- Email: [email protected]
The role of co-activation in strength and force modulation in the elbow of children with unilateral cerebral palsy
Journal of Electromyography and Kinesiology 22 (2012) 137–144
Contents lists available at SciVerse ScienceDirect
Journal of Electromyography and Kinesiology journal homepage: www.elsevier.com/locate/jelekin
The role of co-activation in strength and force modulation in the elbow of children with unilateral cerebral palsy Siri Merete Brændvik a,b,⇑, Karin Roeleveld b a b
Clinical Services, St. Olavs University Hospital, Trondheim, Norway Department of Human Movement Science, NTNU, Norway
a r t i c l e
i n f o
Article history: Received 23 February 2011 Received in revised form 7 September 2011 Accepted 4 October 2011
Keywords: Cerebral palsy Muscle weakness Force modulation Co-activation Spasticity
a b s t r a c t To study the role of coactivation in strength and force modulation in the elbow joint of children and adolescents with cerebral palsy (CP), we investigated the affected and contralateral arm of 21 persons (age 8– 18) with spastic unilateral CP in three tasks: maximal voluntary isokinetic concentric contraction and passive isokinetic movement during elbow flexion and extension, and sub-maximal isometric force tracing during elbow flexion. Elbow flexion–extension torque and surface electromyography (EMG) of the biceps brachii (BB) and triceps brachii (TB) muscles were recorded. During the maximal contractions, the affected arm was weaker, had decreased agonist and similar antagonist EMG amplitudes, and thus increased antagonist co-activation (% of maximal activity as agonist) during both elbow flexion and extension, with higher coactivation levels of the TB than the BB. During passive elbow extension, the BB of the affected arm showed increased resistance torque and indication of reflex, and thus spastic, activity. No difference between the two arms was found in the ability to modulate force, despite increased TB coactivation in the affected arm. The results indicate that coactivation plays a minor role in muscle weakness in CP, and does not limit force modulation. Moreover, spasticity seems particularly to increase coactivation in the muscle antagonistic to the spastic one, possibly in order to increase stability. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction Cerebral palsy (CP) is the most common cause of physical disability in childhood (Koman et al., 2004) with a prevalence around 2.1 per 1000 live births (Andersen et al., 2008). Primary neuromuscular impairments like muscle overactivity, muscle weakness, and decreased selective motor control, and secondary musculoskeletal problems like bony malformations and contractures are multiple motor signs that may contribute to the disability (Koman et al., 2004). The presence of muscle weakness in CP is well documented (Ikeda et al., 1998; Wiley and Damiano, 1998; Engsberg et al., 2000; Vaz et al., 2006), and several reports indicate that this weakness is an important contributor to the activity limitations seen in this group of patients (Damiano et al., 2001; Ross and Engsberg, 2007). There is a general agreement that both neural and muscular factors contribute to this weakness (Mockford and Caulton, 2010). Impaired motor unit recruitment is an important and welldocumented neural factor (Stackhouse et al., 2005; Rose and McGill, 2005), while decreased muscle volume and changes in ⇑ Corresponding author at: Clinical Services, St. Olav University Hospital, Olav Kyrres gt. 17, 7006 Trondheim, Norway. Fax: +47 72571310. E-mail addresses: [email protected], [email protected] (S.M. Brændvik). 1050-6411/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jelekin.2011.10.002
muscle fiber type composition also contribute to the reported muscle weakness (Lieber et al., 2004; Barrett and Lichtwark, 2010). Excessive coactivation is another neural factor potentially contributing to reduced strength in CP. Coactivation is the concurrent activation of agonist and antagonist muscles around a joint and excessive presence of this phenomenon could imply an inherent reciprocal action by the neuromuscular system and can be seen as a form of muscle overactivity (Sheean and Mcguire, 2009). Indeed, reduced strength and increased coactivation are reported in CP compared to typically developing peers (Damiano et al., 2000; Tedroff et al., 2008). Besides its possible strength-limiting effect, coactivation may also represent a motor control strategy in situations with a need of increased joint stability or improved movement accuracy (Gribble et al., 2003). Such stiffness modulation is assumed to vary with different tasks and environmental constraints (Darainy and Ostry, 2008; Silva et al., 2009). Therefore, coactivation may both enhance and put limitations on motor performance. Spasticity, another form of muscle overactivity associated with CP, may play a role in coactivation. One of the most frequently used definitions of spasticity defines it as ‘‘a motor disorder characterized by a velocity-dependent increase in tonic stretch reflexes with exaggerated tendon jerks, resulting from hyper-excitability of the stretch reflex’’ (Lance, 1980), and spasticity is typically measured
138
S.M. Brændvik, K. Roeleveld / Journal of Electromyography and Kinesiology 22 (2012) 137–144
as resistance to passive movement. However, it is generally accepted that the resistance felt during passive muscle stretch includes both passive and active muscular components that are changed in a spastic muscle (Foran et al., 2005). This makes it challenging to differentiate between what is ‘‘true’’ spasticity and what might be structural changes, by studying resistance only. Over activation during stretch can be investigated by electromyography (EMG). Over activation during stretch of a spastic muscle working as antagonist to a concentric contraction (over coactivation) may occur, either due to passive stretch of the spastic muscle, eliciting a tonic stretch reflex or due to a failure of reciprocal inhibition between the two opposing muscle groups (Sheean and Mcguire, 2009). The objective of this study was to investigate the role of coactivation in strength and force modulation in the upper extremities in children/adolescents with unilateral CP. Two questions are addressed: (1) Does antagonist coactivation in CP limit strength and force modulation, and (2) Does spasticity increase coactivation? Coactivation of flexors and extensors of the elbow joint were therefore studied during maximal voluntary concentric contractions and sub-maximal isometric force tracing in the elbow joint. Measures of passive flexion and extension of the elbow are included as this may give indications of the presence of spasticity. In addition to differences between the affected and non-affected arm, differences between flexion and extension were studied. Although spasticity in elbow extensors cannot be excluded, differences between flexion and extension were expected since, clinically, spasticity is mainly observed in the elbow flexors of this patient group (Mayer, 1997). Moreover, the correlation between variables was investigated for the affected and non-affected arms separately. 2. Methods 2.1. Participants Twenty-one children/adolescents (9 males and 12 females), median age 13 years (range 8–18), diagnosed with unilateral CP participated in this study. They all had a functional use of the affected upper extremity corresponding to level III or better on the Manual Ability Classification System (Eliasson et al., 2006) and were able to take instructions verbally. Median active extension deficit in the elbow of the affected arm was 5° (range 0–45). Exclusion criteria were treatment with botulinum toxin A in the last 6 months and/or surgery on the upper extremity in the last 2 years. The study was approved by the Regional Committee for Medical Research Ethics and carried out according to the Declaration of Helsinki. A written informed consent was obtained from the parents and participants before participation. 2.2. Experimental setup Three different tasks were performed in the elbow joint: passive movement, isokinetic concentric maximal voluntary contraction (MVC) and sub-maximal force tracing. All 21 participants performed the first two tasks, 9 participants performed the submaximal force tracing task. A stationary dynamometer (Biodex System 3 Pro) was used to control the movements and measure torque, angle and velocity in all three tasks. The analogue signals were then digitised (sample frequency of 1000 Hz) through a National Instruments card (DAQCard-6036E) by EMG WorksÒ 3.1 (Delsys Inc.). The measurements were first carried out on the affected and then on the non-affected arm. Both the passive movement and the isokinetic MVCs were obtained using the passive mode in the Biodex. Sub-maximal force tracing was recorded using the isometric mode.
Surface EMG (SEMG) was recorded from the M. biceps brachii (BB) and the lateral head of M. triceps brachii (TB), from both the affected and the contralateral arm. Bipolar (1 cm spacing), singledifferential polycarbonate electrodes with a detection area of 10 mm2 were applied on the skin. Preparations and electrode placement above the BB and TB followed the recommendations of the Seniam project (Surface ElectroMyoGraphy for the NonInvasive Assessment of Muscles) (Hermens et al., 2000) and the reference electrode was placed on C7. The signals were amplified (gain of 1000) and band pass filtered (20–450 Hz) with a Bagnoli 16-channel SEMG system (Delsys Inc.) prior to sampling (1000 Hz). The dynamometer and SEMG signals were collected by the same system and therefore synchronized. 2.3. Procedure Details of the procedure are described in a recent paper by our research group (Brændvik et al., 2010). In brief, the participants were positioned in the Biodex chair, with the forearm in neutral position and the shoulder slightly abducted. First, the participants performed three passive trials in flexion and extension at two different velocities; 10° and 180°/sec, with the instruction to relax and just let the machine move the arm. Second, the participants performed three isokinetic MVCs at a velocity of 60°/sec. One test trial was allowed in order to insure that the task was performed correctly. There was a one-minute break between each trial. Finally, the participants were instructed to follow a target force by isometrically flexing the elbow. This task was only performed by nine subjects. For target and performance force normalisation purposes, the participants performed three isometric maximal voluntary elbow flexion contractions, whereof the highest peak torque was used as maximal isometric torque. The target signal consisted of a 15 s linear increase followed by a 15 s linear decrease between 0% and 40% of maximal isometric torque. Online visual feedback of both the target torque and the actual generated torque was provided on a monitor (full screen) placed at a convenient distance in front of the participant. The elbow angle was set to 60° (from full extension). The force tracing task was performed twice with a 10 s break between each trial. Two test trials were allowed prior to the measurements to ensure that the task was understood. 2.4. Data analysis Data analysis was carried out using Matlab (The Mathworks Inc.), version 7.0. Prior to further analyses, torque signals were low-pass filtered with a cut-off frequency of 6 Hz. The torque recorded during the passive movements at 10°/sec was used to obtain the weight of the arm and correct the other torque measures for arm weight. Peak torque and peak resistance torque were obtained during the isokinetic flexion and extension phases of all MVC and passive trials. In the force tracing task, both target and applied torque were presented as percentage of the highest torque generated during the three isometric MVCs. The first and last second of both the increasing and decreasing torque phases were left out from further analyses. For the remaining increasing and decreasing periods, the root mean square (rms) of the difference between the applied torque and the target torque was calculated (DTrms). The isokinetic flexion and extension MVC trials with the highest peak torques were used to reflect flexion and extension strength, respectively. The passive flexion and extension trials with the lowest peak resistance torque were used to reflect muscle tone. Furthermore, the force tracing trial with the lowest DTrms (increasing and decreasing) was used to reflect force modulation ability.
