Exploiting MR venography segmentation for the accurate model estimation of BOLD signal

The amplitude of the BOLD signal depends heavily upon the resting blood volume fraction (V₀). However, most existing haemodynamic data assimilation studies pretermit such concern through assigning arbitrarily the value in a physiological plausible range. In this work, we presented the first exploration of the influence of the fraction V₀ to model estimation, where the exact value of V₀ was calibrated by MR angiography experiment. Our results show this process may greatly increase the accuracy of the data assimilation procedure. Furthermore, they also suggest that the result of traditional intensity-based activation inference would displace spatially from the realistic activated locus, the estimated neuronal efficiency parameter epsiv can accurately label the activated region, thus is a better candidate to evaluate activation level.