Movement classification using ECoG high-gamma powers from human sensorimotor area during active movement

Abstract

Neural activation in high-gamma range (>50 Hz) is robustly observed in sensorimotor area. Previous neurophysiological studies have indicated that there are dominant sensorimotor high-gamma power changes during active movement. Here, we demonstrate that two different movement types (hand grasping and elbow flection) can be discriminated at single-trial conditions with high accuracy using the spatial dynamics of high-gamma features from primary motor cortex. Based on our results, we propose that sensorimotor high-gamma activities during active movement can be a powerful feature for on-going movement classification, and their characteristics mainly represent the instant movement states.