EEG source estimation based on dipole distribution model

To visualize electric brain activities from EEG data, we proposed a method for estimating EEG sources using dipole distribution. Conventional methods for EEG source estimation usually assume a few dipoles as sources of brain activities, but this is not always true for practical brain activities. Therefore, the proposed method estimates the dipole distribution on a specific layer of the cerebral cortex by using sulcus and gyrus region models. These models are based on the anatomical structure of cerebral cortex, particulalrly the pyramidal cell layer. They are also based on a priori knowledge that sensory evoked activity is restricted to the appropriate sensory and association areas of the cerebral cortex such as motor, visual and auditory cortexes. The cerebral cortex model was applied to a three-layer concentric head model consisting of the brain, skull and scalp. The dipole distribution was then estimated by the Moore-Penrose inverse method and by the searching method for all points. In order to evaluate the basic capability of the method, we demonstrated the estimation procedure by computing inverse solutions for sources located within a conducting area. Simulation results and an example EEG analysis indicate that this method and the models were able to discern concentric activities on one layer in the human brain model