Transcription

Kattenstroth et al. BMC Neurology (2018) 18:2DOI 10.1186/s12883-017-1006-zRESEARCH ARTICLEOpen AccessDaily repetitive sensory stimulation of theparetic hand for the treatment ofsensorimotor deficits in patients withsubacute stroke: RESET, a randomized,sham-controlled trialJan C. Kattenstroth1†, Tobias Kalisch1,2†, Matthias Sczesny-Kaiser2, Wolfgang Greulich3, Martin Tegenthoff2and Hubert R. Dinse1,2,4*AbstractBackground: Repetitive sensory stimulation (RSS) adapts the timing of stimulation protocols used in cellular studiesto induce synaptic plasticity. In healthy subjects, RSS leads to widespread sensorimotor cortical reorganizationparalleled by improved sensorimotor behavior. Here, we investigated whether RSS reduces sensorimotor upper limbimpairment in patients with subacute stroke more effectively than conventional therapy.Methods: A single-blinded sham-controlled clinical trial assessed the effectiveness of RSS in treating sensorimotordeficits of the upper limbs. Patients with subacute unilateral ischemic stroke were randomly assigned to receivestandard therapy in combination with RSS or with sham RSS. Patients were masked to treatment allocation. RSSconsisted of intermittent 20 Hz electrical stimulation applied on the affected hand for 45 min/day, 5 days per week,for 2 weeks, and was transmitted using custom-made stimulation-gloves with built-in electrodes contacting eachfingertip separately. Before and after the intervention, we assessed light-touch and tactile discrimination,proprioception, dexterity, grip force, and subtasks of the Jebsen Taylor hand-function test for the non-affected andthe affected hand. Data from these quantitative tests were combined into a total performance index serving asprimary outcome measure. In addition, tolerability and side effects of RSS intervention were recorded.Results: Seventy one eligible patients were enrolled and randomly assigned to receive RSS treatment (n 35) orsham RSS (n 36). Data of 25 patients were not completed because they were transferred to another hospital, resultingin n 23 for each group. Before treatment, sensorimotor performance between groups was balanced (p 0.237). After2 weeks of the intervention, patients in the group receiving standard therapy with RSS showed significantly betterrestored sensorimotor function than the control group (standardized mean difference 0.57; 95% CI -0.013–1.16;p 0.027) RSS treatment was superior in all domains tested. Repetitive sensory stimulation was well tolerated andaccepted, and no adverse events were observed.(Continued on next page)* Correspondence: [email protected]†Equal contributors1Institute for Neuroinformatik, Neural Plasticity Lab, Ruhr-University ofBochum, Bochum, Germany2Department of Neurology, University Hospital Bergmannsheil,Ruhr-University Bochum, Bochum, GermanyFull list of author information is available at the end of the article The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication o/1.0/) applies to the data made available in this article, unless otherwise stated.

Kattenstroth et al. BMC Neurology (2018) 18:2Page 2 of 13(Continued from previous page)Conclusions: Rehabilitation including RSS enhanced sensorimotor recovery more effectively than standard therapyalone. Rehabilitation outcome between the effects of RSS and standard therapy was largest for sensory and motorimprovement; however, the results for proprioception and everyday tasks were encouraging warranting further studiesin more severe patients.Trial registration: The trial was retrospectively registered January 31, 2012 under DRKS00003515 (https://www.drks.de/drks web/navigate.do;jsessionid AEE2585CCB82A22A2B285470B37C47C8?navigationId results).Keywords: Neurorehabilitation, Neuroplasticity, Sensorimotor, Stroke, Repetitive sensory stimulationBackgroundSensorimotor impairment resulting from cerebral dysfunction has substantial physical, psychological, and socialimplications. However, according to a 2014 Cochrane review, no high-quality evidence is available to support interventions currently used as part of routine practice [1].Triggered by substantial progress in understanding the neuroplasticity mechanisms underlying learning and rehabilitation [2, 3], numerous alternative strategies have beensuggested and tested, many of which showed a moderatequality of evidence such as constraint-induced movementtherapy, robot-assisted therapy, mirror therapy, central andperipheral nerve stimulation, and virtual reality approaches[1, 3–6]. There is growing evidence that high doses of intervention are more beneficial than low doses. However, rehabilitation outcome is often limited [1, 7].Recent work in healthy human subjects