Brief Report
EMG-Force Relations During Isometric Contractions of the First Dorsal Interosseous Muscle After Stroke Ping Zhou, PhD, 1,2,3 Xiaoyan Li, PhD,2,3 and William Zev Rymer, MD, PhD2,3 1 Institute of Biomedical Engineering, University of Science and Technology of China, Hefei, China; 2Sensory Motor Performance Program, Rehabilitation Institute of Chicago, Chicago, Illinois, USA; 3Department of Physical Medicine and Rehabilitation, Northwestern University, Chicago, Illinois, USA
Objective: This study examines the electromyogram (EMG)–force relations observed in the first dorsal interosseous (FDI) muscle of hemiparetic stroke survivors. Methods: Fourteen stroke subjects were instructed to perform different levels of index finger abduction using their paretic and contralateral hands, respectively. Surface EMG and force signals were recorded from the FDI muscle. The EMG-force relation was constructed using linear regression of the EMG amplitude and force measurements. Results: We found that there were diverse changes in the slope of the EMG-force relations in paretic muscles compared with contralateral muscles, with significant increases and decreases being observed relative to the contralateral side. Regression analysis did not verify strong correlations between the ratio of paretic and contralateral muscle EMG-force slopes and any clinical parameters. Conclusions: These findings suggest that there appear to be different types of processes (eg, motor unit control property changes, muscle fiber atrophy, spinal motoneuron degeneration, muscle fiber reinnervation, etc) at work post stroke that may impact the EMG-force relations and that may be present in varying degree in any given stroke survivor. Key words: EMG-force relation, FDI muscle, hemiparetic stroke
T
he nature of the relation between isometric muscle force and surface electromyogram (EMG) amplitude is very complex and is dependent on a variety of factors, both muscular and neural.1 Nonetheless, the EMG-force relation can be a useful marker of changes in spinal and/ or muscular components of the motor unit during voluntary or reflex activation of the muscle, or it can reflect changes in motoneuron pool activation. For example, in a reduced animal preparation with a neuraxis injury (such as the decerebrate cat with added cord dorsal hemisection), a significant increase in the regression slope of the EMG-force relation was observed in plantar flexor muscles after the cord section was imposed.2 The EMG-force relation can also be profoundly affected in patients with neurological disorders, such as hemiparetic stroke. In 2 previous studies, it was reported that the slope of the relation between biceps brachii EMG and elbow joint torque in approximately half of the examined stroke subjects was significantly greater in paretic side compared with contralateral or neurologically intact muscles, while the other subjects showed lower or almost unchanged slopes.3,4
A systematic analysis of EMG-force relations can provide valuable data enabling an informed interpretation of the mechanisms underlying neural and muscular pathological changes. Previous studies on EMG-force relations in stroke were mainly performed in proximal muscles.3,4 It is presently unknown whether or how the EMGforce relations may be altered in distal muscles. Accordingly, the objective of this study was to characterize whether the EMG-force relation observed in the affected first dorsal interosseous (FDI) muscle of hemiparetic stroke survivors deviates systematically from that observed in the contralateral muscle. We found that similar to previous reports with the proximal muscles,3,4 there were diverse changes in the slope of the EMGforce relations in paretic muscles compared with contralateral muscles, with significant increases and decreases being observed relative to the contralateral side. These findings, in combination
Top Stroke Rehabil 2013;20(6):537–544 © 2013 Thomas Land Publishers, Inc. www.strokejournal.com doi: 10.1310/tsr2006-537
537
538
TOPICS IN STROKE REHABILITATION/NOV-DEC 2013
with our previous simulation studies,5 suggest that there appear to be different types of processes (eg, motor unit control property changes, muscle fiber atrophy, spinal motoneuron degeneration, muscle fiber reinnervation, etc) at work post stroke that may impact the EMG-force relations and that may be present in a varying degree in any given stroke survivor. Methods Subjects
Fourteen subjects (8 males, 6 females; 58.3 ± 8.0 years) who sustained hemiparetic stroke as reported in Li et al6 participated in this study. All our stroke subjects were recruited from the Clinical Neuroscience Research Registry at the Rehabilitation Institute of Chicago (Chicago, IL). The study was approved by the Institutional Review Board of Northwestern University (Chicago, IL). All subjects gave their written consent before the experiment. A screening examination and clinical assessment were performed by a physical therapist. Among the 14 stroke subjects, the left limb was affected in 5 subjects, and the right limb was affected in 9 subjects. The duration between the stroke onset and the experiment time was 8.2 ± 6.7 years (range, 1.5-24.1 years). The 14 stroke subjects showed a hand portion of Chedoke-McMaster scale7 score of 4.8 ± 1.4 (range, 2-7) and an Upper Extremity Fugl-Meyer score8 of 45.7/66 ± 16.5/66 (range, 16/66 to 66/66). Experiments
Subjects were seated upright in a mobile Biodex chair (Biodex, Shirley, NY) with a standard 6 degrees-of-freedom load cell (ATI Inc, Apex, NC) setup used to accurately record the isometric contraction force of the FDI muscle during index finger abduction.6 The shoulders and waist of the subject were tightly strapped to the chair to limit trunk and shoulder movements. The forearm and wrist were mounted on a plastic platform inside a fiberglass cast. A ring-mount interface was used to strap the wrist in a partial pronation position. The proximal phalanx of the index finger was casted
