Edited by: Bruce H. Dobkin, University of California Los Angeles, USA
Reviewed by: Sandra K. Hunter, Marquette University, USA; George Wittenberg, University of Maryland School of Medicine, USA
This article was submitted to Stroke, a section of the journal Frontiers in Neurology.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Voluntary movement is contingent on the ability to generate and control muscle force. To optimize motor performance commands from the cortex must be transmitted with high fidelity to the peripheral neuromusculature via the descending tracts. Damage to any component of this pathway from stroke can impair voluntary movement (
Voluntary activation can be assessed using two different sites of stimulation: peripheral or cortical. The traditional peripheral method applies a percutaneous electrical stimulus to the motor nerve of the muscle during a maximal voluntary contraction (MVC) (
Peripheral nerve stimulation after stroke has demonstrated reduced voluntary activation on the more-affected side of 48–86% for the lower limb (
The aim of this study was to investigate the contribution of peripheral and supraspinal mechanisms to post-stroke muscle weakness in the elbow flexors using voluntary activation techniques. This is the first study to assess the level of voluntary activation in both elbow flexors of stroke patients and healthy age-appropriate control subjects, using both peripheral nerve and cortical stimulation. The contractile properties and EMG characteristics of the stimulated muscles were also investigated. We hypothesized that patients would have reduced cortical voluntary activation on both sides and that this would reflect changes in descending cortical drive and the consequent EMG response.
The elbow-flexor muscles of 10 stroke patients were studied (aged 39–75 years, 4–123 months post-stroke, see Table
Stroke patients | Healthy subjects | |
---|---|---|
10 | 6 | |
Age | 61.2 ± 12.3 | 61.3 ± 14.0 |
Sex (male:female) | 8:2 | 4:2 |
Months post-stroke | 17.6 ± 17.4 | n/a |
Dominant side (right:left) | 8:2 | 6:0 |
More-affected side (right:left) | 5:5 | n/a |
Dominant side affected | 7 | n/a |
Stroke type ischemic:hemorrhagic | 7:3 | n/a |
Wolf Motor Function Test (s) | 11.2 ± 20.0 | n/a |
Fugl-Meyer assessment | 56.5 ± 5.3 | n/a |
MALQOM | 89.2 ± 42.4 | n/a |
Modified Ashworth scale | ||
Elbow ( |
1.5 [0–2.0] | n/a |
Shoulder ( |
1.0 [1–1.5] | |
Arm circumference (mm) (ma:la, dom:nd) | 293:301 | 292:281 |
The experimental protocols used in this study were designed to replicate the work of Todd and colleagues in healthy subjects (
Elbow flexion torque (Newton meter) was measured using a 2 kN isometric load cell (Xtran, Applied Measurements Australia), low-pass filtered at 20 Hz, amplified 550–1760 times and sampled at 2000 Hz. EMG was recorded with surface electrodes from the biceps brachii and triceps brachii in a muscle belly tendon arrangement. Surface EMG signals were amplified 200–300 times, filtered from 10 to 1000 Hz using a 1902 amplifier (CED, UK) or IP511 amplifier (Grass, USA) and sampled at 5000 Hz. Torque and EMG were recorded and digitized using a 1401 digital analog converter and Spike2 software (CED, UK).
After familiarization with the experiment protocols, tests were completed in the order described below. The less-affected arm of stroke patients or the dominant arm of healthy subjects was tested first to gain familiarity with the protocol on the better performing side. Maximal elbow-flexor torque for each side was determined as the peak torque produced during two to three brief (2–3 s) MVCs performed with strong verbal encouragement and visual feedback. The torque target for contractions at 100, 75, and 50% MVC was displayed on-screen for subsequent trials. Care was taken to ensure the MVC amplitude occupied ~1/3 of the screen regardless of the absolute torque. Participants were instructed to pull as hard as they could for the MVC, but to match the target display for the 75 and 50% efforts.
Three forms of stimulation were used throughout the study: peripheral electrical stimulation at the brachial plexus and at the motor point of the biceps; and stimulation of the motor cortex using TMS. The stimulation protocols are illustrated in Figures
During the experimental setup single pulse electrical stimuli, 0.1 ms pulse width (DS7AH, Digitimer, UK), were delivered to Erb’s point via a cathode over the brachial plexus in the supraclavicular fossa. The anode was placed over the acromion. A stimulus–response curve was undertaken to determine the maximal compound muscle action potential (Mmax) of the resting biceps and triceps muscles. The electrical stimulus was increased from 2.5 to 5 mA and thereafter in 10 mA increments until Mmax was reached. The amplitude of Mmax was required to normalize peripheral and cortically evoked EMG measures (see Data Analysis).
