Current density reconstruction from EEG based on a time varying nonlinear physiological model

A new electroencephalographic current density reconstruction method is introduced using a physiologically based nonlinear modeling that describes better the dynamic behavior of the neural activity. In addition, time-variant parameters are considered into the model to capture the dynamics for normal and pathological states measured from signals. The method is implemented by Unscented Kalman filtering approach. The performance of the new method is evaluated (in terms of mean square error) by application to simulated EEG data over several noise conditions, and a considerable improvement over linear estimation approaches is found.