and fixed to a small ring-mount interface attached to the load cell. This standard position served to minimize spurious force contributions from unrecorded muscles. Surface EMG signals were recorded from the FDI muscle using a 5-pin (0.5 mm in diameter) EMG sensor array (Delsys, Boston, MA). The array is arranged in a Laplace configuration and can generate 4 channels of differential EMG signals.9 A 16-channel Bagnoli amplifier (Delsys Inc, Boston, MA) with a gain of 1,000 was used to record surface EMG and force signals. Both surface EMG and force signals were sampled at 20 kHz using EMGWorks (Delsys, Boston, MA), with a bandpass filter (bandwidth 20-2,000 Hz) for surface EMG and a low pass filter (cutoff frequency 450 Hz) for force measurement, respectively. At the beginning of the experiment, the isometric maximum voluntary contraction (MVC) of each subject’s paretic FDI muscle for index finger abduction was determined. Target force was then assigned as a percentile of the paretic MVC force: from 20% to 80% of the paretic MVC at 10% MVC increments. With a matched force protocol, these target force levels were used as a standard for both paretic and contralateral muscles; for each subject, the same contraction levels as performed by the paretic muscle were used for the contralateral muscle. In each trial, the target force was displayed as a red trapezoid trajectory, and the trace of the voluntary contraction force was updated in realtime in a different color. Subjects were instructed to match the target force trajectory using their visual feedback and to maintain the force as stable as possible for up to 10 seconds. The protocol for each target force was performed at least twice. The order of target forces was randomized. Practice was given to help familiarize the subjects with the task and to make appropriate adjustments to the force. Substantial rest period between trials were provided to prevent potential fatigue during the experiment. Data analysis
Surface EMG and force signals were exported and analyzed offline in Matlab (MathWorks, Natick, MA). A notch filter was applied to surface
EMG-Force Relations During Isometric Contractionss
EMG to remove power line interference. For each trial, a constant force period for at least 1 second was selected by visual assessment of the force signal. Then the average force magnitudes in abduction (Fx) and flexion (Fy) were calculated. The force deviation from pure abduction was assessed by the inverse tangent of the ratio between flexion and abduction forces (Fy/Fx). The trials with a deviation of more than 30° were excluded for further analysis. For surface EMG, the signal to noise ratio (SNR) was calculated for each of the 4 recording channels. The average rectified value (ARV) of the surface EMG was calculated for the channels with SNR greater than 15 dB while the other channels were excluded from further analysis. The EMG-force relation was constructed using linear regression of the individual data set, and the slope of the linear fit was calculated. This was performed for all the surface EMG channels, and the average slope was determined for paretic and contralateral muscles of each stroke subject. Student t tests were used to examine whether the slope of the EMG-force relation reached statistical difference between paretic and contralateral muscles. Regression analysis was performed to examine the correlations between the ratio of EMGforce slope (between paretic and contralateral muscles) and clinical assessment. Statistical significance was defined as P < .05. Results Force measurement
The MVC force in abduction was recorded for both paretic and contralateral FDI muscles in each stroke subject, and their ratio was calculated. As we expected, the paretic muscle MVC force was systematically lower than that of the contralateral muscle. Across subjects, the average MVC force was 22.8 ± 13.3 N for the paretic FDI muscle and 40.2 ± 12.0 N for the contralateral muscle. The average ratio of the MVC force between paretic and contralateral muscles was 0.55 ± 0.25. Across all force levels, there was no significant difference in force deviation from pure abduction between the paretic and contralateral muscles (P > .2), although the average force deviation of paretic
539
muscle (11.1 ± 9.0°) was larger than contralateral muscle (7.3 ± 6.0°). EMG-force relation
Across all subjects, the average slope of the EMGforce relation was 6.19 ± 5.01 µV/N for the paretic muscle, which was slightly higher than the slope of 5.48 ± 4.23 µV/N for the contralateral muscle, but no significance was observed (P > .5). When examining individual stroke subjects, we observed clear increases and decreases in the slope of the EMG-force relation in paretic muscles compared with contralateral muscles. In 9 of the 14 subjects, we found a clear increase in the slope of the EMG-force relation in the paretic muscles. Figure 1 shows an example of the EMG signals recorded at similar contraction levels of the paretic and contralateral muscles of the same stroke subject, and the linear regression of the EMG-force relation, representative of these 9 subjects. In these 9 subjects, for a given force level, surface EMG amplitude was always higher for the paretic muscle than for the contralateral muscle. The average slope for these 9 subjects was 7.50 ± 5.15 µV/N for the paretic muscle, which was significantly higher than the slope of 4.41 ± 3.43 µV/N for the contralateral muscle (P < .05). The mean ratio of slope between paretic and contralateral muscles was 1.95 ± 0.95. The remaining 5 stroke subjects displayed a decreased slope of EMG-force relation for the paretic muscle when compared with the contralateral muscle. Figure 2 shows an example of the EMG signals recorded at similar contraction levels of the paretic and contralateral muscles of the same stroke subject, and the linear regression of the EMG-force relation, representative of these 5 subjects. In these 5 subjects, for a given force level, surface EMG amplitude was always smaller for the paretic muscle than for the contralateral muscle. The average slope for these 5 subjects was 3.83 ± 4.22 µV/N for the paretic muscle, which was significantly lower than the slope of 7.41 ± 5.23 µV/N for the contralateral muscle (P < .05). The mean ratio of slope between paretic and contralateral muscles was 0.45 ± 0.22. Comparison of the clinical data between the 9 stroke subjects (age, 58.7 ± 9.5 years; duration
540
TOPICS IN STROKE REHABILITATION/NOV-DEC 2013
(a)
(a) EMG
EMG
0.5 mV
0.5 mV
5s 5s
2N
Force
Force
2N
Paretic
Paretic (b)
Contralateral
Paretic Contralateral
Surface EMG (mV)
Surface EMG (mV)
Paretic Contralateral 0.08 0.06 0.04 0.02 0
0.08 0.06 0.04 0.02 0
0
1
2
3
4
5
Contralateral
(b)
6
7
8
9
Force (N)
0
5
10 15 Force (N)
20
Figure 1. Representative electromyogram (EMG) and force recordings of the 9 stroke subjects whose linear regression of the EMGforce relation demonstrates higher slope in paretic muscle than in contralateral muscle. (a) A comparison of the EMG signals from paretic and contralateral muscles at the similar force levels (dashed line represents the target force). In this example, the EMG from the paretic muscle was 0.075 mV at the force level of 3.9 N, and the EMG from the contralateral muscle was 0.027 mV at the contraction level of 4.3 N. (b) A linear regression of the EMG-force relation. The slope of the EMG-force relation was 0.0087 mV/N for paretic muscle (R2 = 0.98, P < .001) and 0.0052 mV/N (R2 = 0.80, P < .001) for the contralateral muscle.