S.M. Brændvik, K. Roeleveld / Journal of Electromyography and Kinesiology 22 (2012) 137–144
A digital band-pass filter of 20–450 Hz was applied to the raw SEMG signal prior to further data analysis. Subsequently, RMS of the SEMG signals (BBrms and TBrms) was calculated over the whole isokinetic period of the best MVC flexion and extension trials and the best passive flexion and extension trials. BBrms and TBrms were also obtained during the best force increasing and decreasing periods of the force tracing task. SEMG noise levels, defined as the lowest BBrms and TBrms obtained during rest in periods of 500 ms were subtracted from all other BBrms and TBrms parameters. Median noise level ranged from 1.37 to 7.1 (lV) in BB and TB in the two arms. In the force tracing task, there was no difference between the increasing and decreasing periods for DTrms, BBrms and TBrms. Therefore, for these variables, the means of the two periods were used for further analysis. In order to study antagonist coactivation, the BBrms during isokinetic MVC extension, and the TBrms during MVC flexion and the force tracing task, were normalised to the BBrms during MVC flexion (BBco) and TBrms during MVC extension (TBco), respectively. BBrms during passive extension and TBrms during passive flexion were used as a measure of stretch response. Velocity dependent increase in resistance torque and the corresponding BBrms during passive elbow flexion and extension, was used as measures of spasticity, and was calculated according to the following formulas:
Velocity dependent increase in resistance ð%Þ ! Resistance Torque 180 =sec 1 100 ¼ Resistance Torque 10 =sec Velocity dependent increase in BB amplitude ð%Þ ! BBrms 180 =sec ¼ 1 100 BBrms 10 =sec
ð1Þ
ð2Þ
Moreover, as an indication of reflex activity during passive stretch, the within muscle ratio of TBrms or BBrms during passive stretch and during passive shortening at 180°/sec was determined. 2.5. Statistics Statistical analysis was carried out using PASW statistics, version 17. Non-parametric statistics was used due to small sample size and non-normal distribution of the data. Wilcoxon signed rank test was used to detect difference between affected arm (AA) and contralateral arm (CA). Spearman’s rho (rs) was used to assess relation between variables. A sample size of 21 and 9, will need to have correlation coefficients above .37 and .58, respectively, in order to reach significance (Fisher, 1995). An effect size similar to the standard deviation can be detected with a power of .99 and .75, respectively when paired two-tailed statistics with a probability of 0.05 are applied with a sample size of 21 and 9 (Lenth, 2006-9). 3. Results 3.1. MVCs Fig. 1 shows the results of the MVC task. The affected arm, AA, had significantly lower peak torque than the contralateral arm, CA, in both flexion and extension (p < .001 for both) (Fig. 1a). Both the AA and CA had significantly higher peak torque in elbow extension than in elbow flexion (p = .004 for AA and .001 for CA). There was no difference between AA and CA in relative flexion–extension peak torque (p = .848). During the isokinetic MVCs, the AA had significantly lower agonist SEMG activity compared to CA, indicated by lower BBrms during flexion and TBrms during extension
139
(p < .002 for both) (Fig. 1b). Although the average TBrms and BBrms as antagonists during MVCs were slightly higher in the CA compared to the AA, there was no significant difference between the arms (p = .517 for TBrms during elbow flexion and p = .973 for BBrms during elbow extension) (Fig. 1c). The antagonist coactivation during the MVCs, were significantly higher in the AA than in CA (Fig. 1d) both for BBco during elbow extension (p = .026) and TBco during elbow flexion (p = .002). In addition, this TBco was significantly higher than BBco in both the AA (p = .004) and CA (p < .001), but the relative TBco–BBco was not different between the two arms (p = .272). 3.2. Passive movements Fig. 2 shows results from passive elbow flexion and extension. A significantly higher extension resistance torque was found in the AA compared to CA for both 10 and 180°/sec (p < .001 for both), while no difference was found in flexion resistance torque (p = .179 for 10°/sec and p = .715 for 180°/sec). A velocity dependent increase in resistance torque was found for both flexion and extension in both arms (p < .001), but there was no difference between flexion and extension, or between the arms in this respect (Fig. 2a). As concerns the corresponding muscle stretch response, no difference between the arms in absolute muscle activity was found in flexion or extension at 180°/sec (Fig. 2b). In contrast to resistance torque, a velocity dependent increase in muscle activity was found for the AA during both flexion (p = .016) and extension (p = .001) compared to the CA (Fig. 2c). Moreover, the within muscle ratio BB extension/BB flexion in AA was above one, showing that BB activity was higher when stretched than when shortened during passive movement (180°/sec) (Fig. 2d). 3.3. Force tracing Only nine of the 21 participants completed this task. In the force tracing task, there was no significant difference in DTrms (p = .164), BBrms (p = .139) or TBrms (p = .250) between the AA and CA (Fig. 3a and b). However, TBco was significantly higher in the AA than in CA (p = .038) with a median of 15% of TBrms during MVC extension in AA and 9% in CA (Fig. 3c), while the median BBrms during the force tracing task was about 21% of BBrms during MVC flexion in both AA and CA. 3.4. Correlations Table 1 shows correlations between MVC, coactivation and passive movements. In both arms, there was a positive correlation between peak torque and agonist SEMG RMS during both maximal voluntary elbow flexion and extension. The slope of this relation was also similar in both arms. There was also a positive correlation between peak torque and antagonist SEMG RMS in CA during both flexion and extension and in AA during extension. A negative correlation was found between peak flexion torque and TBco in the AA, while no significant correlations were found between these variables in AA during extension and in CA neither during flexion nor extension. In the AA, but not the CA, a negative correlation was found between peak extension and flexion torque and passive elbow extension resistance torque. None of the strength variables correlated significantly with passive elbow flexion resistance torque. In the AA, a negative correlation was found between BBrms during passive extension and extension peak torque, and there was a trend between BBrms during passive extension and flexion peak torque. None of the strength variables correlated significantly with the TBrms during passive flexion.
140
S.M. Brændvik, K. Roeleveld / Journal of Electromyography and Kinesiology 22 (2012) 137–144
Fig. 1. MVC results. Median (95% confidence interval of mean) for (a) elbow flexion and extension peak torque, (b) agonist and (c) antagonist rms (root mean square), and (d) antagonist co-activation in affected arm (AA) and contralateral arm (CA). TB: triceps brachii, BB: biceps brachii, Nm: Newtonmeter, mV: millievolt. ⁄Significant difference between arms at .05 level.
No significant correlations were found between DTrms, TBco and extension resistance torque in the force tracing task. However, there was a trend for a negative correlation (rs = .55, p = .062) between DTrms and TBco in the CA; increased coactivation is associated with decreased difference between the target and applied force.
4. Discussion In the present study, the role of coactivation in maximal force generation and sub-maximal force tracing was investigated in both the affected and the contralateral elbow in children/adolescents with spastic unilateral CP. The affected arm was found to be weaker and to have more coactivation than the contralateral arm in both flexion and extension. However, the absolute antagonist activity and therefore probably the braking effect of the antagonist, was not different between the arms during maximal voluntary contraction. Indication of reflex activity was found in BB, but not in TB of the affected arm, while the TB had higher coactivation than BB in both arms. Both arms performed equally well in the force tracing task despite increased coactivation in the affected arm.
