Evidence of Neuroplasticity: A Robotic Hand Exoskeleton Study for Post-Stroke Rehabilitation

Background: A novel electromechanical robotic-exoskeleton was designed in-house for rehabilitation of wrist joint and Metacarpophalangeal (MCP) joint. Objective: The objective was to compare the rehabilitation effectiveness (clinical-scales and neurophysiological-measures) of robotic-therapy training-sessions with dose-matched control in patients with stroke. Methods: An observational pilot study was designed with patients within 2 years of chronicity. Patients received an intervention of 20 sessions of 45-minutes each, ve days a week for four-weeks) in Robotic-therapy Group (RG) (n=12) and conventional upper-limb rehabilitation in Control-Group (CG) (n=11). Clinical-scales– Modied Ashworth Scale, Active Range of Motion, Barthel-Index, Brunstrom-stage and Fugl-Meyer scale (Shoulder/Elbow and Wrist/Hand component), and neurophysiological-measures of cortical-excitability (using Transcranial Magnetic Stimulation) –Motor Evoked Potential and Resting Motor-threshold, were acquired pre and post-therapy. Results: RG and CG showed signicant improvement in all clinical motor-outcomes (p<0.05) except Modied Ashworth Scale in CG. RG showed signicantly higher improvement over CG in Modied Ashworth Scale, Active Range of Motion and Fugl-Meyer (FM) scale and FM Wrist-/Hand component) (p<0.05). Increase in cortical-excitability in ipsilesional-hemisphere was found to be statistically signicant in RG over CG, as indexed by decrease in Resting Motor-Threshold and increase in amplitude of Motor Evoked Potential (p<0.05). No signicant changes were shown by the contralesional-hemisphere. Interhemispheric RMT-asymmetry evidenced signicant changes in RG over CG (p<0.05) indicating increased cortical-excitability in ipsilesional-hemisphere along with interhemispheric changes. Conclusion: Neurophysiological-changes in RG could be most likely a consequence of plastic-reorganization and use-dependent plasticity. Robotic-exoskeleton training could signicantly improve motor-outcomes and cortical-excitability in patients with stroke. improvement in quantitative motor clinical-outcomes in patients with chronic stroke (19). however, CG showed a decrease from 1.33±0.30 to 1.30±0.28 (mean decrease of 0.03, p=0.59), indicating a trend of normalization of RMT-asymmetry (RMT asymm should decrease as ipsilsesional RMT should be decreased from pre-to-post) over the duration of intervention in RG. RG also manifested statistically signicant changes in intergroup-comparison over CG (p=0.028). The relative change in interhemispheric-RMT asymmetry-ratio ( ∆ RMTasymm-ratio) changed with RG having a mean increase of 0.12±0.14 and CG a mean increase of 0.011±0.1 (p=0.028), indicating the extent of normalization of RMT-asymmetry over the duration of intervention in RG as compared to CG 100%) (30),(31),(32),(33),(34). In those patients with no MEP recorded, RMT is taken as value of 100, as suggested in literature (35),(36). This study investigated specic impact of therapeutic-interventions on TMS neurophysiological-parameters for cortical-pathways in the context of functional-gains of the hand motor-function. TMS neurophysiological-measures (improvement in RMT and normalization) with functional improvement (35),(50),(51). This also suggests the usefulness of TMS-measures as an index of recovery in the ipsilesional-hemisphere (35). These neurophysiological-measures were obtained specically from cortical-representation of performance-feedback, counteracting exor-hypertonia against gravity, mimicking functional motion, user-friendly, simple design and patient-specic customizable features with different amount of sensory inputs (proprioceptive, visual, tactile) (19). In combination with the above features, the other unique feature of the device is that it allows the facilitation of a specic-pattern of movements mirroring complex inter-joint coordination of hand with a patient-specic impairment, currently not integrated in the available other devices with isolated-joint movements (52). The devices were able to simulate the movement pattern maintaining joint-coordination, especially at the distal-joints which could aid in translating the motor-improvements into ADL.


