This is a Phase II proposal. The PI aims to further develop an integrated software package for spatiotemporal human brain imaging. First, magnetic resonance imaging (MRI) is used to identify possible dipoles generating the magnetoencephalogram (MEG) and/or electroencephalogram (EEG) within the cortex and perpendicular to its surface. A linear estimation approach that employs normalization by estimated noise provides dynamic statistical parametric maps with relatively uniform resolution across the cortical surface. Optionally, the solution can be biased toward those brain areas found to be activated using functional MRI (fMRI) acquired in the same situation. The PI intends to extend the forward and inverse operators to be more accurate and automated, and to include options for comparison with other approaches. He will provide additional tools for interpreting the accuracy and reliability of individual dynamic maps. Finally, the PI proposes numerous changes intended to make the software more robust, compatible and easy-to-use. Special emphasis will be placed on developing a reliable and convenient tool that calculates accurate spatiotemporal maps of cerebral activity from the EEG alone, as well as in combination with MEG and/or fMRI. This software could become widely used in scientific as well as clinical applications that currently use EEG but provide limited spatial localization.
The proposed product would be an effective and validated method for combining the high-spatial resolution fMRI with the high temporal-resolution MEG/EEG using spatial constraints derived from the structural MRI. This method will be useful to cognitive neuroscientists as well as clinical neurologists and neurosurgeons. An initial specific application is in the planning of surgical interventions for epilepsy.
