Reaction and movement times in patients with hemiparesis for unilateral and bilateral elbow flexion.

Key Words: Exercise therapy; Hemiplegia, general; Movement; Reaction time; Upper extremity, elbow.

Application of bilateral and often simultaneous exercises to the upper extremities (UEs) of patients with hemiparesis derives from both pragmatic and theoretical reasons. From a pragmatic standpoint, the relatively slow recovery of the paretic UE precludes early unilateral exercises. Thus, activities assisted by the nonparetic limb are encouraged. Because many patients are discharged from the rehabilitation setting before gaining full motor control over the affected limb, paretic limbs are not necessarily used in unilateral active exercises during the hospitalization period.

From a theoretical point of view, the importance of bilateral (simultaneous) exercises for rehabilitation of the UE is emphasized in some approaches to therapeutic exercises.[1] Contraction of muscles on the nonparetic side appears to facilitate muscular activity on the paretic side of the body (bilateral transfer).[2-5] Accordingly, transfer of muscular activity from the nonparetic muscle to the homologous muscles on the contralateral (paretic) side is frequently utilized for gaining strength and endurance in the affected limb.[6-8] Finally, recent findings that show functional deterioration of the nonparetic UE of patients with hemiparesis provide additional justification for training of both the paretic and nonparetic limbs.[9-11]

Some evidence, however, indicates that bilateral exercises of the UEs may have adverse effects on each limb's function. Ohtuski[12] and Vandervoort et al[13] demonstrated that in healthy subjects, force production decreases when switching from unilateral to bilateral flexion and extension. Di Stefano et al[14] noted that movement velocity in a unilateral manual task was higher than when the same task was performed bimanually.

In monkeys with unilateral deafferentation, usage of the intact limb was found to interfere with functional recovery of the deafferented limb.[15] Rehabilitation of the deafferented limb was assisted by restraint of the intact limb, which forced usage of the impaired limb.[15] In a similar way, functional improvement in the affected UE of patients with chronic hemiparesis was achieved by restraint rather than by coactivation of the unaffected limb.[16] Black et al[17] studied monkeys after unilateral surgical ablation of the cortical precentral forelimb area. They found that combined training of the weak forelimb together with the normal forelimb did not achieve better results than training of the weak forelimb alone. They concluded, therefore, that the intact limb plays a negligible role in rehabilitation of the weak forelimb.

Two variables often used to quantify motor performance are reaction time (RT) and movement time (MI). Reaction time is the time that transpires between a presentation of a stimulus and the initiation of a response. Movement time is the time taken to complete a task after it has been initiated. The summated time of RT and MT is often called the response time.[18] In patients with hemiparesis, both RT and MT of the paretic UE are longer than those of the nonparetic UE. Furthermore, RT and MT of the nonparetic UE are longer than those of age-matched control subjects.[19,20] Cohn[21] found a slowing of the nonparetic limb in bimanual simultaneous movements of patients with hemiparesis. Similarly, in healthy subjects, an increase in RT and MT was noted during bilateral activities as compared with unilateral activities.[14,22,23]

Optimal control of movement initiation and movement speed is required for normal functional performance. Therefore, enhancement of delayed RT and slowed MT in patients with hemiparesis should be a major goal of physical therapy intervention. Review of the literature, however, reveals that neither RT nor MT have specifically been addressed in evaluation and treatment of patients with hemiparesis, nor have the possible effects of separated or combined activity of the UEs on movement initiation and speed been considered in the context of physical therapy intervention.

The main goal of this study was to determine the effects of bilateral activity of the UEs on the RT and MT of patients with hemiparesis. The time of movement initiation and movement speed were controlled. Therefore, requirements for selection of movement type as well as demands for movement accuracy were removed. Accordingly, the task used could reveal the effects of bilateral activity on time parameters, without interaction with other movement qualifiers. The effects of bilateral action on RT and MT of patients with hemiparesis were compared with those obtained in age-matched control subjects to determine whether any of the effects are specific to the condition of hemiparesis.

Method

Subjects

The subjects comprised an experimental group of patients with hemiparesis (n=25) who were residents of a rehabilitation hospital and a control group of age-matched healthy subjects (n=26). Common admission criteria for the two groups were absence of cognitive impairments, unimpaired hearing, and absence of movement disorders in both UEs (control subjects) or in one UE (patients). Absence of movement disorders was confirmed by the investigators by asking the subjects to perform active free shoulder movements. In addition, the patients had to demonstrate their ability to voluntarily flex the elbow on the paretic side, from a partially extended 150') position, over at least 30 degrees, within the constraints of the testing device (see "Instrumentation" section). Patients with bilateral brain damage were not used as subjects.

The patient group consisted of 14 men and 11 women with a mean age of 73 years (SD=1.45, range=61-89). Thirteen patients had left hemiparesis, and 12 patients had right hemiparesis. Sites of the brain lesions were confirmed by computed tomography scans in only 11 patients. According to the medical charts, however, in 24 of the patients, hemiparesis was due to a thromboembolic brain infarction in the territory of the internal carotid artery. One patient's hemiparesis was caused by head trauma. The time interval between the insult and the beginning of the study averaged 2.5 months (SD=2.22, range=1-9). In 24 patients, the right UE was dominant prior to hemiparesis.

The patients' activities of daily living were assessed by the Barthel Index.[24] Scores on the Barthel Index range from 0 (complete dependency) to 100 full independence in activities of daily living). The mean Barthel Index score of the patients was 87 points (SD=11.26, range=45-100). Assessment of UE function was conducted using the Fugl-Meyer evaluation protocol for the UE, which has a range of scores of 0 to 66.25 The mean Fugl-Meyer score of the patients was 56.36 (SD=11.36, range=19-66).

