Efficacy of EMG-triggered electrical arm stimulation in ... - IOS Press

189

Restorative Neurology and Neuroscience 27 (2009) 189–197 DOI 10.3233/RNN-2009-0469 IOS Press

Efficacy of EMG-triggered electrical arm stimulation in chronic hemiparetic stroke patients Friederike von Lewinski a, Sabine Hoferb , Jürgen Kausc, Klaus-Dietmar Merboldtb , Holger Rothkegela, Renate Schweizer b , David Liebetanz a, Jens Frahmb and Walter Paulusa,∗ a

Abteilung Klinische Neurophysiologie, Georg-August Universit a¨ t Göttingen, Go¨ ttingen, Germany Biomedizinische NMR Forschungs GmbH am Max-Planck-Institut f u¨ r biophysikalische Chemie, G o¨ ttingen, Germany c Otto Bock HealthCare Deutschland GmbH, Duderstadt, Germany b

Abstract. Purpose: EMG-triggered electrostimulation (EMG-ES) may improve the motor performance of affected limbs of hemiparetic stroke patients even in the chronic stage. This study was designed to characterize cortical activation changes following intensified EMG-ES in chronic stroke patients and to identify predictors for successful rehabilitation depending on disease severity. Methods: We studied 9 patients with severe residual hemiparesis, who underwent 8 weeks of daily task-orientated multi-channel EMG-ES of the paretic arm. Before and after treatment, arm function was evaluated clinically and cortical activation patterns were assessed with functional MRI (fMRI) and/or transcranial magnetic stimulation (TMS). Results: As response to therapy, arm function improved in a subset of patients with more capacity in less affected subjects, but there was no significant gain for those with Box & Block test values below 4 at inception. The clinical improvement, if any, was accompanied by an ipsilesional increase in the sensorimotor cortex (SMC) activation area in fMRI and enhanced intracortical facilitation (ICF) as revealed by paired TMS. The SMC activation change in fMRI was predicted by the presence or absence of motor-evoked potentials (MEPs) on the affected side. Conclusions: The present findings support the notion that intensified EMG-ES may improve the arm function in individual chronic hemiparetic stroke patients but not in more severely impaired individuals. Functional improvements are paralleled by increased ipsilesional SMC activation and enhanced ICF supporting neuroplasticity as contributor to rehabilitation. The clinical score at inception and the presence of MEPs have the best predictive potential. Keywords: EMG-triggered electrostimulation, stroke, neuroplasticity, TMS, fMRI

1. Introduction EMG-triggered electrostimulation (EMG-ES) of the affected arm muscles can improve motor performance not only in the acute phase, but even in the chronic stage after stroke. The effect is based on the ampli∗ Corresponding

author: Prof. Dr. med. Walter Paulus, Abteilung klinische Neurophysiologie, Universität Göttingen, Robert-Koch Str. 40, 37075 Göttingen, Germany. Tel.: +49 (0)551 39 6650; Fax: +49 (0)551 39 8126; E-mail: [email protected].

fication of the agonist muscle power and a simultaneous reduction of spasticity in the antagonist (Chae & Yu, 2000; Francisco et al., 1998; Kraft et al., 1992). The putative contribution of neuronal reorganization has been addressed only recently. In a study on chronic stroke subjects, improvement following EMG-ES was accompanied by changes in somatosensory cortex activation as measured by fMRI (Kimberley et al., 2004). Other rehabilitative methods such as constraintinduced movement therapy, intensive finger-tracking, or task-oriented training were associated with changes

0922-6028/09/$17.00  2009 – IOS Press and the authors. All rights reserved

190

F. von Lewinski et al. / Electrostimulation in chronic stroke

in size and location of motor output areas (Carey et al., 2002; Jang et al., 2003; Liepert et al., 2000; Richards et al., 2008; Tarkka et al., 2008). The corticospinal tract integrity in addition to the presence or absence of motor-evoked potentials (MEP) seem to play a role in determination of the reorganization pattern and the individual rehabilitative potential (Hamzei et al., 2006; Stinear et al., 2007). The aim of the current project was to further characterize EMG-ES-related neuroplastic changes in the motor system and to identify possible predictors for successful treatment by using combined TMS and fMRI.

series of the stimulator unit. The trigger level was set at 75% of the maximum EMG response of each muscle. Stimulation was carried out under the supervision of an experienced physiotherapist (J.K.) using a stimulator unit (STIWELL med4, Otto Bock Healthcare, Duderstadt, Germany) with biphasic symmetric pulses (35 Hz, single pulsewidth 200 µs, 10 s ON, approximately 20 s OFF, time delay: 3 s ramp up, 2 s ramp down) and round surface electrodes (50 mm ValuTrode , AXELGAARD, Fallbrook, CA, USA). The electrical stimulation was delivered at an intensity of 15–30 mA. Within this range, the current necessary for sufficient muscle contraction was adapted individually for every stimulation point.

