Kinematic analysis of lower-limb movement during ergometer pedaling in hemiplegic and nonhemiplegic subjects.

Rehabilitation programs for patients with hemiplegia resulting from cerebrovascular accidents (CVAs) are designed to return patients to the highest possible functional level. Traditionally, the hemiplegic patient's level of function has been assessed through various categorical schemes. Assessment tools, such as that developed by Fugl-Meyer et al, [1] are available to assist in an objective evaluation of the patient's function. These functional tests may quantify an overall level of function; however, they do not identify or precisely measure characteristics of the movement dysfunction.

Benecke and Conrad [2] suggested that an analysis of the lower extremity during stationary bicycling could be used to quantify motor disturbances. This type of analysis may provide an objective method for measuring movement dysfunction and for evaluating the effectiveness of therapeutic programs. Brown and DeBacher [3] suggested that an exercise program consisting of bicycle ergometer pedaling combined with electromyographic (EMG) biofeedback may improve muscle-activation patterns in patients with hemiplegia. Bicycle pedaling and walking display common motor requirements for their successful execution. Both tasks are characterized by alternating flexion and extension of the hip, knee, and ankle, and both tasks require reciprocal movements between limbs. Bicycle pedaling, as is

[TABULAR DATA OMITTED]

true of locomotion, is a highly stereotyped, reproducible movement with small intraindividual differences. [2] An analysis of lower-extremity function during bicycling has an advantage over gait analysis in that it can be used with patients who are unable to walk.

The purpose of this study was to determine whether the task of stationary cycling can be used as a model for measuring lower-limb movement dysfunction. To accomplish this goal, we compared the lower-extremity kinematic characteristics of hemiplegic subjects with those of normal subjects during pedaling on a stationary bicycle. We investigated lower-extremity movement in the sagittal plane at two pedaling rates (32 and 50 rpm) in all subjects. The three primary research questions were:

1. What are the kinematic characteristics of the involved lower extremity in hemiplegic subjects during bicycle pedaling?

2. How do the kinematic characteristics during pedaling in hemiplegic subjects compare with those of normal subjects?

3. Does the change in rate of pedaling change these characteristics for normal and hemiplegic subjects?

Method

Subjects

Ten normal subjects and 10 subjects with hemiplegia secondary to a CVA were selected to participate in this study. The criteria for selection of the normal subjects were (1) age range of 40 to 75 years; (2) structural leglength discrepancy no greater than 2.54 cm (1 in), when measured from the greater trochanter to the planar surface of the heel during standing with the nakle in a neutral anatomical position; (3) passive hip, knee, and ankle range of motion (ROM) within normal limits (ie, [+ or -] 10% of normal) in the sagittal plane; (4) no current orthopedic disabilities of the lower extremity that could cause abnormal movement (eg, painful arthritis, orthopedic deformities, prosthesis); (5) alert and able to follow simple instructions; and (6) not currently taking medications that were prescribed primarily to increase muscle tone. Passive ROM was measured with a universal goniometer using standard procedures, and the ROM measurements were then compared with normative ROM values. [4]

The criteria for selection of the hemiplegic subjects included those listed for the normal subjects and (1) medical diagnosis of a CVA with an onset greater than 1 year from time of data collection and (2) an observable movement dysfunction in the lower extremity during locomotion. An observable movement dysfunction was operationally defined as a deviation in either displacement, velocity, or acceleration of the lower extremity that was observed by two experienced physical therapists (JCR and CAG). Medical authorization to participate in the study was obtained from the physician of each hemiplegic subject. Each subject signed an informed consent statement prior to participation in the study. Subject characteristics are presented in Table 1.

Instrumentation

High-speed cinematography (64 frames/s) was used to record the subjects' lower-extremity movement during pedaling. A motor-driven 16-mm movie camera (*) with a 25-mm lens was used to film the subjects. Camera speed (frame rate) was calibrated with a millisecond timer. The lens-to-subject distance was 5.49 m, with the lens positioned at a height of 80 cm, level and orthogonal to the sagittal plane of the subject.