The control group consisted of 12 men and 14 women with a mean age of 72 years (SD=6.47, range=62-90), all with right-hand dominance.

Instrumentation

The experimental apparatus consisted of a joint support for each UE. Each arm was rested against a fixed, half-cylindrical, 18-cm-long support at a 45-degree downward-pointing angle, perpendicular to the shoulder's horizontal axis. A single horizontal joint, appositioned to the elbow, connected an aluminum rod with the arm rest. When in the nearly horizontal position (5[degrees] of elevation), the 60-cmX 12-mm rod lay against a fixed metal stopper. Free rotational movements of this rod were allowed within a quarter cycle, bounded by the horizontal and vertical (upward) directions. An adjustable horizontal handle was attached to the aluminum rod and was held by the subject in a pronated forearm position. The distance of the handle from the elbow joint was adjusted to enable free flexion movements. A microswitch, which was incorporated in the horizontal stopper, was activated as the aluminum rod was lifted from it and signaled elbow flexion onset. A second microswitch was activated when the rod passed through 120 degrees of elbow flexion (equivalent to 30[degrees] of flexion from the elbow's initial position). Activation of that microswitch signaled attainment of the movement goal. The four microswitches (two on each side) were connected to a Commodore 64 microcomputer,(*) through which the experiment was controlled. All switch closures were timed, with 1-millisecond accuracy, relative to each trial onset.

Procedure

Subjects were seated in front of the experimental apparatus holding the handles of the horizontally placed rods, with their arms placed comfortably in their supports. Each subject performed four sets of 16 elbow flexion movements (trials). Each set consisted of only one type of movement (ie, unilateral right, unilateral left, or bilateral elbow flexions). The first set was unilateral, applied to the nonparetic UE and treated as familiarization; the results were not saved. The next three sets were applied to the paretic UE, the nonparetic UE, and simultaneously to both UEs in a random order. This design of a simple RT task, in which all trials in a block were of the same type, was preferred over a design of choice RT task, in which a three-type cue would indicate the type of movement to be made in each trial. This procedure reflects an effort to eliminate all factors, other than direct motor control, that could influence the time needed for movement initiation and execution.

Subjects were instructed to initiate movement immediately on hearing an auditory signal and to move the rod(s) as fast as possible past the end point, which was marked by a mechanical pointer(s). A free return movement to the rod's horizontal position followed. The auditory signal, a 200-millisecond, 1-kHz tone, was contingent on the levers being held down horizontally. A random intertrial interval of 0.5, 1.5, 2.5, or 3 seconds preceded each trial.

Reaction time was measured as the time interval between the auditory signal and the release of the first microswitch. Movement time was calculated as the time interval between release of the first microswitch and activation of the second microswitch. No attempt was made to assess the reliability of measurements of RT and MT for either group.

Data Analysis

An analysis of variance (ANOVA) for repeated measures was used to evaluate the RT and MT differences (1) between the patients and the control subjects, (2) between the unilateral and bilateral tasks, and (3) between the paretic and nonparetic UEs of the patients and between the left and right UEs of the control subjects. The repeated-measures factors were (1) task (unilateral versus bilateral) and (2) UE (paretic versus nonparetic in the patients, left versus right in the control subjects). These factors were analyzed across the group factor (patients versus control subjects). The dependent variables were RT and MT.

Correlation analysis was used to study associations between the dependent variables and age, time elapsed since the stroke, Barthel Index scores, and Fugl-Meyer index scores of the paretic UE.

Results

Reaction Time in Unilateral and Bilateral Elbow Flexion

The mean RT values for each UE in unilateral and bilateral elbow flexion are presented in Table 1 for both patients and control subjects. In the unilateral task, the mean RT of the patients' nonparetic UE (436.8 milliseconds) was about 120% longer than the mean RT of the control subjects' slower hand (363.1 milliseconds). This difference was even larger for the paretic UE, in which the mean RT of 459.9 milliseconds amounted to nearly 127% of the control subjects' mean RT. Performance of the bilateral task caused increased RTs in both the patients and the control subjects. in this task, the patients' mean RT of the nonparetic UE (466.7 milliseconds) and that of the paretic UE (502.6 milliseconds) exceeded that of the control subjects' slower UE (382.4 milliseconds) by 122% and 131%, respectively.

[TABULAR DATA 1 OMITTED]

These effects of group and task were found to have highly significant effects on RT (F= 13.58; df= 1,49; P =.0006 for group and F= 16.43; df= 1,49; P =.0002 for task). No significant interaction between these two factors was found. The mean RTs in the paretic UE of the patients were significantly longer than the mean RTs of the nonparetic UE (F= 5.04; df= 1,49; P =.03). The RT differences between the right and left extremities of the control subjects were negligible. These group differences were manifested in the finding of a significant interaction between grouping and UE (F= 6.82; df= 1,49; P=.01). A summary of the ANOVA results is provided in Table 2.

[TABULAR DATA 2 OMITTED]

Movement Time in Unilateral and Bilateral Elbow Flexion

The mean MT values for the two tasks are given in Table 3 for both patients and control subjects. In the unilateral task, the mean MT of the patients' nonparetic UE (680 milliseconds) exceeded that of the control subjects' slower UE (562.8 milliseconds) by 121%. This difference was larger (137%) when the comparison was made with the mean MT of the patients' paretic UE (769.4 milliseconds). Similarly, in the bilateral task, the mean MTs of the patients were longer than those of the control subjects. Accordingly, the MT for the patient group was different from that of the control group (F=11.69; df = 1,49; P=.001).

