HMC Physics Colloquium

Tuesdays at 16:30 in Shanahan Center for Teaching and Learning, Room B460

Stanley Klein

University of California at Berkeley

New Methods for Imaging Brain Activity in Space and Time

March 29, 2005

When one looks at tiny flashing light, electrical currents are induced in the brain’s cortex and blood flow is altered. The altered blood flow can be measured using functional MRI, giving spatial information of the brain activity, but valuable information about the temporal dynamics is lost due to the sluggish nature of the blood flow. Our focus is on making sense of the electrical currents.

These currents can be thought of as being generated by a tiny dipole (like a little battery) embedded in cortex. In principle, given the location, magnitude, and orientation of the current dipole and given the conductivities of the brain, skull and scalp one can calculate the electrical scalp potentials (EEG), and the magnetic fields (MEG) that are measurable above the scalp. This calculation is called the forward problem. However, we are interested in the inverse problem where one is given electric potential measurements at say, 128 scalp locations and magnetic field measurements at 275 locations and one must determine the current dipoles. For a single current dipole in the presence of noise the inverse problem is easily solved using non-linear regression. However, when two closely spaced dipoles are present, as when both visual areas V1 and V2 are active, the inverse problem had been insoluble.

We have developed a relatively simple method that should allow us to solve this long-standing problem. This method will be discussed together with all the challenges that still must be overcome in solving the forward problem, the inverse problem and the validation problem of comparing the EEG/MEG source locations with the predictions from MRI/fMRI.