Figure 2. Representative electromyogram (EMG) and force recordings of the 5 stroke subjects whose linear regression of the EMGforce relation demonstrates lower slope in paretic muscle than in contralateral muscle. (a) A comparison of the EMG signals from paretic and contralateral muscles at the similar force levels (dashed line represents the target force). In this example, the EMG from the paretic muscle was 0.010 mV at the force level of 5.2 N, and the EMG from the contralateral muscle was 0.024 mV at the contraction level of 5.6 N. (b) A linear regression of the EMG-force relation. The slope of the EMG-force relation was 0.0020 mV/N for paretic muscle (R2 = 0.42, P < .001) and 0.0035 mV/N (R2 = 0.82, P < .001) for the contralateral muscle.
of stroke, 6.5 ± 3.7 years; paretic to contralateral MVC ratio, 0.50 ± 0.27; Chedoke score, 5.1 ± 1.2; Fugl-Meyer score, 48.7 ± 16.9) demonstrating an increased slope of the EMG-force relation and the 5 subjects (age, 57.6 ± 5.1 years; duration of stroke, 11.2 ± 10.0 years; paretic to contralateral MVC ratio, 0.64 ± 0.22; Chedoke score, 4.2 ± 1.8; Fugl-Meyer score, 40.4 ± 16.1) demonstrating a
decreased slope did not show significant differences between the 2 groups (P > .3). Regression analysis did not verify strong correlations between the ratio of paretic and contralateral muscle EMGforce slopes and any clinical parameters (Chedoke score, R2 = 0.032, P > .8; Fugl-Meyer score, R2 = 0.08, P > .3; duration of stroke, R2 = 0.1, P > .7; muscle strength, R2 = 0.14, P > .2).
EMG-Force Relations During Isometric Contractionss
Effect of force normalization
The slope of the EMG-force relation was also re-examined by normalizing the force to each muscle’s maximum strength. Comparing the slope ratios of EMG-force relations between paretic and contralateral muscles, we found that the slope ratio was increased across all subjects after force normalization (because of paretic muscle weakness). For the 9 subjects demonstrating increased slope in paretic muscle (without force normalization), after the force normalization the average EMG-force slope was 8.05 ± 7.79 µV/10%MVC for paretic muscle and 1.64 ± 1.18 µV/10%MVC for contralateral muscle, respectively. For the 5 subjects demonstrating deceased slope in paretic muscle (without force normalization), after the force normalization the average EMG-force slope was 1.86 ± 1.27 µV/10%MVC for paretic muscle and 2.94 ± 1.97 µV/10%MVC for contralateral muscle, respectively. After normalizing the force to each muscle’s maximum strength, 13 subjects showed consistent slope ratios (between paretic and contralateral muscles) with respect to greater or less than unity, while 1 subject demonstrated a slope ratio change from 0.45 (without force normalization) to 1.24 (with force normalization).
Discussion We examined the alterations in the EMG-force relations of the FDI muscle after hemispheric stroke. FDI is a multifunctional muscle that generates torque about the second metacarpophalangeal (MCP) joint. The abduction force produced at the joint is generated largely by the FDI muscle, whereas the flexion force of the index finger is generated by a portion of the FDI muscle and by other synergists (eg, the flexor digitorum profundus and flexor digitorum superficialis muscles) as well as by other interosseus and lumbrical muscles. Thus, isometric abduction of the index finger was used in this study as a measure of FDI muscle activation. We chose to compare EMG-force relation of paretic muscles to contralateral muscles of the same stroke subjects rather than muscles from neurologically intact subjects. Considering the
541
similarities in muscle mechanical moment arms, muscle strength, and skeletal dimensions between the 2 limbs of the same subject, such a comparison may allow a more straightforward assessment of EMG-force relation characteristics than would be possible between muscles drawn from different subjects, who would have different limb sizes, work histories, and muscle usage patterns.4 We found that for individual stroke subjects, the slope of the EMG-force relation for paretic muscle showed diverse changes when compared with contralateral muscle, potentially due to different neural and/or muscular changes. In 9 of the 14 stroke subjects, we observed a clear increase in the slope in EMG-force relation on the paretic FDI muscle compared to the contralateral muscle. These may be due to different factors, including muscle fiber reinnervation, impairment of motor unit control properties (ie, reduction of motor unit firing rates and compression of motor unit recruitment threshold), and (nonselective) loss of motor units.4,10,11 In such cases, to reach a given force, additional relatively large motor units have to be recruited. With large action potentials from these relatively big motor units, the resultant surface EMG is higher. At high force levels, this effect (ie, increase in EMG) is especially evident, resulting in a clear slope increase of the EMG-force relation. For the other 5 stroke subjects, the slope of the EMG-force relation from the paretic muscles demonstrated a smaller value compared with the contralateral muscles. This could be an indication of either muscle fiber atrophy or a selective degeneration of the large and superficial motor units.12,13 In such cases, to reach a given force, the resultant surface EMG of the paretic muscle involves summation of a relatively large number of action potentials with small amplitudes compared with fewer action potentials with relatively large amplitudes in the contralateral muscle. Due to the effect of action potential cancellation, the size of action potentials plays a more dominant role in determining the overall EMG amplitude than do the number of action potentials.5 Therefore, the surface EMG amplitude is lower for paretic muscle at a given force. At higher force levels, this effect (ie, decrease in EMG) is more evident, generating a slope decrease in the EMG-force relation.
