Combining ENG and EEG integrated analysis for better sensitivity and specificity of neuroprosthesis operations

摘要

Combining non-invasive monitoring of action-related brain signals with the invasive recordings of the nerve motor output could provide robust natural and bidirectional multimodal Brain-Machine interfaces. One 26 years old, right-handed male who had suffered traumatic trans-radial amputation of the left arm was connected in a bidirectional way with a robotic hand prostheses. Cortical signals related with movement programming, execution, and feed-back were recorded by non-invasive scalp electrodes to detect high-level information (i.e. onset of movement intention), while the efferent neural activity containing the low-level commands towards the missing limb was recorded from the amputated nerves by multipolar intra-neural electrodes. The aim of this article is to report advanced experiences aiming to investigate whether information on "hand-related" activities can be decoded by the combined analysis of motor-related signals simultaneously gathered via intraneural electrodes implanted into the peripheral nervous system and scalp recorded electroencephalography signals to govern a dexterous hand prosthesis using the natural neural "pathway".