fNIRS motion artifact correction for overground walking using entropy based unbalanced optode decision and wavelet regression neural network

Abstract

Functional near-infrared spectroscopy (fNIRS) can be employed to investigate brain activation by measuring the absorption of near-infrared light through an intact skull. fNIRS can measure hemoglobin signals, which are similar to functional magnetic resonance imaging (fMRI) blood-oxygen-level-dependent (BOLD) signals. The general linear model (GLM), which is a standard method for fMRI imaging, has been applied for fNIRS imaging analysis. However, when the subject moves, the fNIRS signal can contain artifacts during the measurement. These artifacts are called motion artifacts. However, the GLM has a drawback of failure because of motion artifacts. Recently, wavelet and hemodynamic response function based algorithms are popular detrending methods of motion artifact correction for fNIRS signals. However, these methods cannot show impressive performance in harsh environments such as overground walking tasks. This paper suggests a new motion artifact correction method that uses an entropy based unbalanced optode decision rule and a wavelet regression based back propagation neural network. Through the experiments, the performance of the proposed method was proven using graphic results, a brain activation map, and an objective performance index when compared with conventional detrending algorithms.