demonstratedthat intensive training may not be necessary to inducebehavioral improvement; however, it can be effectivelyacquired using a complementary approach in whichplasticity processes are driven in response to exposure torepetitive sensory stimulation (RSS) [8, 9]. RSS is an approach that targets the cortical areas that represent thesite of sensory stimulation to facilitate the developmentof neuroplastic processes. For example, a few hours ofRSS in healthy participants has been demonstrated toalter cortical maps and cortical excitability representingthe sites of the RSS stimulation [10, 11]. This becomespossible through the use of long-term potentiation-likesensory stimulation protocols [12, 13].Despite different backgrounds and rationales foruse, the concept of sensory stimulation protocols toinduce neuroplasticity processes has attracted substantial interest and is currently being investigatedand applied in many laboratories; however, differentlaboratories use different terms and different stimulation protocols [9, 13–19]. From our perspective, therationale behind RSS is using the broad knowledge ofbrain plasticity to design specific sensory stimulationprotocols in humans to induce synaptic plasticity toalter perception and behavior. The concept is totranslate protocols that induce plasticity at the cellular level, such as long-term potentiation (LTP) andlong-term depression (LTD), into sensory stimulationprotocols [8, 20]. Central to using RSS is its ability todrive and facilitate neuroplasticity processes [2, 3], aproperty shared by central stimulation methods suchas intracortical microstimulation, transcranial directcurrent stimulation, and transcranial magnetic stimulation [21–24].In cellular research, high-frequency stimulation is used toinduce LTP, whereas low-frequency stimulation evokesLTD [25, 26]. In the present study, we used a LTP-likeprotocol consisting of intermittent high-frequency tactilestimulation, which has been used before in healthy subjectsto drive improvements in tactile perceptual abilities [12]parallel to cortical reorganization. To explain the changesevoked by RSS, this specific form of stimulation was suggested to evoke LTP-like plasticity processes in the corticalregions representing the stimulated skin sites [8, 27]. As aresult, synaptic transmission is altered and cortical processing is remodeled, facilitating the reactivation of the corticaltissue that has preserved some functionality. The behavioraloutcome of these processes is reflected in behavioralrecovery. Evidence from studies in healthy subjects demonstrated far-reaching cortical remodeling including changesin cortical excitability, expansion of cortical representationalareas, and enhanced functional connectivity betweenthe somatosensory and motor cortex [10, 11, 28–30].Accordingly, the background behind RSS as used in thisstudy differs from electrical mesh-glove stimulation orelectrical therapeutic stimulation, although these procedures have been reported to have a beneficial outcome oncortical excitability or muscular strength. In contrast towhole hand stimulation [31, 32], stimulating the tips ofthe fingers, which are the most densely innervated, allowsa very specific targeting of somatosensory cortical representational areas. In fact, available imaging and EEG (electroencephalography) data from healthy participantsprovide supporting evidence that RSS selectively activatesareas in somatosensory and motor areas representing thefingers and the hand [10, 11, 28, 33].Various forms of electrical stimulation exist thatare currently used in rehabilitation with mixedresults [1, 34–38]. For each of these approaches, a wide rangeof stimulation parameters are in use, and the underlying

Kattenstroth et al. BMC Neurology (2018) 18:2mechanisms mediating beneficial effects remain largely to beclarified. FES (functional electrical stimulation) is applied toinduce contraction of muscles to support motoraction. On the other hand, so-called therapeutic electrical stimulation methods are applied to improve performance after the termination of stimulation, such asNMES (neuromuscular electrical stimulation), EMG(electromyography) -triggered electrical stimulation(EMG-ES), and TENS (transcutaneous electrical nervestimulation). While TENS was introduced for paintreatment, effects observed after NMES and EMG-ESare assumed to be related to repetitive muscle contractions. Accordingly, as described above, the underlying principle and the aimed target of RSS differsfrom that of FES, NMES, or TENS.The effects of RSS protocols have been extensively explored in both healthy young and elderly adults. Whenapplied to the fingers, substantial improvements