Cerebral Low Frequent Oscillation Signals Analysis Based On Optical Imaging

Optical imaging is a functional mapping technique based on intrinsic signals. In this paper, cortical image sequences in the condition of non-stimuli and stimuli are captured by the OI system. The separation of artery and vena is performed by utilizing the feature of intrinsic signals. The haemodynamic response function is identified by the use of correlation analysis method and event-related selective averaging method.There are abundant physiological signals contained in the cortex. We tried to separate artery from vena based on spectrum features of the heartbeat and respiration oscillations. Vessel network was firstly extracted from a frame of cortical image by threshold segmentation and region growing method, then we computed the spectral powers of heartbeat and respiration signals separately according to an oscillation -related physiological feature that arterial oscillation distributes between 5Hz and 6Hz and venous oscillation distributes between 1Hz and 2Hz. Separation of artery and vena was successfully achieved by utilizing the spectral power-percentage of heartbeat and respiration signals. The potential value of 0.1Hz oscillation in artery and vena separation was discussed at last. In the condition of stimuli, cortical image sequences are collected with the illumination of 546nm green light and 610nm red light. Influence to vessels introduced by the stimuli is compared.We use an anti-correlative pseudo-random event sequence to stimuli the special functional cortical architecture and collect the image sequences. The haemodynamic response function is identified via correlation analysis method and selective averaging separately by using the stimuli data. It is found that selective averaging is better than correlation analysis method in identification. The main reason may be that selective averaging has no need for the hypothesis of Gauss white background noise whereas the correlation analysis has that. Hence the selective averaging method is more adaptive in the identification of haemodynamic response function under the condition of complicated circumstance.