2. Methods

2.3. Clinical outcome measures

2.1. Patients

Box & Block test (Mathiowetz et al., 1985) and Action research arm test (ARAT) (Lyle, 1981) were used to assess task-orientated arm function. In the Box & Block test, the dependent variable was the average number of blocks transported from one side of a portioned box to the other in three timed (60 s) trials. Both tests have proven valid and reliable for assessment of arm function in post-stroke patients in previous studies (Platz et al., 2005a). The tests were performed at the beginning and end of the baseline period as well as within 2 days after the end of treatment, but not at the same day of the last treatment session. Because patients showed constant performance in either of these tests during the baseline period these data were pooled.

Nine patients (see Table 1) with residual arm paresis following an ischemic stroke in the territory of the middle cerebral artery (MCA) and onset 10 months ago were included. Prerequisite was a stable motor performance over a baseline observational period of 5 to 9 weeks. Exclusion criteria were stroke due to intracranial hemorrhage, brainstem ischemia, heart pacemaker, age 80 years, severe aphasia, severe psychiatric illness, dementia and history of migraine. Patients with a history of seizures or cranial metal implants were excluded from the TMS part, those with any form of metal implant from fMRI. Approval was obtained by the Ethics committee of the University of G o¨ ttingen. Patients gave written informed consent to participate. Previous regular physiotherapy was continued during time of the study. 2.2. EMG-triggered electrostimulation (EMG-ES) All patients received 8 weeks of daily 40 minutes multi-channel EMG-ES of the affected arm with training of task-orientated movements. They were asked to put plastic cups of different sizes into each other and move away from the body. This required wrist and finger extension, finger flexion, arm abduction and elbow extension, which was supported by consecutive stimulation of forearm wrist and finger extensor muscles (M. extensor carpi radialis, M. extensor digitorum, M. extensor pollicis longus and brevis), finger flexor muscles, M. deltoideus and M. triceps brachii. Voluntary activity in at least one muscle group of the upper extremity was necessary to trigger the electrical impulse

2.4. Functional magnetic resonance imaging (fMRI) fMRI studies were conducted at 2.9 T (TIM Trio, Siemens Healthcare, Erlangen, Germany) using a CP head coil. Anatomic images were based on a T 1 weighted 3D fast low angle shot (FLASH) MRI sequence (repetition time TR = 11 ms, echo time TE = 4.9 ms, flip angle 15 ◦ , GRAPPA technique R = 2, 1 mm isotropic resolution). To identify possible activation patterns mainly in the motor areas of the brain, functional mapping by blood oxygenation level dependent (BOLD) MRI started with a multislice BOLD-sensitive echo-planar imaging (EPI) sequence (TR = 2000 ms, TE = 36 ms, flip angle 70 ◦ ) at 2.0 × 2.0 mm 2 resolution using 20 transverse sections at 2 mm thickness. The field of view was 256x160. To assess putative differences in brain activation before and after treatment, a fMRI protocol including electrically provoked hand movements (passive wrist


191

Table 1 Patients’ details and outcome in arm function tests Nr. 1 2 3 4 5 6 7 8 9 Mean S.D. P

Age (years) 60 58 59 18 74 65 51 59 73 57.4 16.5

Gender f m m f f m m f f

Stroke localization IC MCA MCA MCA BG MCA MCA MCA MCA

side L R R L L R L R R

duration (months) 69 147 30 59 13 33 49 10 39 49.9 41.3

TMS y y y y y n n y y

fMRI y y n y n y y y y

Box & Block baseline 4 16.5 0 3 0 0 0 0 1 2.7 5.4

change 4 5.5 0 1 1 0 0 0 −1 1.2 2.2 0.14

ARAT baseline 13 26 4 9 0 3 4 5 8.5 8.1 7.7

change 9 3 0 2 3 0 0 2 −0.5 2.1 2.9 0.07

F = female, m = male; IC = internal capsule, BG = basal ganglia, MCA = extensive involvement of the middle cerebral artery territory; L = left, R = right; TMS = transcranial magnetic stimulation; fMRI = functional magnetic resonance imaging; y = yes, n = no; Box & Block Test (number of blocks); ARAT = action research arm test (max. points 57); change = difference to baseline after treatment; S.D. = standard deviation; P = t-test comparing clinical performance pre and post EMG-ES.

