Functional brain imaging: extracting temporal responses of multiple cortical areas from multi-focal visual evoked potentials
Shahram Dastmalchi, Stanley A. Klein · 2003
Electro/magneto-encephalographic (E/MEG) source analysis has the potential to provide researchers and medical practitioners with timing information about brain activity in the millisecond range. The high-resolution temporal information can be combined with the high-resolution spatial information from other imaging techniques in order to build a spatio-temporal image of neural activity in the brain. However, when multiple neural generators are closely positioned, as with the early visual cortical areas, the E/MEG based solutions become unreliable due to the rotation problem. The rotation problem results from many linear combinations of simultaneously active sources being able to account for the same spatio-temporal surface potential topography. Chapter one of this thesis describes several signal-to-noise optimized variations of the Laplacian derivation. The Laplacian operator reduces rotation ambiguity by mitigating the cross-contamination of sources that are distant to the measurement sensors. Chapter two presents and examines, via simulations, a novel source analysis approach for overcoming the rotation problem caused by sources in the visual cortical areas V1 and V2. The new approach solves the rotation ambiguity by first isolating the signal subspace of interest and then finding a solution by placing anatomical and physiological constraints, specific to the visual cortex, on the subspace. Chapter three tests the source analysis approach via visual evoked potentials. Pattern-reversal evoked potentials were collected using multi-stimulus m-sequence technology. The temporal solutions from the left and right hemispheres were found to be very similar thus validating the methodology.