If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Chow JW, Yablon SA, Stokic DS. Electromyogram–lengthening velocity relation in plantar flexors during stance phase of gait in patients with hypertonia after acquired brain injury.
Objective
To examine the velocity-dependent change in medial gastrocnemius (MG) activity during the stance phase of gait in patients with moderate to severe resting hypertonia after stroke or traumatic brain injury (TBI).
Design
Cohort study.
Setting
Motion analysis laboratory in a tertiary-care rehabilitation hospital.
Participants
Convenience sample of patients with chronic TBI and stroke (n=11 each), and age- and sex-matched healthy controls (n=22).
Intervention
Not applicable.
Main Outcome Measures
Frequency and gain (steepness) of positive (>0) and significant positive (>0 and goodness of fit P≤.05) electromyogram–lengthening velocity (EMG-LV) linear regression slope in MG during the stance phase of gait.
Results
Positive and significant positive slopes were found significantly more often on the more affected (MA) than less affected (LA) side in patients with TBI but not stroke. Both the frequencies of positive and significant positive slopes on the MA side in patients with TBI were also significantly higher than in controls. However, neither the gain of positive nor significant positive EMG-LV slope was different between the MA and LA sides or in comparison with controls. Positive slope parameters were not related to Ashworth score on the MA side.
Conclusions
The frequency and gain of positive EMG-lengthening slope did not effectively differentiate patients from controls, nor were they related to the resting muscle hypertonia. Motor output during MG lengthening in the stance phase of gait is apparently not exaggerated or related to resting hypertonia in patients with chronic TBI and stroke. Thus, changes in gait during stance cannot be ascribed to increased stretch reflex activity in MG muscle after acquired brain injury.
SPASTICITY IS COMMONLY defined as “a motor disorder characterized by a velocity-dependent increase in tonic stretch reflexes (muscle tone) with exaggerated tendon jerks, resulting from hyperexcitability of the stretch reflex as one component of the upper motor neuron syndrome.”
(p485) In addition to hyperactive stretch reflexes, other positive features of the upper motor neuron syndrome include muscle spasms, abnormal muscle coactivation, clonus, and spastic dystonia.
Opposing views, however, exist relative to the contribution of spasticity in its strict sense to gait dysfunction. Early electromyogram (EMG) studies during treadmill walking only revealed greater activity in the tibialis anterior (TA) muscle during swing in patients with spinal spasticity compared with healthy controls.
simulated a single support phase of gait in the sitting position by rotating the foot at different frequencies with the knee extended while the subject was asked to activate the gastrocnemius muscle at 10% of the maximum EMG level. They reported no difference in stretch-induced EMG between 14 stroke patients (5–20mo postonset) and healthy controls, and concluded that stretch reflex activity during walking is not increased after stroke. However, interventions that reduce stretch reflex excitability, such as intrathecal administration of baclofen, may improve,
Thus, further studies to delineate features of spasticity during gait are warranted to guide selection of patients, interventions, and outcome measures.
(p572) introduced the concept of “dynamic assessment of spasticity during walking” on the premise that “the higher the effectiveness of velocity-sensitive excitatory reflex components, the tighter would be the correlation between motor output and velocity of muscle lengthening.”
This was determined by plotting EMG against muscle lengthening velocity during walking in patients with cerebral palsy. Based on this concept, Lamontagne et al
(p1696) developed a “locomotor-specific measure of spasticity” in ankle plantar flexors during the stance phase of gait, with EMG-lengthening velocity (EMG-LV) slope as an outcome measure (called “spasticity index” in Lamontagne
When averaged over all gait cycles, the gain of EMG-LV slope in the plantar flexors during stance was found positive (>0) on the paretic side in 20 of 30 subjects within 6 months of stroke, but not on the nonparetic side or in controls. The average gain of EMG-LV slope correlated positively with the Ashworth score (AS) in the paretic plantar flexors and negatively with gait speed.
found that EMG-LV slope in the paretic leg was positive in 44% of gait cycles from 68 stroke subjects (7d to 31y poststroke). The spasticity index, calculated as the average gain of positive slopes only, was found significantly correlated to most temporospatial parameters of gait. Both Lamontagne
interpreted positive EMG-LV slope as the evidence of hyperactive stretch reflexes during lengthening of plantar flexors in the stance phase of gait.