and finger extension) and voluntary hand movements (fist opening/closure) was used to stimulate activity in the affected sensorimotor system of each individual patient. 7 patients (see Table 1) successfully performed the given tasks. Stimulation of hand and finger extension was carried out as described above (2.2) but without EMG triggering (4 s stimulation, 12 s rest, 24 cycles). Voluntary movements of the paretic hand were repeated as follows: 12 s movement, 18 s rest, 8 cycles. Visual instructions (Start – Stop) were projected onto a display system mounted atop the head coil within the magnet. Their timing was computer controlled and synchronized with the MRI system. Hand movements were visually controlled. Activation maps were determined with a correlation analysis using in-house software (Baudewig et al., 2003). The signal intensity time course of every voxel was cross-correlated with a boxcar reference function derived from the stimulation protocol, adding a delay of two images (= 4 s) to account for the hemodynamic response latency. Correlation coefficients were then rescaled as percentile ranks of the underlying noise distribution and thresholded in a two-step procedure. First, voxels with a correlation coefficient above the 99.99 percentile range were identified as activated and second, directly neighboring voxels to the previously identified were added as activated, as long as their correlation coefficients exceeded a second lower threshold at the 95 percentile rank. Areas of investigation included the primary sensorimotor areas (M1 and S1) in each patient. The number of activated voxels in these areas (extending over up to 6 sections) was determined before and after treatment.

2.5. Transcranial magnetic stimulation (TMS) A subgroup of seven patients (see Table 1) underwent TMS at the same time points as functional testing. TMS was delivered using a MAGSTIM 200 stimulator (Magstim Company Ltd., Whitland, Dyfed, UK) with a 7 cm figure-of-eight coil. The coil was held tangential to the skull with the coil handle perpendicular to the assumed line of the central sulcus. Motorevoked potentials (MEP) were recorded with surface electromyographic (EMG) electrodes after stimulation of the hotspot of the corresponding muscle. The TMS recordings were made from the extensor digitorum muscle, which was stimulated and trained the most extensively. Only if this was not possible due to incomplete relaxation (patients 4, 5 and 9) recordings were made from the abductor pollicis brevis (apb) muscle. Electrical stimulation of the median nerve was applied to measure M-response latencies, F-wave latencies and M-response amplitudes. Recordings were obtained with surface electrodes from the apb muscle. The maximal MEP amplitude (peak-to-peak) was expressed as percentage of the maximal M-response amplitude. The central motor conduction time (CMCT) was calculated as follows: total latency minus (M-response latency plus F-wave latency minus 1)/2. Recordings from the unaffected side were made using the same electrode positions as contralaterally. EMG signals were amplified (1000 ×) and band-pass filtered between 2 Hz and 3 kHz before being digitized at 5 kHz. The resting motor threshold (RMT) was defined as the stimulus intensity needed to produce MEPs with an amplitude of >50 µV in 3 of 6 consecutive trials during complete muscle relaxation, active motor threshold (AMT) to

192


Fig. 1. Clinical outcome after EMG-ES. Box & Block test values at baseline correlate with the change after treatment (B & B delta). The linear relationship between B & B at baseline and B & B delta (R2 = 0.75) is illustrated by the regression line.

produce MEPs with ∼200 µV amplitude during slight voluntary contraction. When no MEP was recordable from the paretic side RMT was assumed 100%. Stimulus response curves were tested at rest using single TMS-pulses with intensities of 100%, 110%, 120%, 130% and 140% RMT. For each stimulus intensity 10 trials were performed. Paired pulse protocols were used to investigate short latency intracortical inhibition and facilitation (SICI/ICF) (Nakamura et al., 1997; Ziemann et al., 1996). For this purpose, two magnetic stimulators were connected to one coil through a Bistim Device (Magstim). A subthreshold conditioning stimulus with 90% AMT intensity preceded a suprathreshold test stimulus. Its intensity was adjusted to produce a MEP of ∼0.5–1 mV in peak-topeak amplitude or, if this was not possible, to 120% RMT. Interstimulus intervals (ISIs) were set at 2, 3, 5, 7, 10, and 15 ms. Non-conditioned (control) stimuli and paired TMS were intermixed randomly, for each condition 12 responses were collected. Thereafter, the mean amplitude of the conditioned response was expressed relative to the mean amplitude of the control response (amplitude ratio). For evaluation of therapeutic effects, responses at ISIs with inhibitory (2 and 3 ms), neutral (5 and 7 ms), and facilitating (10 and 15 ms) character were pooled. 2.6. Statistics T-tests were used to compare clinical scores and TMS measures (MT, MEP amplitude, CMCT) pre and post treatment as well as to evaluate side differences of the latter. For the fMRI data a non-normal distribu-