The current findings indicate that antagonist coactivation in the elbow joint plays a minor role in muscle weakness in CP, and increased antagonist coactivation is likely to increase stability in a movement involving a spastic muscle in CP. 4.1. The role of coactivation in strength Coactivation, the simultaneous activation of agonist and antagonist muscles, implies an inherent reciprocal action by the neuromuscular system, potentially hampering force production capacity. Our results are in line with studies reporting decreased strength and increased coactivation in patients with CP (Ikeda et al., 1998; Elder et al., 2003; Tedroff et al., 2008). Moreover, our participants were weaker in elbow flexion than extension, combined with higher coactivation in TB than in BB, but there was no difference between the affected and the contralateral arm according to this. However, our findings did also include a significant lower agonist EMG activity in the affected arm compared to the contralateral arm during both flexion and extension. As coactivation in our study is operationalised as percentage of maximal agonist contraction, this will result in a relatively higher coactivation level in the affected arm compared to the contralateral arm. Thus, the increased
S.M. Brændvik, K. Roeleveld / Journal of Electromyography and Kinesiology 22 (2012) 137–144
141
Fig. 2. Passive movement results. Median (95% confidence interval of mean) for (a) velocity dependent change in resistance torque (b) muscle stretch response rms (root mean square) at a velocity of 180°/sec, (c) velocity dependent change in muscle stretch response and (d) ratio BB extension/BBflexion and TB flexion/TB extension during passive elbow flexion and extension at 180°/sec. AA: affected arm, CA: contralateral arm. TB: triceps brachii, BB: biceps brachii, Nm: Newtonmeter, mV: millievolt. ⁄Significant difference between arms at .05 level.
coactivation reflects agonist weakness rather than being a direct cause. Moreover, the absolute antagonist activity was not different between the arms, indicating that a possible braking effect was not higher in the affected than the contralateral arm. Neither did we find any systematic relation between strength and coactivation, except for a significant negative relation between elbow flexion strength and TB coactivation in the affected arm. In all, this indicates that the cause of the muscle weakness lies in the agonist itself, and this would be in agreement with studies showing that persons with CP have a reduced ability to use the full potential of the muscle (Rose et al., 2005; Stackhouse et al., 2005). We hypothesised that spastic activity could increase coactivation. Therefore, we evaluated spasticity. In the current study, the affected arm had an increased muscle tone in the elbow flexors in both slow and fast passive extension compared to the contralateral arm. However, the velocity dependent increase in muscle tone was not different between the arms or between flexors and extensors (Fig. 2a). Velocity dependent increase in resistance to movement is not exclusive for spasticity (Pandyan et al., 2005). Also the viscoelastic properties of the muscle are velocity dependent
(Dietz, 1995), and this could explain the lack of difference in velocity dependent increase in muscle tone. Another notion to be made is the large variability in velocity dependent response found in passive extension of the contralateral arm which also could attribute to the lack of difference. However, all these results indicate that non-reflexive changes may cause the increased mechanical stiffness during extension in the affected arm. The literature supports such an assumption as secondary structural and functional changes in spastic muscles are reported although the primary lesion is neural (Lieber et al., 2004). However, the SEMG results indicate a neural mechanism. There was a velocity dependent increase in muscle stretch response during both passive extension and flexion in the affected arm compared to the contra lateral one, while only the elbow flexors of the affected arm showed more activity during passive stretching than shortening. This indicates the presence of spasticity, as defined by Lance, in the affected BB and a possibly non-reflexive increased TB activation during stretch. The inverse relation between BB passive stretch response and extension strength, and between flexor muscle tone and extension strength, could easily lead to a preliminary conclusion that
142
S.M. Brændvik, K. Roeleveld / Journal of Electromyography and Kinesiology 22 (2012) 137–144
Fig. 3. Force tracing results. Median (95% confidence interval of mean) of (a) Dtorque rms (the root mean square of the difference between the applied force and the target force), (b) SEMG rms in submaximal force modulation task (between 0% and 40% of maximal voluntary contraction) and (c) TB co-activation. BB: biceps brachii, TB: triceps brachii, AA: affected arm, CA: contralateral arm. ⁄Significant difference between arms at .05 level.
Table 1 Correlation (Spearman’s rho) between peak elbow extension/flexion torque in the affected (AA) and contralateral (CA) arm, and agonist/antagonist root mean square (rms) of biceps brachii (BB) and triceps brachi (TB), antagonist co-activation (as% of MVC), extension/flexion resistance torque and antagonist stretch response during the passive flexion and extension.
Agonist rms Antagonist rms Antagonist co-activation Extension resistance torque Flexion resistance torque BBrms during passive extension TBrms during passive flexion * **
Peak extension torque AA
Peak extension torque CA
Peak flexion torque AA
Peak flexion torque CA
.68 (p < .001)** .436 (p = .024)* .20(p = .192) .71 (p < .001)** .08 (p = .358) .38 (p = .045)* .25 (p = .140)
.59 (p = .003)** .396 (p = .038)* .30 (p = .098) .28 (p = .108) .003 (p = .496) .126 (p = .293) .087 (p = .345)
.78 (p < .001)** .264 (.124) .43 (p = .027)* .52 (p = .008)** .26 (p = .124) .30 (p = .089) .24 (p = .151)
.53 (p = .007)** .596 (p = .002)** .05 (p = .414) .105 (p = .325) .245 (p = .142) .10 (p = .341) .08 (p = .358)
Significant at .05 level. Significant at .01 level.
spasticity induced coactivation may cause the reduced extension strength in the affected arm. However, the BB coactivation during the strength task was not especially high and it is at least not likely that it caused a higher braking effect in the affected arm than in the non-affected arm and in flexion than in extension. Moreover, a trend for an inverse relation between the BB stretch response
and flexion strength, and an inverse relation between flexor muscle tone (and not extensor muscle tone) and flexion strength indicates that the impaired function in the flexor muscles reflects its force generating capabilities. One important notion according to the effect of spasticity has to be made; while the muscle tone measurements are carried out on muscles at rest, the strength
S.M. Brændvik, K. Roeleveld / Journal of Electromyography and Kinesiology 22 (2012) 137–144
measurements are voluntary actions of an active motor system. Thus, we do not know how the response during passive conditions reflects active situations. Moreover, if spasticity as the involuntary activation of a muscle during stretch would be similar during passive and active contractions, its contribution to force production would be negligible, since the antagonist EMG amplitude during passive movements (Fig. 2b) is much smaller than during maximal voluntary contractions (Fig. 1c). Another point to be made is the difference in velocity between the two tasks; 180°/sec versus 60°/sec, which may affect the muscle response. In contrast to BB coactivation, no indication of spasticity was found in TB coactivation. The significant negative relation between elbow flexion strength and TB coactivation in the affected arm, combined with a relative high coactivation and antagonist activation during concentric MVCs compared to BB during equivalent tasks, suggests that TB coactivation reflects a different aspect of motor performance than BB coactivation. TB coactivation may be part of a motor control strategy to improve motor control in a weak and spastic BB. That would be in agreement with increased coactivation in order to increase joint stability or improve movement accuracy in other situations (Gribble et al., 2003) and in early phases of learning (Darainy and Ostry, 2008). Specifically, it is hypothesized that the antagonistic muscles in coactivation might act as a kind of mechanical filter by suppressing the effects of force variability in the agonist (Selen et al., 2005). 4.2. The role of coactivation in sub maximal force tracing No difference in performance between the arms was found in the force tracing task. This indicates that the affected arm performs equally well as the contralateral arm providing it is allowed to act relative to its capacity. This is in line with Smits-Engelsman et al (2005) who studied force control ability in the hand of hemiplegic children, also scaled to MVC of the affected hand. Despite equal performance, higher TB coactivation was found in the affected arm. However; we did not detect any relation between performance and coactivation. This is in contrast to Bandholm et al (2009) who reported a relation between ankle dorsiflexion torque steadiness and plantarflexion coactivation level. This could be explained by differences in force levels; their submaximal level was scaled to torque in typically developing children. Thus, the significance of coactivation as means of motor control may increase with increasing force levels. This assumption is supported by increasing variability in force control ability at higher force levels (SmitsEngelsman et al., 2005) and also corresponds with our results from the maximal strength task. The small sample size in the force tracing task (n = 9) could affect the power of these results. However, since significant and nearly significant differences were found between the two arms, it is not likely that a larger sample size would have changed the result regarding the relation between performance and coactivation.
143
lower extremities (Damiano et al., 2001). Although no indication of reflex activity in the triceps brachii was found in the current study, a velocity dependent increase in muscle stretch response was indeed found. Maximal force production in children with CP is complicated by the reduced ability to fully activate the muscles according to their potential. Thus, maximal EMG activity is underestimated. This has to be accounted for in the interpretations, but will not change our conclusions since they are based on relative comparisons and not on an absolute estimation of force. Moreover, skin fold thickness is a factor that may infer with inter-individual variance in SEMG amplitude (Nordander et al., 2003) and information about this was not obtained in this study sample. However, the normalization procedure reduces this effect and moreover, the slope of the relation between strength and agonist amplitude of SEMG of the two arms was equal and thus gave no indications of such an effect. The power calculations made in this study are based on parametric assumptions, and the use of non-parametric statistics reduces this power. However, the calculated power was high, .99 and .75, so it is not likely that this will weaken the results. 5. Concluding remarks The results of this study do not support that antagonist coactivation is a significant contributor to the muscle weakness observed in children with CP, or limits the ability to modulate force. Rather, the results indicate that antagonist coactivation reflects different aspects of impaired motor control. More specific, coactivation of the BB during maximal elbow extension may have a spastic component, however relatively low, while coactivation of the TB during elbow flexion most likely is a motor control strategy in tuning impaired BB function during both maximal and submaximal tasks. These results need to be verified for children with different MACS levels and compared to and an age and gender matched reference group. Additional use of electrostimulation will give a direct measure of the force generating capability in the muscle, and thus a more accurate measure of coactivation. However, such a method may be problematic in the upper arm in children. Moreover, intervention studies addressing spasticity treatment, i.e. Botulinum toxin A and strength training, would give more insight into the role of coactivation with respect to the activity limitations in CP. Acknowledgments We would like to thank the children for their participation in the study, Ann-Kristin Gunnes Elvrum for her contribution during study planning and data acquisition, and Rannei Sæther and Torarin Lamvik for providing study participants. We would also like to thank Paul Jarle Mork for valuable comments on the manuscript. This study was funded by the Norwegian Directorate of Health and Unimed Innovation AS.