[TABULAR DATA 3 OMITTED]

In both groups, the mean MT of each UE in bilateral elbow flexion was significantly longer than that of the corresponding UE in the unilateral task (F= 11.83; df = 1,49; P=.001). This increase in MT was similar for both UEs in the control group, being 37.9 milliseconds (7%) for the left UE and 54 milliseconds (10%) for the right UE. In the patient group, the mean MT of the nonparetic UE was prolonged by 147.6 milliseconds (22%) in the bilateral task, whereas the mean MT of the paretic UE was prolonged by only 85.7 milliseconds (11%). Accordingly, mean MT differences between the paretic and nonparetic UEs was 13% for the unilateral task, but only 3% in the bilateral task. In the control subjects, the corresponding differences between the right and left UEs were 8% and 4%, respectively. The statistical significance of the UE effect on MT was marginal (F=3.79; df=1,49; P=.057). The probability of detecting existing differences in MT between the two extremities of the patients, however, was low, especially for the bilateral task. This can be learned from the power analysis performed on the difference in MT between the two elbows of the patients (Figs. 1 and 2). High variability in values of MT, especially in the bilateral task, is the most obvious explanation for this low power. A complete listing of the ANOVA test results is given in Table 4.

[TABULAR DATA 4 OMITTED]

Relationships Between Descriptors of the Patients and the Dependent Variables

Reaction time and MT differences between women and men were not significant. No associations were found between either RT or MT and age, time elapsed since the stroke, and Barthel Index scores. Prolongation of both RT and MT in the bilateral task occurred in patients with right hemiparesis as well as in patients with left hemiparesis.

A significant negative correlation r= -.48, P=.008) was found between the Fugl-Meyer UE score and the paretic/nonparetic UE mean RT ratio in the bilateral task. Similarly, a significant negative correlation (r= -.40, P=.02) was found between the FM score and the paretic/nonparetic UE mean MT ratio in the bilateral activity. A comparable negative correlation between the paretic/nonparetic UE mean MT ratio and the FM score r= -.34, P=.05) was also found in the unilateral task. Such a correlation was not obtained for the RT variable under this task condition.

Discussion

The time taken by the patients with hemiparesis to initiate and execute fast elbow flexion unilaterally and bilaterally was studied in the present work. The findings obtained for the patients with hemiparesis were compared with those obtained for the age-matched healthy control subjects, who performed the same motor tasks. The well-known slowing of the paretic UE was confirmed. In addition, we found that bilateral performance of a simple elbow flexion task, even when devoid of any requirements for interlimb cooperation, caused a significant slowing of both UEs in both patients and control subjects. The possible mechanisms of these findings and their implications are discussed below.

Reaction time and MT represent different aspects of motor function. Reaction time is a single quantity that reflects the level of central nervous system arousal, sensory-perceptual processing time, and delays imposed by response selection and movement programming. It is considered a nonspecific capacity that is not well localized.[26] Movement time represents a specific executive capability of the neuromuscular system and is thus a more concrete function.[18]

The mean RT for the patients' paretic UE was longer than the mean RT for their nonparetic UE. The mean RT for the patients' nonparetic UE, however, was still longer than the control subjects' mean RT. These findings are in agreement with those of prior reports.[14,19,20] Our experimental procedure did not involve a process of response selection (ie, subjects did not have to choose between different types of elbow flexion in each trial according to the meaning of a sensory cue). In addition, the required movement was simple, stereotyped, and thus not likely to involve preprogramming in each trial. Accordingly, the increase in the patients' RT probably reflects a nonspecific deficiency, which hindered their readiness to initiate the required movement. This conclusion is strengthened by the fact that an increase in RT was found in both UEs of the patients. The change from the unilateral task to the bilateral task resulted in a similar prolongation of RT in both groups of subjects. This RT increase thus appears to reflect additional central processing for simultaneous movement initiation,[22] rather than a specific effect of hemiparesis.

Movement times of the patients' nonparetic and paretic UEs in the unilateral task were lengthened by 21% and 37%, respectively, relative to the slower UE (left) of the control subjects. During bilateral movements, slowing of MT was noted in both UEs of subjects in both groups. The relative slowing of the patients' MT in comparison with that of the control subjects was even more accentuated in the bilateral task.

Interestingly, in switching from unilateral to bilateral elbow flexion, the nonparetic UE underwent greater movement slowing than the paretic UE. Consequently, an almost equal speed of the patients' UEs was attained by compromising performance of the nonparetic limb. It is known that when the UEs are simultaneously confronted with tasks of unequal difficulties, performance speed is determined by the limb that faces the more difficult task and that moves more slowly.[22] In our study, both UEs performed the same task, but the reduced motor capability of the damaged hemisphere must have rendered the task more difficult for the paretic UE. Consequently, control of this UE probably became the rate-limiting factor in bilateral performance. As Kelso et al[22] Suggested, simultaneity of a bilateral action is achieved by functional grouping of muscles on both limbs within a central program, rather than by controlling each limb independently.

The results of our study suggest that the program formed in patients with hemiparesis for the nonparetic UE is adjusted according to the limitations of control over the paretic limb. The large standard deviations of MTs in the bilateral task (Tab. 3) and the low probability of disclosing differences in MT between the patients' UEs (Fig. 2), however, point to the need for further inquiry into the effect of bilateral tasks on performance rate in patients with hemiparesis. A likely explanation for an increased MT in the bilateral task in both extremities of patients as well as in age-matched control subjects is that in a bilateral task, the two hemispheres interact to coordinate action of both limbs. Presumably, this interaction is a time-consuming process that can account for an increased MT in bilateral movements. Hence, an increased MT in bilateral elbow flexion is not unique to patients with hemiparesis. Its extent in patients with hemiparesis, however, is most probably determined by the poorer control mechanisms of the paretic UE rather than by the normal control mechanism of the nonparetic UE.