Fig. 2. Brain activations (colored areas) superimposed onto original T2*-weighted echo-planar images elicited by electrically provoked passive hand movement before and after EMG-ES of patient 1. After EMG-ES, expanded activations in the sensorimotor area (SMC) of the affected hemisphere could be observed in 4 out of 7 subjects.

tion was assumed and pre and post therapy conditions were compared using the Wilcoxon signed rank test. Repeated measures analysis of variance (ANOVA) was employed for analysis of stimulus response curves and SICI/ICF with side, time point and, where applicable, intensity or interstimulus interval as within subject factors. As level of significance we regarded p < 0.05. Unless indicated otherwise, values are given as mean ± standard deviation; error bars in figures also represent standard deviation. Linear regression analysis was conducted to determine the variables that predict motor improvement after EMG-ES or activation increase in fMRI. Statistics were performed using SigmaStat 3.5 or StatView 5.0 (for ANOVAs) software.

3. Results 3.1. Clinical outcome After 8 weeks of EMG-ES the functional arm tests ARAT and Box & Block revealed an improvement after therapy only in a subset of patients and the change for the entire group failed to reach significance (Table 1). In the Box & Block Test the score at inception predicted the change after EMG-ES with higher gain for less affected patients (R2 = 0.75, p = 0.003, Fig. 1). For those patients with Box & Block scores of less than 4 at inception there was no significant gain (0 or 1 point) after EMG-ES therapy. Although the correlation between both test scores at inception was high (r = 0.96), the ARAT score at inception had no predictive

F. von Lewinski et al. / Electrostimulation in chronic stroke Table 2 Changes in fMRI and TMS parameters after EMG-ES treatment Pat. Nr. 1 2 3 4 5 6 7 8 9

fMRI pre post 15 37 0 18 0

0

0 2 53 0

4 0 46 10

SMC activation Increase Increase No change Increase Decrease Decrease Increase

TMS MEP Yes Yes Absent Absent Yes

ICF change Increase Increase n.d. n.d. Increase

Absent Yes

n.d. Increase

Sensorimotor cortex (SMC) activation given as number of activated voxels pre and post treatment. ICF = intracortical facilitation, MEP = motor-evoked potentials, n.d. = not detectable.

193

As shown in Fig. 3, there was a relationship between the change in SMC activation and response to motor practice. Only patients with a large (more than 10 points, patients 1 and 2) increase in SMC activation showed a significant improvement in the Box & Block test (>1 point), whereas those with less SMC activation increase had no functional benefit. However, the prediction of response to motor practice (Box & Block Test) by SMC activation change just failed to reach significance (linear regression analysis, R 2 = 0.55, p = 0.06). There were no consistent activation changes in areas other than SMC, i.e. motor association cortices. The voluntary hand movements were mostly accompanied by variable unintentional mirror movements or movements of the feet, which at least partially overlapped the expected response. Therefore, we excluded fMRI changes according to self-contained movements from the analysis. 3.3. Transcranial magnetic stimulation (TMS)

Fig. 3. Relationship between fMRI changes and response to motor practice. Relevant motor improvements following EMG-ES as revealed by the Box & Block test (B & B delta) were only seen for patients with a large increase in SMC activation (>10 points) in fMRI.

potential for the change after EMG-ES (R 2 = 0.15, p = 0.31). 3.2. Functional magnetic resonance imaging (fMRI) Regarding the activation pattern during passive wrist/finger extension, fMRI showed an increase in the spatial extent of activation in the SMC in 4 of 7 patients (patient 1, 2, 6 and 9, Table 2) after EMG-ES therapy (Fig. 2). The remaining 3 patients showed no change or a slight decrease in the number of activated voxels in SMC. In line with the clinical outcome, there were no statistically significant differences for the entire group (p = 0.22, Wilcoxon test).

