ABSTRACT
The ability to compute the distribution of current on the cortex that underlies observed extracranial magnetic fields will determine the ultimate usefulness of magnetic source imaging (MSI). Assuming that a current dipole is located within a sphere filled with a uniform conducting solution, 17 and only 1, external field pattern can be produced by that source. Thus, in principle, the so-called forward problem of computing the field has a unique solution. However, there is no unique solution to the inverse problem: that of locating the source of a field solely on the basis of knowledge of the properties of that field. This chapter briefly describes some efforts toward resolving the barriers posed by the nonuniqueness of inverse solutions. It focuses on 1 approach, the minimum norm least squares inverse (MNLS), which explicitly incorporates knowledge of the geometry of the cortical surface and the assumption that currents giving rise to the field are on that surface. Such constraints make it possible to go beyond simple equivalent current dipole sources and describe the extended patterns of current that give rise to external fields. The MNLS is then extended to account for locating on the cortical surface the currents that give rise to regions of enhanced or depressed incoherent activity This approach makes use of measures of field power rather than field per se. In principle, it feasible to localize regions of the cortex that are more or less active than other regions, thus mirroring modalities such as positron emission tomography and functional magnetic resource imaging, but examining altered electrical activity rather than blood flow.