A Schwinn Ergo Metric stationary bicycle ergometer was pedaled by all subjects. The bicycle pedals were equipped with toe clips and leather straps to secure the subject's feet (Fig. 1). The bicycle seat height was adjusted for each subject and corresponded to 100% of the subject's trochanteric leg length (the distance from the greater trochanter of the femur to the sole of the foot with the knee fully extended and the ankle in neutral).

Subjects were filmed wearing shorts, shoes, and short socks that allowed accurate placement and viewing of joint identification markers. The joint markers were constructed of black and white adhesive tape and were placed on the skin over the following anatomical landmarks of the lower limbL greater trochanter, lateral femoral condyle, lateral malleolus, lateral border of the heel, and fifth metatarsal head. An experienced physical therapist (JCR) used manual palpation to identify the anatomical landmarks.

Procedure

Kinematic data were recorded during pedaling at a work load of 350 kg.m/min at 32 and 50 rpm. The work-load and pedaling rates selected for the subjects were based on current literature [5] and pilot data collected in our laboratory prior to this study. The subjects were allowed a warm-up period of 30 seconds at each speed prior to testing. During testing, the subjects pedaled three times at each speed, alternating speed with each trial (a total of six trials). The initial pedaling rate was assigned randomly. Each trial lasted approximately 45 seconds.

The subjects were instructed to grip the handlebars with both hands. If upper-extremity hemiparesis prevented a volitional grip, the hand was held on the handlebar by an assistant.

Cinematographic data collection began when the subject maintained the specified pedaling rate (approximately 15 seconds after the initiation of pedaling). A speedometer mounted on the handlebars provided subjects with feedback on the pedaling rate. An assistant monitored the pedaling rate and stood beside the subjects to ensure their safety. Three pedaling revolutions during each trial were recorded on film. One limb for each subject was evaluated. The involved limb of the hemiplegic subjects was evaluated, and a limb selected at random was evaluated for the normal subjects.

Data Analysis

A Vanguard motion analyzer was used to project the film data and to measure the angular displacements of the hip, knee, ankle, and pedal-crank. One cycle from each trial was analyzed. Of these three trials for each pedaling speed, the trial closest to the criterion speed was selected for statistical analysis. Angular-displacement measurements were recorded every 20 degrees and every 16 degrees of the crank position during the 50- and 32-rpm pedaling rates, respectively. These sampling rates were based on the frame rate (64 frames/s). Thus, at 50 rpm, 76.8 frames were recorded on film for each revolution and every fourth frame was measured, yielding 19.2 samples. At 32 rpm, 120 frames comprised one revolution and every sixth frame (18[degrees]) was measured, which yielded 20 samples. Because of the limited frame rate and subject variability in cycle time, not every cycle was initially measured at exactly top dead center (TDC). Thus, some cycles may have been within one standard deviation of the calculated sampling rate. The resolution of our pedal-crank measurement at 50 rpm was 5 degrees (measurement error of 1.38% of full scale). Angular data were compiled and stored through an interface between the motion analyzer and a personal computer. A customized computer program calculated and smoothed angular-displacement velocities using a three-point moving-average routine.

Angular-displacement measurements for the hip, knee, ankle, and pedal-crank angles are illustrated in Figure 2. The hip angle was defined as the angle formed by the long axis of the thigh and a vertical line drawn in space through the hip-joint axis. Knee angle was measured by the angle formed between the long axes of the thigh and the shank. The ankle angle was the angle formed by the long axes of the shank and the foot. The pedal-crank angle was the position of the bicycle crank arm relative to TDC. Top dead center is the point at which the crank arm is positioned vertically, with the pedal superior to the crank center. This point is also referred to as 0 or 360 degrees. Bottom dead center (BDC) is a crank angle of 180 degrees.

Statistical analyses of the angular-displacement values for the normal and the hemiplegic subjects were based on one representative pedal cycle (from TDC to TDC) for each subject at each speed. This representative cycle was chosen from the trial in which the subject pedaled at a rate nearest to the specified pedaling rate (32 and 50 rpm). All representative cycles met a criterion of [+ or -]2 rpm of the specific rate.

