The aim of this study was to define whether a gesture similarity index (GSI) based on muscle synergies extracted by a wearable device (REMO, Morecognition Ltd, Turin, Italy) provides the same information of clinical outcome measures, after specific training of hand function. The secondary aim was to define whether specific hand tasks could be considered as target tasks for hand function assessment, after stroke.
REMO is a wearable composed of 8 surface-electromyography (sEMG) sensors able to detect forearm muscles activation in real-time. The acquisition protocol consisted of bilateral execution of wrist and hand movements of the Fugl-Meyer Assessment (FMA-Hand, maximum score 24 points) using REMO. Then, we applied the Non-Negative Matrix Factorization (NMF) algorithm to extract muscle synergies from the sEMG signals. We calculated the GSI which represents the mathematical distance between the muscle synergies extracted from a group of elderly-healthy subjects (as reference model) and from the patient's performance (the lowest distance represented the maximum similarity, GSI=100%). Stroke survivors were assessed before (T0) and after (T1) a specific hand training of 15 sessions (1 hour each, 5 time/week, 3 weeks) (ClinicalTrial.gov ID NCT05815368). After randomization, the patients underwent to a task-specific training (control group, CG) or biofeedback training using REMO (experimental group, EG). We compared the FMA-Hand and GSI results before and after the training, and we calculated Spearman's correlation to investigate the relationship between the level of FMA-Hand and the GSI index.
We collected data from a control group of 4 elderly-healthy subjects (age 74±6.3) and a group of 14 people with diagnosis of stroke (age 64.7±13.8). Overall patients showed a significant hand motor recovery (FMA-Hand: T0=11.8 ± 3.0; T1=15.7 ± 3.7; p=0.002). Based on GSI results, data showed a statistically significant improvement in wrist circumduction movements (T0= 82% ± 4.5%; T1= 85.4% ± 4.3%; p=0.048). No differences between groups were found both in FMA-Hand and in GSI analysis. Then, at the baseline, correlation analysis showed a positive significant correlation between FMA-Hand and GSI value in circumduction movement (r=0.54; p=0.046).
The results showed that a specific hand motor training induced a significant improvement in patients’ functional level. GSI results showed that wrist circumduction could be a target movement to describe motor recovery after stroke. Future perspectives will include a wider data collection to confirm the results.
REMO device allowed to record real-time muscle activation (sEMG) during tasks execution. The study of neurophysiological parameters underlying the ability to perform complex tasks, such as wrist circumduction, could provide significant information on hand motor impairment after stroke, depending on the clinical outcome.
Muscle Synergies
Hand motor impairment