Introduction
Stroke is one of the leading causes of mortality and morbidity worldwide (1). The ability to actively initiate extension movements at wrist and ngers against exor-hypertonia is one of the key indicators of motor recovery (2), (3). Regaining hand-function and Activities of daily-living (ADL) is particularly impervious to therapy or rehabilitation pertaining to the complexity of motor-control needed for distal-joints (4). Conventional rehabilitation-therapy is time taking, labour-intensive and subjective, which with high clinical-load and absence of skilled resources gets di cult for the present medical and healthcare-system to provide appropriate or effective rehabilitation services (5).
Although rehabilitation with neuro-rehabilitation robots has shown encouraging clinical-results (5,6,15,(7)(8)(9)(10)(11)(12)(13)(14), it is currently limited to a very few hospitals and not widely used because of associated high-cost and an infrastructural-requirement to station, size, complexity, set-up time, safety and usability restricting its success (16), (17),(18). Rehabilitation-strategies need to take into account the multifaceted nature of disability, which itself changes with time elapsed post-stroke and address with a multimodal-approach. Hence, the device needs to be exible enough to accommodate a large patient-population. An effective rehabilitation device for hand should be able to facilitate a speci c pattern of movements mirroring complex inter-joint coordination of hand with a patient-speci c impairment, currently not integrated by the available devices.
In our previous work, we designed a robotic-hand exoskeleton for rehabilitation of the wrist and MCP (Metcarpo-phallengeal) joint, to synchronize wrist-extension with nger-exion and wrist-exion with nger-extension, mimicking ADL (19). With simple and easy-to-operate exoskeleton for low-resource settings, the exoskeleton targets spasticity through a synergy-based rehabilitation approach while also maintaining patient-initiated therapy through residual muscle-activity for maximizing voluntary effort. The lightweight and portable device has shown evidence of improvement in quantitative motor clinical-outcomes in patients with chronic stroke (19).
The aim of the present study was twofold. The rst objective was to assess the clinical effectiveness of the novel robotic-exoskeleton device (19) and the second is comparison of its clinical-effectiveness with conventional upper-limb rehabilitation. We hypothesized that the exoskeleton could show higher improvement of distal-function and cortical-excitability in patients with stroke as compared to conventional-rehabilitation.

Materials And Methods
More than 300 patients (n > 300) were screened in the out-patient clinic of the Department of Neurology, AIIMS, New-Delhi over 3 years from July-2016 to January-2019. Stroke diagnosis was established clinically in all patients. All clinical-assessments and standard-care were given to the patients with stroke by a trained physiotherapist.

Patient Enrolment
Twenty-seven patients (n = 27) were enrolled for the therapy and twenty-three patients (n = 23) completed the therapy successfully (all righthanded patients with stroke, age = 41.9 ± 11.1 years, Male:Female = 19:4), details in Figure-1 and

Cortical-excitability measures using TMS
Patients were allowed to sit comfortability in the chair, kept forearm pronated, elbow-joint at 90-120° exion, wrist-joint at a neutral position and ngers at rest. Single-pulse Transcranial Magnetic Stimulation (TMS) was given to evoke the Motor Evoked Potential (MEP) signal, using a at 70 mm gure-of-eight coil (type D70 (AC), serial no. 0326, Magstim Rapid2, Magstim, UK), at the cortical-representation of the Extensor Digitorum Communis (EDC) muscle (on contralateral motor-cortex with reference to the EEG cap) of the ipsilesional and contralesional-hemisphere.
Cortical-excitability was measured in terms of Resting motor-threshold (RMT) and Motor Evoked Potential (MEP)-amplitude using TMS over ipsilesional and contralesional-hemisphere according to the standard-protocol (20). Resting Motor-threshold was de ned as the minimum intensity of TMS required to elicit a MEP in target contralateral-muscle in 5/10 trials, recorded in EMG, over the muscle cortical-representation in the primary motor-cortex. MEP encapsulates information relevant to the cortical-excitability of the brain, conduction, and functional-integrity of the corticospinal-tract (21). MEP should be ≥ 50µv peak-to-peak amplitude at the hotspot in 5/10 consecutive trials.

