Abstract:
is a cerebrovascular disease (CVD) that results in hemiplegia. The most severe effect of stroke is a full or partial loss of functional (motor) control of one side (half) of the body which causes patients to have difficulty in performing routine activities. In most cases, the patient needs to be institutionalized to receive physiotherapy or robot-assisted therapy for the improvement of motor function. Commercially, various home-based devices are available for upper limb rehabilitation, but to date, no clinically effective device is available for motor function rehabilitation of the lower limb. In this pilot study, we explored the potential use of surface electromyography (sEMG) as a control input for the motor/functional rehabilitation of lower limbs in stroke patients. To achieve this, we decoded the movements originating at the ankle joint using sEMG and explored their correlation with the level of motor impairment. Motor impairment level was assessed via Fugl-Meyer Assessment (FMA) scale. Three channels of sEMG data from 11 stroke patients were recorded during the performance of the following motion classes: Dorsiflexion (DF), Plantarflexion (PF), Eversion (E), and Inversion (I). In the current study, time-domain (TD) features (Hudgins) were extracted from the signal, and then they were classified with linear discriminant analysis (LDA) as well as artificial neural networks (ANN). On average 63.85% of the movements were accurately classified in offline analysis for LDA and 67.10 % for ANN. A positive correlation (Spearman’s) was found between classification accuracy and motor impairment; suggesting that with a moderate increase in classification accuracy a slight raise occurs in lower limb Fugyl Meyer score. The findings of this study suggest that an EMG-based Interactive therapeutic system having the ability of home-based customized therapy can be designed using surface electromyography for the improvement (functional) of the lower limb after stroke.