in thetactile, haptic, proprioceptive, and sensorimotor performance parallel to cortical reorganization were demonstrated [10, 12, 14, 15, 28, 39]. The effectiveness ofthis form of learning is assumed to arise from usingstimulation protocols optimized to alter synaptictransmission and efficacy [8, 9]. While a number ofstudies used RSS or variants of this approach in patients with subacute or chronic stroke, RSS has yet tobe implemented in routine clinical practice. So far,available studies indicate mixed effects in the treatment of upper limb impairment [16–18, 27, 39–43].These mixed results of sensory stimulation come fromthe fact that sensory stimulation as method is poorlydefined. Sensory stimulation approaches employ various forms of stimulation and testing parameters, aswell as quite varying treatment times ranging fromsingle applications to long-term treatment. This almost certainly causes large variability in outcome parameters. From our view, a significant advantage ofrepetitive stimulation is its passive nature, which doesnot require active subject participation, making theintervention substantially easier to implement andmore acceptable to the individual.Therefore, here we aimed to address whether in routine clinical practice, repetitive stimulation reduces sensorimotor deficits following stroke more effectively thanconventional therapy. Based on our previous data aboutthe effectiveness of RSS [10, 14, 15, 28, 39–41], we hypothesized that the standard therapy with RSS is superior to standard therapy with sham RSS. Because strokecan affect diverse aspects of sensorimotor abilities, weundertook a broad objective behavioral assessmentevaluating the sensory, proprioceptive, sensorimotor, andmotor functions. In addition, we aimed to find out tolerability, acceptance, and possible side-effects of the gloveapplied RSS intervention.Page 3 of 13MethodsStudy designThis randomized single-blinded sham-controlled clinicaltrial was designed as a proof-of-concept study evaluatingthe effectiveness, safety, and compliance. We recruited 71patients with subacute ischemic stroke with contralateralsensorimotor impairment from the HELIOS rehabilitationclinic in Hagen Ambrock, Germany. The study was donein accordance with the ethical principles from the Declaration of Helsinki. It was approved by the Ethics Committee of Ruhr-University of Bochum. All patients providedverbal and written informed consent before participating.The trial was retrospectively registered January 31, 2012under DRKS00003515 (https://www.drks.de/drks web/navigate.do;jsessionid AEE2585CCB82A22A2B285470B37C47C8?navigationId results). Recruitment had startedJune 15, 2010, lasting until December 31, 2012).PatientsThe main inclusion criteria were age of 40–70 years, adiagnosis of unilateral subacute ischemic stroke, i.e., aleft or right medial cerebral artery infarction withcontralateral sensorimotor deficits of the upper limbs 3to 4 weeks post-ictus. Also, patients should have lowlevels of spasticity, and stimulation perception thresholds of at least 20 mA. All patients were right-handed.Patients with mild transient ischemic stroke lastingfewer than 24 h, hemorrhagic stroke, and carotid arterydissection, history of cerebrovascular disease, wearing apacemaker, aphasia, or cognitive impairment thatprevented completion of the assessment were excluded.Patients were recruited with the help of physio- andoccupational therapists at the rehabilitation clinic. Themost relevant criteria patients did not meet were spasticity and paresis. Because of difficulties with patient enrollment, we widened the age criterion to patients aged30 to 90 years.Randomization and maskingPatients were randomly assigned to either the target orcontrol group using block randomization. The selectionwas performed using a computer-generated random listof numbers. Randomization sequence was accessible toassessors of the main outcome parameters only. Noattempts were made to balance for gender or the side ofthe stroke. The patients and assessors of clinical tests(modified Rankin Scale, National Institutes of HealthStroke Scale, modified Barthel Index, Medical ResearchCouncil Scale, Frenchay Arm Test, and Wolf MotorFunction Test) were masked to the treatment allocation. There was no masking for assessors of the mainoutcome parameters. Individual RSS intervention wasconducted and monitored by therapists, who hadreceived a detailed instruction in handling RSS

