A number of techniques have been developed for the study of human brain functions in normal and disordered states. Each of them monitors function from a selective point of view. Whereas functional imaging techniques are based on neurovascular coupling (functional magnetic resonance imaging (fMRI), positron emission tomography (PET), nearinfrared spectroscopy (NIRS) and optical imaging), electrophysiological methods directly reflect neuronal activity. Electroencephalography (EEG) and magnetencephalography (MEG) provide information about global as well as regional activity under stationary conditions or for evoked or event related responses. Subdural or epicortical recordings for the identification of epileptogenic foci and microelectrode recordings for functional target identification during stereotactic neurosurgery complement the spectrum of the clinical application of neurophysiological tools.

Taken together, the neurovascular and neurophysiological methods provide largely complementary information that allows better insights in normal and disturbed brain function. The superimposition of MEG data on MR images combines the high spatial and temporal resolution of the respective methods. EEG recordings can now be conducted inside the MR scanner allowing the recognition of epileptogenic foci during EEG-defined seizure episodes.

Towards an integrative MEG and EEG study of cognitive brain functions

EEG and MEG measures arise from the same cortical source, namely from ordered intracellular currents in the pyramidal cells (Okada, 1982). Since the pyramidal cells in a circumscribed area show basically the same orientation orthogonal to the surface, the superimposed neural currents may result in macroscopically measurable electromagnetic fields. The electric and magnetic fields have orthogonal orientation and bear formally equivalent information; however, it may well be that stimulus-triggered event related potentials (ERPs) or magnetic fields (MFs) reveal different aspects of the underlying neural sources depending, for example on the nature of the task and the anatomy of the cortical region of interest. For several reasons, source estimation based on MFs is often superior to ERP based fits. First, opposite to electrical currents MF remain almost unchanged when passing through anatomical structures between the cortex and the sensor sets. Hence, the MFs are less susceptible to inadequate head models, both regarding anatomical shape and tissue conductivities (CSF space, scull, scalp). Secondly, the superficially observed MF originate from the primary neuronal currents (i.e. currents inside the dendrites) whereas volume currents do not contribute for physical reasons. In contrast, ERPs reflect both primary and volume currents.