During the peripheral stimulation protocol single pulse supramaximal electrical stimuli, 0.1 ms pulse width (DS7AH, Digitimer, UK), were delivered to the intramuscular branches of the musculocutaneous nerve innervating the biceps. The stimulating electrode for the biceps was located at approximately two–thirds of the distance between the axillary and elbow creases in the midline of the biceps muscle, with the anode on the biceps tendon. A stimulus–response curve was performed as described above to define the current required to produce a resting twitch of maximal amplitude in the unpotentiated biceps muscle. The supramaximal stimulation intensity was set to 120% of the level required to produce the maximal twitch. Once set for each side, stimulator intensity remained at this level for the duration of the experiment.
Peripheral voluntary activation was assessed from five MVCs performed with a 2-min rest between contractions to avoid fatigue (Figure
Transcranial magnetic stimulation was applied to the motor cortex (Magstim 2002, Magstim Co, UK) to elicit MEPs in the biceps and triceps. A circular coil of 12.5 cm external diameter was positioned and held by the experimenter over the vertex (Figure
As cortical and motoneuronal excitability increases during a voluntary effort, the amplitude of the cortically evoked resting twitch must be estimated rather than measured directly (
The functional ability of stroke patients was assessed on the more-affected side on a separate day prior to voluntary activation testing (Table
Torque signals were smoothed using a 10-ms time constant. Maximal voluntary torque was measured as the peak amplitude prior to the stimulus for 100% contractions, and over 50 ms prior to the stimulus for 75 and 50% contractions. The amplitude of the superimposed and resting twitches was calculated as the difference between the twitch peak and the mean torque for 50 ms immediately prior to the stimulus for all contraction strengths. To allow inter-subject comparisons twitch amplitude was normalized to the amplitude of the largest MVC recorded for that side (% MVC). The time-to-peak of the superimposed twitch and resting twitch were measured (milliseconds) with the onset taken from the stimulus to provide an unambiguous measurement point. This was particularly important when measuring time-to-peak of the superimposed twitch evoked with peripheral nerve stimulation as the exact twitch onset was not always evident during high activation. Half-relaxation time of the resting twitch was measured from the peak of the twitch to the point where the twitch torque was reduced by 50%. The estimated resting twitch was calculated from a linear regression between the amplitude of the superimposed twitch (Newton meter) and the voluntary torque (Newton meter) at 100, 75, and 50% MVC. The
The level of voluntary activation was calculated for both peripheral and cortical stimulation methods using the interpolated twitch technique: voluntary activation (%) = (1 − superimposed twitch/(estimated) resting twitch) × 100.
Electromyographic activity and MEPs were recorded from the biceps and triceps during cortical stimulation and analyzed for participants who completed bilateral assessments. EMG could not be recorded during peripheral voluntary activation due to stimulation at the biceps motor point. The DC offset was removed from the EMG signal before root mean square (RMS) processing using a 10-ms time constant. The mean of the background noise in the EMG signal was measured over 500 ms and subtracted from EMG values. RMS EMG was measured as the mean amplitude over 500 ms immediately prior to the stimulus, and normalized to the amplitude of Mmax (% Mmax). The area of Mmax and the MEPs were measured between set cursors for both the biceps and triceps muscles. The ratios of the triceps to biceps EMG and MEPs were calculated to gauge co-contraction of the agonist and antagonist muscles. Finally, the duration of the silent period following cortical stimulation was measured in the biceps EMG from the stimulus to the return of continuous EMG. MEP responses evoked with cortical stimulation were normalized to Mmax to enable between subject comparisons.
Data were analyzed using separate two-way ANOVAs for torque, EMG, and MEP data. Factors for the torque data were side-tested and stimulation type. Factors for the EMG and MEP data were side-tested and contraction level (% MVC). Results are presented as mean ± standard error of the mean (SEM) with Holm–Sidak
Functional assessments scores for the patient group are reported in Table
Torque demonstrated an effect for the side-tested (
There was an effect of side-tested for biceps Mmax area (
The current required to evoke Mmax during Erb’s point stimulation and biceps motor point stimulation was not significantly different either within or between groups. The mean current at Erb’s point was 46.0 ± 6.7 mA on the more-affected side, 51.0 ± 8.2 mA on the less-affected side, and 56.7 ± 7.5 mA for healthy subjects. The mean current at the biceps motor point was 64.0 ± 4.9 mA on the more-affected side, compared to 62.0 ± 6.3 mA on the less-affected side, and 57.5 ± 4.9 mA for healthy subjects.
