A Filter Approach to Attenuate the Effects of Venous Effects in Task-based fMRI Data

There are biological noises such as blood pressure, respiration and heartbeat in fMRI signals, and some existing methods remove these fundamental noises, but the effect of venous vessels in the brain is often neglected. The venous effect interferes with the extraction of blood oxygen signals from brain activity and can cause peak shift and trailing of the HRF(hemodynamic response function), and improving the spatial resolution of fMRI requires eliminating the effect of venous effect. The existing methods to address the venous effect are still imperfect, and most of the existing methods to address the venous effect are done from source acquisition of fMRI data, which is demanding on equipment and costly in time. The difficulty lies in eliminating the effect on BOLD(blood oxygenation level dependent) signals at the data level without increasing the time cost using signal processing methods. It is proposed to extract data from different spatial layers and at different brain depths, ranging from the surface of the soft cerebrum to the junction of the gray matter white matter, with data covering the location of the large veins. After the onset of external experimental stimulation, the effect of venous effects is attenuated by a data-driven approach with HRF peak correction for data points in the ventricles affected by venous vessels.