Cognitive Psychophysiology: Event-Related Potentials and the Study of Cognition

David Regan
Dalhousie University, Halifax, Nova Scotia

My instructions were, as far as I can recollect, to be reasonably critical, and to try to make it reasonably controversial. My own estimate is that ERP recording is a little more scientifically difficult than either psychophysics (auditory or visual) or single-unit recording. Technically I would say that they are all about equally demanding. It therefore, I think, does not make sense to treat evoked potential recording as though it were easy. That is in case you get the impression I'm, from now on, totally pessimistic. That's not the case.

The discussion is entitled "Early Components in Psychophysics." I don't think the concept of early components has a great deal of excitement to it. I like to classify evoked potentials by stimulus manipulation and not by the ERP waveform or the latency.

Having defined a sensory ERP entirely by stimulus manipulations, you might ask, what use is it? I think that an important application is to investigate sensory information-processing "channels" objectively. The appropriate stimulus fits the channel something like the keyhole and the key. The existence of different channels may offer an objective method for the study in children of the functions of the auditory and visual pathways or different parts of the brain or different neural organizations. Different diseases can affect these various neural organizations differently so that testing sensory channels can lead to more discriminating tests of disease processes.

The question of whether the ERP does or the ERP does not correlate with psychophysical data, which one often sees raised, is not such an interesting question, I would say.

That we do not know the site at which the ERPs are generated is considered by many single-unit investigators a damning criticism of ERP recording. But I think one can think of ERPs and psychophysics as being in a similar position

-303-