The resting twitch was not significantly different either between or within groups for amplitude (Newton meter), time-to-peak, or half-relaxation time (Figure
Stroke patients |
Healthy subjects | ||
---|---|---|---|
More-affected | Less-affected | Pooled data | |
side | side | ||
RT time-to-peak (ms) | 96 ± 3 | 95 ± 5 | 97 ± 2 |
RTHRT (ms) | 83 ± 7 | 71 ± 8 | 75 ± 5 |
RTSD | 1.2 (0.7–1.8)* | 0.9 (0.7–1.2)* | 0.4 (0.2–0.5) |
SIT time-to-peak (ms) | 67 ± 3** | 61 ± 7 | 53 ± 3 |
SITSD | 1.7 ± 3* | 0.9 ± 3 | 0.5 ± 1 |
SIT time-to-peak (ms) | 91 (84–92)* | 77 (62–91) | 76 (44–83) |
SITSD | 2.1 (1.1–3.7)* | 1.2 (1.0–2.1)** | 0.6 (0.4–1.0) |
Biceps | 12.1 ± 0.3* | 10.9 ± 0.3 | 10.9 ± 0.3 |
Triceps | 14.2 ± 0.5* | 13.6 ± 0.5* | 12.6 ± 0.5 |
Silent period biceps (ms) | 250 ± 22 | 170 ± 22* | 220 ± 20 |
There was an effect for side-tested for the amplitude of the superimposed twitch (
Overall there was an effect for side-tested (
Biceps EMG demonstrated an effect of side-tested (
The ratio of triceps to biceps EMG demonstrated an effect for side-tested (
There was no difference in the level of stimulator output required to evoke the optimal biceps and triceps MEP amplitudes between sides or between groups. Stimulator output was 74.0 ± 4.0% on the more-affected side, 73.5 ± 3.7% on the less-affected side, and 75.5 ± 3.6% in healthy subjects.
There was an effect for side-tested during the biceps MEP stimulus–response curve collected during contractions of 50% MVC (
A main effect for side-tested was demonstrated for biceps MEP (
The target amplitude for biceps and triceps MEPs ( >60% Mmax and <10–15% Mmax, respectively) was only achieved by a single patient on the more-affected side. This increased to 4/10 on the less-affected side. The triceps to biceps MEP ratio demonstrated an effect for side-tested (
The biceps MEP latency had an effect for side-tested (
The duration of the silent period evoked during cortical stimulation was measured only in the biceps due to the low level of triceps activity (Table
The median amplitude of the superimposed twitch evoked at 100% MVC with cortical stimulation was different between sides-tested when reported in both Newton meter (
In contrast to healthy subjects, the amplitude of the superimposed twitch for the patient group was not consistently graded to the amplitude of the contraction and there was notable inter-trial variation (Figure
This study is the first to use both peripheral and cortical stimulation techniques to assess voluntary activation in the elbow-flexor muscles on both sides of stroke patients and healthy age-appropriate control subjects. The contractile properties and EMG characteristics of the stimulated muscles were also examined to suggest where in the motor pathway post-stroke impairments occurred. Although impairments in the voluntary activation of the less-affected side have been demonstrated previously in the lower limb post-stroke, the results of this study suggest the bilateral impairments in voluntary activation in the upper limb are the result of distinct differences in neural drive to the more- and less-affected side. Peripheral nerve stimulation revealed significant reductions in voluntary activation of the elbow flexors on both the more- and less-affected sides after stroke compared to healthy age-matched subjects. However, no differences in the contractile properties were evident from twitches (in Newton meter) evoked at rest by this stimulation, suggesting reduced neural drive to the muscle is a greater contributor to post-stroke impairments in this cohort than changes to the muscle itself. Cortical stimulation was completed in all patients, although the inconsistency of the superimposed twitch (Figure
This is the first study to implement both peripheral nerve and cortical stimulation voluntary activation protocols in the same way for both patients and healthy control subjects. The methods and muscles studied were chosen to replicate as closely as possible the work of Todd and colleagues (
The necessary methodological differences that occurred during the cortical stimulation protocol between patients and healthy subjects may also have influenced our results. When stimulating over the vertex during a 50% MVC, the desired biceps and triceps MEP target amplitudes ( >60% Mmax and <15% Mmax, respectively) were not achievable in most patients, despite selection of the stimulation site over the cortex generating the largest biceps and smallest triceps MEPs. Similarly, we may have
Due to the characteristics of the small cohort examined in this study, our results can only be generalized to well-recovered patients after stroke. Although the use of the more-affected upper arm was quantified by the MALQOM as moderate to high (Table
