The aim of this study was to quantitatively analyze the amount of methodological noise and the spatial and temporal variability of laser Doppler flowmetry (LDF) signals mapping cerebrocortical microflow. In an experimental setup with latex beads, the methodological LDF-signal variability was determined (coefficient of variation or CV(method)). The biological variability of the LDF signals was measured in animal experiments using 10 anesthetized rabbits. One stationary reference probe was used to assess temporal heterogeneity (CV(temp)) and a micromanipulator-driven scanning probe was used to assess spatial heterogeneity (CV(spat)) in a cortical area of 3.5 x 4.5 mm with 252 measurement points. CO(2) tests were used to modulate cerebrovascular resistance. CV(method) was found to be 4.94 +/- 1.7. The CV(temp) for the LDF-velocity signal was assessed to be 13.93 +/- 5.9 during normocapnia. Scanning of the brain surface with the scanning probe revealed a CV(spat) for LDF velocity of 65.0 +/- 16.2 during normocapnia. CO(2) modulation (hypocapnia --> normocapnia --> hypercapnia) of the cerebral resistance did not show a significant change in temporal heterogeneity (10.84 +/- 3.1 --> 13.93 +/- 5.9 --> 14.82 +/- 3.9), whereas spatial heterogeneity decreased significantly (81.31 +/- 12.0 --> 65.0 +/- 16.2 --> 54.04 +/- 21.8). Although the spatial and temporal variability of LDF signals evoked by cerebrocortical microflow is in the same range as with other methods and in other organs, LDF cerebrocortical mapping is restricted by the large temporal and spatial heterogeneity of the cerebrocortical vasculature. The definitions of sample volume, scanning step width, probe to brain surface distance, and average time per scanning point are critical concerning reliable LDF cerebrocortical mapping techniques.