A Proposal for Complementary Use of Readiness Potential and NIRS in BMI Development

The purpose of this study is to develop a novel Brain Machine Interface (BMI) algorithm using near-infrared spectroscopy (NIRS) to supplement data on readiness potential (RP) obtained from electroencephalography (EEG) in order to recognize human motions. Congenital and acquired disabilities make it impossible to live comfortably as an able-bodied person. BMI assists people with disabilities in moving prosthetics, computer operations, etc. Our recent research has shown that the hybrid 1DCNN-BiLSTM algorithm, which combines two deep learning algorithms, can more correctly identify between left- and right-handed motions than EEG when total hemoglobin is monitored by NIRS. This is due to the fact that brain waves come in a variety of unique waveforms, such as RP. EEG has a weakness of external and internal noise. NIRS, on the other hand, is more noise-resistant than EEG, despite it lacking the precise indicators of activity readiness that EEG does. In this study, NIRS signals were combined into EEG signals to support EEG in distinguishing left-right hand movements. The EEG-NIRS and EEG dataset were passed hybrid 1DCNN-BiLSTM to discriminate accuracy separately to compare the accuracy and the SD. As a result, EEG-NIRS can more distinguish between left- and right-hand movements than using only-EEG signals. Considering the result, the combination of EEG and NIRS signal is possible to support BMI based on EEG to distinguish the left- and right-hand movement. It may be possible to determine the non-movement and movement signals using RP that appears on EEG.