Applications and analysis strategies of magnetoencephalography imaging

Magnetoencephalography (MEG) is a method used to measure temporal changes in magnetic fields with a sensitivity range of femto-Tesla to nano-Tesla. It is well suited for measuring the very small magnetic fields generated by neuronal brain activity as well as any small varying magnetic fields. MEG is a versatile tool with untapped potential. This thesis explores three aspects of MEG: instrumentation, application, and new directions. Magnetic source phantoms were developed as ground truths to help with instrument standardization as well as model testing. A portable magnetic dipole phantom with a constant current generator was developed that can be independently driven without reliance on system specific hardware. Extending the phantom design to a multi-dipole model, a potential ground truth for MEG source imaging was developed. Inverse analysis of MEG data for dipole and beamformer strategies were studied and applied specifically to the properties and neural generators for facial processing. Event-related activity between the perception of face and non-face object stimuli appears as an increase in amplitude for the face condition peaking around 170 ms over the posterior regions of MEG sensor recordings (M170). The anatomic origins of this face-selective 170 ms peak remains unresolved. A new model for the M170 neural generators was developed through the analysis of four different MEG studies. The M170 response was linked to simultaneous activity in the fusiform face area, occipital face area and superior temporal sulcus rather than any one individual location. Finally, a novel technique was developed to extend the usage of MEG for magnetic nanoparticle imaging. There is potential to utilize the magnetic properties of these particles for internal imaging of living organisms using MEG. Stationary samples of magnetic nanoparticles in liquid suspensions were measured within the MEG helmet. The magnetic field of the particles appeared as a distinct increase in baseline noise in the frequency domain - highest in the low frequencies and decaying to below the noise floor above 500 Hz. A method using beamformers to localize these particles using this rise in frequency provided a robust technique for the potential localization of magnetic nanoparticles in vivo.