3.3.1. Motor thresholds, MEP amplitudes, conduction times Seven patients were included for TMS measurements (see Table 1). Of them, 3 patients had no MEP of the affected side before and/or after treatment. At baseline, motor thresholds (RMT, AMT) were higher and MEP amplitudes were smaller in the affected hemisphere compared to the contralateral side but without statistical significance (RMT affected hem. 61.3 ± 29.1%, unaffected hem. 38.4 ± 6.3%, p = 0.08, paired t-test; AMT affected hem. 54.2 ± 31.8%, unaffected hem. 32.6 ± 5.3%, p = 0.12; maximal MEP amplitude relative to M-response: affected hem. 0.13 ± 0.17, unaffected hem. 0.29 ± 0.29, p = 0.06). CMCTs were significantly longer in the affected hemisphere (8.0 ± 2.3 ms, n = 4) compared to the contralateral side (5.0 ± 1.7 ms, n = 6, p < 0.05, unpaired t-test). No significant alteration was observed in either parameter following treatment with EMG-ES. 3.3.2. Stimulus response curves Stimulus response curves showed significantly less recruitment of the affected hemisphere (n = 4) compared to the unaffected hemisphere (n = 7). Under baseline conditions, the average maximum MEP amplitude of the affected hemisphere at 140% RMT was 0.82 mV versus 1.96 mV of the unaffected hemisphere. Although recruitment was slightly weaker after therapy on the unaffected side the difference as compared to baseline was statistically not significant (repeated mea-

194


Fig. 4. Changes in intracortical excitability following EMG-ES. The diagram displays the effect of subthreshold conditioning TMS on subsequent suprathreshold motor-evoked potentials (MEP) of the affected hemisphere before and after EMG-ES (n = 4). Different interstimulus intervals with inhibitory (ICI), neutral, or facilitating character (ICF) were applied. Data are expressed as ratio of the mean conditioned MEP to mean single pulse MEP. Values below 1.0 (dashed line) represent inhibition. Significantly more intracortical facilitation was seen after treatment with EMG-ES (p < 0.05).

peated measures ANOVA, interaction of time point by group, p = 0.03; paired t-test for ICF p = 0.03). The absolute test MEP amplitudes pre and post intervention were 0.56 ± 0.29 and 0.59 ± 0.54, respectively. In the unaffected hemisphere (n = 7) no significant changes in ICI or ICF were observed (MEP amplitude ratios at baseline: 0.55 ± 0.32 (ICI), 1.51 ± 0.6 (ICF) and post treatment: 0.49 ± 0.33, 1.21 ± 0.33).

3.4. Correlation of fMRI with TMS parameters

Fig. 5. Prediction of SMC activation change by MEP presence. The changes in SMC activation after EMG-ES were predicted by the presence or absence of motor-evoked potentials (MEPs) following TMS with R2 = 0.83, p = 0.03.

sures ANOVA, effect of side, p = 0.127; effect of side by intensity, p < 0.001, no significant effect of time point). 3.3.3. Paired pulse protocols To assess the effect of EMG-ES on intracortical excitability paired-pulse protocols were applied. Figure 4 displays the amplitude ratios of paired TMS at inhibitory (ICI), neutral, and facilitating (ICF) interstimulus intervals of the affected hemisphere (n = 4). The average MEP amplitude ratios at baseline were 0.63 ± 0.18 (ICI), 0.98 ± 0.26 (neutral), and 1.24 ± 0.36 (ICF). EMG-ES therapy led to a significant increase in ICF, whereas ICI was unaltered (MEP amplitude ratios post treatment 0.63 ± 0.29, 1.0 ± 0.23, 1.45 ± 0.32; re-

A subgroup of 5 patients underwent both fMRI and TMS (see Table 2). Three of them showed an increase in SMC activation on the affected side post treatment (pre: mean = 5.0 ± 8.66; post: mean = 21.67 ± 13.87) as well as 18 ± 24% more ICF after therapy. The remaining two patients showed no change or a slight decrease in the area of SMC activation. Interestingly, their MEPs were highly pathologic and mostly absent, which prevented determination of cortical excitability changes with paired TMS. Consistently, the presence or absence of MEPs strongly predicted the SMC activation change (R 2 = 0.83, p = 0.03; see Fig. 5). In addition, there seems to be a correlation between ICF, if determinable, and SMC activation change over the training period (r = 0.75, n = 3). However, the presence or absence of MEPs was predictive for the response to motor practice (change in Box & Block test) only in 4 of 5 patients and the regression analysis failed to reach significance (R2 = 0.21, p = 0.3).