542
TOPICS IN STROKE REHABILITATION/NOV-DEC 2013
In general, our findings in distal muscles were consistent to those from proximal muscles. Although previous studies in proximal muscles seemed to emphasize the EMG force slope increase in paretic muscles compared with contralateral muscles, not all the examined stroke subjects demonstrated such a change. For example, Gemperline et al examined biceps brachii of 6 stroke subjects and reported that 3 subjects showed substantially elevated EMG-force slopes in paretic muscles compared with contralateral muscles; 1 showed lower slopes and the other 2 had a slope ratio close to unity.4 Tang and Rymer reported that in almost half of their examined hemiparetic subjects (8/17) the slope of the EMG-force relation, measured over the biceps brachialis and brachioradialis musle groups, was increased in the paretic limb compared with the unimpaired limb.3 The findings in proximal muscles from the previous studies and distal muscles from the current study both demonstrated a rather inconsistent picture of EMG-force slope changes after stroke, suggesting that there may be different types of processes (eg, motor unit control property changes, muscle fiber atrophy, spinal motoneuron degeneration, muscle fiber reinnervation, etc) at work post stroke that may impact the EMG-force relations.5 Furthermore, the experimentally observed EMGforce slope variations in hemiparetic muscles of stroke subjects could be the overall effects of many different factors. We also note that the slope increase demonstrated in this study was not as dramatic as reported in previous studies.3,4 Such an observation may be due to different processes present in varying degrees in proximal and distal muscles post stroke. It may also relate to the differences in recording characteristics of the EMG systems. Compared with the previous studies using a conventional surface EMG electrode, the current study recorded surface EMG data with a highly selective surface electrode, which was primarily designed or configured for surface EMG decomposition.9 Although we were unable to establish relations between the clinical data and EMG-force slope changes, examination of the EMG-force relation may have potential clinical implications for diagnosis and treatment of stroke sequelae. Using simulation of motoneuron pool activity
(muscle force and surface EMG) as a guide, experimental recording of EMG-force relation can provide valuable knowledge enabling an informed interpretation of the motor unit mechanisms underlying paretic muscle changes after stroke. This is particularly important given the fact that discrimination of single motor unit activity from surface EMG is a very difficult task. Monitoring of EMG-force relation in paretic muscles can help assess the nature of the neural or muscular deficits and the response of these deficits to medication and physical therapies. For example, muscle weakness induced by drop in motor unit firing rates may be treatable through appropriate training and/or drug interventions. The increase in motor unit firing rates induced by such interventions can be evaluated by monitoring the EMG-force slope changes. Finally, in addition to our focus on EMGforce slope changes, we observed that the linear fitting of the EMG-force relation for the paretic muscles tended to introduce an increased ordinate intercept. We observed such a trend in 10 of the 14 stroke subjects. Due to resolution limitation of the load cell and difficulty in holding very low constant force levels, the EMG signals were not routinely recorded at near zero forces. The relatively high intercept for paretic muscles implied production of increased EMG activity at the near-zero force levels compared with the contralateral muscles. This could be ascribed to relatively large background EMG activity, for example, sustained spontaneous motor unit spikes as observed in the resting paretic muscle. Alternatively, there might be a nonlinear recruitment sequence involved for the initial force generation of the paretic muscle. Acknowledgments Financial support/disclosures: This work was supported by the Brinson Stroke Foundation, the National Institute on Disability and Rehabilitation Research of the US Department of Education under grant H133F110033, the National Institutes of Health of the US Department of Health and Human Services under grant 2R24 HD050821, and the 1000 Talents Plan Special Program of China (Recruitment Program of Global Experts). Additional contributions: The authors thank Nina Suresh, PhD, for helping with data collection.
EMG-Force Relations During Isometric Contractionss
543
REFERENCES 1. Zhou P, Rymer WZ. Factors governing the form of the relation between muscle force and the EMG: A simulation study. J Neurophysiol. 2004;92(5):2878-2886. 2. Blaschak MJ, Powers RK, Rymer WZ. Disturbances of motor output in a cat hindlimb muscle after acute dorsal spinal hemisection. Exp Brain Res. 1988;71(2):377-387. 3. Tang A, Rymer WZ. Abnormal force–EMG relations in paretic limbs of hemiparetic human subjects. J Neurol Neurosurg Psychiatry. 1981;44(8): 690-698. 4. Gemperline JJ, Allen S, Walk D, Rymer WZ. Characteristics of motor unit discharge in subjects with hemiparesis. Muscle Nerve. 1995;18(10): 1101-1114. 5. Zhou P, Suresh NL, Rymer WZ. Model based sensitivity analysis of EMG-force relation with respect to motor unit properties: Applications to muscle paresis in stroke. Ann Biomed Eng. 2007;35(9):1521-1531. 6. Li X, Suresh A, Zhou P, Rymer W. Alterations in the spike amplitude distribution of the surface electromyogram post-stroke. IEEE Trans Biomed Eng. 2013;60(3):845-852.
7. Gowland C, Stratford P, Ward M, et al. Measuring physical impairment and disability with the Chedoke-McMaster Stroke Assessment. Stroke. 1993;24(1):58-63. 8. Gladstone, DJ, Danells CJ, Black SE. The Fugl-Meyer assessment of motor recovery after stroke: A critical review of its measurement properties. Neurorehabil Neural Repair. 2002;16:232-240. 9. De Luca CJ, Adam A, Wotiz R, Gilmore LD, Nawab SH. Decomposition of surface EMG signals. J Neurophysiol. 2006;96:1646-1657. 10. Lukacs M, Vecsei L, Beniczky S. Changes in muscle fiber density following a stroke. Clin Neurophysiol. 2009;120:1539-1542. 11. Kallenberg LA, Hermens HJ. Motor unit properties of biceps brachii in chronic stroke patients assessed with high-density surface EMG. Muscle Nerve. 2009;39:177-185. 12. Ryan AS, Dobrovolny CL, Smith GV, Silver KH, Macko RF. Hemiparetic muscle atrophy and increased intramuscular fat in stroke patients. Arch Phys Med Rehabil. 2002;83:1703-1707. 13. Lukacs M, Vecsei L, Beniczky S. Large motor units are selectively affected following a stroke. Clin Neurophysiol. 2008;119:2555-2558.
