Simulative And Experimental Investigations On Time Domain Optical Topography

Currently, positron emission tomography (PET) and functional magneticresonance imaging (fMRI) are two major methods of brain functional imaging. Theformer uses isotope to achieve the diagnosis of brain disease, which might greatlyharm human body and cost expensive; the latter can only measure the total change ofhemoglobin concentration indirectly, and could also suffer from contraindications. Tocompensate the above defects of PET and fMRI, near-infrared (NIR) spectroscopy, asa noninvasive diagnostic technology for monitoring cerebral hemodynamics andoxygenation state, has become one of the most prominent topics in the medicalimaging community. It can not only realize real-time detection and imaging of livingtissue without injury, but also obtain the information on oxy-and deoxy-hemoglobinconcentrations directly. Due to the limitation of penetration depth of NIR light inbrain, NIRS with topological mode, referred to as Optical Topography, is commonlyused to study the stress reaction process in the cerebral cortex.This thesis explores a time-domain optical topography method that combinesNIRS principle with the time-correlated single photon counting (TCSPC) technique.By applying two calculation methods, i.e. the least-squares fitting based on analyticalsolution to the time-domain diffusion equation, the modified Lambert-Beer’s Lawbased on calculated and empirical optical path, to the midway point between eachadjacent source and detection sites, the changes of absorption coefficient can bespatially resolved. We conducted a series of simulation and phantom experiments forthe validation of the methods, using the established multichannel TCSPC system, andcompared the results from the three methods.The results show that the proposed time-domain optical topography method canqualitatively locate the target with reasonable accuracy, and has flexible configurationof the source-detection array. It could provide the richest hemodynamic informationon the cortical activations through multi-wavelength measurement. We believe thismethod could be a considerable supplement to the arsenal of the brain functionalimaging.