Means and standard deviations of the total range of angular displacements and maximum and minimum joint angles were computed for the hip, knee, and ankle for all subjects at each pedaling rate (Tab. 2). A 2x2 analysis of variance (ANOVA) for repeated measures on pedaling rate was performed to assess significant main effects on group, pedaling rate, and interaction. Differences were accepted as significant at the .05 level.

Angular displacement versus crank angle plots were constructed for the thigh, knee, and ankle at each pedaling speed, based on the subject's representative cycle. Angle-angle diagrams for the hip-knee and knee-ankle displacements were constructed to qualitatively analyze subject intralimb coordination. Angular velocity versus angular displacement diagrams were constructed for the ankle joint to analyze individual joint control.

Validity and Reliability of

Measurements

Angular data in this study were measured from 16-mm motion picture film on the motion analyzer. To determine the validity of our motion analyzer, we constructed a model of known angles, photographed them, and then remeasured them with the motion analyzer. The maximum difference between the known values and those measured with the motion analyzer was less than 1 degree. Because of possible errors associated with palpation of the bony landmarks and skin movement over joint axes, however, the error between anatomical and measured angles may have been greater than 1 degree.

To determine the reliability of recording the X,Y coordinates using the motion analyzer, we measured the joint angles of one subject during a randomly selected pedal cycle on two occasions. Four angles (hip, knee, ankle, and pedal-crank) on 21 frames of film (approximately one complete pedal cycle) were measured on two different days. The Pearson Product-Moment Correlation Coefficient (r) was .99 for the hip, knee, ankle, and pedal-crank measurements. All angular data were measured by one experienced film reader (JCR).

Within-subject variability was measured in three normal and three hemiplegic subjects. Variability was assessed by comparing the standard deviations of the angular displacements of three pedaling cycles at each speed. A comparison of between-subject variability among normal and hemiplegic subjects indicated that lower-extremity kinematic characteristics are consistent from trial to trial for both groups. Variability among normal subjects in this study was similar to the variability among subjects reported by Nordeen. [6]

Variability between groups was assessed by comparing the standard deviations of the angular-displacement variables at both speeds across groups. The average difference in standard deviation between the two groups was 2.5 degrees. If the relatively large ankle plantar-flexion differences (9.8 degrees and 9.6 degrees) are excluded, then the mean difference is 1.25 degrees. This difference may be considered functionally insignificant and is close to the resolution of our measurement system. Thus, except for maximum plantar flexion, there was little variability between groups for the angular-displacement variables.

Results

Angular Displacements

The angular-displacement variables analyzed in the two-way repeated-measures ANOVA were maximum and minimum joint angles and total joint excursion for the hip, knee, and ankle in the sagittal plane. The ANOVA revealed no significant differences between group means at either pedaling rate. The variables that were significantly different between pedaling rates for each group were ankle dorsiflexion and ankle excursion (Tab. 3). There were no significant group x pedaling rate interactions for any of the angular-displacement variables.

Typical angular displacement versus crank angle plots for the hip, knee, and ankle of a normal subject at 32 and 50 rpm are presented in Figure 3. The hip and knee displacement patterns for the hemiplegic group were similar to those of the normal group at both pedaling rates. There were, however, some individual differences in hip displacement patterns in the hemiplegic group that appear to be associated with the ankle displacement pattern (Fig. 4A). This pattern shows a fluctuation of both hip and ankle displacement curves between a crank angle of 200 and 300 degrees. Three distinct abnormal patterns of ankle displacement were demonstrated in the hemiplegic group and were consistent across the pedaling rates. Three hemiplegic subjects demonstrated a pattern characterized by an alternating plantar-flexion-extension response during the recovery phase of the pedal cycle (Fig. 4A). Two subjects had a prolonged plantar-flexion response during the recovery phase (Fig. 4B), and three subjects demonstrated a pattern characterized by limited ankle excursion (Fig. 4C). The remaining two hemiplegic subjects had lower-limb patterns similar to those of the normal group and were the hemiplegic subjects with the fewest gait deficits, as determined by qualitative observation.

