Pharmacologically induced changes in the EEG may be related to activity within a limited region of the cortex, as for example in pharmacotherapy of focal epilepsy. Observation of the effects obtained will therefore benefit from improved spatial selectivity in the EEG derivation technique. An approach based on simple matrix algebra concepts can be used to comprehend the potential field in the scalp as one entity, rather than observing single potential differences on the surface. It is shown that existing methods of deriving EEG (bipolar, referential, etc.) can be clearly formulated using such concepts, which can also include interpretive operations. The simplest possible interpretive model utilizes the fact that derived signals refer to a given pattern of electrodes on a surface. This model provides the basis for 'source derivation', as presented in earlier papers. The interpretive model may be further developed to include the transfer of electrical potentials from a driving potential distribution on a subsurface level in a conductive body to the resultant distribution on the surface. By expressing this transfer function as a matrix operator and performing a matrix inversion, coefficients can be determined for a simple analogue network which is capable of deriving on-line signals representative of a 'focal plane' below the surface, thus providing a corresponding improvement in spatial resolution.
