Removing signals from scalp blood flow for functional near-infrared spectroscopy

Abstract

Near-infrared spectroscopy (NIRS) can easily measure brain activation during rehabilitation exercises. However, irrelevant signals mainly from scalp blood flow change are often observed in NIRS measurements. In this study, we simulate optical diffusion approximation in a four-layered tissue head model and focus on the association of the optical attenuation changes by the scalp blood flow in each probe distance, and propose an algorithm to extract signals derived from the brain activities with a pair of probe distance. We show that the proposed method is effective in removing artifacts by applying the method to the simulation data and the measured data of NIRS.