A comparative study of localization approaches to EEG source imaging

摘要

The EEG inverse problem of reconstructing the neural electrical current that produced a given measurement is ill-posed and many different source configurations can yield the same scalp potentials. In this study, we compared localizing capabilities or three EEG inverse algorithms: MUSIC, LORETA and Bayesian MCMC method. Simulations on a realistic head model show that comparing to MUSIC and LORETA, the computational power of MCMC methods offers a flexible and robust tool for EEG source imaging.