Kattenstroth et al. BMC Neurology (2018) 18:2Page 4 of 13procedures. Table 1 shows the baseline characteristicsof the study patients.ProceduresInterventions were commenced 3.5 weeks (median) afterstroke. Patients allocated to the control group receivedthe same standard stroke rehabilitation as patients allocated to the intervention (RSS) group. Patients receivedRSS with standard rehabilitation therapy (physio- andergotherapy) or sham RSS with standard therapy (control). In case of sham RSS, patients assumed a subthreshold treatment, although zero mA was applied. RSStreatment was applied independent of the schedule ofthe standard therapy. Total RSS-treatment time for bothgroups was 2 weeks (10 days).InterventionsRSS and sham RSS were applied for 45 min daily on theaffected hand of the patients from both treatmentgroups. The stimulation sequence was the same as described previously [12] and consisted of 20-Hz bursts for1.4 s with 5-s inter-train intervals, and a ramp/fall timeof 0.3 s and 0.2 ms pulse width. The pulse trains weredelivered with 2-channel stimulation devices. To accountfor innervation of the fingers, the stimulation for thepredominantly median nerve-innervated fingers d1-d3(the thumb, index, and middle finger) and the predominantly ulnar nerve-innervated fingers d4 and d5 (ringand little finger) was separately controlled and delivered.The pulses were transmitted by custom-made stimulation gloves that had built-in electrodes (1 4 cm)Table 1 Demographic and baseline characteristicslocated on the first and third segment of each finger(cathode proximal). In the RSS group, the intensity ofthe stimulation was set individually at the highest valuesthat the patient could easily tolerate for an extendedperiod, but without reaching pain levels. For sham RSS,the same stimulation parameters and the same stimulation gloves were used, except for the stimulation intensity, which was set at zero mA.The standard rehabilitation therapy consisted of individualized programs depending on the degree and natureof patients’ sensorimotor impairment. Occupationaltherapy was applied according to the concepts ofBobath, Affolter, and Perfetti. ADL training (activities ofdaily living) consisted of self-care tasks such as bathingand showering, dressing, self-feeding, food preparation,and personal hygiene. Cognitive therapy and activationtraining were offered to facilitate the recovery for independent living. Special and curative education includedpottery-making to foster the skills and abilities of patients,and to prepare them for coping with confinement.To characterize the upper extremity performance ofthe patients prior to the study, we used several clinicalscales closely related to the performance of tasks ineveryday life: the modified Rankin Scale (ranging 0 to 6,0 no symptoms), National Institutes of Health StrokeScale (ranging 0 to 42, 0 no symptoms), modified BarthelIndex (ranging 0 to 100, 100 no symptoms), MedicalResearch Council Scale (ranging 0 to 5, 5 no symptoms),Frenchay Arm Test (ranging 0 to 5, 5 no symptoms),and 15 tasks of the Wolf Motor Function Test (ranging0 to 15, 15 no symptoms). According to these scales, patients were characterized by slight to moderate to severemoderate disability. All patients had some sensory lossof various degrees.RSS group (n 23)Sham RSS group (n 23) Male18 (78%)16 (70%) Female5 (22%)7 (30%)Age (years)64 (34–86)59 (43–89)White ethnic origin2323Lesion side (Rt./Lt.)9/1413/10Handedness (R/L/A)23//23//mRS3.35 0.713.26 0.75NIHSS5.48 3.225.17 2.23mBI64.8 34.5966.5 34.23Tactile performanceMRCA3.73 2.683.98 1.49FAT3.48 6.193.50 5.99WMTF11.76 16.4311.56 14.91Touch thresholds were evaluated by probing the fingertipsof the left and right index finger with von Frey filaments(Marstocknervtest, Marburg, Germany) [27, 44]. The testkit contained 16 different filaments calibrated to forcesranging from 0.25–294 mN in the logarithmic scale. Weused a staircase procedure during which patients were required to close their eyes and report when they perceivedan indentation of the skin on their fingertips. The appliedSexdepicts demographic and baseline characteristics. Data are mean ( SD) ornumber (%). Abbreviations: Lesion side (Rt./Lt.) right hemisphere/lefthemisphere; Handedness (R/L/A) right handed, left handed, ambidextrous;mRS modified Rankin Scale; NIHSS National Institutes of Health StrokeScale; mBI modified Barthel Index; MRCA Medical Research Council Scale;FAT Frenchay Arm Test; WMTF Wolf Motor Function TestObjective assessment of sensorimotor behaviorTo quantify the behavior objectively, the following testswere performed before and after the end of treatmentfor the affected and the non-affected limbs. The termnon-affected is meant to indicate the limb contralateralto the site of stroke, which does not exclude potentialchanges in performance. Testing was performed in onesession with breaks in between, total time varied between 60 and 90 min.

