Near infrared spectroscopy signal processing for interfacing brain function

Near-infrared spectroscopy (NIRS) which can detect rapid signals and slow signals has been widely used in the detection of brain function. However, due to the near-infrared spectroscopy signal contains some of the inevitable noisel, we need to pre-treat near-infrared spectroscopy signal. In this paper, interference suppression for data preprocessing on the premise in the original signal head movement, combined with the mobile standard deviation and wavelet singular value detection, to determine the exact location of the moving interference, the processed data can inhibit large fluctuations and pulse interferences. Motion artifacts can be suppressed automatically by combined with the mobile standard deviation and wavelet singular value detection to determine the exact location of the movement interferences, the method can restrain the large fluctuations and pulse interferences; Wavelet threshold denoising inhibit biological interferences and high-frequency noise, EMD threshold denoising can inhibit biological interferences, high-frequency interferences and baseline drift, independent component analysis also inhibited biological interferences, high-frequency interferences and baseline drift. The performance of EMD threshold denoising method is best with the highest signal noise ratio and the relative short processing time.