4.3. Limitations of the study
References
Inferences in this study are based on differences between the affected and the contralateral arm in unilateral CP. Statistically, this increases the power of the results and may also be considered a conservative approach. However, it is questionable to what extent the contralateral arm in unilateral CP also is affected (Wiley and Damiano, 1998; Elder et al., 2003; Smits-Engelsman et al., 2005; Steenbergen and Meulenbroek, 2006). Moreover, inferences are made based on differences in resistance torque and muscle stretch response, between flexors and extensors. Although clinical focus with respect to spasticity in the elbow joint mainly is on the flexors, indication of spasticity is also experienced in the opposing muscle group. Indeed, this is also reported with respect to the
Andersen GL, Irgens LM, Haagaas I, Skranes JS, Meberg AE, Vik T. Cerebral palsy in Norway: prevalence, subtypes and severity. Eur J Paediat Neurol 2008;12(1): 4–13. Bandholm T, Rose MH, Sløk R, Sonne-Holm S, Jensen BR. Ankle torque steadiness is related to muscle activation variability and coactivation in children with cerebral palsy. Muscle Nerve 2009;40(3):402–10. Barrett RS, Lichtwark GA. Gross muscle morphology and structure in spastic cerebral palsy: a systematic review. Dev Med Child Neurol. 2010;52(9): 794–804. Brændvik SM, Elvrum AKG, Vereijken B, Roeleveld K. Relationship between neuromuscular body functions and upper extremity activity in children with cerebral palsy. Dev Med Child Neurol 2010;52(2):e29–34. Damiano DL, Martellotta TL, Sullivan DJ, Granata KP, Abel MF. Muscle force production and functional performance in spastic cerebral palsy: Relationship of cocontraction. Arch Phys Med Rehabil 2000;81(7):895–900 [doi: 10.1053/ apmr.2000.5579].
144
S.M. Brændvik, K. Roeleveld / Journal of Electromyography and Kinesiology 22 (2012) 137–144
Damiano DL, Quinlivan J, Owen BF, Shaffrey M, Abel MF. Spasticity versus strength in cerebral palsy: relationships among involuntary resistance, voluntary torque, and motor function. Eur J Neurol 2001;8(s5):40–9. Darainy M, Ostry D. Muscle cocontraction following dynamics learning. Exp Brain Res 2008;190(2):153–63. Dietz V, Berger W. Cerebral palsy and muscle transformation. Dev Med Child Neurol 1995;37:180–4. Elder GCB, Stewart G, Cook K, Weir D, Marshall A, Leahey L. Contributing factors to muscle weakness in children with cerebral palsy. Dev Med Child Neurol 2003;45(8):542–50. Eliasson AC, Krumlinde-Sundholm L, Rösblad B, Beckung E, Arner M, Öhrvall A-M, et al. The Manual Ability Classification System (MACS) for children with cerebral palsy scale development and evidence of validity and reliability. Dev Med Child Neurol 2006;48(7):549–54. Engsberg JR, Ross SA, Olree KS, Park TS. Ankle spasticity and strength in children with spastic diplegic cerebral palsy. Dev Med Child Neurol 2000;42(1):42–7. Fisher RAYF. Statistical Tables for Biological, Agricultural, and Medical Research. Essex, United Kingdom: Pearson Education; 1995. Foran JRH, Steinman S, Barash I, Chambers HG, Lieber RL. Structural and mechanical alterations in spastic skeletal muscle. Dev Med Child Neurol 2005;47(10): 713–7. Gribble PL, Mullin LI, Cothros N, Mattar A. Role of cocontraction in arm movement accuracy. J Neurophysiol 2003;89(5):2396–405. Hermens HJ, Freriks B, Disselhorst-Klug C, Rau G. Development of recommendations for SEMG sensors and sensor placement procedures. J Electromyogr Kinesiol 2000;10(5):361–74. Ikeda AJAM, Granata KP, Damiano DI. Quantification of cocontraction in spastic cerebral palsy. Electromyogr Clin Neurophysiol 1998;38(8):497–504. Koman LA, Smith BP, Shilt JS. Cerebral palsy. The Lancet 2004;363(9421):1619–31. Lance JW. The control of muscle tone, reflexes, and movement: Robert Wartenberg lecture. Neurology 1980;30(12):1303–13. Lenth, RV. (2006-9). Java Applets for Power and Sample Size [Computer software]. Retrieved 08, 31, 2011 from http://www.stat.uiowa.edu/~rlenth/Power. Lieber RL, Steinman S, Barash IA, Chambers H. Structural and functional changes in spastic skeletal muscle. Muscle Nerve 2004;29(5):615–27. Mayer NH, Esquenazi A, Childers MK. Common patterns of clinical motor dysfunction. Muscle Nerve 1997(Suppl. 6):S21–35. Mockford M, Caulton JM. The pathophysiological basis of weakness in children with cerebral palsy. Pediatr Phys Ther 2010;20(4):318–33. Nordander C, Willner J, Hansson GÅ, Larsson B, Unge J, Granquist L, et al. Influence of the subcutaneous fat layer, as measured by ultrasound, skinfold calipers and BMI, on the EMG amplitude. Eur J Appl Physiol 2003;89(6):514–9. Pandyan AD, Gregoric M, Barnes MP, Wood D, Vanwijck F, Burridge J, et al. Spasticity: clinical perceptions, neurological realities and meaningful measurement. Disabil Rehabil 2005;27(1):2–6. Rose J, Mcgill KC. Neuromuscular activation and motor-unit firing characteristics in cerebral palsy. Dev Med Child Neurol 2005;47(5):329–36. Ross SA, Engsberg JR. Relationships between spasticity, strength, gait, and the GMFM-66 in persons with spastic diplegia cerebral Palsy. Arch Phys Med Rehabil 2007;88(9):1114–20. Selen L, Beek P, Dieën J. Can co-activation reduce kinematic variability? A simulation study. Biol Cybern 2005;93(5):373–81. Sheean G, Mcguire JR. Spastic hypertonia and movement disorders: pathophysiology, clinical presentation, and quantification. PM&R 2009;1(9): 827–33.
Silva PL, Fonseca ST, Ocarino JM, Gonã§Alves P, Mancini MC. Contributions of cocontraction and eccentric activity to stiffness regulation. J Motor Behav 2009;41(3):207–28. Smits-Engelsman BC, Rameckers EA, Duysens J. Muscle force generation and force control of finger movements in children with spastic hemiplegia during isometric tasks. Dev Med Child Neurol 2005;47(5):337–42. Stackhouse SK, Binder-Macleod SA, Lee SCK. Voluntary muscle activation, contractile properties, and fatigability in children with and without cerebral palsy. Muscle Nerve 2005;31(5):594–601. Steenbergen B, Meulenbroek RGJ. Deviations in upper-limb function of the lessaffected side in congenital hemiparesis. Neuropsychologia 2006;44(12): 2296–307 [doi: 10.1016/j.neuropsychologia.2006.05.016]. Tedroff K, Knutson LM, Soderberg GL. Co-activity during maximum voluntary contraction: a study of four lower-extremity muscles in children with and without cerebral palsy. Dev Med Child Neurol 2008;50(5):377–81. Vaz DV, Cotta M, Fonseca ST, De Melo Pertence AE. Muscle stiffness and strength and their relation to hand function in children with hemiplegic cerebral palsy. Dev Med Child Neurol 2006;48(9):728–33. Wiley ME, Damiano DL. Lower-extremity strength profiles in spastic cerebral palsy. Dev Med Child Neurol 1998;40(2):100–7.
Siri Merete Brændvik was born in Trondheim, Norway in 1961. She graduated as PT from Bergen University College in 1985 and her clinical work has been in the field of neuro-/orthopedics. In 2003, she received her M.Sc. degree in Human Movement Science at the Norwegian University of Science and Technology (NTNU), Norway. She is now a Ph.D. student at Department of Human Movement Science, NTNU. Her research interest is in the field of motor control, especially motor disorders associated with cerebral palsy (CP). Her Ph.D. thesis (expected 2012) addresses the interrelation between neuromuscular impairments in CP and how these are related to activity limitations.
Karin Roeleveld received the M.Sc. degree in human movement sciences from the Vrije Universiteit in Amsterdam, The Netherlands in 1992. Thereafter, she joined the University Hospital in Nijmegen and received her Ph.D. degree in 1997 on the fundaments of surface electromyography using multi electrode surface electromyography. Thereafter she spent two years working as a post doc at the Friedrich-Schiller-University Jena, Germany. Since 1998 she is working at the human movement sciences programme, at the Norwegian University of Science and Technology (NTNU), Norway, where she is currently working as professor. Her research interests are related to neuromuscular function and how this adapts with activity, disease and age.