Robotic therapy-sessions
An electromechanical robotic-exoskeleton was developed for rehabilitation of wrist-joint and ngers-joint (19) (Figure-2). It synchronizes wristextension with nger-exion and wrist-exion with nger-extension, a common pattern encountered in ADL. Device is actively-initiated by Electromyogram (EMG) and provides interactive adaptive performance-biofeedback in real-time. Device was safe, user-friendly and patientcentric with customizable motion-parameters (as per the clinical-presentation: i) initial-position for range of motion (ROM), ii) nal-position for ROM, iii) speed, iv) residual muscle-activity and v) height of nger-support. All sessions were given at the hospital set-up under the supervision of an expert clinician. Therapy-protocol consisted of total 20 sessions, each of 45-min over 4-weeks. Each 45-minute robotic-therapy session consisted of approximately 250-trials of 10-seconds each. Patients were advised to take a 5-min break for rest in between the therapy if there is a feeling of pain or fatigue. For details on exoskeleton design and protocol, please refer (19).

Data analysis
Data analysis was done in MATLAB R2018a (MATHWORKS). The data were tested for normality using Shapiro-Wilk test and was found that clinical-measures were not normally distributed in CG. Hence, non-parametric Wilcoxon signed-rank were used for intragroup-comparison of differences in post-pre-therapy within the group and non-parametric Man-Whitney tests were used for intergroup-comparison of RG and CG. Interhemispheric-asymmetry for pre and post-therapy measures were calculated and was tested using Wilcoxon signed-rank test. Two-way repeated measure ANOVA was applied to assess the effect of time (two levels-pre and post) and side (two levels-ipsilesional and contralesional) on RMT. Regression and correlation-analyses were performed to investigate the relationship of recovery parameters TMS neurophysiologicalparameter with clinical-outcome. A p-value < 0.05 was considered as signi cant. MAS score of 1, 1+, 2, 3, 4 was mapped as 1, 1.5, 2, 3, 4 for all statistical calculation purpose, respectively as by Rong et. al (13).

Results
All patients well tolerated the robotic therapy-sessions and completed in 30-34 days. There were no differences in the pre-therapy measures in terms of clinical-scales and MEP signal among both the intervention-groups (p > 0.05). One patient (n=1) in RG and three patients (n=3) in CG could not complete the therapy, thus data were excluded from further analysis. At pre-therapy measurements, MEP was evoked only for 9 patients (RG=4, CG=5) out of total 23 in ipsilesional-hemisphere, and for all patients in contralesional-hemisphere. FMU/L scores measure sensorimotor-control gain in both groups after the intervention. FMU/L for RG changed from 36±7.78 to 50.25±6.59 (p=0.0004) and from 37.45±9.1 to 45.45±9.7 for CG (p=0.0009). RG manifested statistically signi cant improvement in sensorimotor-scores as compared to CG with signi cant differences in intergroup-comparison (p=0.039) ( Table-2). For the proximal-part-Shoulder/Elbow component of FMS/E, both groups showed statistically signi cant increase, RG changing from 26.2±5.6 to 33.5±3.8 (p=0.0009) and from 26±7.07 to 29.8±7.08 in CG (p=0.002). However, the intergroup-comparison did not show any signi cant differences (p=0.13). For the distal-part Wrist/Hand component of FM (FMW/H), both groups showed statistically signi cant increase, in RG changing from 9.7±2.7 to 16.6±4.3 (p=0.0004) and in CG changing from 11.45±2.9 to 15.18±3.6 (p=0.0009). RG manifested statistically signi cant sensorimotor-improvement in intergroupcomparison over CG (p=0.012) ( Table-2). In this study, ~54% of patients in CG did not evoke MEP. Approximately ~67% of patients in RG too did not evoke MEP at the pre-therapy measurements, most of which later evoked post-therapy showing the therapeutic-effectiveness of the exoskeleton. It is worth-noting that in RG, measurable MEP was evoked only in 4/12 patients at the pre-therapy measurements, however, post-therapy MEP was observed in 9/12 patients, indicating a considerable increase in cortical-excitability in 5 patients. In CG, MEP was evoked only in 5/11 patients at the pre-therapy measurements and was observed in 6/11 patients post-therapy.

