Application of synergy analysis for motion decoding using EEG signals

Relatore: NATSUE YOSHIMURA, Ph.D., Biointerfaces Unit, FIRST - Institute of Innovative Research Tokyo Institute of Technology

Seminario organizzato dal Laboratorio di Fisiologia Neuromotoria

Abstract

Muscle synergy analysis has drawn wide attention in physiological and rehabilitation fields. By allowing for the identification of patterns of muscle co-activation in response to a single neural command, and thereby eliminating redundant degrees of freedom, muscle synergy analysis can provide valuable insight into how we control our bodies.

Methods for decoding motion from brain signals have also gained popularity as signal acquisition and computational techniques have advanced. Applications include brain-machine interfaces for prosthetic control, tele-manipulation, and communication, as well as neurorehabilitation methods for restoring function via mechanical or visual feedback. Decoder analyses can also provide critical information on which areas are relevant to a given motion.

In this talk, I will introduce our effort to apply the same synergy analysis to electroencephalography (EEG) signals to identify brain activity synchronizations that we refer to this activity as “brain activity synergy”. Using EEG and anatomical magnetic resonance imaging, we estimated cortical current sources (CS) signals and performed synergy analysis to decode finger movement in terms of not only movement directions (i.e., external coordinate frame such as upward and downward) but also finger movements (i.e., internal coordinate frame such as flexion and extension).

Our results showed that decoding performance was drastically increased when applying the synergy analysis to CS signals. Moreover, owing to the high temporal resolution of EEG, a quantitative analysis of features selected by the decoders revealed temporal transitions among motor areas, which may reflect transitions in motor planning and execution. A quantitative analysis revealed that difference in the transitions between internal and external coordinate movement decodings. The methodology and findings may be useful to understanding motor control and provide a proof of concept for brain activity synergy estimation using CS signals.

CENTRO CONGRESSI - FONDAZIONE SANTA LUCIA IRCCS