Hybrid cortical imaging techniques and their fusion with high-resolution medical imaging

This dissertation presents a visualization scheme for integrating EEG data with structural information derived from MRI imaging. This data fusion method executes cortical imaging techniques, alone or in hybrid fashion, which are computational algorithms for non-invasively reconstructing the potential or radial current flow near the cortical surface. These techniques produce spatially enhanced potential or current density maps. EEG data or artificially generated data are used as “scalp” potential boundary conditions to be met by one of these algorithms, the cortical imaging technique (CIT). It is co-registered with a MRI-derived scalp surface and with the digitized electrode coordinates from EEG recordings.; In the visualization pipeline, a mesh surface is obtained from MRI volume data via the marching cubes algorithm. Separate scalp and cortical surfaces are rendered by color mapping the respective potential computed by CIT at each mesh vertex. In one hybrid imaging technique, the surface Laplacian of the CIT-reconstructed potential at the level of the cortex is computed, giving a measure of radial current flow through the cortex. The resulting current density maps show greater spatial resolution compared to either the CIT-reconstructed cortical potential alone or compared to the surface Laplacian computed at the level of the scalp alone. This technique is applied to artificially generated data and to visual evoked potential recordings. In a second hybrid imaging technique, the CIT algorithm is executed using phase encoded voltages that represent brain electrical activity at a selected frequency over an epoch of time during an EEG recording. This technique was applied to clinical recordings at the onset of an epileptic seizure in one subject and the resulting potential maps support other analyses of this data performed at the Mayo Clinic.; The computational simplicity of CIT allowed for the creation of an animated sequence of potential maps which serves to reveal complex dynamics at the high time resolution available by EEG as compared to other imaging modalities. Whether viewed dynamically via animations or as a single time slice, the method presented in this dissertation results in highly intuitive visualizations of clinical and cognitive brain activity as renderings on the underlying anatomy.