Contralesional-hemisphere
There were no signi cant changes shown by the contralesional-hemisphere. Both RG and CG evidenced minimal differences in RMT (mean increase of ~2% in both groups) (

Inter-hemispheric differences and asymmetries
The effect of robotic-exoskeleton training on cortical-excitability was assessed within both hemispheres. RG showed statistically signi cant differences between ipsilesional and contralesional-sides as one factor and time points-pre and post-therapy as another factor on RMT (p=0.049, F=4.08), evidencing the dependence of time and hemisphere sides on each other. However, CG did not show any statistical differences (p=0.06, F=3.68).
RG also evidenced a statistically signi cant reduction in interhemispheric-RMT asymmetry as measured by the ratio of RMT for two hemispheres (RMT asymm = RMT Ipsilesional / RMT contralesional) from pre-to-post-therapy. RG showed a decrease in RMT asymm from 1.43±0.21 to 1.25±0.31 (mean decrease of 0.18, p=0.012), however, CG showed a decrease from 1.33±0.30 to 1.30±0.28 (mean decrease of 0.03, p=0.59), indicating a trend of normalization of RMT-asymmetry (RMT asymm should decrease as ipsilsesional RMT should be decreased from pre-to-post) over the duration of intervention in RG. RG also manifested statistically signi cant changes in intergroup-comparison over CG (p=0.028). The relative change in interhemispheric-RMT asymmetry-ratio (∆RMTasymm-ratio) changed with RG having a mean increase of 0.12±0.14 and CG a mean increase of 0.011±0.1 (p=0.028), indicating the extent of normalization of RMT-asymmetry over the duration of intervention in RG as compared to CG (Table-3).

Relationship between TMS neurophysiological-measures and clinical-outcome
The recovery parameters from TMS-measures denoting the change from pre-to-post-therapy were observed to be strongly correlated with the relative change/improvement in distal motor-outcome (∆FMW/H). The rst parameter, relative change in RMT in the ipsilesional-hemisphere (∆RMTipsi) was signi cantly different for both the groups (p=0.0235) with a mean increase of 0.16±0. 12

Discussion
The study demonstrated clinical and neurophysiological-changes in response to the robotic-exoskeleton (19) training compared to the conventional-rehabilitation. The clinical-scales showed improved changes in both RG and CG, however, increased cortical-excitability in the ipsilesional-hemisphere was shown only in RG. Five patients in RG, with the absence of MEPs at the pre-therapy measurements, showed the appearance of MEPs in the ipsilesional-hemisphere post-therapy. The improvement in RMT in ipsilesional-hemisphere showed a trend of normalization over the intervention and were also correlated with sensorimotor-functional improvement.

Comparison of Clinical-scales of Robotic-therapy group with control-group
The robotic-therapy was effective at releasing spasticity at the wrist-joint with ~ 26% (p = 0.03) improvement over ~ 14% in CG. The regaining normal muscle-tone is considered as a predictor of recovery or the rst-step in recovery (22) followed by an increase in muscle-strength and improvement in functional movements or ADL, nally leading to muscle-strength. In this study, AROM and Barthel-Index have been measured as the indicators of ADL (Table-2). Both groups showed signi cant improvement of AROM, however, RG showed signi cantly higher improvement of 130% over 47% in CG (p = 0.02). AROM is one of an important parameters in evaluating ADL and increase in AROM of wrist could lead to greater participation in ADL (23). For Barthel-Index, both groups showed similar (~ 20%) improvement (p = 0.82). BI, non-speci c crude discrete measure of ADL, measuring discretely for independence, partial-dependence and fully-dependence, and hence, even minimal improvement counts and increases the score by 5. A signi cant improvement in BI was observed for both the groups (RG & CG). BI is not much reliable as scores might get affected depending on dominant/non-dominant side, thus, is not alone predictor for therapeutic-outcomes (24). Both groups showed signi cant improvement for BS, however, RG showed ~ 32% improvements over ~ 20% in CG (intergroup p = 0.31) ( Table-2