Contents lists available at SciVerse ScienceDirect
Journal of Electromyography and Kinesiology journal homepage: www.elsevier.com/locate/jelekin
The role of co-activation in strength and force modulation in the elbow of children with unilateral cerebral palsy Siri Merete Brændvik a,b,⇑, Karin Roeleveld b a b
Clinical Services, St. Olavs University Hospital, Trondheim, Norway Department of Human Movement Science, NTNU, Norway
a r t i c l e
i n f o
Article history: Received 23 February 2011 Received in revised form 7 September 2011 Accepted 4 October 2011
Keywords: Cerebral palsy Muscle weakness Force modulation Co-activation Spasticity
a b s t r a c t To study the role of coactivation in strength and force modulation in the elbow joint of children and adolescents with cerebral palsy (CP), we investigated the affected and contralateral arm of 21 persons (age 8– 18) with spastic unilateral CP in three tasks: maximal voluntary isokinetic concentric contraction and passive isokinetic movement during elbow flexion and extension, and sub-maximal isometric force tracing during elbow flexion. Elbow flexion–extension torque and surface electromyography (EMG) of the biceps brachii (BB) and triceps brachii (TB) muscles were recorded. During the maximal contractions, the affected arm was weaker, had decreased agonist and similar antagonist EMG amplitudes, and thus increased antagonist co-activation (% of maximal activity as agonist) during both elbow flexion and extension, with higher coactivation levels of the TB than the BB. During passive elbow extension, the BB of the affected arm showed increased resistance torque and indication of reflex, and thus spastic, activity. No difference between the two arms was found in the ability to modulate force, despite increased TB coactivation in the affected arm. The results indicate that coactivation plays a minor role in muscle weakness in CP, and does not limit force modulation. Moreover, spasticity seems particularly to increase coactivation in the muscle antagonistic to the spastic one, possibly in order to increase stability. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction Cerebral palsy (CP) is the most common cause of physical disability in childhood (Koman et al., 2004) with a prevalence around 2.1 per 1000 live births (Andersen et al., 2008). Primary neuromuscular impairments like muscle overactivity, muscle weakness, and decreased selective motor control, and secondary musculoskeletal problems like bony malformations and contractures are multiple motor signs that may contribute to the disability (Koman et al., 2004). The presence of muscle weakness in CP is well documented (Ikeda et al., 1998; Wiley and Damiano, 1998; Engsberg et al., 2000; Vaz et al., 2006), and several reports indicate that this weakness is an important contributor to the activity limitations seen in this group of patients (Damiano et al., 2001; Ross and Engsberg, 2007). There is a general agreement that both neural and muscular factors contribute to this weakness (Mockford and Caulton, 2010). Impaired motor unit recruitment is an important and welldocumented neural factor (Stackhouse et al., 2005; Rose and McGill, 2005), while decreased muscle volume and changes in ⇑ Corresponding author at: Clinical Services, St. Olav University Hospital, Olav Kyrres gt. 17, 7006 Trondheim, Norway. Fax: +47 72571310. E-mail addresses: [email protected], [email protected] (S.M. Brændvik). 1050-6411/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jelekin.2011.10.002
muscle fiber type composition also contribute to the reported muscle weakness (Lieber et al., 2004; Barrett and Lichtwark, 2010). Excessive coactivation is another neural factor potentially contributing to reduced strength in CP. Coactivation is the concurrent activation of agonist and antagonist muscles around a joint and excessive presence of this phenomenon could imply an inherent reciprocal action by the neuromuscular system and can be seen as a form of muscle overactivity (Sheean and Mcguire, 2009). Indeed, reduced strength and increased coactivation are reported in CP compared to typically developing peers (Damiano et al., 2000; Tedroff et al., 2008). Besides its possible strength-limiting effect, coactivation may also represent a motor control strategy in situations with a need of increased joint stability or improved movement accuracy (Gribble et al., 2003). Such stiffness modulation is assumed to vary with different tasks and environmental constraints (Darainy and Ostry, 2008; Silva et al., 2009). Therefore, coactivation may both enhance and put limitations on motor performance. Spasticity, another form of muscle overactivity associated with CP, may play a role in coactivation. One of the most frequently used definitions of spasticity defines it as ‘‘a motor disorder characterized by a velocity-dependent increase in tonic stretch reflexes with exaggerated tendon jerks, resulting from hyper-excitability of the stretch reflex’’ (Lance, 1980), and spasticity is typically measured
138
S.M. Brændvik, K. Roeleveld / Journal of Electromyography and Kinesiology 22 (2012) 137–144
as resistance to passive movement. However, it is generally accepted that the resistance felt during passive muscle stretch includes both passive and active muscular components that are changed in a spastic muscle (Foran et al., 2005). This makes it challenging to differentiate between what is ‘‘true’’ spasticity and what might be structural changes, by studying resistance only. Over activation during stretch can be investigated by electromyography (EMG). Over activation during stretch of a spastic muscle working as antagonist to a concentric contraction (over coactivation) may occur, either due to passive stretch of the spastic muscle, eliciting a tonic stretch reflex or due to a failure of reciprocal inhibition between the two opposing muscle groups (Sheean and Mcguire, 2009). The objective of this study was to investigate the role of coactivation in strength and force modulation in the upper extremities in children/adolescents with unilateral CP. Two questions are addressed: (1) Does antagonist coactivation in CP limit strength and force modulation, and (2) Does spasticity increase coactivation? Coactivation of flexors and extensors of the elbow joint were therefore studied during maximal voluntary concentric contractions and sub-maximal isometric force tracing in the elbow joint. Measures of passive flexion and extension of the elbow are included as this may give indications of the presence of spasticity. In addition to differences between the affected and non-affected arm, differences between flexion and extension were studied. Although spasticity in elbow extensors cannot be excluded, differences between flexion and extension were expected since, clinically, spasticity is mainly observed in the elbow flexors of this patient group (Mayer, 1997). Moreover, the correlation between variables was investigated for the affected and non-affected arms separately. 2. Methods 2.1. Participants Twenty-one children/adolescents (9 males and 12 females), median age 13 years (range 8–18), diagnosed with unilateral CP participated in this study. They all had a functional use of the affected upper extremity corresponding to level III or better on the Manual Ability Classification System (Eliasson et al., 2006) and were able to take instructions verbally. Median active extension deficit in the elbow of the affected arm was 5° (range 0–45). Exclusion criteria were treatment with botulinum toxin A in the last 6 months and/or surgery on the upper extremity in the last 2 years. The study was approved by the Regional Committee for Medical Research Ethics and carried out according to the Declaration of Helsinki. A written informed consent was obtained from the parents and participants before participation. 2.2. Experimental setup Three different tasks were performed in the elbow joint: passive movement, isokinetic concentric maximal voluntary contraction (MVC) and sub-maximal force tracing. All 21 participants performed the first two tasks, 9 participants performed the submaximal force tracing task. A stationary dynamometer (Biodex System 3 Pro) was used to control the movements and measure torque, angle and velocity in all three tasks. The analogue signals were then digitised (sample frequency of 1000 Hz) through a National Instruments card (DAQCard-6036E) by EMG WorksÒ 3.1 (Delsys Inc.). The measurements were first carried out on the affected and then on the non-affected arm. Both the passive movement and the isokinetic MVCs were obtained using the passive mode in the Biodex. Sub-maximal force tracing was recorded using the isometric mode.
Surface EMG (SEMG) was recorded from the M. biceps brachii (BB) and the lateral head of M. triceps brachii (TB), from both the affected and the contralateral arm. Bipolar (1 cm spacing), singledifferential polycarbonate electrodes with a detection area of 10 mm2 were applied on the skin. Preparations and electrode placement above the BB and TB followed the recommendations of the Seniam project (Surface ElectroMyoGraphy for the NonInvasive Assessment of Muscles) (Hermens et al., 2000) and the reference electrode was placed on C7. The signals were amplified (gain of 1000) and band pass filtered (20–450 Hz) with a Bagnoli 16-channel SEMG system (Delsys Inc.) prior to sampling (1000 Hz). The dynamometer and SEMG signals were collected by the same system and therefore synchronized. 2.3. Procedure Details of the procedure are described in a recent paper by our research group (Brændvik et al., 2010). In brief, the participants were positioned in the Biodex chair, with the forearm in neutral position and the shoulder slightly abducted. First, the participants performed three passive trials in flexion and extension at two different velocities; 10° and 180°/sec, with the instruction to relax and just let the machine move the arm. Second, the participants performed three isokinetic MVCs at a velocity of 60°/sec. One test trial was allowed in order to insure that the task was performed correctly. There was a one-minute break between each trial. Finally, the participants were instructed to follow a target force by isometrically flexing the elbow. This task was only performed by nine subjects. For target and performance force normalisation purposes, the participants performed three isometric maximal voluntary elbow flexion contractions, whereof the highest peak torque was used as maximal isometric torque. The target signal consisted of a 15 s linear increase followed by a 15 s linear decrease between 0% and 40% of maximal isometric torque. Online visual feedback of both the target torque and the actual generated torque was provided on a monitor (full screen) placed at a convenient distance in front of the participant. The elbow angle was set to 60° (from full extension). The force tracing task was performed twice with a 10 s break between each trial. Two test trials were allowed prior to the measurements to ensure that the task was understood. 2.4. Data analysis Data analysis was carried out using Matlab (The Mathworks Inc.), version 7.0. Prior to further analyses, torque signals were low-pass filtered with a cut-off frequency of 6 Hz. The torque recorded during the passive movements at 10°/sec was used to obtain the weight of the arm and correct the other torque measures for arm weight. Peak torque and peak resistance torque were obtained during the isokinetic flexion and extension phases of all MVC and passive trials. In the force tracing task, both target and applied torque were presented as percentage of the highest torque generated during the three isometric MVCs. The first and last second of both the increasing and decreasing torque phases were left out from further analyses. For the remaining increasing and decreasing periods, the root mean square (rms) of the difference between the applied torque and the target torque was calculated (DTrms). The isokinetic flexion and extension MVC trials with the highest peak torques were used to reflect flexion and extension strength, respectively. The passive flexion and extension trials with the lowest peak resistance torque were used to reflect muscle tone. Furthermore, the force tracing trial with the lowest DTrms (increasing and decreasing) was used to reflect force modulation ability.
