EMG based biofeedback with the smarting system

Abstract

Learning of a skill is a practice (task related exercise) in which feedback provides information about the performance. If the feedback signal comes from physiological activity then it is termed “biofeedback”. We present a new algorithm for real time classification of muscle activities from several muscles that can be used for feedback that is motivating for the student to learn. We used the “Smarting” system that is light (40 g), self-standing, has a 24-channels digital amplifier, and communicates via Bluetooth with an Android or Windows based platform/monitor. The Smarting system can record voltages above about 1 μV in the frequency range from 0 to 250 Hz (sampling rate at 500 Hz). The algorithm operates on the receiving platform in the Matlab environment. We present implementation of the algorithm for the recognition/distinction of four movements: fingers extension and flexion, and radial and ulnar flexion. The feedback that was used is a custom designed game on the computer (car race) where the car is controlled by four distinct signals recognized from muscle activities recorded with 18 points on the skin (monopolar configuration). The system can be implemented for other games which require four inputs since it operates as the computer peripheral. The system was designed for neurorehabilitation of humans after brain injury or disease but with the intention to be used for personal computer control, dedicated system control, and gaming.