Comparison of Cortical-excitability of Robotic-therapy with the control-group
Cortical-excitability in pre-therapy measurements was found to be lower in patients with stroke as observed by higher RMT and lower MEP (30), (31), (32), (33), (34). In some patients due to low cortical-excitability, MEP is not recordable even after delivering TMS-stimuli at the highest possible intensity (Maximum Stimulator Output at 100%) (30),(31),(32), (33), (34). In those patients with no MEP recorded, RMT is taken as value of 100, as suggested in literature (35),(36). This study investigated speci c impact of therapeutic-interventions on TMS neurophysiologicalparameters for cortical-pathways in the context of functional-gains of the hand motor-function.
Cortical-excitability measures are used as an objective investigative tool to measure the treatment responsiveness as it provide insights into membrane-excitability of neurons, conduction, and functional-integrity of corticospinal-tract and neuromuscular-junctions (37). The signi cant decrease in RMT and increase in MEP-amplitude in the ipsilesional-hemisphere demonstrates signi cant amount of increase in corticalexcitability (38), as was demonstrated in the RG versus CG. It can be interpreted that recovery of motor-function could most likely be a consequence of plastic-reorganization and use-dependent plasticity (38). Cortical-excitability and corticospinal-tract integrity have also been shown to be correlated with functional recovery potential in patients with chronic stroke (31) and exoskeleton-training appears to be bene cial in activating the ipsilesional-hemisphere for chronic patients (13.8 ± 9.1 months). Activation of ipsilesional-hemisphere could indicate either vicariation of the loss of neural circuits or unmasking of pre-existing synapses or recruitment of perilesional areas in ipsilesional-hemisphere or exploitation of the preserved functional recovery reservoir in ipsilesional-hemisphere (35),(39),(40), (41). Further, a ~ 30% decrease in MEP-amplitude in contralesional-hemisphere over the duration of intervention might indicate a decrease in cortical-excitability, evidencing a trend towards restoring the Inter-Hemisphere Inhibition (IHI) balance in the motor-network between the two hemispheres(39), (40), however, it needs to be further evaluated in a larger cohort.
The cortical-excitability measures, usually, are acquired in pre and post-intervention mostly involving brain-stimulation studies. Examples are repetitive TMS, Transcranial Direct-Current Stimulation (tDCS) (42), (43), etc. or in a combination of brain-stimulation with other neurorehabilitation strategies like Constrain Induced Movement-Therapy (CIMT) (44) or mirror-therapy (45) or training(46), (47). However, studies evaluating the therapeutic-changes in the cortical-excitability using robotic-training intervention are very rare. To best of our knowledge, only two studies attempt to evaluate the effect of active robotic-training on changes in cortical-excitability, using commercially available devices, such as Lokomat-robot (lower-limb) (48) and ARMEO (upper-limb) (49).

Speci c ve-patients in RG
A very critical outcome of the therapy was that in RG, MEP was evoked in ipsilesional-hemisphere only for 4/12 patients at the pre-therapy measurements, however, MEP was later evoked for 9/12 patients after robotic-therapy. However, in CG, MEP was evoked only for 5/11 patients and was later evoked for 6/11 patients at post-therapy. Five speci c patients in RG who did not evoke MEP at pre-therapy (0 µv) and later evoked MEP (mean = 136.6 ± 38.48 µv), showed a decrease of RMT in ipsilesional-hemisphere (mean = 27 ± 9.64), relative change RMT-ratio (mean = 0.22 ± 0.13) and mean increase in clinical-scales (FMW/H: 7.8 ± 2.38, BI: 22 ± 11.72, AROM: 22 ± 2.73). These changes were relatively much higher than the changes in patients who had MEP evoked at pre-therapy measures. The appearance of MEP in ve patients indicates that the robotic-therapy (of just four-weeks with 20 therapy-sessions) has an immense potential of training and reorganization of brain based on usedependent plasticity, for patients with chronic stroke. The increase in cortical-excitability and normalization of TMS neurophysiological-makers on the ipsilesional-side are also accompanied by greater recovery of hand-function, indexed by sensorimotor and functional recovery (by clinicalscales FMW/H, BI & AROM). Indeed, the appearance of MEPs in ipsilesional-side could be a critical recovery marker in stroke recovery.