Phase-Plane Analysis

Intralimb coordination characteristics of the subject's lower limb are revealed in the phase-plane analyses of simultaneous hip-knee and knee-ankle movements during the pedal cycle. The rate of pedaling did not affect the coordination patterns for either group. A representative hip-knee displacement plot for normal subjects at 32 rpm, which displays smooth flexion and extension movement of the hip and knee during one pedal cycle from TDC to BDC, is shown in Figure 5. The smoothness of the perimeter indicated well-coordinated movement between the joints, and the area of the plot indicated the joint excursion.

Eight hemiplegic subjects demonstrated coordinated hip-knee movement patterns that were similar to those of the normal group. Two hemiplegic subjects, however, displayed slight hip-knee coordination problems (Fig. 5B). This pattern displays smooth coordinated motion between the hip and knee during the power phase and a sudden increase in hip flexion during the recovery phase, which corresponds to the same position in the cycle at which the subjects had an alternating plantar-flexion-extension pattern.

The phase-plane analysis of the knee-ankle revealed intralimb coordination problems in hemiplegic subjects as compared with normal subjects (Fig. 6). The abnormal ankle patterns identified were consistent across pedaling rates for each subject. The alternating plantar-flexion-extension pattern was characterized by smooth, coordinated knee-ankle movement during most of the power phase and several direction reversals in ankle motion during the recovery phase and at TDC (Fig. 6B). One of the three hemiplegic subjects with limited ankle excursion displayed a knee-ankle pattern, as shown in Figure 6C. The lack of joint excursion is indicated by the small area of the diagram, and poor intralimb control is illustrated by the irregularity of the perimeter.

A phase-plane analysis of the ankle velocity versus ankle displacement was performed to further assess control of ankle movement. The ankle patterns for all normal subjects were similar, demonstrating an orbital shape (Fig. 7A). The subjects with alternating plantar-flexion response during the recovery phase had distorted velocity patterns. These patterns were characterized by major perimeter deflections, self-intersecting loops, and a general deterioration of the orbital shape (Fig. 7B). The hemiplegic subjects with limited ankle ROM had abnormal ankle velocity-ankle patterns characterized by moderate deflections in the perimeter and by the small area circumscribed by the plot (Fig. 7C).

Discussion

The average values of angular-displacement variables of the hip, knee, and ankle for the normal and hemiplegic groups are similar to those reported for normal subjects in other studies. [6,7] Houtz and Fischer, [8] however, reported angular displacements that differ from those demonstrated in this study. These inconsistencies may be explained by differences in methodology. The most significant difference in methodology was the normalization of seat height to leg length, which was performed in our study and in other studies, [6,7] but not by Houtz and Fischer. It appears that when seat height is adjusted to leg length, angular-displacement variables are relatively constant among subjects and between groups.

The consistency in angular displacement between pedaling speeds, among subjects, and between groups suggests that excursion of the lower limb during bicycle pedaling is partially constrained by the mechanics of the bicycle-rider system. Nordeen and Cavanagh [9] suggested that, although the bicycle-rider system is constrained to some degree, an infinite number of movement patterns are possible as the hip, knee, and ankle joints collectively rotate the pedal-crank. In addition, the pelvis may move as the cyclist redistributes his or her weight during the pedal cycle. [9] Although the kinematic variables appeared to be partially constrained during pedaling, the movements of the lower extremity are still indeterminate.

During pedaling, the normal subjects displayed smooth movement reversals in flexion and extension at all three joints. The pedaling task did not reveal gross dysfunction of hip or knee movement in the hemiplegic group. It did, however, reveal three abnormal patterns of ankle movement in the hemiplegic group.

The alternating plantar-flexion-extension pattern occurred when the contralateral limb appeared to push the hemiparetic limb through the latter half of the pedal cycle. This response appeared to create a rapid dorsiflexion movement, followed by alternating ankle flexion and extension that continued through the recovery phase. Two subjects with this response also showed clonic-like movements at the ankle when their foot rested on the bicycle pedal. The alternating plantar-flexion-extension response may have resulted from weak lower-leg musculature that was unable to control ankle joint movement or from a hypersensitive stretch response of the triceps surae muscle group. This plantar-flexion response may be similar to what Knutsson and Richards [10] described in EMG analysis of hemiplegic gait. They reported a premature activation of the calf muscle during the stance phase of gait, which appeared to be due to hypersensitive stretch reflexes. They also suggested that the hypersensitive stretch reflex of the calf muscles caused a compensatory response of knee hyperextension during midstance. [10] Similarly, the inappropriate plantar flexion during pedaling may have caused the sudden increase in hip flexion (Fig. 4A).