The results of our study have several implications. Temporal variables such as RT and MT need to be incorporated into the assessment protocols of patients with hemiparesis. Unlike other variables that are regularly evaluated by physical therapists, such as reflex activity, sensation, and stage of recovery,[24] RT and MT are fully quantitative and can constitute objective measurements of functional recovery. Because of their inherent variability, protocols of repeated measures will probably be required. Bitenski and colleagues[27] have recently used such measurements and reported improvement in RT over time in patients with hemiparesis. Our results also suggest that physical therapy interventions should address the variables RT and MT specifically. Exercises in which the speed of movement initiation and execution is controlled, and for which feedback is given, should be included in the treatment regimen. Efforts to minimize delays in movement initiation and speed of the paretic UE should emphasize unilateral exercises of this UE, rather than bilateral activity involving the nonparetic UE.

Conclusions

The RT and MT of simple elbow flexion in each UE of the patients with hemiparesis were prolonged relative to those of the age-matched control subjects, with the paretic limb being slower than the nonparetic limb. Simultaneous flexion of the two elbows resulted in additional slowing of movement. This temporal prolongation of MT in bilateral movements was greater in the nonparetic UE than in the paretic UE, presumably because of the rate limitation imposed on the nonparetic limb by the MT of the paretic limb. Therefore, bilateral exercises cannot be expected to increase the speed at which the paretic UE is moved. Furthermore, bilateral exercises appear to compromise performance of the intact UE.

Acknowledgment

We thank Ilana Gelernter, Department of Mathematics and Statistics, Tel Aviv University, for the statistical analysis of the data. (*) Commodore Business Machines, 1200 Wilson Dr. West Chester, PA 19380.

References

[1] Flanagan E. Methods of facilitation and inhibition of motor activity. Am J Phys Med. 1967; 46:789-798 [2] Mills VM, Quintana L. Electromyography results of exercise overflow in hemiplegic patients. Phys Ther. 1985;65:1041-1045. [3] Pink M. Contralateral effects of upper extremity proprioceptive neuromuscular facilitation patterns. Phys Ther. 1981;61:1158-1162. [4] Brunnstrom S. Movement Therapy in Hemiplegia. New York, NY: Harper & Row, Publishers Inc; 1970:22-23. [5] Knott M, Voss D. Proprioceptive Neuromuscular Facilitation. 2nd ed. New York, NY: Harper & Row, Publishers Inc; 1968:90. [6] Moritani T, DeVries HA. Neural factors vs hypertrophy in the time course of muscle strength gain. Am J Phys Med. 1979;58:115-130. [7] Wellock LM. Development of bilateral strength through ipsilateral exercise. Phys Ther Rev. 1957;38:671-675 [8] Hellebrandt FA, Houtz J. influence of bimanual exercise on unilateral work capacity. Arch Phys Med Rehabil. 1950;2:446-452. [9] Colebatch JG, Gandevia SC. The distribution of muscular weakness in upper motor neuron lesions affecting the arm. Brain. 1987; 1 12:749-763. [10] Halaney ME, Carey JR. Tracking ability of hemiparetic and healthy subjects. Phys Ther. 1989;69:342-348. [11] Jones RD, Donaldson IM, Parkin PJ. Impairment and recovery of ipsilateral sensory-motor function following unilateral cerebral infarction. Brain. 1989;112:113-132. [12] Ohtsuki T. Decrease in human voluntary isometric arm strength induced by simultaneous bilateral exertion. Behav Brain Res. 1983; 7:165-178. [13] Vandervoort AA, Sale DC, Moroz J. Comparison of motor unit activation during unilateral and bilateral leg extension. J Appl Physiol. 1984;56:46-51. [14] Di Stefano M, Morelli M, Marzi CA, Berlucchi G, Hemispheric control of unilateral and bilateral movements of proximal and distal parts of the arm as inferred from simple reaction time to lateralized light stimuli in man. Exp Brain Res. 1980;38:197-204. [15] Taub E. Somatosensory deafferentation research with monkeys: implications for rehabilitation medicine. In: Ince LP, ed. Behavioral Psycbology in Rehabilitation Medicine. Baltimore, Md: Williams & Wilkins; 1980: chap 11. [16] Wolf SL, LeCraw DE, Barton IA, Jann BB. Forced use of hemiplegic upper extremily to reverse the effect of nonuse among chronic stroke and head-injured patients. Exp Neurol. 1989;104:125-134. [17] Black P, Markowitz RS, Cianci S. Recovery of motor function after lesions in the motor cortex of monkey. In: Outcome of Severe Damage to the CNS: Ciba Foundation Symposium. Amsterdam, the Netherlands: Elsevier Science Publishers BV; 1975:65-83. [18] Sage GH. Motor Learning and Control: A Neuropsychological Approach. Dubuque, Iowa: Wm C Brown Cos; 1984:24-27. [19] Kaizer F, Bitenski KN, May N, et al. Response time of stroke patients to a visual stimulus. Stroke. 1988;19:335-359. [20] Benton AL. Reaction time in brain disease: some reflections. Cortex. 1986;22:129-140. [21] Cohn R. Interaction in bilaterally simultaneous voluntary motor function. AMA Arch Neurol Psychiatry. 1951;65:472-476. [22] Kelso JAS, Southard DL, Goodman D. On the nature of interlimb coordination. Science. 1979;203:1029-1031. [23] Jeeves MA, Dixon NF. Hemispheric differences in response rate to visual stimuli. Psychon Sci. 1970;20:249-251. [24] Mahoney FI, Barthel DW. Functional evaluation: the Barthel Index. Maryland State Med J. 1965;14:61-65. [25] Fugl-Meyer AR, Jaasko L, Leyman I, et al. The post-stroke hemiplegic patient, I: a method for evaluation of physical performance. Scand J Rehabil Med. 1975;7:13-20. [26] Coslett HB, Bowers D, Heilman KM. Reduction in cerebral activation after right hemispheric stroke. Neurology. 1987;37:957-962. [27] Bitenski NK, Mayo NE, Kaizer F. Changes in response time of stroke patients and controls during rehabilitation. Am J Phys Med Rehabil. 1990;69:32-38.