Inter-hemispheric differences and asymmetries
The diaschisis between ipsilesional-areas and intact neuronal-networks of contralesional-areas may disturb the cortical-excitability and connectivity-patterns of connected, remote, or primary-motor areas of contralesional-hemisphere (via transcallosal-bers). The effect of roboticexoskeleton training on cortical-excitability of both hemisphere shows remodelling of the bilateral primary-motor areas in RG (time*sides p = 0.049, F = 4.08) which is not shown in CG (time*sides p = 0.06, F = 3.68). The effect of exoskeleton-training shows the potential of the exoskeleton to accelerate the cortical-plasticity phenomena in favour of functional-restoration with changes in both ipsilesional and contralesional-hemispheres.
For cortical excitability to be increased in ipsilesional-hemisphere for patients with stroke, the ipsilesional-RMT should be decreased from pre-topost-therapy and hence, RMT asymm (RMT Ipsilesional/RMT contralesional) should decrease/approach normalization (35). Signi cant differences were observed between the groups when TMS-neurophysiological changes over the intervention was expressed in terms of the interhemisphericasymmetry ratio RMT asymm indicating a signi cantly greater trend towards the normalization of asymmetry of TMS-measures in RG in response to exoskeleton-training than CG (p = 0.028). Also, the extent of normalization i.e. ∆RMTasymm-ratio showed a mean increase of 0.12 ± 0.14 in RG and CG an increase of mean 0.011 ± 0.1 (Table-3). Normalization might indicate the recruitment of peri-lesional areas in the ipsilesionalhemisphere or exploitation of the preserved functional-recovery reservoir in the ipsilesional-hemisphere (35), (39), (40), (41). Normalization in response to therapy, in terms of TMS measures on the ipsilesional-side, has been shown to have a greater recovery of arm and hand function in acute, sub-acute and chronic stages (35).

TMS neurophysiological improvement correlating the motor-outcome of both groups
The amount of change in TMS neurophysiological-measures of corticomotor-pathways (∆RMT ipsi and ∆RMTasymm-ratio) were found to be associated with the amount of improvement in functional motor-outcome during rehabilitation of the distal-part of upper-limb (∆FMW/H) (Figure-3). These parameters were signi cantly different for RG and CG (∆RMT ipsi p = 0.0235, ∆RMTasymm-ratio p = 0.028 and ∆FMW/H p = 0.012). Greater improvement (decrease) in motor-threshold tend to show greater increases with clinical-outcome and was found to have strong positive statistical correlation with ∆FMW/H in RG (∆RMT ipsi r = 0.64, p = 0.022 and not in CG r = 0.47, p = 0.13 and ∆RMTasymm-ratio r = 0.6, p = 0.03 and not in CG (r = 0.29, p = 0.38) (Figure-3). The improvement in RMT were most likely due to increased cortical-excitability of preserved motor-pathways with earlier studies in sub-acute and chronic stroke demonstrating correlation of improvement in TMS neurophysiologicalmeasures (improvement in RMT and normalization) with functional improvement (35), (50), (51). This also suggests the usefulness of TMSmeasures as an index of recovery in the ipsilesional-hemisphere (35). These neurophysiological-measures were obtained speci cally from cortical-representation of EDC muscle, a clinically affected muscle, with a speci c function which was involved in training with a roboticexoskeleton, whereas most clinical-measures do not necessarily require a particular muscle-group and measures motor-function in a broader sense.
Also, these neurophysiological-parameters individually establishes as a signi cant predictor (∆RMT ipsi r = 0.64, F = 7.24, p = 0.022 and ∆RMTasymm-ratio r = 0.6, F = 5.77, p = 0.03) of functional rehabilitation-outcome of hand (∆FMW/H) in RG, indicating that changes in corticalexcitability of ipsilesional-hemisphere could be used to predict the clinical-outcome, hence, emerging as critical recovery parameters to be considered and evaluated in larger data-samples. This might possibly be the plasticity markers predicting the responsiveness of chronic poststroke patients (49). Hence, the correlation and prediction of improvement in ∆FMW/H component by these muscle-speci c neurophysiologicalmeasures comes as an evidence of task-speci c rehabilitation of speci c-muscle.