4. Discussion In this study we tested the clinical performance as well as measures of neuroplasticity before and after 8 weeks of daily intense EMG-ES in a population of chronic hemiparetic stroke patients. Although the training intensity applied here, in particular the number of stimulation channels, combination with task orientation and number of sessions, was higher as compared to many previous studies, (i.e. Cauraugh & Kim, 2002; Jang et al., 2003; Kimberley et al., 2004; Kraft et al., 1992), the functional testing revealed only marginal improvements of arm function for the entire group. This finding can be attributed to the high level of disability in our study population. As revealed by the Box & Block test the capacity for improvement in arm function was dependent on the level of baseline performance and no benefit was observed if the score at inception was below 4. Clearly, the level of handicap in our study population was higher compared to other studies, where more benefit after EMG-ES was reported (Cauraugh & Kim, 2002; Kimberley et al., 2004; Kraft et al., 1992). It was thus not possible to improve function in severely affected patients by this technique even with a daily treatment over 8 weeks. In accordance with previous statements (Chae & Yu, 2000) our data therefore support the notion that EMG-ES is most effective for patients with mild to moderate motor impairment. EMG-ES triggered activation changes in the sensorimotor cortex during wrist joint movements were evaluated using fMRI. As reported earlier, passive hand movements, either motor or examiner driven, as well as pure sensory stimulation induce almost identical SMC activation patterns compared to active movements in healthy subjects and stroke patients (Lee et al., 1998; Tombari et al., 2004; Weiller et al., 1996). Here we evaluate fMRI scans during passive, electrically stimulated hand and finger extension. Unfortunately, scans while performing active movements had to be excluded from the analysis (see 3.2), therefore the overlap of patterns compared to electrically provoked movements could not be determined. However, due to the above mentioned reports, it has to be assumed that fMRI patterns during electrically provoked hand extension are representative for active movements. After EMG-ES treatment, there was no significant change of the size of sensorimotor representation as revealed by fMRI in the overall analysis. This missing group effect is in line with the non-significant benefit in clinical function. Nevertheless, fMRI indicated an increased SMC activation area of the affected motor

195

cortex after treatment in a subgroup of patients (n = 4), which was paralleled by an increased intracortical facilitation (ICF) as revealed by TMS (correlation coefficient r = 0.75). Most importantly, this subgroup included the 2 patients with the good clinical response. In contrast, the subgroup with no change or a decrease in SMC activation showed overall poor clinical performance and mostly absent MEPs on the affected side. The fact that clinical improvements after EMG-ES, if any, are paralleled by cortical activation and excitability changes supports the assumption that treatment-driven clinical improvements in chronic stroke patients are directly related to a reorganization at the cortical level as reported earlier (Carey et al., 2002; Koski et al., 2004; Liepert et al., 2000; Liepert et al., 1998; Platz et al., 2005b). Interestingly, the presence or absence of MEPs in transcranial magnetic stimulation on the affected side strongly predicted the SMC activation change in fMRI with an increase in SMC activation for patients with MEPs and no change or decrease for those without MEPs. This indicates that the neuroplastic potential of the affected brain area in patients without MEPs is severely limited. On the other hand, the presence/absence of MEPs was not sufficient to predict the response to motor practice. Here, the observation may be considered that the benefit from treatment was confined to patients with a large increase in SMC activation area (compare Fig. 3). Therefore a threshold might exist, above which increases in SMC activation result in clinical improvement. Such patients with present MEPs but only small SMC activation increase may therefore not benefit from treatment. In addition, other factors such as the corticospinal tract integrity may play a role in prediction of the response to motor practice as reported recently (Stinear et al., 2007). Hamzei et al. (2006) observed two different patterns of reorganization after constraint-induced movement therapy in chronic stroke patients. In their study, all subjects improved clinically, but the ipsilesional SMC activation (intensity and area) either increased in parallel with a decrease in intracortical excitability, or decreased paralleled by an increase in intracortical excitability. The group with increased SMC activation showed an affection of M1 or outgoing fibers, which is also the case for our patients. In agreement with these findings, we observe an increased SMC activation area indicative of an extension of neural activity after EMG-ES arm training. In contrast to Hamzei et al. we found this to be paralleled by increased intracortical excitability (increased ICF). Therefore, EMG-ES seems

