Inverse Problems Insights

A window to the working brain. The communication between neurons in the brain occurs through the transmission of electrochemical signals across synaptic clefts, triggering a small ionic current in the receiving neuron, creating a very weak associated magnetic field. When a bundle of thousands of neurons fire together, the net electric current may be strong enough to induce a magnetic field that can be measured outside the head. Magnetoencephalography (MEG) maps the brain neuronal activity from the measurements of this magnetic field. The scarcity of data and the spurious magnetic signals from the external environment and within the brain itself make the task extremely difficult. A particular challenge for MEG is to map brain activity in the deep brain areas, which are the location of the thalamus and the limbic system controlling the mood, feelings, instincts and behavior. There are important clinical applications of MEG, including localization of epileptic focii, cognitive studies and investigations of brain connectivity. The combination of Bayesian statistical methodology, that takes advantage of anatomical information about the orientation of the neurons in different parts of the brain, and iterative methods for solving large linear systems have opened a window on neuronal activity for the brain using data collected in a totally noninvasive manner. The resulting MEG inverse solvers identify with high resolution focal cortical activity and localize deep cerebral activity. The computational efficiency of the method makes it suitable for the processing of the large time series data, the price to pay for the excellent time resolution of MEG. Once the methodology is in place, we will use it to understand if there are activation patters typical of different brain states, including resting state and meditation, and to find clues for altered moods, such as depression, which manifest themselves in an identifiable manner in the brain activity patterns.