The purpose of this study was to further characterize the velocity-dependent increase in medial gastrocnemius (MG) activity in the stance phase of gait in patients with acquired brain injury (ABI) with moderate to severe lower limb muscle hypertonia. We replicated the previously used approach
for comparison purposes, but put special emphasis on the strength of the linear association between EMG and lengthening velocity, in accordance with the concept of Crenna.
Thus, our specific aim was to compare the frequency and gain of positive EMG-LV slope in MG muscle during stance between patients with ABI and age- and sex-matched healthy controls. We hypothesized that the EMG-LV slope would be more often positive and steeper on the more affected (MA) side of patients compared with their less affected (LA) side and healthy controls. The associations with gait speed and resting muscle hypertonia were explored in secondary analyses.
Methods
Participants
A convenience sample of 11 stroke and 11 TBI patients was recruited from a spasticity and motor disorders clinic (table 1). The inclusion criteria were a significant increase in lower limb muscle hypertonia that impairs function or care, and the ability to walk safely faster than 10cm/s for at least 10m with or without assistive devices. Wearing a short, nonrigid polypropylene ankle-foot orthosis to prevent foot drop during gait was permitted because it allowed passive dorsiflexion during the stance phase, as verified by ankle range of motion. Patients with evidence of ankle clonus during the stance phase were excluded to eliminate the confounding effect of the rhythmic EMG pattern to MG activity. Because of the age difference between stroke and TBI patients, 2 age- and sex-matched healthy control groups were recruited (n=11 each). Control subjects in the stroke-matched group (mean ± SD: age, 41±9y; height, 170±14cm; body mass, 76±22kg) and TBI-matched group (mean ± SD: age, 29±10y; height, 174±14cm; body mass, 76±22kg) reported no orthopedic and neurologic disorders at the time of testing. Each subject attended 1 data collection session and signed the informed consent approved by our institutional review board for human research.
Gait data were collected at 60Hz with 8 digital Hawk camerasa surrounding a 7-m-long walkway. Five forceplatesb concealed and flush in the middle of the walkway were used to determine critical instants of gait (sample rate, 1200Hz). Passive spherical reflective markers were affixed to body landmarks according to the Helen Hayes marker system.
Bipolar surface EMG electrodes with built-in preamplificationc (model MA411; gain 20, 2-cm center-to-center distance, input impedance >1010Ω, common-mode rejection ratio [CMRR] >100dB) were attached to MG and TA muscles bilaterally, using electrode placement guidelines described in Cram and Kasman.
Signals were further amplified by an EMG systemc (model MA300; input impedance, 31KΩ, CMRR >50dB) before 12-bit analog-to-digital conversion (sampling rate, 1200Hz). Video, EMG, and ground reaction force (when available) data were collected synchronously on the EVaRT data acquisition system.a
Experimental Protocol
The subject walked 8 to 10 times along the walkway with short pauses in between with customary shoes and assistive devices, if any. Patients were instructed to walk at a self-selected free speed, and controls at a self-selected very slow speed, which provides more appropriate comparison between patients with ABI and controls
and differentiates EMG-LV features that are disorder-specific from walking at a slow speed. Very slow rather than patient-matched speed was chosen for controls to ensure natural walking. No instruction was given about stepping on the forceplates. Data acquisition started after the subject took a few steps and terminated before the end of the walkway. Thus, each trial included a steady-state gait.
Before recording gait in patients, muscle hypertonia was assessed by a physical therapist in the bilateral hip flexors and extensors, knee flexors and extensors, and ankle plantar flexors, using the Modified Ashworth Scale
Physical therapists were not available to evaluate 2 patients (1 TBI, 1 stroke). The average AS served to identify MA and LA sides. If the difference was less than 0.4 or data were not available, the side with the longer stance time was considered the LA side.
Coordinates of the markers were smoothed with a low-pass, bidirectional, fourth-order Butterworth filter (cutoff 6Hz). Marker location and forceplate data were processed with OrthoTrak Gait Analysis software.a The software preferentially uses ground reaction force to determine initial foot contact and toe-off events. In the absence of ground reaction force, the software relies on marker kinematics.
A gait cycle was defined by 2 consecutive initial contacts of the same foot. The software also provided joint angle data for the computation of MG length. Raw EMG signals were baseline-adjusted, filtered (recursive digital Matlab Elliptic filter,d band pass 10–500Hz), and full-wave rectified.