Commentaries

Following are two commentaries on "Reaction and Movement Times in Patients With Hemiparesis for Unilateral and Bilateral Elbow Flexion."

Physical therapists continue to pursue their interests in movement control with increasing regularity and persistence. One such example of applying a very basic movement paradigm to patients with stroke has been presented by Dickstein and colleagues. They are to be commended on making an important contribution to our literature. Although slowing of total limb or joint segmental movements in a hemiparetic upper extremity is intuitively obvious, the delays clearly quantified in reaction time (RT) and movement time (MT) now provide objective verification for movement delays that clinicians have recognized for many decades. Specifically, the authors conclude from the analysis of their data that increasing RT during bilateral elbow flexion may reflect additional central processing times necessary for simultaneous movements to be initiated rather than specific events related to hemiparesis itself. This important speculation, which is a logical deduction from the data provided, offers us an opportunity to further test the degree to which this explanation is justified. Equally as important, however, is the need to examine the premises on which this study was initiated as well as the data provided. Accordingly, the comments offered in this discussion fall within three areas: premise for the work, mechanisms to explain observations, and relevance of findings to the practice of physical therapy.

Premise for the Work

The effort to which this study directs itself is bilateral simultaneous elbow flexion. There is no resistive work involved in these movements. Hence, we need to recognize that mechanisms such as "bilateral transfer" and overflow," which rely heavily on unilateral resistance to precipitate contralateral movement, are probably irrelevant to even rendering a premise for why the work should be done or for how these "older" techniques are related to this study.

The observations cited by Dickstein and co-workers, that healthy subjects will have a reduction in force outputs and a slowing in movement velocity when unilateral tasks are done bilaterally, is important. Many studies seeking to examine speed and coordination of movement use RT and MT as important dependent variables. The operating assumption here is, of course, that somehow faster reactions and quicker movement, performed either unilaterally or bilaterally by patients with stroke, are appropriate behaviors to initiate or complete a functional task. We need to assess the relevance of such an assumption. In fact, the degree of intersegmental coordination and the circumstances on which an upper extremity is engaged may be more important than the speed generated in efforts to complete a task.

Under normal circumstances, the elbow participates in a movement paradigm that requires proximal stabilization of the shoulder rather than relative quiescence in upper-extremity activities, as was used in this study. Therefore, why temporal variables such as RT and MT "need to be incorporated into the assessment protocols of patients with hemiparesis" must be examined and debated among physical therapists treating patients who have neurological disorders. A kinematic analysis of how the elbow joint or the entire upper extremity moves through space from one target to another may be a more important assessment tool than measuring speed and reactions. The only readily defensible need for monitoring and "improving" hemiparetic upper-extremity joint movement speeds would be in ensuring protective responses due to sudden body perturbations. This consideration was an important screening criterion in our study on forced use,[1] cited by Dickstein and colleagues. In forced use applications that require immobilization of uninvolved (less involved) upper extremities of patients with neurologic deficits, the need to move the involved upper extremity in a rapid and timely manner to reduce bodily trauma on falling is an obvious consideration.

Mechanisms to Explain Observations

The authors spend considerable time explaining the results in terms of factors that affect RT and MT. Although one cannot deny the importance of such issues as arousal and central processing time after receiving exteroceptive cues, there is a body of literature totally neglected by the authors. That literature has as its basis an understanding of the anatomy and physiology of bilateral descending motor control systems. Such an understanding may explain why bilaterally induced activities are "slower" than unilateral activities in either patients with central nervous system deficits or in healthy individuals.

We have known for some time that efferent pathways in subhuman primates, originating within the core of the ventromedial brain stem, project bilaterally and can profoundly influence activation of motoneurons to proximal muscles in both the upper and lower extremities.[2] Brinkman and Kuypers[3] have shown that each half of the monkey brain has full control over the contralateral extremity and may also control proximal and complex movements of the ipsilateral extremity. The multiplicity of synapses receiving bilateral projections from the precentral gyrus at interneurons and motoneurons of wrist, finger, and arm muscles during bilateral voluntary movements in monkeys has been further supported by Matsunami and Hamada.[4]

Collectively, these studies suggest that brain-stem and cortical neurons can influence upper-extremity musculature bilaterally. if an individual performs a bilateral activity and additional or "redundant" pathways activate homologous sets of interneurons and motoneurons, the "synaptic noise" occurring at cervical cord levels is bound to be increased. There is no rule stating that the increased bilateral drive to such motoneurons would enhance synaptic efficacy. Accordingly, if there is more "traffic" for interneurons to sort out, synaptic delays, manifested as "reaction times" prolonged by 10 to 100 milliseconds, would not be unusual in healthy individuals. Given the inherent pathology to efferent systems caused by internal carotid arterial lesions, increased synaptic delays in the order of 0.5 second during bilaterally initiated tasks would not seem unreasonable. Training to enhance the "speed" of movement that might possibly reduce synaptic noise at the spinal cord level, and thus lead to functionally relevant changes, would need to be explored by the authors in subsequent work. Furthermore, the influence of other factors, such as coactivation of antagonist muscles, must be assessed before ascribing changes in RT or MT to any one mechanism.

