Magnetic Field Tomography: theory, applications and examples from the visual system

Abstract

Magnetic field tomography (MFT) is based on a non-linear solution of the ill-posed biomagnetic inverse problem and it is applied independently to each single snapshot (timeslice) of the average or single trial magnetoencephalographic (MEG) signal. MFT produces tomographic estimates of the electrical activity in the brain, cerebellum and brainstem from the non-invasive, non-contact MEG data. The localization capability of MFT is demonstrated with computer-generated data and with reconstructions of V1 activity elicited by quadrant sinusoidal gratings. Instantaneous tomographic MFT solutions demonstrate precise retinotopy in single timeslices of single trials, but only at specific early latencies. A quantitative measure of changes of timelocked and non-timelocked activity can be obtained from single-trial MFT solutions. Methods developed for the delineation of regional increases in activity over seconds from hemodynamic measures are just as effective with MFT solutions at millisecond time resolution. Such comparisons reveal increase and decrease in V1 activity in the contralateral and ipsilateral side of the stimulated visual fields, respectively. Other methods for post-MFT analysis emphasize the relations between regional activations and their changes with time or stimulus properties.