Changes due to the device
The exoskeleton training in RG induced an evident modulation in ipsilesional and contralesional-hemispheres. However, (signi cant) changes in CG were found to be limited only to the clinical-scales, and the changes in brain was speci cally found only in the RG. An increase in corticalexcitability in the ipsilesional-hemisphere along with interhemispheric-normalization of RMT asymm could point towards the recruitment of perilesional areas in ipsilesional-hemisphere or exploitation of the preserved functional recovery reservoir in the ipsilesional-hemisphere (35), (39), (40), (41). The decrease of RMT and change in RMT asymmetry from distal-muscle was also accompanied by functional markers-FMW/H evidencing sensorimotor-plasticity, functional recovery along with task-dependent rehabilitation. Greater magnitude of the neurophysiologicalchanges observed in RG, as compared to CG, may be attributed to the unique features of the device e.g. easy donning and do ng of the device for repositioning of hand, maximum nger-extension at baseline position to provide maximum stretch to reduce spasticity; and diverse attributes of device e.g bio-triggered, real-time adaptive performance-feedback, counteracting exor-hypertonia against gravity, mimicking functional motion, user-friendly, simple design and patient-speci c customizable features with different amount of sensory inputs (proprioceptive, visual, tactile) (19). In combination with the above features, the other unique feature of the device is that it allows the facilitation of a speci c-pattern of movements mirroring complex inter-joint coordination of hand with a patient-speci c impairment, currently not integrated in the available other devices with isolated-joint movements (52). The devices were able to simulate the movement pattern maintaining joint-coordination, especially at the distal-joints which could aid in translating the motor-improvements into ADL.

Limitations
Even though data are promising, the study had few limitations such as small sample-size and no long term follow-up of patients. As most of the patients at our quaternary hospital came from far places across India and it was not possible to follow-up with them once they have left New-Delhi.

Conclusion
The robotic-exoskeleton illustrated considerable improvement in motor clinical-outcomes and increased cortical-excitability in the ipsilesionalhemisphere in patients with stroke as compared to the control-group. This technology-based intervention has the potential to enhance the recovery of stroke neuro-rehabilitation.

Declarations
Ethical Statement: Institutional Review Board (IRB) at All India Institute of Medical Sciences (AIIMS), New-Delhi, India, approved the study under protocol-number IEC/NP-99/13.03.2015. All the patients signed the written informed consent before enrolment.

Consent for Publication:
All the patients signed the written informed consent for publication before enrolment.
Availability of Data and Material: All data requests from the scienti c community should be directed to the corresponding author for consideration.
Competing Interest: The Author(s) declare(s) that there is no con ict of interest. Funding: This work was nancially supported by SERB, DST India (YSS/2015/000697).

Authors Contributors:
NS and AM conceptualized and designed the study. AM led the study and provided the scienti c inputs. MS performed patient recruitment, physiotherapy and data collection. NK and PS provided the scienti c inputs, clinical support and clinical resources for experiments. NS performed literature survey, coordinated experiments, data analysis, data interpretation, wrote the manuscript. AM reviewed the manuscript at multiple iteration with NS. All autors reviewed and approved the manuscript. Whole set-up of exoskeleton with performance biofeedback, voluntary cue and PCB in the black control-box which also works as user-interface (refer Singh et al (19))