S.M. Brændvik, K. Roeleveld / Journal of Electromyography and Kinesiology 22 (2012) 137–144
A digital band-pass filter of 20–450 Hz was applied to the raw SEMG signal prior to further data analysis. Subsequently, RMS of the SEMG signals (BBrms and TBrms) was calculated over the whole isokinetic period of the best MVC flexion and extension trials and the best passive flexion and extension trials. BBrms and TBrms were also obtained during the best force increasing and decreasing periods of the force tracing task. SEMG noise levels, defined as the lowest BBrms and TBrms obtained during rest in periods of 500 ms were subtracted from all other BBrms and TBrms parameters. Median noise level ranged from 1.37 to 7.1 (lV) in BB and TB in the two arms. In the force tracing task, there was no difference between the increasing and decreasing periods for DTrms, BBrms and TBrms. Therefore, for these variables, the means of the two periods were used for further analysis. In order to study antagonist coactivation, the BBrms during isokinetic MVC extension, and the TBrms during MVC flexion and the force tracing task, were normalised to the BBrms during MVC flexion (BBco) and TBrms during MVC extension (TBco), respectively. BBrms during passive extension and TBrms during passive flexion were used as a measure of stretch response. Velocity dependent increase in resistance torque and the corresponding BBrms during passive elbow flexion and extension, was used as measures of spasticity, and was calculated according to the following formulas:
Velocity dependent increase in resistance ð%Þ ! Resistance Torque 180 =sec 1 100 ¼ Resistance Torque 10 =sec Velocity dependent increase in BB amplitude ð%Þ ! BBrms 180 =sec ¼ 1 100 BBrms 10 =sec
ð1Þ
ð2Þ
Moreover, as an indication of reflex activity during passive stretch, the within muscle ratio of TBrms or BBrms during passive stretch and during passive shortening at 180°/sec was determined. 2.5. Statistics Statistical analysis was carried out using PASW statistics, version 17. Non-parametric statistics was used due to small sample size and non-normal distribution of the data. Wilcoxon signed rank test was used to detect difference between affected arm (AA) and contralateral arm (CA). Spearman’s rho (rs) was used to assess relation between variables. A sample size of 21 and 9, will need to have correlation coefficients above .37 and .58, respectively, in order to reach significance (Fisher, 1995). An effect size similar to the standard deviation can be detected with a power of .99 and .75, respectively when paired two-tailed statistics with a probability of 0.05 are applied with a sample size of 21 and 9 (Lenth, 2006-9). 3. Results 3.1. MVCs Fig. 1 shows the results of the MVC task. The affected arm, AA, had significantly lower peak torque than the contralateral arm, CA, in both flexion and extension (p < .001 for both) (Fig. 1a). Both the AA and CA had significantly higher peak torque in elbow extension than in elbow flexion (p = .004 for AA and .001 for CA). There was no difference between AA and CA in relative flexion–extension peak torque (p = .848). During the isokinetic MVCs, the AA had significantly lower agonist SEMG activity compared to CA, indicated by lower BBrms during flexion and TBrms during extension
139
(p < .002 for both) (Fig. 1b). Although the average TBrms and BBrms as antagonists during MVCs were slightly higher in the CA compared to the AA, there was no significant difference between the arms (p = .517 for TBrms during elbow flexion and p = .973 for BBrms during elbow extension) (Fig. 1c). The antagonist coactivation during the MVCs, were significantly higher in the AA than in CA (Fig. 1d) both for BBco during elbow extension (p = .026) and TBco during elbow flexion (p = .002). In addition, this TBco was significantly higher than BBco in both the AA (p = .004) and CA (p < .001), but the relative TBco–BBco was not different between the two arms (p = .272). 3.2. Passive movements Fig. 2 shows results from passive elbow flexion and extension. A significantly higher extension resistance torque was found in the AA compared to CA for both 10 and 180°/sec (p < .001 for both), while no difference was found in flexion resistance torque (p = .179 for 10°/sec and p = .715 for 180°/sec). A velocity dependent increase in resistance torque was found for both flexion and extension in both arms (p < .001), but there was no difference between flexion and extension, or between the arms in this respect (Fig. 2a). As concerns the corresponding muscle stretch response, no difference between the arms in absolute muscle activity was found in flexion or extension at 180°/sec (Fig. 2b). In contrast to resistance torque, a velocity dependent increase in muscle activity was found for the AA during both flexion (p = .016) and extension (p = .001) compared to the CA (Fig. 2c). Moreover, the within muscle ratio BB extension/BB flexion in AA was above one, showing that BB activity was higher when stretched than when shortened during passive movement (180°/sec) (Fig. 2d). 3.3. Force tracing Only nine of the 21 participants completed this task. In the force tracing task, there was no significant difference in DTrms (p = .164), BBrms (p = .139) or TBrms (p = .250) between the AA and CA (Fig. 3a and b). However, TBco was significantly higher in the AA than in CA (p = .038) with a median of 15% of TBrms during MVC extension in AA and 9% in CA (Fig. 3c), while the median BBrms during the force tracing task was about 21% of BBrms during MVC flexion in both AA and CA. 3.4. Correlations Table 1 shows correlations between MVC, coactivation and passive movements. In both arms, there was a positive correlation between peak torque and agonist SEMG RMS during both maximal voluntary elbow flexion and extension. The slope of this relation was also similar in both arms. There was also a positive correlation between peak torque and antagonist SEMG RMS in CA during both flexion and extension and in AA during extension. A negative correlation was found between peak flexion torque and TBco in the AA, while no significant correlations were found between these variables in AA during extension and in CA neither during flexion nor extension. In the AA, but not the CA, a negative correlation was found between peak extension and flexion torque and passive elbow extension resistance torque. None of the strength variables correlated significantly with passive elbow flexion resistance torque. In the AA, a negative correlation was found between BBrms during passive extension and extension peak torque, and there was a trend between BBrms during passive extension and flexion peak torque. None of the strength variables correlated significantly with the TBrms during passive flexion.
140
S.M. Brændvik, K. Roeleveld / Journal of Electromyography and Kinesiology 22 (2012) 137–144
Fig. 1. MVC results. Median (95% confidence interval of mean) for (a) elbow flexion and extension peak torque, (b) agonist and (c) antagonist rms (root mean square), and (d) antagonist co-activation in affected arm (AA) and contralateral arm (CA). TB: triceps brachii, BB: biceps brachii, Nm: Newtonmeter, mV: millievolt. ⁄Significant difference between arms at .05 level.
No significant correlations were found between DTrms, TBco and extension resistance torque in the force tracing task. However, there was a trend for a negative correlation (rs = .55, p = .062) between DTrms and TBco in the CA; increased coactivation is associated with decreased difference between the target and applied force.
4. Discussion In the present study, the role of coactivation in maximal force generation and sub-maximal force tracing was investigated in both the affected and the contralateral elbow in children/adolescents with spastic unilateral CP. The affected arm was found to be weaker and to have more coactivation than the contralateral arm in both flexion and extension. However, the absolute antagonist activity and therefore probably the braking effect of the antagonist, was not different between the arms during maximal voluntary contraction. Indication of reflex activity was found in BB, but not in TB of the affected arm, while the TB had higher coactivation than BB in both arms. Both arms performed equally well in the force tracing task despite increased coactivation in the affected arm.