With the use of the approach proposed by Lamontagne,
the EMG-LV relationship of the MG during stance was determined for each gait cycle. Muscle-tendon length (LMG) was used to approximate fiber length because the latter could not be determined without simultaneous muscle imaging
where α and β are the knee flexion angle (0° for full knee extension) and ankle angle (>90° for dorsiflexion and <90° for plantar flexion), respectively, with LMG expressed as a fraction of shank length (LS).
Such approximation was deemed appropriate since the profiles of gastrocnemius muscle-tendon length and fascicle length are largely similar during the stance phase of gait.
Moreover, any systematic error introduced by the chosen approach was unlikely to impact the slope of the velocity-dependent increase in MG activity during stance.
The muscle shortening/lengthening velocity was computed using the central difference method
for the MA and LA limbs in patients and the dominant limb in controls. Raw EMG was amplitude normalized to its peak value during stance (%MAX). For each gait cycle analyzed, both muscle shortening/lengthening velocity and normalized EMG were time normalized to 100% gait cycle (fig 1A). The periods of MG lengthening during each stance phase were identified, and their durations were summed up into a total lengthening duration expressed as a percent of the gait cycle duration. The peak lengthening velocity (in LS/s) was also determined. Corresponding EMG and lengthening velocity (LS/s) data points were extracted and aggregated into an EMG-LV plot for the stance phase of each gait cycle (fig 1B). A linear regression line was fitted to the data, and the slope direction (positive or negative) and slope steepness (gain, %/[LS/s]) were derived. Both positive and negative slopes were found across multiple gait cycles in most subjects.
Fig 1Computation of MG EMG-LV slope and goodness of fit during stance phase of a gait cycle. EMG (solid line) and lengthening velocity (dashed line) data pairs during the lengthening portion of stance (A) are used for regression analysis (B). The gait cycle on the MA side is shown for a TBI patient with the AS of 4 in the MA plantar flexors. Abbreviation: LS, shank length; %MAX, percent of the maximum EMG.
The hypothesis was that characterization of positive slopes may better differentiate patients from controls than the average EMG-LV gain derived from both positive and negative slopes
Additional emphasis was put on the goodness of fit, which was considered significant if the linear regression returned P≤.05 (see fig 1B). Therefore, 4 positive slope parameters were calculated across all gait cycles for each limb of patients and the dominant limb of controls:
•
Frequency of positive slope: proportion of slopes >0 irrespective of the goodness of linear fit (fig 2, step 2A)
Fig 2Procedure for deriving different EMG-LV parameters in a representative subject: the overall average gain from positive and negative slopes (step 1), frequency and average gain of positive slopes (step 2, A and B), and frequency and average gain of significant positive slopes (step 3, C and D). Each cross represents a gain of the slope during stance of a gait cycle, calculated as in figure 1. Abbreviations: LS, shank length; %s/LS, percent of the maximum EMG per LS/s.
Average gain of positive slope: average steepness of positive slopes (fig 2, step 2B)
•
Frequency of significant positive slope: proportion of slopes >0 with significant linear fit (fig 2, step 3C)
•
Average gain of significant positive slope: average steepness of significant positive slopes (fig 2, step 3D)
Statistical Analysis
For descriptive purposes, the AS, gait speed, stance duration, duration of lengthening periods, and peak lengthening velocity were submitted to the Wilcoxon rank-sum test (MA vs LA comparison) and the Mann-Whitney U test (MA/LA vs control comparison) in each patient group. To evaluate the effects of the ankle orthosis, we pooled TBI and stroke data and compared patients who did (n=12) with those who did not (n=10) wear the orthosis and found no significant differences for any of the positive slope parameters on each side (Mann-Whitney U tests, .059≤P≤.974). Furthermore, the same comparisons within groups indicated no difference in ankle range of motion (Mann-Whitney U tests, P=.53 for TBI and P=.28 for stroke) and peak MG lengthening velocity (Mann-Whitney U tests, P=.23 for TBI and P=1.0 for stroke) for the MA limb (table 2), and any of the EMG-LV parameters (Mann-Whitney U tests, P≥.41 for TBI and P≥.50 for stroke). This justifies pooling data of patients who did and did not wear an ankle orthosis for the main statistical analysis.