Relevance of the Findings to the Practice of Physical Therapy

The detailed numerical analyses provide an excellent example of the generation of statistically significant findings that may be of restricted importance within a clinical context. We must note that, based on scores provided by the Barthel Index and from the Fugl-Meyer evaluation protocol, these patients with stroke were comparatively unimpaired. This factor alone may limit the relevance of these results to more seriously impaired patients with stroke.

Equally as important, the influence of limb position and voluntary task requirements would profoundly affect speed of movement and reaction to a visual or auditory stimulus to move. One wonders whether performing such movement tasks under conditions requiring voluntary shoulder stabilization or elbow extension or flexion against gravity would influence the results and their interpretation. Clearly, these types of movements are far more clinically relevant than the ones described in this research article.

Very little can be said about the degree to which MTs were influenced by co-contraction during the 30-degree range of elbow motion. in a study that compared bilateral upper-extremity movements in patients with stroke with a targeted biofeedback training paradigm, we5 were able to show that bilateral elbow flexion could be enhanced in terms of range of motion and biceps brachii muscle electromyographic (EMG) output when the EMG output from the "uninvolved" biceps muscle was used to guide the EMG output from the homologous "involved" muscle. Importantly, the degree of co-contraction was relatively uninfluenced by this training paradigm. On the other hand, homologous training of the triceps brachii muscle in either gravity-eliminated or antigravity positions yielded EMG activity that could be enhanced via this "motor copy" technique and was accompanied by an increased active range of elbow extension. During such a training paradigm, however, biceps brachii muscle EMG activity persisted; that is, co-contraction was still evident. The relative importance of antagonist muscles to movement initiation in patients with neurological disorders cannot be underestimated.[6] Additionally, an understanding of why RT and MT are delayed in unilateral or bilateral tasks must also include an analysis of progressive changes in viscoelastic properties about joints in which the actions are governed by spastic muscles.[7,8]

Several more observations made by the authors were not discussed. These observations were most intriguing and perhaps require a response. For example, given that all control subjects were right-handed, one would be interested in knowing why RTs were faster for the left elbow but MTs were faster for the right elbow during bilateral elbow flexion efforts (compare Tabs. 2 and 3). Also, the authors developed a comparison between functional scores and a paretic-to-nonparetic upper extremity RT-TO-MT ratio. As could be expected, given the comparatively high functional scores for this patient complement, the correlations were very negative. An explanation for why this correlational analysis was undertaken and speculation about its functional relevance would be of interest.

We should not dismiss the importance of temporal variables in the assessment of patients with neurological disorders. Many movements we request of our patients are time or distance dependent. If, on the other hand, we seek objective measures of recovery, then there is a need to relate speed and conditions for movement to functional tasks. Hopefully, future studies will address these concerns and integrate the importance of these interrelationships to the practice of physical therapy.

Steven L Wolf, PhD, PT, FAPTA Professor and Director of Research, Department of Rehabilitation Medicine Professor, Division of Geriatrics, Department of Internal Medicine Associate Professor, Department of Anatomy and Cell Biology Emory University School of Medicine 1441 Clifton Rd NE Atlanta, GA 30322

References

[1] Wolf SL, LeCraw DE, Barton IA, Jann BB. Forced use of hemiplegic upper extremily to reverse the effect of nonuse among chronic stroke and head-injured patients. Exp Neurol. 1989;104:125-134. [2] Brooks VB. The Neural Basis of Motor Control New York, NY: Oxford University Press Inc; 1986:93-104. [3] Brinkman J, Kuypers HGJM, Cerebral control of contralateral and ipsilateral arm, hand and finger movements in the split-brain rhesus monkey. Brain. 1973;95:653-674. [4] Matsunami K, Hamada I. Characteristics of the ipsilateral movement-related neuron in the motor cortex of the monkey. Brain Res. 1981; 205:29-42. [5] Wolf SL, LeCraw DE, Barton LA. Comparison of motor copy and targeted biofeedback training techniques for restitution of upper extremity function among patients with neurologic disorders. Phys Ther. 1989;69:719-735. [6] Gowland C, deBruin H, Basmajian JV, et al. Agonist and antagonist activity during voluntary upper-limb movement in patients with stroke. Phys Ther. 1992;72:624-633. [7] Powers RK, Marder-Meyer J, Rymer WZ. Quantitative relations between hypertonia and stretch reflex threshold in spastic hemiparesis. Anal Neurol. 1988;23:115-124. [8] Katz R, Rymer WZ. Spastic hypertonia; mechanisms and measurement. Arch Phys Med Rehabil. 1989;70:144-155.

Using a simple reaction time (RT) paradigm, Dickstein and colleagues showed that bilateral rapid elbow flexion results in prolonged RT and movement time (MT) relative to unilateral rapid elbow flexion in patients with hemiparesis and in subjects without neurologic impairment. The authors' rationale for choosing RT and MT to quantify motor performance was that "optimal control of movement initiation and movement speed is required for normal functional performance." What is unclear about this rationale is the meaning of "optimal control." The authors suggested that the prolongation of RT and MT in the bilateral task is indicative of less-than-optimal control. In this commentary, we focus on two issues. First, we offer an alternative interpretation of these findings and argue that "optimal control" is task specific and that the bilateral temporal coupling appears to be quite normal in these patients with hemiparesis. Second, we discuss a methodological option that could provide critical information related to the locus of RT effects.