The current findings indicate that antagonist coactivation in the elbow joint plays a minor role in muscle weakness in CP, and increased antagonist coactivation is likely to increase stability in a movement involving a spastic muscle in CP. 4.1. The role of coactivation in strength Coactivation, the simultaneous activation of agonist and antagonist muscles, implies an inherent reciprocal action by the neuromuscular system, potentially hampering force production capacity. Our results are in line with studies reporting decreased strength and increased coactivation in patients with CP (Ikeda et al., 1998; Elder et al., 2003; Tedroff et al., 2008). Moreover, our participants were weaker in elbow flexion than extension, combined with higher coactivation in TB than in BB, but there was no difference between the affected and the contralateral arm according to this. However, our findings did also include a significant lower agonist EMG activity in the affected arm compared to the contralateral arm during both flexion and extension. As coactivation in our study is operationalised as percentage of maximal agonist contraction, this will result in a relatively higher coactivation level in the affected arm compared to the contralateral arm. Thus, the increased
S.M. Brændvik, K. Roeleveld / Journal of Electromyography and Kinesiology 22 (2012) 137–144
141
Fig. 2. Passive movement results. Median (95% confidence interval of mean) for (a) velocity dependent change in resistance torque (b) muscle stretch response rms (root mean square) at a velocity of 180°/sec, (c) velocity dependent change in muscle stretch response and (d) ratio BB extension/BBflexion and TB flexion/TB extension during passive elbow flexion and extension at 180°/sec. AA: affected arm, CA: contralateral arm. TB: triceps brachii, BB: biceps brachii, Nm: Newtonmeter, mV: millievolt. ⁄Significant difference between arms at .05 level.
coactivation reflects agonist weakness rather than being a direct cause. Moreover, the absolute antagonist activity was not different between the arms, indicating that a possible braking effect was not higher in the affected than the contralateral arm. Neither did we find any systematic relation between strength and coactivation, except for a significant negative relation between elbow flexion strength and TB coactivation in the affected arm. In all, this indicates that the cause of the muscle weakness lies in the agonist itself, and this would be in agreement with studies showing that persons with CP have a reduced ability to use the full potential of the muscle (Rose et al., 2005; Stackhouse et al., 2005). We hypothesised that spastic activity could increase coactivation. Therefore, we evaluated spasticity. In the current study, the affected arm had an increased muscle tone in the elbow flexors in both slow and fast passive extension compared to the contralateral arm. However, the velocity dependent increase in muscle tone was not different between the arms or between flexors and extensors (Fig. 2a). Velocity dependent increase in resistance to movement is not exclusive for spasticity (Pandyan et al., 2005). Also the viscoelastic properties of the muscle are velocity dependent
(Dietz, 1995), and this could explain the lack of difference in velocity dependent increase in muscle tone. Another notion to be made is the large variability in velocity dependent response found in passive extension of the contralateral arm which also could attribute to the lack of difference. However, all these results indicate that non-reflexive changes may cause the increased mechanical stiffness during extension in the affected arm. The literature supports such an assumption as secondary structural and functional changes in spastic muscles are reported although the primary lesion is neural (Lieber et al., 2004). However, the SEMG results indicate a neural mechanism. There was a velocity dependent increase in muscle stretch response during both passive extension and flexion in the affected arm compared to the contra lateral one, while only the elbow flexors of the affected arm showed more activity during passive stretching than shortening. This indicates the presence of spasticity, as defined by Lance, in the affected BB and a possibly non-reflexive increased TB activation during stretch. The inverse relation between BB passive stretch response and extension strength, and between flexor muscle tone and extension strength, could easily lead to a preliminary conclusion that
142
S.M. Brændvik, K. Roeleveld / Journal of Electromyography and Kinesiology 22 (2012) 137–144
Fig. 3. Force tracing results. Median (95% confidence interval of mean) of (a) Dtorque rms (the root mean square of the difference between the applied force and the target force), (b) SEMG rms in submaximal force modulation task (between 0% and 40% of maximal voluntary contraction) and (c) TB co-activation. BB: biceps brachii, TB: triceps brachii, AA: affected arm, CA: contralateral arm. ⁄Significant difference between arms at .05 level.
Table 1 Correlation (Spearman’s rho) between peak elbow extension/flexion torque in the affected (AA) and contralateral (CA) arm, and agonist/antagonist root mean square (rms) of biceps brachii (BB) and triceps brachi (TB), antagonist co-activation (as% of MVC), extension/flexion resistance torque and antagonist stretch response during the passive flexion and extension.
Agonist rms Antagonist rms Antagonist co-activation Extension resistance torque Flexion resistance torque BBrms during passive extension TBrms during passive flexion * **
Peak extension torque AA
Peak extension torque CA
Peak flexion torque AA
Peak flexion torque CA
.68 (p < .001)** .436 (p = .024)* .20(p = .192) .71 (p < .001)** .08 (p = .358) .38 (p = .045)* .25 (p = .140)
.59 (p = .003)** .396 (p = .038)* .30 (p = .098) .28 (p = .108) .003 (p = .496) .126 (p = .293) .087 (p = .345)
.78 (p < .001)** .264 (.124) .43 (p = .027)* .52 (p = .008)** .26 (p = .124) .30 (p = .089) .24 (p = .151)
.53 (p = .007)** .596 (p = .002)** .05 (p = .414) .105 (p = .325) .245 (p = .142) .10 (p = .341) .08 (p = .358)
Significant at .05 level. Significant at .01 level.
spasticity induced coactivation may cause the reduced extension strength in the affected arm. However, the BB coactivation during the strength task was not especially high and it is at least not likely that it caused a higher braking effect in the affected arm than in the non-affected arm and in flexion than in extension. Moreover, a trend for an inverse relation between the BB stretch response
and flexion strength, and an inverse relation between flexor muscle tone (and not extensor muscle tone) and flexion strength indicates that the impaired function in the flexor muscles reflects its force generating capabilities. One important notion according to the effect of spasticity has to be made; while the muscle tone measurements are carried out on muscles at rest, the strength
S.M. Brændvik, K. Roeleveld / Journal of Electromyography and Kinesiology 22 (2012) 137–144
measurements are voluntary actions of an active motor system. Thus, we do not know how the response during passive conditions reflects active situations. Moreover, if spasticity as the involuntary activation of a muscle during stretch would be similar during passive and active contractions, its contribution to force production would be negligible, since the antagonist EMG amplitude during passive movements (Fig. 2b) is much smaller than during maximal voluntary contractions (Fig. 1c). Another point to be made is the difference in velocity between the two tasks; 180°/sec versus 60°/sec, which may affect the muscle response. In contrast to BB coactivation, no indication of spasticity was found in TB coactivation. The significant negative relation between elbow flexion strength and TB coactivation in the affected arm, combined with a relative high coactivation and antagonist activation during concentric MVCs compared to BB during equivalent tasks, suggests that TB coactivation reflects a different aspect of motor performance than BB coactivation. TB coactivation may be part of a motor control strategy to improve motor control in a weak and spastic BB. That would be in agreement with increased coactivation in order to increase joint stability or improve movement accuracy in other situations (Gribble et al., 2003) and in early phases of learning (Darainy and Ostry, 2008). Specifically, it is hypothesized that the antagonistic muscles in coactivation might act as a kind of mechanical filter by suppressing the effects of force variability in the agonist (Selen et al., 2005). 4.2. The role of coactivation in sub maximal force tracing No difference in performance between the arms was found in the force tracing task. This indicates that the affected arm performs equally well as the contralateral arm providing it is allowed to act relative to its capacity. This is in line with Smits-Engelsman et al (2005) who studied force control ability in the hand of hemiplegic children, also scaled to MVC of the affected hand. Despite equal performance, higher TB coactivation was found in the affected arm. However; we did not detect any relation between performance and coactivation. This is in contrast to Bandholm et al (2009) who reported a relation between ankle dorsiflexion torque steadiness and plantarflexion coactivation level. This could be explained by differences in force levels; their submaximal level was scaled to torque in typically developing children. Thus, the significance of coactivation as means of motor control may increase with increasing force levels. This assumption is supported by increasing variability in force control ability at higher force levels (SmitsEngelsman et al., 2005) and also corresponds with our results from the maximal strength task. The small sample size in the force tracing task (n = 9) could affect the power of these results. However, since significant and nearly significant differences were found between the two arms, it is not likely that a larger sample size would have changed the result regarding the relation between performance and coactivation.
143
lower extremities (Damiano et al., 2001). Although no indication of reflex activity in the triceps brachii was found in the current study, a velocity dependent increase in muscle stretch response was indeed found. Maximal force production in children with CP is complicated by the reduced ability to fully activate the muscles according to their potential. Thus, maximal EMG activity is underestimated. This has to be accounted for in the interpretations, but will not change our conclusions since they are based on relative comparisons and not on an absolute estimation of force. Moreover, skin fold thickness is a factor that may infer with inter-individual variance in SEMG amplitude (Nordander et al., 2003) and information about this was not obtained in this study sample. However, the normalization procedure reduces this effect and moreover, the slope of the relation between strength and agonist amplitude of SEMG of the two arms was equal and thus gave no indications of such an effect. The power calculations made in this study are based on parametric assumptions, and the use of non-parametric statistics reduces this power. However, the calculated power was high, .99 and .75, so it is not likely that this will weaken the results. 5. Concluding remarks The results of this study do not support that antagonist coactivation is a significant contributor to the muscle weakness observed in children with CP, or limits the ability to modulate force. Rather, the results indicate that antagonist coactivation reflects different aspects of impaired motor control. More specific, coactivation of the BB during maximal elbow extension may have a spastic component, however relatively low, while coactivation of the TB during elbow flexion most likely is a motor control strategy in tuning impaired BB function during both maximal and submaximal tasks. These results need to be verified for children with different MACS levels and compared to and an age and gender matched reference group. Additional use of electrostimulation will give a direct measure of the force generating capability in the muscle, and thus a more accurate measure of coactivation. However, such a method may be problematic in the upper arm in children. Moreover, intervention studies addressing spasticity treatment, i.e. Botulinum toxin A and strength training, would give more insight into the role of coactivation with respect to the activity limitations in CP. Acknowledgments We would like to thank the children for their participation in the study, Ann-Kristin Gunnes Elvrum for her contribution during study planning and data acquisition, and Rannei Sæther and Torarin Lamvik for providing study participants. We would also like to thank Paul Jarle Mork for valuable comments on the manuscript. This study was funded by the Norwegian Directorate of Health and Unimed Innovation AS.
