Thanks to recent technological improvements, perfusion measurements by laser Doppler imaging is becoming an increasingly important tool in research an clinics. Typically, a series of perfusion images at 2Hz are acquired during a occlusion paradigm (normal blood flow, occlusion, recovery). The time-course signal during recovery reveals important information of the underlying microvascular structure. The aim of of this project is to analyze the time-course signal using general linear models; and in particular to exploit the spatial correlations by the wavelet transform.