Development of NIRS-BCI system using perceptron

Abstract

Brain Computer Interface (BCI) is a system that controls machines and devices by extracting neural information from human brain activity and it is expected for nursing robot such as an artificial hand. The present study focuses on BCI that uses near-infrared spectroscopy (NIRS). But signal processing methods for NIRS signals have not yet been established. For these reasons, it is difficult obtain highly accurate control operations. This study proposes BCI systems that can control a robot using NIRS which measures neural information from human brain activity. In this study propose a new brain activity judgment method that uses dispersion ratio and perseptron, to achieve highly accurate on/off operations is developed. Brain activity during actual grasping tasks and imagined grasping tasks are measured to demonstrate the validity of the proposed BCI system. Experimental results showed that the operation results can be greatly improved with the proposed method by both tasks.