Developing an Optimal Brain Computer Interface Model using Functional Near Infrared Spectroscopy: A Review

Abstract

Brain-Computer Interfaces (BCIs) are promising in advancing numerous applications. Although many functional near-infrared spectroscopy (fNIRS)-based BCIs have been studied, the development of an optimal fNIRS-based BCI model remains unclear. This study aims to review recent methodologies that used to optimize fNIRS-BCI models in four aspects i.e. signal acquisition, pre-processing, feature extraction, and machine learning. Besides, the differences, strengths, and limitations of various algorithms are discussed and highlighted. By comprehensively examining the recent trends and challenges in fNIRS BCI model development, this study proposes and discusses potential techniques in advancing fNIRS-based BCIs model development. The results suggest that future fNIRS-based BCI studies should focus on addressing cross-subject classification challenges and real-world fNIRS-BCI applications.