Functional near infrared spectroscopy (fNIRS) is a popular technology for studying human brain function. It can provide information about changes in oxy-, deoxygenated and total hemoglobin (Hb) concentration. fNIRS has several advantages compared to functional magnetic resonance imaging (fMRI), such as higher temporal resolution, ease of use, and less expensiveness. However, fNIRS has disadvantages such as lower spatial resolution and decrease in signal-to-noise ratio (SNR) by artifacts. Studies have shown that superficial hemodynamic changes in the scalp can affect the fNIRS hemoglobin signals, and in recent years several methods such as principal component analysis (PCA), independent component analysis (ICA) and the hemodynamic modality separation method are proposed to separate hemodynamic change in the brain from the fNIRS signals. To investigate the relation between these fNIRS signals and brain functional hemodynamics, we conducted simultaneous fNIRS and fMRI measurements at human parietal area during a finger tapping task, and compared the hemoglobin signals of fNIRS and blood-oxygen-level-dependent (BOLD) signals of fMRI spatio-temporally. We also applied the brain signal separation methods to the measured fNIRS signals and compared the separated brain signals and the BOLD signals to elucidate the effectiveness of the separation methods.