EMG-Force Relations During Isometric Contractions of the First Dorsal Interosseous Muscle After Stroke Ping Zhou, PhD, 1,2,3 Xiaoyan Li, PhD,2,3 and William Zev Rymer, MD, PhD2,3 1 Institute of Biomedical Engineering, University of Science and Technology of China, Hefei, China; 2Sensory Motor Performance Program, Rehabilitation Institute of Chicago, Chicago, Illinois, USA; 3Department of Physical Medicine and Rehabilitation, Northwestern University, Chicago, Illinois, USA
Objective: This study examines the electromyogram (EMG)–force relations observed in the first dorsal interosseous (FDI) muscle of hemiparetic stroke survivors. Methods: Fourteen stroke subjects were instructed to perform different levels of index finger abduction using their paretic and contralateral hands, respectively. Surface EMG and force signals were recorded from the FDI muscle. The EMG-force relation was constructed using linear regression of the EMG amplitude and force measurements. Results: We found that there were diverse changes in the slope of the EMG-force relations in paretic muscles compared with contralateral muscles, with significant increases and decreases being observed relative to the contralateral side. Regression analysis did not verify strong correlations between the ratio of paretic and contralateral muscle EMG-force slopes and any clinical parameters. Conclusions: These findings suggest that there appear to be different types of processes (eg, motor unit control property changes, muscle fiber atrophy, spinal motoneuron degeneration, muscle fiber reinnervation, etc) at work post stroke that may impact the EMG-force relations and that may be present in varying degree in any given stroke survivor. Key words: EMG-force relation, FDI muscle, hemiparetic stroke
T
he nature of the relation between isometric muscle force and surface electromyogram (EMG) amplitude is very complex and is dependent on a variety of factors, both muscular and neural.1 Nonetheless, the EMG-force relation can be a useful marker of changes in spinal and/ or muscular components of the motor unit during voluntary or reflex activation of the muscle, or it can reflect changes in motoneuron pool activation. For example, in a reduced animal preparation with a neuraxis injury (such as the decerebrate cat with added cord dorsal hemisection), a significant increase in the regression slope of the EMG-force relation was observed in plantar flexor muscles after the cord section was imposed.2 The EMG-force relation can also be profoundly affected in patients with neurological disorders, such as hemiparetic stroke. In 2 previous studies, it was reported that the slope of the relation between biceps brachii EMG and elbow joint torque in approximately half of the examined stroke subjects was significantly greater in paretic side compared with contralateral or neurologically intact muscles, while the other subjects showed lower or almost unchanged slopes.3,4
A systematic analysis of EMG-force relations can provide valuable data enabling an informed interpretation of the mechanisms underlying neural and muscular pathological changes. Previous studies on EMG-force relations in stroke were mainly performed in proximal muscles.3,4 It is presently unknown whether or how the EMGforce relations may be altered in distal muscles. Accordingly, the objective of this study was to characterize whether the EMG-force relation observed in the affected first dorsal interosseous (FDI) muscle of hemiparetic stroke survivors deviates systematically from that observed in the contralateral muscle. We found that similar to previous reports with the proximal muscles,3,4 there were diverse changes in the slope of the EMGforce relations in paretic muscles compared with contralateral muscles, with significant increases and decreases being observed relative to the contralateral side. These findings, in combination
Top Stroke Rehabil 2013;20(6):537–544 © 2013 Thomas Land Publishers, Inc. www.strokejournal.com doi: 10.1310/tsr2006-537
537
538
TOPICS IN STROKE REHABILITATION/NOV-DEC 2013
with our previous simulation studies,5 suggest that there appear to be different types of processes (eg, motor unit control property changes, muscle fiber atrophy, spinal motoneuron degeneration, muscle fiber reinnervation, etc) at work post stroke that may impact the EMG-force relations and that may be present in a varying degree in any given stroke survivor. Methods Subjects
Fourteen subjects (8 males, 6 females; 58.3 ± 8.0 years) who sustained hemiparetic stroke as reported in Li et al6 participated in this study. All our stroke subjects were recruited from the Clinical Neuroscience Research Registry at the Rehabilitation Institute of Chicago (Chicago, IL). The study was approved by the Institutional Review Board of Northwestern University (Chicago, IL). All subjects gave their written consent before the experiment. A screening examination and clinical assessment were performed by a physical therapist. Among the 14 stroke subjects, the left limb was affected in 5 subjects, and the right limb was affected in 9 subjects. The duration between the stroke onset and the experiment time was 8.2 ± 6.7 years (range, 1.5-24.1 years). The 14 stroke subjects showed a hand portion of Chedoke-McMaster scale7 score of 4.8 ± 1.4 (range, 2-7) and an Upper Extremity Fugl-Meyer score8 of 45.7/66 ± 16.5/66 (range, 16/66 to 66/66). Experiments
Subjects were seated upright in a mobile Biodex chair (Biodex, Shirley, NY) with a standard 6 degrees-of-freedom load cell (ATI Inc, Apex, NC) setup used to accurately record the isometric contraction force of the FDI muscle during index finger abduction.6 The shoulders and waist of the subject were tightly strapped to the chair to limit trunk and shoulder movements. The forearm and wrist were mounted on a plastic platform inside a fiberglass cast. A ring-mount interface was used to strap the wrist in a partial pronation position. The proximal phalanx of the index finger was casted
