Hand motion reconstruction using EEG and EMG

Motion reconstruction of continuous hand movement is a problem that can be solved in different ways. Using invasive technologies has showed great results. However obtaining similar precision with non-invasive methods is something that has not been achieved. In particular we focus our attention on prosthetic devices for trans-humeral amputees. In this study we use two different non-invasive acquisition systems (EEG and EMG) in combination with two different predictor architectures to find the most appropriate to solve the problem. In addition, the importance of each one of the systems was studied along with the importance of the previously reconstructed positions. Data were collected from 16 healthy subjects, reaching correlation values up to 0.893.