The goal of this project focuses on addressing the trade-off between temporal resolution and accuracy in deep tissue blood flow imaging in human brain using diffuse correlation spectroscopy (DCS). The light is delivered to the brain via the forehead using optical fibers. The CBF in the prefrontal cortical region can be measured using single photon avalanche detectors (SPAD). In order to obtain better SNR one needs more data which in turn results in poor temporal resolution.We address this using deep learning based methods like data imputation.