and fixed to a small ring-mount interface attached to the load cell. This standard position served to minimize spurious force contributions from unrecorded muscles. Surface EMG signals were recorded from the FDI muscle using a 5-pin (0.5 mm in diameter) EMG sensor array (Delsys, Boston, MA). The array is arranged in a Laplace configuration and can generate 4 channels of differential EMG signals.9 A 16-channel Bagnoli amplifier (Delsys Inc, Boston, MA) with a gain of 1,000 was used to record surface EMG and force signals. Both surface EMG and force signals were sampled at 20 kHz using EMGWorks (Delsys, Boston, MA), with a bandpass filter (bandwidth 20-2,000 Hz) for surface EMG and a low pass filter (cutoff frequency 450 Hz) for force measurement, respectively. At the beginning of the experiment, the isometric maximum voluntary contraction (MVC) of each subject’s paretic FDI muscle for index finger abduction was determined. Target force was then assigned as a percentile of the paretic MVC force: from 20% to 80% of the paretic MVC at 10% MVC increments. With a matched force protocol, these target force levels were used as a standard for both paretic and contralateral muscles; for each subject, the same contraction levels as performed by the paretic muscle were used for the contralateral muscle. In each trial, the target force was displayed as a red trapezoid trajectory, and the trace of the voluntary contraction force was updated in realtime in a different color. Subjects were instructed to match the target force trajectory using their visual feedback and to maintain the force as stable as possible for up to 10 seconds. The protocol for each target force was performed at least twice. The order of target forces was randomized. Practice was given to help familiarize the subjects with the task and to make appropriate adjustments to the force. Substantial rest period between trials were provided to prevent potential fatigue during the experiment. Data analysis
Surface EMG and force signals were exported and analyzed offline in Matlab (MathWorks, Natick, MA). A notch filter was applied to surface
EMG-Force Relations During Isometric Contractionss
EMG to remove power line interference. For each trial, a constant force period for at least 1 second was selected by visual assessment of the force signal. Then the average force magnitudes in abduction (Fx) and flexion (Fy) were calculated. The force deviation from pure abduction was assessed by the inverse tangent of the ratio between flexion and abduction forces (Fy/Fx). The trials with a deviation of more than 30° were excluded for further analysis. For surface EMG, the signal to noise ratio (SNR) was calculated for each of the 4 recording channels. The average rectified value (ARV) of the surface EMG was calculated for the channels with SNR greater than 15 dB while the other channels were excluded from further analysis. The EMG-force relation was constructed using linear regression of the individual data set, and the slope of the linear fit was calculated. This was performed for all the surface EMG channels, and the average slope was determined for paretic and contralateral muscles of each stroke subject. Student t tests were used to examine whether the slope of the EMG-force relation reached statistical difference between paretic and contralateral muscles. Regression analysis was performed to examine the correlations between the ratio of EMGforce slope (between paretic and contralateral muscles) and clinical assessment. Statistical significance was defined as P < .05. Results Force measurement
The MVC force in abduction was recorded for both paretic and contralateral FDI muscles in each stroke subject, and their ratio was calculated. As we expected, the paretic muscle MVC force was systematically lower than that of the contralateral muscle. Across subjects, the average MVC force was 22.8 ± 13.3 N for the paretic FDI muscle and 40.2 ± 12.0 N for the contralateral muscle. The average ratio of the MVC force between paretic and contralateral muscles was 0.55 ± 0.25. Across all force levels, there was no significant difference in force deviation from pure abduction between the paretic and contralateral muscles (P > .2), although the average force deviation of paretic
539
muscle (11.1 ± 9.0°) was larger than contralateral muscle (7.3 ± 6.0°). EMG-force relation
Across all subjects, the average slope of the EMGforce relation was 6.19 ± 5.01 µV/N for the paretic muscle, which was slightly higher than the slope of 5.48 ± 4.23 µV/N for the contralateral muscle, but no significance was observed (P > .5). When examining individual stroke subjects, we observed clear increases and decreases in the slope of the EMG-force relation in paretic muscles compared with contralateral muscles. In 9 of the 14 subjects, we found a clear increase in the slope of the EMG-force relation in the paretic muscles. Figure 1 shows an example of the EMG signals recorded at similar contraction levels of the paretic and contralateral muscles of the same stroke subject, and the linear regression of the EMG-force relation, representative of these 9 subjects. In these 9 subjects, for a given force level, surface EMG amplitude was always higher for the paretic muscle than for the contralateral muscle. The average slope for these 9 subjects was 7.50 ± 5.15 µV/N for the paretic muscle, which was significantly higher than the slope of 4.41 ± 3.43 µV/N for the contralateral muscle (P < .05). The mean ratio of slope between paretic and contralateral muscles was 1.95 ± 0.95. The remaining 5 stroke subjects displayed a decreased slope of EMG-force relation for the paretic muscle when compared with the contralateral muscle. Figure 2 shows an example of the EMG signals recorded at similar contraction levels of the paretic and contralateral muscles of the same stroke subject, and the linear regression of the EMG-force relation, representative of these 5 subjects. In these 5 subjects, for a given force level, surface EMG amplitude was always smaller for the paretic muscle than for the contralateral muscle. The average slope for these 5 subjects was 3.83 ± 4.22 µV/N for the paretic muscle, which was significantly lower than the slope of 7.41 ± 5.23 µV/N for the contralateral muscle (P < .05). The mean ratio of slope between paretic and contralateral muscles was 0.45 ± 0.22. Comparison of the clinical data between the 9 stroke subjects (age, 58.7 ± 9.5 years; duration
540
TOPICS IN STROKE REHABILITATION/NOV-DEC 2013
(a)
(a) EMG
EMG
0.5 mV
0.5 mV
5s 5s
2N
Force
Force
2N
Paretic
Paretic (b)
Contralateral
Paretic Contralateral
Surface EMG (mV)
Surface EMG (mV)
Paretic Contralateral 0.08 0.06 0.04 0.02 0
0.08 0.06 0.04 0.02 0
0
1
2
3
4
5
Contralateral
(b)
6
7
8
9
Force (N)
0
5
10 15 Force (N)
20
Figure 1. Representative electromyogram (EMG) and force recordings of the 9 stroke subjects whose linear regression of the EMGforce relation demonstrates higher slope in paretic muscle than in contralateral muscle. (a) A comparison of the EMG signals from paretic and contralateral muscles at the similar force levels (dashed line represents the target force). In this example, the EMG from the paretic muscle was 0.075 mV at the force level of 3.9 N, and the EMG from the contralateral muscle was 0.027 mV at the contraction level of 4.3 N. (b) A linear regression of the EMG-force relation. The slope of the EMG-force relation was 0.0087 mV/N for paretic muscle (R2 = 0.98, P < .001) and 0.0052 mV/N (R2 = 0.80, P < .001) for the contralateral muscle.