Another abnormal ankle pattern consisted of prolonged plantar flexion in the recovery phase associated with inadequate dorsiflexion and a continuation of plantar flexion during the power phase. The continuation of plantar flexion in these hemiplegic subjects may be associated with a prolonged recruitment of the triceps surae muscle group. A similar abnormal pattern of muscle activity in hemiplegic subjects during locomotion has been reported. [10] Sahrmann and Norton [11] reported that prolonged recruitment of agonist muscles may be an impediment to smooth, voluntary reversals of movement in hemiplegic patients.

The third abnormal ankle pattern was decreased joint excursion throughout the pedal cycle. Benecke et al [12] also reported limited ankle joint excursion during bicycle pedaling in patients with neuromuscular disorders. The decreased ankle joint excursion during pedaling may have resulted from coactivation of agonist-antagonist muscle pairs. Simultaneous contraction of plantar flexors and dorsiflexors during cycling would increase stiffness about the ankle joint and result in decreased joint ROM. Two subjects in our study had limited ankle plantarflexion ROM during pedaling. These subjects appeared to compensate for the apparent limb shortening by increasing hip and knee extension when the pedal-crank position was between 90 and 270 degrees. Coactivation of lower-extremity muscle pairs was reported as a cause of movement dysfunction during hemiplegic gait. [13]

The abnormal ankle control during pedaling was more evident in the phase-plane analyses. Ankle velocity versus ankle displacement diagrams revealed poor ankle control in some hemiplegic subjects, which might have been undetected if only displacement patterns and values had been analyzed. Poor ankle control was revealed in the velocity versus displacement diagrams by frequent and rapid changes in velocity and inappropriate reversals of joint movement.

The abnormal kinematic patterns observed in subjects with hemiplegia may be due to the intersegmental dynamics of the bicycle-rider system and abnormal patterns of muscle activity. The contribution of these control variables to the patterns observed will require further investigation using both EMG and intersegmental kinematic analyses.

Conclusions

Average kinematic values were similar for normal and hemiplegic groups, yet patterns of displacement and velocity were different. This finding emphasizes the importance of performing evaluations that measure not only isolated kinematic variables but also movement control and coordination through dynamic analyses.

A change in pedaling rate did not statistically alter hip or knee angular-displacement variables, although ankle joint excursion and maximum dorsiflexion were significantly different across pedaling rates. Three distinct abnormal ankle patterns were revealed in the hemiplegic group: (1) an alternating plantar-flexion-extension response, (2) a prolonged plantar-flexion response during the recovery phase, and (3) limited ankle joint excursion. Phase-plane analyses depicted poor isolated joint control and poor intralimb coordination at the ankle in most hemiplegic subjects. These patterns of movement dysfunction were independent of pedaling rate.

This preliminary study provided evidence that analysis of bicycle pedaling may be used to document motor control deficits in subjects with lower-limb movement dysfunction. Additional studies investigating muscle activity patterns during cycling may help explain the abnormal kinematic patterns found in the hemiplegic subjects. Further research in movement dysfunction is needed and should focus on quantifying the characteristics of the movement dysfunction. Through quantitative analyses, we can demonstrate the efficacy of treatment intervention and contribute to our patients' rehabilitation.

[TABULAR DATA OMITTED]

[TABULAR DATA OMITTED]

Table 3. Results of Two-Way Analysis of Variance for Repeated Measures on Pedaling Rate

Dependent
Variable            df  SS       MS    F         P
Ankle dorsiflexion
 Rate                1  88.21     88.21 5.79     .0271
 Group               1   116.28  116.28 1.32(a)  .2663
 Subjects (group)   18  1590.25   88.35 5.80     .0003
 Rate x group        1     1.08    1.08 0.07     .7923
 Error              18   274.32   15.24
 Corrected total    39  2070.16
Ankle excursion
 Rate                1   487.20  487.20 13.94    .0015
 Group               1   487.20  487.20  2.50(a) .1309
 Subjects (group)   18  3501.35  194.52 16.05    .0001
 Rate x group        1     6.72    6.72  0.55    .4660
 Error              18   218.18   12.12
 Corrected total    39  4382.38
  (a) Using Type III sum of shares.

