Comparison of single current dipole and Magnetic Field Tomography analyses of the cortical response to auditory stimuli

Summary

Measurements of the magnetic field elicited by a 50 ms long auditory stimulus, from three normal subjects and one head injured subject, are used to estimate the three dimensional distribution of generators in the brain. The resulting images are compared with point source solutions obtained with the usual single current dipole fitting procedures, over a latency range which includes the extrema in the (average) measured signal. In all cases considered, 100 or so epochs time-locked to the stimulus were magnetically recorded. These were averaged, and then analyzed using two techniques; a new distributed current model known as Magnetic Field Tomography (MFT), and the standard single current dipole (SCD) model. Both methods provide estimates of the current generators in the brain. In two of the normal subjects, the MFT solutions are super-imposed onto Magnetic Resonance Images (MRI) of the relevant cortical area. The results show that when the SCD model provides a reasonable description of the data, the MFT estimate shows one dominant localized region in agreement with the current dipole position. In the MFT sequence of solutions the activity evolves smoothly; multiple areas of activity often arise as the focal activity in one region declines while focal activity in another region grows. In contrast the SCD solutions during these intermediate periods fit the data poorly, and may move erratically from one locale to another. We conclude that MFT seems to provide a reasonable description of the activity through cortical and subcortical regions. The evolution of activity, as derived from the average signal, can be traced continuously from the onset of the stimulus, not just at the peaks. This could be particularly important for clinical applications where injury or other pathology produce a response with no clear dipolar pattern.