196


to induce not only recruitment of additional sensorimotor networks but also increased synaptic efficiency. However, the comparison of results may be limited considering that, as Hamzei et al. point out, excitability measurements by TMS only represent one point of the activation area and regional differences might therefore not be detected. Irrespective of the exact mechanism, the apparent involvement of intracortical excitability changes in EMG-ES induced stroke recovery might offer the possibility to enhance the effects by methods known to alter cortical excitability levels, i.e. repetitive TMS or transcranial direct current stimulation (tDCS). These methods have recently been shown to improve arm pareses in acute or chronic stages after stroke (Fregni et al., 2005; Khedr et al., 2005; Takeuchi et al., 2005). Therefore a combination with EMG-ES for potentiation of therapeutic effects seems promising and should be subject to investigation in further studies. There are two limitations of our study, which should be acknowledged. First, there is no control group. However, this is offset by having multiple baseline measurements demonstrating a stable motor performance as well as TMS conditions before start of the arm training. Therefore, the reported changes in clinical and functional parameters are unlikely to result from spontaneous fluctuations or recovery, although this cannot be completely ruled out. Also we cannot exclude nonspecific effects of the applied therapy such as the intensified emotional involvement (placebo effect). Second, the number of patients is small, in particular those who underwent both fMRI and TMS (n = 5). In addition, only 4 patients had MEPs after TMS of the affected pathways. Although statistically significant results were obtained for subgroup analyses, the confidence in these results is therefore limited and confirmation of the results in larger samples is warranted. In conclusion, EMG-ES may improve the arm function of chronic hemiparetic stroke patients, in particular if motor deficits are mild to moderate (Box & Block test baseline score 4). If successful, the underlying mechanism seems to involve an increase in the sensorimotor cortex activation area and enhanced intracortical facilitation on the affected side. The clinical score at inception and the presence of MEPs have the best predictive potential.

the Bernstein Center for Computational Neuroscience (BCCN) Göttingen under Grant No. 01GQ0432. Competing Interests J. K. is employed by the Otto Bock Company, which provided the stimulator unit.

References Baudewig, J., Dechent, P., Merboldt, K. D., & Frahm, J. (2003). Thresholding in correlation analyses of magnetic resonance functional neuroimaging. Magn Reson Imaging, 21, 1121-30. Carey, J. R., Kimberley, T. J., Lewis, S. M., Auerbach, E. J., Dorsey, L., Rundquist, P., & Ugurbil, K. (2002). Analysis of fMRI and finger tracking training in subjects with chronic stroke. Brain, 125, 773-88. Cauraugh, J. H., & Kim, S. (2002). Two coupled motor recovery protocols are better than one: electromyogram-triggered neuromuscular stimulation and bilateral movements. Stroke, 33, 1589-94. Chae, J., & Yu, D. (2000). A critical review of neuromuscular electrical stimulation for treatment of motor dysfunction in hemiplegia. Assist Technol, 12, 33-49. Francisco, G., Chae, J., Chawla, H., Kirshblum, S., Zorowitz, R., Lewis, G., & Pang, S. (1998). Electromyogram-triggered neuromuscular stimulation for improving the arm function of acute stroke survivors: a randomized pilot study. Arch Phys Med Rehabil, 79, 570-5. Fregni, F., Boggio, P. S., Mansur, C. G., Wagner, T., Ferreira, M. J., Lima, M. C., Rigonatti, S. P., Marcolin, M. A., Freedman, S. D., Nitsche, M. A., & Pascual-Leone, A. (2005). Transcranial direct current stimulation of the unaffected hemisphere in stroke patients. Neuroreport, 16, 1551-5. Hamzei, F., Liepert, J., Dettmers, C., Weiller, C., & Rijntjes, M. (2006). Two different reorganization patterns after rehabilitative therapy: an exploratory study with fMRI and TMS. Neuroimage, 31, 710-20. Jang, S. H., Kim, Y. H., Cho, S. H., Lee, J. H., Park, J. W., & Kwon, Y. H. (2003). Cortical reorganization induced by task-oriented training in chronic hemiplegic stroke patients. Neuroreport, 14, 137-41. Khedr, E. M., Ahmed, M. A., Fathy, N., & Rothwell, J. C. (2005). Therapeutic trial of repetitive transcranial magnetic stimulation after acute ischemic stroke. Neurology, 65, 466-8. Kimberley, T. J., Lewis, S. M., Auerbach, E. J., Dorsey, L. L., Lojovich, J. M., & Carey, J. R. (2004). Electrical stimulation driving functional improvements and cortical changes in subjects with stroke. Exp Brain Res, 154, 450-60.