Table 2Gait Characteristics and Parameters of EMG-LV Slope in the MG Muscle
the same tests were used on the overall EMG-LV gain. Associations of different positive slope parameters on the MA side with the corresponding plantar flexor AS and gait speed were explored using Spearman rank correlation. In the latter analyses, a gain of zero was assigned to a given limb in the absence of a positive or significant positive slope. A P value ≤.05 indicated statistical significance, and no adjustments were made for multiple tests because of the exploratory nature of the study.
Results
As expected, the average AS and the plantar flexor AS were significantly higher on the MA than LA side in both patient groups (.004≤P≤.011, see table 1). Only 3 patients with TBI had higher-than-normal resting hypertonia on the LA side. The self-selected free speed of TBI and stroke patients was on average 10 to 15cm/s slower than the very slow speed of the corresponding healthy controls (see table 2), but the differences were not significant (P≥.10). The durations of stance and MG lengthening were significantly longer on the LA side in both patient groups compared with the MA side (P≤.008) and controls (P≤.003). However, the peak lengthening velocity was not significantly different between the 2 sides of patients or in comparison with controls (P≥.33).
The overall EMG-LV gain (see fig 2, step 1), calculated as an average of both positive and negative slopes, was highly positive on the MA side and highly negative on the LA side in TBI patients, whereas it was on average around zero on both sides of stroke patients (see table 2). As a result, the overall EMG-LV gain significantly differed between MA and LA sides only in TBI (P=.013) but not stroke patients (P=.86) or in comparison with controls (P≥.06).
The overall EMG-LV slope was positive on the MA side in 8 (73%) of 11 TBI patients and in 6 (55%) of 11 stroke patients (see table 2). All limbs of patients and controls had at least 1 positive slope, except the LA limb in TBI (7/11, 64%). Nearly all MA limbs of both TBI and stroke patients had at least 1 significant positive slope (9/11, respectively), but so did the LA limb of stroke patients (see table 2).
Further analysis (see fig 2, steps 2 and 3) revealed that the frequencies of positive and significant positive slopes across all gait cycles were on average significantly higher on the MA side compared with the LA side of TBI but not stroke patients (fig 3A). Also, the frequencies of positive and significant positive slopes on the MA side of TBI patients were significantly higher than in controls (see fig 3A). However, the average gains of positive and significant positive slopes were not statistically different between the 2 sides of patients or in comparison with controls (fig 3B). No significant correlations were found between positive slope parameters and the AS in MA plantar flexors (P≥.18) or gait speed (P≥.16) in both patient groups.
Fig 3Comparison of frequencies (A) and gains (B) of positive (left) and significant positive (right) EMG-LV slopes in MG during stance in the study samples.
The purpose of this study was to further characterize the velocity-dependent increase in MG activity during the stance phase of gait in stroke and TBI patients with moderate to severe resting hypertonia. The results indicate an inconsistent ability of different EMG-LV slope parameters to distinguish the MA from LA side of patients, or patients from controls. This implies that a velocity-dependent increase in MG activity during stance is not exaggerated in patients with chronic ABI.
The background for our hypothesis was the work of Lamontagne,
The proposed “locomotor measure of spasticity” represents the overall gain of the EMG-LV slope during MG lengthening in the stance phase of gait. The overall EMG-LV gain is derived by averaging both positive and negative slopes, whereas only positive slopes conform to the concept of Crenna.
To follow this concept more closely, we hypothesized that the EMG-LV slope during MG lengthening will be more often positive and steeper on the MA than LA side of patients and in comparison with the dominant side of controls. This hypothesis was largely refuted based on several converging results from an in-depth analysis of positive slopes. A partial support of our hypothesis comes from bilaterally higher frequencies of positive and significant positive slopes found in TBI patients only and in comparison with their controls. The corresponding gains, however, were not statistically different between the 2 sides of patients or in comparison with controls. These findings may be ascribed to unexpectedly high frequencies of positive and significant positive slopes and high average gains found in controls. The paucity of significant differences between patients and controls raises questions about the discriminative utility of EMG-LV slope as a measure of dynamic spasticity during the stance phase of gait once the concept of Crenna
when both positive and negative EMG-LV slopes were analyzed. The overall EMG-LV slope was positive on the MA side in 73% of our TBI and 55% of our stroke patients, which approximates the rates previously reported for the paretic side after stroke (66%–73%
). Also, the overall EMG-LV gain was higher on the MA than LA side in patients with TBI. The key differences, however, are in the LA side of patients and among controls, since the overall EMG-LV gain was mostly negative in the latter 2 studies, whereas it was positive in more than 25% of our patients and controls. Selection bias may be the reason for the discrepancy; previous studies included 30 patients within 6 months of stroke with a broad range of resting muscle hypertonia compared with our sample of chronic TBI and stroke patients limited to moderate to severe hypertonia. Early reorganization in both hemispheres
Selection bias may also explain the relatively higher frequencies of positive slope on the MA side of our TBI (69%) and stroke (55%) patients compared with the 44% rate in a group of acute and chronic stroke patients reported by Lin.