Temporal interlimb interactions have been observed in a variety of studies in which one of two simultaneously moving limbs is unexpectedly blocked,[1] weighted differently,[2] or required to perform a more difficult task.[3] The perturbed limb affects the unperturbed limb by slowing the response,[3] prolonging the electromyographic (EMG) burst,[1] or altering the kinematic trajectory.[4]

Interlimb interactions also have been observed in patients with hemiparesis during repetitive upper or lower limb movements and during locomotion.[5-7] Cohn[6] found an interlimb interaction in subjects with hemiparesis performing bilateral supination-pronation movements of the forearms. He noted that when bilateral movements were performed, the paretic limb decreased the movement facility of the nonparetic limb. In Belmont and colleagues'[5] experiment, adults with left-sided hemiparesis performed unilateral and bilateral repetitive ankle flexion and extension movements as fast as possible for 10 seconds. A marked rate reduction occurred for each limb in the bipedal condition in contrast to the unipedal condition, and the nonparetic limb slowed to a much greater extent than did the paretic limb. These findings have now been replicated by Dickstein and colleagues. Similarly, during walking, Peat7 observed a bilateral alteration in the relative durations of stance and swing periods, with the greatest alteration seen in the nonparetic limb.

Thus, as Dickstein et al and others have observed, the interlimb interaction in the individual with hemiparesis is such that the nonparetic limb assumes a temporal pattern similar to that of the paretic limb. These findings are analogous to the results of Kelso and colleagues'[3] bimanual aiming experiment with subjects without neurologic impairment in which the limb with the easy task (ie, low index of difficulty) assumed an MT compatible with that of the limb performing the more difficult task (ie, high index of difficulty). These robust findings in both subjects with and without brain damage suggest that the optimal control of bilateral movements is quite different from the optimal control of unilateral movements. It appears that the timing of bilateral movements depends on the timing of the slowest limb, suggesting that there is one internal clock controlling these kinds of movements. Interlimb coordination appears to be constrained by the slowest element.

The authors stated that "the relative slowing of the patients' MT in comparison with that of the control subjects was even more accentuated in the bilateral task." A significant groupXtask interaction would have warranted such a conclusion; however, this finding was not reliable (P=.1410 for MT). There was a reduction in the MT difference between limbs for afl subjects during bilateral performance of the task compared with unilateral performance. Interestingly, the RT difference between limbs increased during bilateral performance of the task compared with unilateral performance. This finding suggests two different control parameters, with MT being more sensitive than RT to the temporal constraints of bilateral coordination. Dickstein and colleagues' results reveal that the nervous systems' solution to the control of bilateral upper extremity (UE) movements is governed by the same temporal constraint as that of subjects without neurologic impairment. Many bilateral UE functional tasks depend on this temporal coupling. It could be argued, therefore, that the preservation of this ability in the individual with brain damage is beneficial rather than detrimental to functional performance.

The specificity of training literature suggests that bilateral exercises cannot be expected to improve unilateral UE performance.[8] Although the authors concluded that "bilateral exercises cannot be expected to increase the speed at which the paretic UE is moved," it should be emphasized that practice of bilateral UE movements has been shown to reduce the MT of each limb during bilateral movements in subjects without neurologic impairment.[9] We would expect a similar finding for patients with hemiparesis; however, to our knowledge, this hypothesis remains to be empirically tested.

Simple RT and MT are variables of motor control that reflect critical movement-related central processing capabilities that have a long and well-established foundation in psychological science.[10] The authors argued that such variables of motor performance are almost absent when it comes to the evaluation and treatment of patients with hemiparesis. We would agree that the focus of most clinical physical therapy evaluations of subjects with hemiparesis seems to be on the more peripheral effector problems exhibited by the paretic limbs. Thus, current evaluation findings provide little insight into critical central processing deficits to which therapy must be targeted. From this perspective, it is refreshing to see an approach that provides some insight into these critical information-processing capabilities in individuals with brain damage.

The authors stated that the simple RT paradigm obviates the need to preprogram the movement during the RT interval, and thus the increase in RT in the patients with hemiparesis probably reflects a "nonspecific deficiency." The RT interval is known to consist of both central (ie, premotor RT) and peripheral (ie, motor RT) components. The premotor RT is the interval from the stimulus to the onset of EMG activity and reflects central processes of perception and decisions related to movement parameters. In contrast, the motor RT is the interval from the first change in the EMG activity to the onset of movement and reflects processes associated with the musculature itself.[11] From a methodological perspective, had the authors fractionated RT into these two components using EMG activity of the elbow flexor muscles, a clearer picture of the locus of slowing in the subjects with hemiparesis might have been revealed.

The authors should be commended for a study that used definitive measures of central movement-related processes and that certainly adds to the knowledge pertaining to motor control of individuals with brain damage. The clinician is cautioned, however, to interpret these findings in light of principles of motor control, including those of interlimb coordination, specificity of training, and effects of practice.