Figure 2. Representative electromyogram (EMG) and force recordings of the 5 stroke subjects whose linear regression of the EMGforce relation demonstrates lower slope in paretic muscle than in contralateral muscle. (a) A comparison of the EMG signals from paretic and contralateral muscles at the similar force levels (dashed line represents the target force). In this example, the EMG from the paretic muscle was 0.010 mV at the force level of 5.2 N, and the EMG from the contralateral muscle was 0.024 mV at the contraction level of 5.6 N. (b) A linear regression of the EMG-force relation. The slope of the EMG-force relation was 0.0020 mV/N for paretic muscle (R2 = 0.42, P < .001) and 0.0035 mV/N (R2 = 0.82, P < .001) for the contralateral muscle.
of stroke, 6.5 ± 3.7 years; paretic to contralateral MVC ratio, 0.50 ± 0.27; Chedoke score, 5.1 ± 1.2; Fugl-Meyer score, 48.7 ± 16.9) demonstrating an increased slope of the EMG-force relation and the 5 subjects (age, 57.6 ± 5.1 years; duration of stroke, 11.2 ± 10.0 years; paretic to contralateral MVC ratio, 0.64 ± 0.22; Chedoke score, 4.2 ± 1.8; Fugl-Meyer score, 40.4 ± 16.1) demonstrating a
decreased slope did not show significant differences between the 2 groups (P > .3). Regression analysis did not verify strong correlations between the ratio of paretic and contralateral muscle EMGforce slopes and any clinical parameters (Chedoke score, R2 = 0.032, P > .8; Fugl-Meyer score, R2 = 0.08, P > .3; duration of stroke, R2 = 0.1, P > .7; muscle strength, R2 = 0.14, P > .2).
EMG-Force Relations During Isometric Contractionss
Effect of force normalization
The slope of the EMG-force relation was also re-examined by normalizing the force to each muscle’s maximum strength. Comparing the slope ratios of EMG-force relations between paretic and contralateral muscles, we found that the slope ratio was increased across all subjects after force normalization (because of paretic muscle weakness). For the 9 subjects demonstrating increased slope in paretic muscle (without force normalization), after the force normalization the average EMG-force slope was 8.05 ± 7.79 µV/10%MVC for paretic muscle and 1.64 ± 1.18 µV/10%MVC for contralateral muscle, respectively. For the 5 subjects demonstrating deceased slope in paretic muscle (without force normalization), after the force normalization the average EMG-force slope was 1.86 ± 1.27 µV/10%MVC for paretic muscle and 2.94 ± 1.97 µV/10%MVC for contralateral muscle, respectively. After normalizing the force to each muscle’s maximum strength, 13 subjects showed consistent slope ratios (between paretic and contralateral muscles) with respect to greater or less than unity, while 1 subject demonstrated a slope ratio change from 0.45 (without force normalization) to 1.24 (with force normalization).
Discussion We examined the alterations in the EMG-force relations of the FDI muscle after hemispheric stroke. FDI is a multifunctional muscle that generates torque about the second metacarpophalangeal (MCP) joint. The abduction force produced at the joint is generated largely by the FDI muscle, whereas the flexion force of the index finger is generated by a portion of the FDI muscle and by other synergists (eg, the flexor digitorum profundus and flexor digitorum superficialis muscles) as well as by other interosseus and lumbrical muscles. Thus, isometric abduction of the index finger was used in this study as a measure of FDI muscle activation. We chose to compare EMG-force relation of paretic muscles to contralateral muscles of the same stroke subjects rather than muscles from neurologically intact subjects. Considering the
541
similarities in muscle mechanical moment arms, muscle strength, and skeletal dimensions between the 2 limbs of the same subject, such a comparison may allow a more straightforward assessment of EMG-force relation characteristics than would be possible between muscles drawn from different subjects, who would have different limb sizes, work histories, and muscle usage patterns.4 We found that for individual stroke subjects, the slope of the EMG-force relation for paretic muscle showed diverse changes when compared with contralateral muscle, potentially due to different neural and/or muscular changes. In 9 of the 14 stroke subjects, we observed a clear increase in the slope in EMG-force relation on the paretic FDI muscle compared to the contralateral muscle. These may be due to different factors, including muscle fiber reinnervation, impairment of motor unit control properties (ie, reduction of motor unit firing rates and compression of motor unit recruitment threshold), and (nonselective) loss of motor units.4,10,11 In such cases, to reach a given force, additional relatively large motor units have to be recruited. With large action potentials from these relatively big motor units, the resultant surface EMG is higher. At high force levels, this effect (ie, increase in EMG) is especially evident, resulting in a clear slope increase of the EMG-force relation. For the other 5 stroke subjects, the slope of the EMG-force relation from the paretic muscles demonstrated a smaller value compared with the contralateral muscles. This could be an indication of either muscle fiber atrophy or a selective degeneration of the large and superficial motor units.12,13 In such cases, to reach a given force, the resultant surface EMG of the paretic muscle involves summation of a relatively large number of action potentials with small amplitudes compared with fewer action potentials with relatively large amplitudes in the contralateral muscle. Due to the effect of action potential cancellation, the size of action potentials plays a more dominant role in determining the overall EMG amplitude than do the number of action potentials.5 Therefore, the surface EMG amplitude is lower for paretic muscle at a given force. At higher force levels, this effect (ie, decrease in EMG) is more evident, generating a slope decrease in the EMG-force relation.
