA RC (A-2002)IAIMID G XIII, part B, 1994. 16. Kelso JA, Tuller B: Toward a theory of apractic syndromes. Bruin Lang 12:224-245, 1981.
ASSESSMENT OF MOTOR RECOVERY 17. Laurent M, Thomson JA: The role of visual information in control of a constrained locomotor task. J Mot Behav 20: 17-37, 1988. 18. Massion J: Postural control system. Cut-r Opin Neurobiol 4: 877-887, 1994. 19. McFadyen BJ, Winter DA: Anticipatory locomotor adjustments during obstructed human walking. Neurosci Res Commm 9:3744, 1991. 20. Messenger N, Bowker P: The role of gait analysis in clinical medicine: a survey of UK centres. Engng Med 16:221227, 1987. 21. Mulder Th: A process-oriented model of motor behavior: towards a theory-based rehabilitation auproach. Phvs Ther __ 71:157-164, 1991. 22. Mulder Th: Current ideas on motor control and learning: imulications for theraov. In: Suinal Cord Dvshmctions: Vol. 1I:‘lmplications for thk;apy, ed by Illis L. Oxford University Press, Oxford, pp 187-209, 1992. 23. Mulder Th: Current topics in motor control: implications for rehabilitation. In: Neurological Rehabilitation, ch 11, ed by Greenwood RJ, Barnes MP, McMillan TM, Ward CD. Churchill Livingstone, Edinburgh; pp 125-133, 1993a. 24. Mulder Th: The learning machine: ideas about adaptation and learning following nervous system damage. In: L&rung
29.
30. 31. 32.
33. 34. 35.
Motor Skills: Proceedings of the Third Svmoosium of the Graduate Institute of Human “Movement. ch 2, kd by Do&en-
25. 26. 27.
28.
bosch CAM, Out L, Commissaris DACM, Oudejans, RRD, Wimmers RH, Roszek B, Stins JF. Free University Press, Amsterdam, pp 29-47, 1993b. Mulder Th, Geurts ACH: The assessmentof motor dysfunctions: preliminaries to a disability-oriented approach. Hum Movem Sci 10565-574, 1991. Mulder Th, Geurts ACH: Recovery of motor skill following nervous system disorder: a behavioral approach. Bailliere’s Clin New-01 2:1-13, 1993. Nashner LM, Shumway-Cook A, Marin 0: Stance posture control in select groups in select groups of children with cerebral palsy: deficits in sensory integration and muscular coordination. Exp Brain Res 49:393409, 1983. Nashner LM, Shupert CL, Horak FB, Black FO: Organization
36. 37.
38.
39.
145
of posture controls: an analysis of sensory and mechanical constraints. In: Progress in Brain Research, ed by Allum JHJ, Hulliger M. Vol. 80, Elsevier Science, Amsterdam, pp 411423, 1989. Neumann 0: Automatic processing: a review of recent findings and a plea for an old theory. In: Cognition and Motor Processes, ed by Prim W, Sanders AE, Springer, Berlin, pp 255-293, 1984. Patla AE: In search of laws for the visual control of locomotion: some observations. .I exp Psychol: Hum Percept Pet-f 15:624-628, 1989. Patla AE: Adaptability of Human Gait: Implications for the Control of Locomotion. North Holland, Amsterdam, 1991. Patla AE, Robinson C, Samways M, Armstrong CJ: Visual control of step length during overground locomotion: task specific modulation of the locomotor synergy. J exp Psychol: Hum Percent Perf 15:603-617. 1989. Rosenbaum‘ DA: “Human Motor Control. Academic Press, New York, 1991. Rozendal RH: Gait analysis and ICIDH. J Rehabil Sci 2: 89-93, 1989. Teasdale N, Bard C, Dadouchi F, Fleury M, Lame J, Stelmath GE: Posture and elderly persons: evidence for deficits in the central integrative mechanisms. In: Tutorials in Motor Behavior II, ed by Stelmach GE, Requin, J. North Holland Publishing, Amsterdam, pp 917-931, 1992. Teasdale N, Bard C, Lame J, Fleury M: On the cognitive penetrability of posture control. Exp Aging Res 19:1-13, 1993. Teasdale N, Lajoie J, Bard C, Fleury M, Courtemanche R: Cognitive processes involved for maintaining postural stability while standing and walking. In: Sensorimotor Impairment in the Elderly, ed by Stelmach GE, Homberg V. Kluwer Publishing, Dordrecht, pp 157-168, 1993. Van Galen GP, Van Doom RRA, Schomaker LR: Effects of motor programming on power spectral density function of finger and wrist movements. J exp Psychol: Hum Percept Perf 16:755-765, 1990. Wickens CD: Engineering Psychology and Human Perhormante. Harper Collins Publishing, New York, 1992.