Neural source localization using advanced sensor array signal processing techniques

This paper aims to describe a hybrid technique that combines the feasibility of our recently developed Multiresolution Analysis of Signal Subspace Invariance Technique (MASSIT) [K. Oweiss and D. Anderson, 2001] with the Finite Element Method (FEM) analytic model developed in [D. Anderson et al., 2001] to obtain an accurate localization scheme for neural sources in an extracellular recording environment. The power of the proposed method stems from the fact that robust array signal processing approach is fused with the FEM analysis yielding the closest scenario to practical experimental situations. Results from experimental signal and noise simulated composite are summarized and the overall performance is evaluated.