4.3. Limitations of the study
References
Inferences in this study are based on differences between the affected and the contralateral arm in unilateral CP. Statistically, this increases the power of the results and may also be considered a conservative approach. However, it is questionable to what extent the contralateral arm in unilateral CP also is affected (Wiley and Damiano, 1998; Elder et al., 2003; Smits-Engelsman et al., 2005; Steenbergen and Meulenbroek, 2006). Moreover, inferences are made based on differences in resistance torque and muscle stretch response, between flexors and extensors. Although clinical focus with respect to spasticity in the elbow joint mainly is on the flexors, indication of spasticity is also experienced in the opposing muscle group. Indeed, this is also reported with respect to the
Andersen GL, Irgens LM, Haagaas I, Skranes JS, Meberg AE, Vik T. Cerebral palsy in Norway: prevalence, subtypes and severity. Eur J Paediat Neurol 2008;12(1): 4–13. Bandholm T, Rose MH, Sløk R, Sonne-Holm S, Jensen BR. Ankle torque steadiness is related to muscle activation variability and coactivation in children with cerebral palsy. Muscle Nerve 2009;40(3):402–10. Barrett RS, Lichtwark GA. Gross muscle morphology and structure in spastic cerebral palsy: a systematic review. Dev Med Child Neurol. 2010;52(9): 794–804. Brændvik SM, Elvrum AKG, Vereijken B, Roeleveld K. Relationship between neuromuscular body functions and upper extremity activity in children with cerebral palsy. Dev Med Child Neurol 2010;52(2):e29–34. Damiano DL, Martellotta TL, Sullivan DJ, Granata KP, Abel MF. Muscle force production and functional performance in spastic cerebral palsy: Relationship of cocontraction. Arch Phys Med Rehabil 2000;81(7):895–900 [doi: 10.1053/ apmr.2000.5579].
144
S.M. Brændvik, K. Roeleveld / Journal of Electromyography and Kinesiology 22 (2012) 137–144
Damiano DL, Quinlivan J, Owen BF, Shaffrey M, Abel MF. Spasticity versus strength in cerebral palsy: relationships among involuntary resistance, voluntary torque, and motor function. Eur J Neurol 2001;8(s5):40–9. Darainy M, Ostry D. Muscle cocontraction following dynamics learning. Exp Brain Res 2008;190(2):153–63. Dietz V, Berger W. Cerebral palsy and muscle transformation. Dev Med Child Neurol 1995;37:180–4. Elder GCB, Stewart G, Cook K, Weir D, Marshall A, Leahey L. Contributing factors to muscle weakness in children with cerebral palsy. Dev Med Child Neurol 2003;45(8):542–50. Eliasson AC, Krumlinde-Sundholm L, Rösblad B, Beckung E, Arner M, Öhrvall A-M, et al. The Manual Ability Classification System (MACS) for children with cerebral palsy scale development and evidence of validity and reliability. Dev Med Child Neurol 2006;48(7):549–54. Engsberg JR, Ross SA, Olree KS, Park TS. Ankle spasticity and strength in children with spastic diplegic cerebral palsy. Dev Med Child Neurol 2000;42(1):42–7. Fisher RAYF. Statistical Tables for Biological, Agricultural, and Medical Research. Essex, United Kingdom: Pearson Education; 1995. Foran JRH, Steinman S, Barash I, Chambers HG, Lieber RL. Structural and mechanical alterations in spastic skeletal muscle. Dev Med Child Neurol 2005;47(10): 713–7. Gribble PL, Mullin LI, Cothros N, Mattar A. Role of cocontraction in arm movement accuracy. J Neurophysiol 2003;89(5):2396–405. Hermens HJ, Freriks B, Disselhorst-Klug C, Rau G. Development of recommendations for SEMG sensors and sensor placement procedures. J Electromyogr Kinesiol 2000;10(5):361–74. Ikeda AJAM, Granata KP, Damiano DI. Quantification of cocontraction in spastic cerebral palsy. Electromyogr Clin Neurophysiol 1998;38(8):497–504. Koman LA, Smith BP, Shilt JS. Cerebral palsy. The Lancet 2004;363(9421):1619–31. Lance JW. The control of muscle tone, reflexes, and movement: Robert Wartenberg lecture. Neurology 1980;30(12):1303–13. Lenth, RV. (2006-9). Java Applets for Power and Sample Size [Computer software]. Retrieved 08, 31, 2011 from http://www.stat.uiowa.edu/~rlenth/Power. Lieber RL, Steinman S, Barash IA, Chambers H. Structural and functional changes in spastic skeletal muscle. Muscle Nerve 2004;29(5):615–27. Mayer NH, Esquenazi A, Childers MK. Common patterns of clinical motor dysfunction. Muscle Nerve 1997(Suppl. 6):S21–35. Mockford M, Caulton JM. The pathophysiological basis of weakness in children with cerebral palsy. Pediatr Phys Ther 2010;20(4):318–33. Nordander C, Willner J, Hansson GÅ, Larsson B, Unge J, Granquist L, et al. Influence of the subcutaneous fat layer, as measured by ultrasound, skinfold calipers and BMI, on the EMG amplitude. Eur J Appl Physiol 2003;89(6):514–9. Pandyan AD, Gregoric M, Barnes MP, Wood D, Vanwijck F, Burridge J, et al. Spasticity: clinical perceptions, neurological realities and meaningful measurement. Disabil Rehabil 2005;27(1):2–6. Rose J, Mcgill KC. Neuromuscular activation and motor-unit firing characteristics in cerebral palsy. Dev Med Child Neurol 2005;47(5):329–36. Ross SA, Engsberg JR. Relationships between spasticity, strength, gait, and the GMFM-66 in persons with spastic diplegia cerebral Palsy. Arch Phys Med Rehabil 2007;88(9):1114–20. Selen L, Beek P, Dieën J. Can co-activation reduce kinematic variability? A simulation study. Biol Cybern 2005;93(5):373–81. Sheean G, Mcguire JR. Spastic hypertonia and movement disorders: pathophysiology, clinical presentation, and quantification. PM&R 2009;1(9): 827–33.
Silva PL, Fonseca ST, Ocarino JM, Gonã§Alves P, Mancini MC. Contributions of cocontraction and eccentric activity to stiffness regulation. J Motor Behav 2009;41(3):207–28. Smits-Engelsman BC, Rameckers EA, Duysens J. Muscle force generation and force control of finger movements in children with spastic hemiplegia during isometric tasks. Dev Med Child Neurol 2005;47(5):337–42. Stackhouse SK, Binder-Macleod SA, Lee SCK. Voluntary muscle activation, contractile properties, and fatigability in children with and without cerebral palsy. Muscle Nerve 2005;31(5):594–601. Steenbergen B, Meulenbroek RGJ. Deviations in upper-limb function of the lessaffected side in congenital hemiparesis. Neuropsychologia 2006;44(12): 2296–307 [doi: 10.1016/j.neuropsychologia.2006.05.016]. Tedroff K, Knutson LM, Soderberg GL. Co-activity during maximum voluntary contraction: a study of four lower-extremity muscles in children with and without cerebral palsy. Dev Med Child Neurol 2008;50(5):377–81. Vaz DV, Cotta M, Fonseca ST, De Melo Pertence AE. Muscle stiffness and strength and their relation to hand function in children with hemiplegic cerebral palsy. Dev Med Child Neurol 2006;48(9):728–33. Wiley ME, Damiano DL. Lower-extremity strength profiles in spastic cerebral palsy. Dev Med Child Neurol 1998;40(2):100–7.
Siri Merete Brændvik was born in Trondheim, Norway in 1961. She graduated as PT from Bergen University College in 1985 and her clinical work has been in the field of neuro-/orthopedics. In 2003, she received her M.Sc. degree in Human Movement Science at the Norwegian University of Science and Technology (NTNU), Norway. She is now a Ph.D. student at Department of Human Movement Science, NTNU. Her research interest is in the field of motor control, especially motor disorders associated with cerebral palsy (CP). Her Ph.D. thesis (expected 2012) addresses the interrelation between neuromuscular impairments in CP and how these are related to activity limitations.
Karin Roeleveld received the M.Sc. degree in human movement sciences from the Vrije Universiteit in Amsterdam, The Netherlands in 1992. Thereafter, she joined the University Hospital in Nijmegen and received her Ph.D. degree in 1997 on the fundaments of surface electromyography using multi electrode surface electromyography. Thereafter she spent two years working as a post doc at the Friedrich-Schiller-University Jena, Germany. Since 1998 she is working at the human movement sciences programme, at the Norwegian University of Science and Technology (NTNU), Norway, where she is currently working as professor. Her research interests are related to neuromuscular function and how this adapts with activity, disease and age.