542
TOPICS IN STROKE REHABILITATION/NOV-DEC 2013
In general, our findings in distal muscles were consistent to those from proximal muscles. Although previous studies in proximal muscles seemed to emphasize the EMG force slope increase in paretic muscles compared with contralateral muscles, not all the examined stroke subjects demonstrated such a change. For example, Gemperline et al examined biceps brachii of 6 stroke subjects and reported that 3 subjects showed substantially elevated EMG-force slopes in paretic muscles compared with contralateral muscles; 1 showed lower slopes and the other 2 had a slope ratio close to unity.4 Tang and Rymer reported that in almost half of their examined hemiparetic subjects (8/17) the slope of the EMG-force relation, measured over the biceps brachialis and brachioradialis musle groups, was increased in the paretic limb compared with the unimpaired limb.3 The findings in proximal muscles from the previous studies and distal muscles from the current study both demonstrated a rather inconsistent picture of EMG-force slope changes after stroke, suggesting that there may be different types of processes (eg, motor unit control property changes, muscle fiber atrophy, spinal motoneuron degeneration, muscle fiber reinnervation, etc) at work post stroke that may impact the EMG-force relations.5 Furthermore, the experimentally observed EMGforce slope variations in hemiparetic muscles of stroke subjects could be the overall effects of many different factors. We also note that the slope increase demonstrated in this study was not as dramatic as reported in previous studies.3,4 Such an observation may be due to different processes present in varying degrees in proximal and distal muscles post stroke. It may also relate to the differences in recording characteristics of the EMG systems. Compared with the previous studies using a conventional surface EMG electrode, the current study recorded surface EMG data with a highly selective surface electrode, which was primarily designed or configured for surface EMG decomposition.9 Although we were unable to establish relations between the clinical data and EMG-force slope changes, examination of the EMG-force relation may have potential clinical implications for diagnosis and treatment of stroke sequelae. Using simulation of motoneuron pool activity
(muscle force and surface EMG) as a guide, experimental recording of EMG-force relation can provide valuable knowledge enabling an informed interpretation of the motor unit mechanisms underlying paretic muscle changes after stroke. This is particularly important given the fact that discrimination of single motor unit activity from surface EMG is a very difficult task. Monitoring of EMG-force relation in paretic muscles can help assess the nature of the neural or muscular deficits and the response of these deficits to medication and physical therapies. For example, muscle weakness induced by drop in motor unit firing rates may be treatable through appropriate training and/or drug interventions. The increase in motor unit firing rates induced by such interventions can be evaluated by monitoring the EMG-force slope changes. Finally, in addition to our focus on EMGforce slope changes, we observed that the linear fitting of the EMG-force relation for the paretic muscles tended to introduce an increased ordinate intercept. We observed such a trend in 10 of the 14 stroke subjects. Due to resolution limitation of the load cell and difficulty in holding very low constant force levels, the EMG signals were not routinely recorded at near zero forces. The relatively high intercept for paretic muscles implied production of increased EMG activity at the near-zero force levels compared with the contralateral muscles. This could be ascribed to relatively large background EMG activity, for example, sustained spontaneous motor unit spikes as observed in the resting paretic muscle. Alternatively, there might be a nonlinear recruitment sequence involved for the initial force generation of the paretic muscle. Acknowledgments Financial support/disclosures: This work was supported by the Brinson Stroke Foundation, the National Institute on Disability and Rehabilitation Research of the US Department of Education under grant H133F110033, the National Institutes of Health of the US Department of Health and Human Services under grant 2R24 HD050821, and the 1000 Talents Plan Special Program of China (Recruitment Program of Global Experts). Additional contributions: The authors thank Nina Suresh, PhD, for helping with data collection.
EMG-Force Relations During Isometric Contractionss
543
REFERENCES 1. Zhou P, Rymer WZ. Factors governing the form of the relation between muscle force and the EMG: A simulation study. J Neurophysiol. 2004;92(5):2878-2886. 2. Blaschak MJ, Powers RK, Rymer WZ. Disturbances of motor output in a cat hindlimb muscle after acute dorsal spinal hemisection. Exp Brain Res. 1988;71(2):377-387. 3. Tang A, Rymer WZ. Abnormal force–EMG relations in paretic limbs of hemiparetic human subjects. J Neurol Neurosurg Psychiatry. 1981;44(8): 690-698. 4. Gemperline JJ, Allen S, Walk D, Rymer WZ. Characteristics of motor unit discharge in subjects with hemiparesis. Muscle Nerve. 1995;18(10): 1101-1114. 5. Zhou P, Suresh NL, Rymer WZ. Model based sensitivity analysis of EMG-force relation with respect to motor unit properties: Applications to muscle paresis in stroke. Ann Biomed Eng. 2007;35(9):1521-1531. 6. Li X, Suresh A, Zhou P, Rymer W. Alterations in the spike amplitude distribution of the surface electromyogram post-stroke. IEEE Trans Biomed Eng. 2013;60(3):845-852.
7. Gowland C, Stratford P, Ward M, et al. Measuring physical impairment and disability with the Chedoke-McMaster Stroke Assessment. Stroke. 1993;24(1):58-63. 8. Gladstone, DJ, Danells CJ, Black SE. The Fugl-Meyer assessment of motor recovery after stroke: A critical review of its measurement properties. Neurorehabil Neural Repair. 2002;16:232-240. 9. De Luca CJ, Adam A, Wotiz R, Gilmore LD, Nawab SH. Decomposition of surface EMG signals. J Neurophysiol. 2006;96:1646-1657. 10. Lukacs M, Vecsei L, Beniczky S. Changes in muscle fiber density following a stroke. Clin Neurophysiol. 2009;120:1539-1542. 11. Kallenberg LA, Hermens HJ. Motor unit properties of biceps brachii in chronic stroke patients assessed with high-density surface EMG. Muscle Nerve. 2009;39:177-185. 12. Ryan AS, Dobrovolny CL, Smith GV, Silver KH, Macko RF. Hemiparetic muscle atrophy and increased intramuscular fat in stroke patients. Arch Phys Med Rehabil. 2002;83:1703-1707. 13. Lukacs M, Vecsei L, Beniczky S. Large motor units are selectively affected following a stroke. Clin Neurophysiol. 2008;119:2555-2558.
