Electrophysiological Brain Source Imaging via Combinatorial Search with Provable Optimality

Electrophysiological Source Imaging (ESI) refers to reconstructing the underlying brain source activation from non-invasive Electroencephalography (EEG) and Magnetoencephalography (MEG) measurements on scalp. Estimating locations their extents is a fundamental tool in clinical neuroscience applications. However, estimation challenging because of ill-posedness high coherence leadfield matrix as well noise EEG/MEG data. In this work, we proposed combinatorial search framework address ESI problem with provable optimality guarantee. Specifically, by exploiting graph neighborhood information space, converted into designed algorithm under A* solve it. The guaranteed give an optimal solution problem. Experimental results both synthetic data real epilepsy EEG demonstrated that could faithfully reconstruct brain.