Maximal voluntary torque was significantly reduced on both the more- and less-affected sides of stroke patients compared to healthy subjects. Maximal Mwave amplitudes were also reduced in the stroke patients and this suggests a deficit in the muscles. However, this was not apparent in evoked force responses. Surprisingly, there were no differences in the contractile properties of the biceps resting twitch between groups, despite considerable variation in the amplitude of the five twitches in some patients (Figure
Despite the similarities in the contractile properties of the muscle, an increase in muscle resistance or passive stiffness as measured with the modified Ashworth scale (Table
Voluntary activation scores calculated from peripheral nerve stimulation were significantly reduced on both the more- and less-affected sides of stroke patients compared to healthy subjects. Bilateral voluntary activation deficits have been reported previously in leg muscles after stroke (
The reduced voluntary activation identified through peripheral nerve stimulation indicates that the motoneurones, and therefore the muscle, are not being driven adequately on either the more- or less-affected side post-stroke. This may be due to reductions in the number of neural connections, or it could be that the connections are intact but are not being fully driven (
The reduced Mmax amplitude on both sides post-stroke suggests there may be some peripheral deficit, but the use of peripheral nerve or Erb’s point stimulation alone cannot distinguish between changes in the functional connectivity and anatomy. Thus voluntary EMG and muscle responses to cortical stimulation were examined in detail to investigate these mechanisms after stroke and in healthy controls. EMG was reduced post-stroke on the more-affected but not the less-affected side compared to healthy subjects. The reduction in voluntary activation on the more-affected side was consistent with decreased EMG activity compared to both the less-affected side and healthy subjects. This reduction in voluntary EMG presumably reflects not only reduced descending drive but also post-stroke impairments in the firing rate, rate modulation, and recruitment of motoneurones (
Superimposed twitches were evoked for all participants using cortical stimulation (Figure
We examined MEPs to ascertain possible differences in the connectivity of the descending corticospinal tract. The latencies of the MEPs on the more-affected side were prolonged and the area of the maximal MEP on the more-affected side was reduced compared to the less-affected side and healthy subjects. Differences were apparent during both the cortical stimulus–response curve at 50% MVC (Figure
Similar to the voluntary EMG, the area of the biceps MEPs on the less-affected side of patients was not different to that of healthy subjects (Figure
This is the first study to demonstrate reductions in upper-limb voluntary activation and maximal torque on both the more- and less-affected sides in patients with relatively high motor-function post-stroke, compared to healthy age- and sex-matched control subjects. Comparisons between the results of the peripheral and cortical stimulation methods suggest the bilateral voluntary activation impairments measured are caused by distinct impairments in neural drive. Peripheral nerve stimulation did not reveal differences in the contractile properties of the elbow-flexor muscles at rest. In contrast, the voluntary activation score could not be calculated with certainty with cortical stimulation due to the inconsistency of the superimposed twitches. Irrespective, examination of the EMG results from the cortical stimulation showed significant bilateral, yet distinct differences in the neural drive after stroke, and in the connectivity of the descending motor pathway on the more-affected side. We suggest that the expected muscle weakness on the more-affected side of stroke patients was due to impairments of the descending corticospinal connections, coupled with an inability to drive through those connections. In contrast, descending connections to the less-affected side appeared to be intact, but the weakness was due to an inability to voluntarily drive the connections to the muscle. This study suggests quantification of voluntary activation with cortical stimulation is not possible in patients after stroke. Despite this finding, cortical stimulation revealed changes in the neural drive and descending tracts in this cohort with high motor-function that could not be measured when using peripheral nerve stimulation alone. These findings are presumably substantially greater in patients with less motor-function. The impairments in cortical networks and descending pathways identified in this study highlight the importance of neurorehabilitation strategies that target both sides of the body.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The authors gratefully acknowledge funding from the National Health and Medical Research Council, Australia; New South Wales Office of Science and Medical Research; and the Faculty of Medicine, UNSW Australia.