Comparison About EEG Signals Processing in BCI Applications

TL;DR

This paper compares two EEG signal processing methods for BCI-based arm rehabilitation in stroke patients, focusing on identification accuracy and computational efficiency to determine the most suitable real-time approach.

Contribution

It provides a comparative analysis of EEG processing techniques specifically for real-time BCI applications in stroke rehabilitation, highlighting their performance and feasibility.

Findings

Identifies the more efficient EEG processing method for real-time BCI

Shows the method with higher accuracy has acceptable computational complexity

Supports selection of optimal processing technique for clinical BCI use

Abstract

In the context of a Brain Computer Interface platform implemented for the arm rehabilitation of mildly impaired stroke patients, two methods of EEG signals processing are compared in terms of (i) their identification performance rate and (ii) their computational complexity with the overall goal to select the most efficient and feasible real-time procedure. An effective signal processing is, indeed, one of the most critical issue for such kind of technology which aims to establish a real-time communication between the subject's brain and a machine, i.e. a computer, a robotic arm or another device, that should implement his/her intention to move in place of his/her impaired arm.