Kattenstroth et al. BMC Neurology (2018) 18:2forces, starting with a noticeable stimulus, were decreasedin a stepwise manner until the subjects no longer perceived the stimulus (lower boundary) and then increaseduntil the stimulus was perceived again (upper boundary).This procedure was repeated 5 times resulting in 10 valuesthat were averaged to provide the touch threshold.The grating orientation task (GOT) was tested in 2 alternative forced-choice paradigms [27, 45]. A set of ninecustom-made hemispherical plastic domes with gratingscut into their surfaces, i.e., parallel ridges and grooves ofequal widths for each dome, were applied to the tip ofthe index finger using a holder with a calibrated spring(150 mN) to enable constant application force. Thewidth of the ridges and grooves (spatial frequency) varied from 0.5 to 9.5 mm. Each dome was presented 20times. Immediately after touching the plastic domes, patients were asked to report the perceived orientation.The grating discrimination threshold was defined as thelevel at which 75% of the responses were correct and wasdetermined by interpolating between the groove widthswith 75% correct responses. The performance at this levelwas midway between chance and perfect performance.Nine-hole peg test (9-HPT)To measure upper extremity fine motor performance(dexterity), we used the 9-HPT, a brief standardizedquantitative test of the upper extremity function [46].We measured the time needed separately for placing thepegs in and out of holes.Grip strengthGrip strength was measured 3 consecutive times foreach hand with a Jamar hand dynamometer (SammonsPreston Inc., Bolingbrook, IL). Subjects were asked tostand up and hold the dynamometer with the arms parallel to the body [47]. The final results were the averageacross 3 trials.Jebsen-Taylor hand function test (JTHFT)For the assessment of the functional hand motor skills,we used JTHFT [48]. Three of the 7 JTHFT subtestswere performed: (1) picking up small objects and placingthem in a can (SOP); (2) picking up small objects with ateaspoon and placing them in a can (FEED); and (3)stacking checkers (STACK). The performance was evaluated based on the time needed to complete each subtest.Joint position sense (JPS) assessmentThe JPS assessment was conducted using the “BochumJoint Position Sense Assessment” (BJPSA) as reportedpreviously [39]. Patients were asked to compare lightweight polystyrene balls of different diameters held inthe affected hand to a reference ball held in their nonaffected hand, and to report, without visual information,Page 5 of 13if the tested ball located in the affected hand was larger,smaller, or equal in volume. In 3 consecutive subtests,the complete set of polystyrene balls (diameters: 3 cm,5 cm, 6 cm, 7 cm, 8 cm, 10 cm, and 12 cm) were compared to a small reference (diameter 5 cm), mid-sizedreference (diameter 7 cm), and large-sized reference(diameter 12 cm). The performance was assessed by calculating the number of errors (ERRnumb, a total of 21decisions) and the weight of errors (ERRweight, calculated as the volume difference between the referenceand test object). Previous studies showed that ERRweight and ERRnumb are independent parameters characterizing the joint position sense [39].OutcomesThe primary outcome was the sensorimotor performance (total performance index – TPI) as obtained by acombination of 10 different quantitative objective testsshown in Tables 3 and 4, which were performed beforeand after the end of treatment for both the affected andthe non-affected limb: touch thresholds, acuity threshold(grating orientation task thresholds), dexterity using 9HPT, grip strength, and proprioceptive functions according to JPS assessment and subtests of JTHFT: picking upsmall objects and placing them in a can (SOP), pickingup small objects with a teaspoon and placing them in acan (FEED), and stacking checkers (STACK).To compare and average performances across all testsand all subjects in both groups, we calculated the normalized performance indices (IP) [49]. This approachhas been used by Engeneer et al. some years ago to combine performance data across different tasks that havedifferent dimensions which cannot be averaged [50]. Inaddition, pooling the performance obtained for each single task into a single total performance index increasesstatistical power. IPs were calculated for each subjectand each test. IP was calculated using the formula(wp-ip)/(wp-bp), where for a given test independent of aspecific time point of measurement wp is the worst performance of all subjects, ip is the individual performance, and bp is the best performance of all subjects. IPsranged between 1 and 0, where the best IP was 1, andthe worst was 0. The total performance index (TPI) wascalculated by averaging IP data across all 10 testsperformed.The secondary outcome measures were the performances in the 4 domains covering similar functional domains. For the “Sensory” domain, IP data from the touchthreshold and 2-point discrimination tests were averaged. For the “Motor” domain, the IP data from the gripstrength and the 9-HPT were averaged. For the “Proprioception” domain, IP data from the number andweight of errors of the JPS tests were averaged. For the