Reference

[1] Fugl-Meyer AR, Jaasko L, Leyman I, et al. The post-stroke hemiplegic patient: a method J Rebabil Med. 1975;7:13-31.

[2] Benecke R, Conrad B. Evaluation of motor deficits in patients suffering from multiple sclerosis. In: Bauer JJ, Poser CM, Ritter W, eds. Progress on Multiple Sclerosis. Berlin, Germany: Springer-Verlag; 1980:590-595.

[3] Brown DA, DeBacher GA. Bicycle ergometer and electromyographic feedback for treatment of muscle imbalance in patients with spastic hemiparesis. Phys Ther. 1987;67:1715-1719. Suggestion from the Field.

[4] Joint Motion: Method of Measuring and Recording. Chicago, Ill: American Academy of Orthopaedic Surgeons; 1965.

[5] Nordeen-Snyder KS. The effect of bicycle seat height variation upon oxygen consumption and lower limb kinematics. Med Sci Sports. 1977;9:113-117.

[6] Nordeen KS. The Effect of Bicycle Seat Height Variation upon Oxygen Consumption in Both Experimental and Simulated Lower Limb Kinematics. University Park, Pa: The Pennsylvania State University; 1976. Master's thesis.

[7] Ericson MB, Nisell R, Nemeth G. Joing motions of the lower limb during ergometer cycling. Journal of Orthopaedic and Sports Physical Therapy. 1988;9:273-278.

[8] Houtz SJ, Fisher FJ. An analysis of muscle action and joint excursion on a stationary bicycle. J Bone Joint Surg [Am]. 1959;41:123-131.

[9] Nordeen KS, Cavanagh PR. Simulation of lower limb kinematics during cycling. In: Komi PV, ed. International Jounral of Biomechanics. Baltikmore, Md: University Park Press; 1975;5:26-33.

[10] Knutsson E, Richards C. Different types of distrubed motor control in gait of hemiplegic patients. Brain. 1979; 102-405-430.

[11] Sahrmann SA, Norton BJ. The relationship of voluntary movement to spasticity in the upper motor neuron syndrome. Ann Newrol. 1977-2;2:60-465.

[12] Benecke R, Conrad B, Meinch HM, Hohne J. Electromyographic analysis of bicycling on an ergometer for evaluation of spasticity of lower limbs in man. In: Desmedt JE, ed. Motor Control Mechanisms in Health and Disease. New York, NY: Raven Press; 1983;39:1035-1046.

[13] Knutsson E, Martensson A. Dynamic motor capity in spastic paresis and its relation to prime mover dysfunction, spactic reflexes, and antagonist co-activation. Scand Rebabil Med. 1980; 12:93-106.

(*) Bolex, Div of Paillard Inc, 1990 Lower Rd, Linden, NJ 07036.

J Rosecrance, MS, PT, is a doctoral candidate in the Department of Physical Therapy Education, The University of Iowa, 2600 Steindler Bldg, Iowa City, IA 52242 (USA). This study was completed in partial fulfillment of the requirement for Mr Rosecrance's master of science degree, Division of Physical Therapy, The University of North Carolina at Chapel Hill, Address all correspondence to Mr Rosecrance.

C Giuliani, PhD, PT is Assistant Profeessor, Division of Physical Therapy, The University of North Carolina at Chapel Hill, CB #7135, Medical School Wing E 22II, Chapel Hill, NC 27599.

This study was approved by the Committee on the Protection of the Rights Human Subjects, School of Medicine, The University of North Carolina at Chapel Hill.

This study was supported in part by the Foundation for Physical Therapy Inc.

This article was submitted January 18, 1990, and was accepted November 14, 1990.