Machine-Learning-Based Prediction of Gait Events From EMG in Cerebral Palsy Children

Machine-learning techniques are suitably employed for gait-event prediction from only surface electromyographic (sEMG) signals in control subjects during walking. Nevertheless, a reference approach is not available cerebral-palsy hemiplegic children, likely due to the large variability of foot-floor contacts. This study designed investigate machine-learning-based approach, specifically developed binary classify gait events and predict heel-strike (HS) toe-off (TO) timing sEMG hemiplegic-child To this objective, acquired five hemiplegic-leg muscles nearly 2500 strides 20 acknowledged as Winters’ group 1 2. signals, segmented overlapping windows 600 samples (pace = 5 samples), used train multi-layer perceptron model. Intra-subject inter-subject experimental settings tested. The best-performing intra-subject able provide population mean classification accuracy (±SD) 0.97±0.01 suitable HS TO events, terms average absolute error (MAE, 14.8±3.2 ms 17.6±4.2 TO) F1-score (0.95±0.03 0.92±0.07 TO). These results outperform previous sEMG-based attempts populations comparable with outcomes achieved by approaches populations. In conclusion, findings prove feasibility neural networks predicting two main using EMG also condition high signal cerebral palsy.