Kattenstroth et al. BMC Neurology (2018) 18:2“Everyday” domain, IP data from the 3 JTHFT subtestswere averaged.The tertiary outcomes were self-reported assessmentsof patients from both groups on adverse effects were administered after the completion of treatment includingthe questions: “was RSS pleasant or unpleasant?” and“what was the type of sensation evoked by RSS?” Theuser feedback obtained by a custom-made questionnaireincluded the ease of use, perceived sensations duringRSS, positive and negative aspects of RSS, willing to continue using after release from the hospital, and willing torecommend to others.Post-assessments were conducted within the week directly after completing the 10 days of treatments (mean2.9 1.4 days).Statistical analysisData were checked for normal distribution usingShapiro-Wilk test. Descriptive statistics were compiled.Normally distributed data were reported as the meanand standard deviation (SD). We used Student’s t-test todetect the differences between the 2 groups after theintervention. For statistical evaluation of the demographic and baseline characteristics, we used Student’s tand Chi-square tests. Moreover, we computed effectsizes according to Cohen’s d, and confidence intervals.As early intervention with RSS following a stroke hasnot yet been tested, sample size and power calculationwas therefore based on effect sizes in the range of 0.4 to0.8 obtained in previous studies on sensory and dexterityperformance [27], unpublished data, as well as data fromelderly age-matched healthy participants [51], whichusing an alpha of .05 and a power of 80% resulted in asample size of 32 per group. To accommodate theFig. 1 Consort flow diagramPage 6 of 13possible drop outs, we selected a sample size of 70. Allcalculations were performed using Microsoft Excel 2010and SPSS version 18.0 and higher.ResultsParticipantsWe screened 143 patients with ischemic stroke for eligibility. Seventy-one of these patients were enrolled andrandomly assigned to either the standard therapy withRSS (35 patients) or the control group receiving standard therapy with sham RSS (36 patients; see Fig. 1).Seventy-two were excluded (54 not meeting inclusioncriteria, 18 refused to participate). In the RSS group, 23of 35 patients completed the treatment; a similar proportion (23 of 36 patients) completed the treatment inthe sham RSS group. Twelve (RSS group) and 13 patients (sham RSS group) did not complete their assignedtreatment because they were transferred to another hospital or daycare. This transfer was solely caused byorganizational reasons or personal reasons of the patients. The patients transferred showed no particularitiesand were within the range documented for the successfully treated patients according to the age, type of infarct, and severity of behavioral impairment. The data of46 patients (n 23 each group) were assessed for furtherstatistical analysis. All patients received all interventionsper protocol.The baseline demographic and clinical characteristicswere not statistically significantly different in bothgroups (Table 1). Both groups were balanced for age andgender. Most importantly, prior to the intervention,there were no differences in the clinical tests: the modified Rankin Scale, National Institutes of Health StrokeScale, modified Barthel Index, Medical Research Council

Kattenstroth et al. BMC Neurology (2018) 18:2Scale, Frenchay Arm Test, and Wolf Motor FunctionTest. Furthermore, no significant differences were foundin the averaged total performance at baseline.Stimulation-glove: Acceptance and side effectsThe glove-based RSS treatment was well tolerated, andno negative side-effects were recorded. The averagestimulation intensity was 10.4 3.87 mA for the mediannerve-innervated fingers, and 6.7 3.46 mA for theulnar nerve-innervated fingers. The sensations perceivedduring RSS were rated neutral to pleasant. The ease ofuse was reported as uncomplicated, possibly because inall cases, therapists assisted putting on the stimulationgloves. About two-third of the patients group were interested in continuing using RSS glove-treatment later intheir homes.Average sensorimotor performance: Total performancescoreAs the first step in providing an overview of the effectsof RSS versus sham RSS, we calculated the total performance index based on the performance of all 10 testsapplied. An IP of 0 indicates the worst performance,characterized by the patient not being able to performthe task. An IP of 1 indicates the best performance observed for the non-affected limb. Accordingly, an IP of0.5 indicates that the performance of the patient was atabout 50%. The patients in both groups showed improvement; however, the beneficial effects were 2-foldhigher in the RSS group compared to the sham RSSgroup (RSS: 22.9% gain from 0.56 to 0.68, p 0.00005;sham RSS: 10.0% gain from 0.64 to 0.70, p 0.05; thepre-post differences were significant between bothgroups using Student’s t-test; one-sided p 0.027). While17 out of 23 patients (74%) in the RSS group improvedtheir performance by more than 0.05 points on the TPIscale, only 10 out of 23 patients (43%) in the sham RSSgroup showed such an improvement. Calculating theStandardized Mean Difference (SMD) revealed an effectsize of 0.54 for RSS versus sham RSS therapy. The differential effectiveness was also apparent in the effect sizecalculated separately for each group, wh

therapy, robot-assisted therapy, mirror therapy, central and peripheral nerve stimulation, and virtual reality approaches [1, 3–6]. There is growing evidence that high doses of inter-vention are more beneficial than low doses. However, re-habilitation outcome is often limited [