5. Funding

Koski, L., Mernar, T. J., & Dobkin, B. H. (2004). Immediate and long-term changes in corticomotor output in response to rehabilitation: correlation with functional improvements in chronic stroke. Neurorehabil Neural Repair, 18, 230-49.

We acknowledge financial support by the German Ministry for Education and Science (BMBF) via

Kraft, G. H., Fitts, S. S., & Hammond, M. C. (1992). Techniques to improve function of the arm and hand in chronic hemiplegia. Arch Phys Med Rehabil, 73, 220-7.

F. von Lewinski et al. / Electrostimulation in chronic stroke Lee, C. C., Jack, C. R., Jr., & Riederer, S. J. (1998). Mapping of the central sulcus with functional MR: active versus passive activation tasks. AJNR Am J Neuroradiol, 19, 847-52. Liepert, J., Bauder, H., Wolfgang, H. R., Miltner, W. H., Taub, E., & Weiller, C. (2000). Treatment-induced cortical reorganization after stroke in humans. Stroke, 31, 1210-6. Liepert, J., Miltner, W. H., Bauder, H., Sommer, M., Dettmers, C., Taub, E., & Weiller, C. (1998). Motor cortex plasticity during constraint-induced movement therapy in stroke patients. Neurosci Lett, 250, 5-8. Lyle, R. C. (1981). A performance test for assessment of upper limb function in physical rehabilitation treatment and research. Int J Rehabil Res, 4, 483-92.

197

Richards, L. G., Stewart, K. C., Woodbury, M. L., Senesac, C., & Cauraugh, J. H. (2008). Movement-dependent stroke recovery: A systematic review and meta-analysis of TMS and fMRI evidence. Neuropsychologia, 46, 3-11. Stinear, C. M., Barber, P. A., Smale, P. R., Coxon, J. P., Fleming, M. K., & Byblow, W. D. (2007). Functional potential in chronic stroke patients depends on corticospinal tract integrity. Brain, 130, 170-80. Takeuchi, N., Chuma, T., Matsuo, Y., Watanabe, I., & Ikoma, K. (2005). Repetitive transcranial magnetic stimulation of contralesional primary motor cortex improves hand function after stroke. Stroke, 36, 2681-6.

Mathiowetz, V., Volland, G., Kashman, N., & Weber, K. (1985). Adult norms for the Box and Block Test of manual dexterity. Am J Occup Ther, 39, 386-91.

Tarkka, I. M., Kononen, M., Pitkanen, K., Sivenius, J., & Mervaala, E. (2008). Alterations in cortical excitability in chronic stroke after constraint-induced movement therapy. Neurol Res, in press.

Nakamura, H., Kitagawa, H., Kawaguchi, Y., & Tsuji, H. (1997). Intracortical facilitation and inhibition after transcranial magnetic stimulation in conscious humans. J Physiol, 498(3), 817-23.

Tombari, D., Loubinoux, I., Pariente, J., Gerdelat, A., Albucher, J. F., Tardy, J., Cassol, E., & Chollet, F. (2004). A longitudinal fMRI study: in recovering and then in clinically stable subcortical stroke patients. Neuroimage, 23, 827-39.

Platz, T., Pinkowski, C., van Wijck, F., Kim, I. H., di Bella, P., & Johnson, G. (2005a). Reliability and validity of arm function assessment with standardized guidelines for the Fugl-Meyer Test, Action Research Arm Test and Box and Block Test: a multicentre study. Clin Rehabil, 19, 404-11.

Weiller, C., Juptner, M., Fellows, S., Rijntjes, M., Leonhardt, G., Kiebel, S., Muller, S., Diener, H. C., & Thilmann, A. F. (1996). Brain representation of active and passive movements. Neuroimage, 4, 105-10.

Platz, T., van Kaick, S., Moller, L., Freund, S., Winter, T., & Kim, I. H. (2005b). Impairment-oriented training and adaptive motor cortex reorganisation after stroke: a fTMS study. J Neurol, 252, 1363-71.

Ziemann, U., Rothwell, J. C., & Ridding, M. C. (1996). Interaction between intracortical inhibition and facilitation in human motor cortex. J Physiol, 496(3), 873-81.