The EMG-LV results differed between our TBI and stroke patients. Specifically, positive and significant positive slopes across all gait cycles were on average significantly more prevalent on the MA side compared with the LA side of TBI but not stroke patients (see fig 3A). In the absence of evidence about different pathophysiology of muscle hypertonia after TBI compared with stroke, the discrepancy between our 2 groups may be attributable to the often bilateral impairment after TBI, age difference, or small sample size. We can only speculate whether contralesional neuroplastic changes after stroke
can cause the LA limb to exhibit EMG-LV characteristics similar to those in the MA limb, resulting in less apparent differences between the 2 sides. On the other hand, a larger proportion of stroke than TBI patients wore a nonrigid ankle orthosis on the MA side, which may have limited the MG lengthening (see table 2). The argument against this is the lack of significant differences in the ankle range of motion and peak MG lengthening velocity on the MA side during stance, as well as positive slope parameters, between patients with and without orthosis.
The clinical significance of a positive EMG-LV slope remains unclear because of conflicting results. It has been previously suggested that hyperactive stretch reflexes during lengthening of plantar flexors in early stance may disturb kinematics of the impaired leg and compromise the efficiency of ankle push-off in late stance.
concluded that a short-latency stretch reflex does not significantly contribute to premature calf muscle activity in the early stance. The large number of positive slopes with nonsignificant fit reported here also suggests that coupling between the velocity of muscle lengthening and the associated EMG is not as tight during the stance phase of gait and likely involves other mechanisms than the stretch reflex. Because the positive EMG-LV correlation by itself is not sufficient to conclude that dynamic sensitivity of the stretch reflex is increased,
the contribution of altered supraspinal and spinal mechanisms to MG lengthening activity during stance should also be considered. Considerable contribution of the supraspinal input is supported by our finding of a significant correlation between the frequency of a positive EMG-LV slope and coactivation between the TA and MG muscles found in TBI patients during stance.
A weak correlation between the AS and positive slope parameters on the MA side suggests that resting hypertonia is unrelated to “dynamic spasticity” assessed in this way. However, the inclusion of patients with moderate to severe hypertonia may have confounded the results because of the narrow range of AS scores. Also, assigning zero values when a positive or significant positive slope was not present may have skewed the results of correlation analyses.
Study Limitations
The potential impact of selection bias has already been addressed, since all subjects were drawn from an ABI population evaluated for antispasticity treatment. Patients were allowed to use customary walking aids to avoid unnatural walking and ensure stability.
Nevertheless, wearing a nonrigid ankle orthosis did not substantially affect the results. Although the use of walking aids might have altered the pattern of muscle activation, not allowing walking aids would likely lead to compensatory strategies that would further confound the results of this study. Movement of the soft tissues between skin markers and bones is a source of error in skin marker–based motion analysis; however, the agreement between bone-based and skin marker–based knee flexion/extension and ankle plantar flexion/dorsiflexion angles during walking is generally good.
Residual effects of a traumatic brain injury on locomotor capacity: a first study of spatiotemporal patterns during unobstructed and obstructed walking.
it is still relatively small and limits generalization of findings.
Conclusions
The results suggest that the velocity-dependent increase in MG activity is not exaggerated during the stance phase of gait or related to resting muscle hypertonia and gait speed in chronic TBI and stroke patients. Thus, the associated changes in gait in these patients cannot be predominantly ascribed to increased stretch reflex activity during stance.
Residual effects of a traumatic brain injury on locomotor capacity: a first study of spatiotemporal patterns during unobstructed and obstructed walking.
Supported in part by the Wilson Research Foundation, Jackson, MS.
No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated.
In-press corrected proof published online on May 15, 2012, at www.archives-pmr.org.
P<.05,
00239-0&id=gr1.jpg){kind=link}
00239-0&id=gr2.jpg){kind=link}
00239-0&id=gr3.jpg){kind=link}