부산대학교 도서관

Functional electrical stimulation (FES) consists of applying low-intensity electrical currents to affected tissues to generate muscle contractions and movements in a controlled manner. FES-based therapy has shown notable benefits in the reeducation of movements and in the recovery of motor functions. We developed a non-invasive FES device to perform FES therapy. Being a medical device, a set of additional testing and validations is required to be used in a clinical setting. We performed a feasibility and safety test of the FES device to assist ankle dorsiflexion movement by stimulating the tibialis anterior muscle of a rabbit, a widely used animal model in muscle stimulation research. The aim of this work is to present a movement analysis methodology to assess the movements induced by the FES device in the tibiotarsal joint of the animal model. The protocol consists of 1) determining stimulation thresholds and applying 20 repetitions of FES while videorecording, 2) Videos are processed using a mesh to find vectors among frames that describe the trajectory and magnitude of the movement, 3) A Convolutional Neural Network is trained to classify individual video frames corresponding to dorsiflexion or rest, obtaining an average of 80% accuracy. This validates that FES generated by our device is effective and can evoke a functional movement.