Event-Related Potentials and Selective Attention
Steven A. Hillyard
University of California, San Diego
With many others at this conference, I share the optimistic outlook that the event-related electric fields we record on the scalp are related to important aspects of information processing in the brain. Yet, the question recurs as to whether or not the ERP is an appropriate level to study the encoding and transmission of neural information. It seems to me there is no a priori reason for believing that studies of single-unit activities will provide more immediate insight into complex brain functions than investigations of the field properties of large neural populations. Because we know so little about the codes used by the brain, it may well be the case that measures of ensemble activity such as the ERPs will reveal critical parameters of brain function. My optimism is fueled by the increasing numbers of interesting correlations that are being revealed between the scalp-recorded ERPs and cognitive functions in man.
Figure 3.1 expresses the optimism that some of us feel. These are data from a mouse experiment. The ERPs are being recorded from two church mice as they spy their friend Samson the cat looking in at them. At the precise moment they notice the cat, the ERP has a specific configuration that is related to each mouse's perceptual experience. However, even this elegant experiment has some ambiguity, because the poor mice also "cheered up a bit" when they saw Samson. Thus, we really can't be sure whether this ERP reflects their perception of Samson or their delight at meeting him. Such ambiguities are often encountered in ERP research.
Figure 3.2 may present a slightly more realistic view, though I'm told it's a little outdated. Actually, it is a perversion of a figure from Lindsay and