Carolee J Winstein, PT, PhD Assistant Professor Department of Biokinesiology and Physical Therapy University of Southern California 2250 Alcazar St, CSA 208 Los Angeles, CA 90033

Patricia S Pohl, PT Research Assistant Department of Biokinesiology and Physical Therapy University of Southern California

References

[1] Shapiro DC, Walter CB. Control of rapid bimanual aiming movements: the effect of a mechanical block. Society for Neuroscience Abstracts. 1982;8:733. Abstract. [2] Marteniuk RG, MacKenzie CL, Baba DM. Bimanual movement control: information processing and interaction effects. Q J Exp Psychol [A]. 1984;36:335-365. [3] Kelso JAS, Southard DL, Goodman D. On the coordination of two-handed movements. J Exp Psychol [Hum Percept]. 1979;5:229-238. [4] Kelso JAS, Putnam CA, Goodman D. On the space-time structure of human interlimb coordination. Q J Exp Psychol [A]. 1983;35: 347-375. [5] Belmont I, Karp E, Birch HB. Hemispheric incoordination in hemiplegia. Brain. 1971;94: 337-348. [6] Cohn R. Interaction in bilaterally simultaneous voluntary motor function. AMA Arch Neurol Psychiatry. 1951;65:472-476. [7] Peat M. Temporal characteristics of abnormal gait. In: Proceedings of the 19th International Congress of World Confederation for Physical Therapy, May 23-28, 1982, Stockholm, Sweden. [8] Winstein CJ. Designing practice for motor learning: clinical implications. In: Lister MJ, ed. H Step: Contemporary Management of Motor Control Problems. Washington, DC: Foundation for Physical Therapy inc; 1991:65-76. [9] Sherwood DE, Canabal MY. The effect of practice on the control of sequential and simultaneous actions. Human Performance. 1988;1:237-260. [10] Posner MI. Chronometric Explorations of Mind. Hillsdale, NJ: Lawrence Erlbaum Associates Inc; 1978. [11] Fischman MG. Programming time as a function of number of movement parts and changes in movement direction. Journal of Motor Behavior 1984;16:405-423.

Author Response

We would like to thank Dr Wolf, Dr Weinstein, and Ms Pohl for their thoughtful commentaries on our study. We fully agree with the opinion that more than answering its specific question, the study is raising additional problems regarding the contribution of bilateral exercises to rehabilitation of the paretic upperextremity functions of patients with hemiparesis.

Two basic assumptions were underlying this work. The first was the notion that one goal of physical treatment is improvement of function of the paretic upper limb. The second was the fact that performance of activities with the paretic limb is slower than normal. This slowness could be a major impediment to normal performance of either bilateral or unilateral activities. It is thus conceivable that exercise therapy should address the rate of performance of the paretic upper extremity as a specific treatment target.

We agree with the comments that the activity we monitored was not functional. We have chosen such a simple movement in order to be able to follow a rigid methodology of data collection. We hope that a similar line of research, which will look into more functional activities, will be pursued in the future. As suggested, a full kinematic analysis of these activities is desired. The same is true with regard to electromyographic (EMG) recordings from the activated muscles. In a currently ongoing study, we have added EMG measurements. We also agree that slowing of reaction time (RT) and movement time (MT) in bilateral tasks is a normal phenomenon. We argue, however, that for this reason, bilateral exercises cannot be expected to improve speed of the impaired limb. Presumably, the merits of bilateral exercises lie mainly in enhancement of postural tasks that were highlighted by the traditional neurophysiological approaches.[1,2] Based on our results and on prior information, they are not the right choice for increasing speed of performance of the affected limb. We think that the temptation to use the nonparetic limb to act simultaneously with the affected limb is unjustified when optimal unilateral activity is the ultimate goal. This is not contradictory to the concept of practicing a bilateral functional task if its mastering is desired.

It was correctly mentioned that our patients ranked high on the Barthel index and the Fugl-Meyer test. Overtly, however, they were not performing normally with their affected limb, probably because their movements were slow and cumbersome. Their high scores were therefore a reflection of insufficient sensitivity of the chosen tests to rate of performance. Nevertheless, the negative correlations found between the Fugl-Meyer upper-extremity scores and the paretic extremity/nonparetic extremity RT and MT ratios show that the degree to which the upper extremity was impaired was related to the patient's ability to recruit and move it rapidly.

Last, two interesting findings that were noted by the reviewers and were not discussed in the article were (1) the faster RTs of the left elbow and the faster MTs of the right elbow of the control subjects in the bilateral task and (2) the fact that, compared with unilateral performance, there was a reduction in the MT difference of the limbs in the bilateral task, but an increase in the difference between their RTs. Because of space constraints, we tried to focus the current discussion on those findings that have relevance to exercise therapy applied to patients with stroke. If repeated, we will elaborate on these additional results in a future report. [1] Flanagan E. Methods of facilitation and inhibition of motor activity. Am J Phys Med. 1967; 46:789-798. [2] O'Sullivan SB. Strategies to improve motor control. In: O'Sullivan SB, Schmilz TJ, eds. Physical Rehabilitation: Assessment and Treatment. 2nd ed. Philadelphia, Pa: FA Davis Co; 1988:253-280.

This article was submitted April 5, 1991, and was accepted January 29, 1993.

R Dickstein, DSc, PT, Department of Physical Therapy, Sackler Faculty of Medicine, Tel Aviv Univer Ramat Aviv, Tel Aviv 69978, Israel, and Flieman Geriatric Rehabilitation Hospital, POB 2263, Haifa, Israel. Address all correspondence to Dr Dickstein at the first address.

S Hocherman, Phd, Department of Physiology and Biophysics, The Technion-Israel Institute of Technology, Haifa, Israel.

G Amdor was a physical therapy student, Department of Physical Therapy, Sackler Faculty of Medicin Tel Aviv University, when this study was conducted.

T Pillar, MD, Flieman Geriatric Rehabilitation Hospital.