Size-specific information channels and selective attention: visual evoked potential and behavioral measures.

Subjects were shown a stimulus (diffuse light or a checkerboard pattern with checks subtending 12' or 35' of arc) and were instructed to remember and attend (give a reaction time response to or count) that stimulus during the course of an approximately 6-min trial. A trial consisted of the random presentation of 8 stimulus flashes (diffuse light and checkerboards with 9, 12, 18, 24, 35, 45, 94 and 95 min checks) at a rate of 1/555--930 msec. Visual evoked potentials and reaction times were averaged to each of the 8 stimuli. The purpose of the experiment was to assess how specifically the relevant stimulus could be attended, as indicated by VEPs and behavioral responses to the relevant and 7 irrelevant stimuli. Attention to one check size resulted in greater amplitude VEP components to that size than to other sizes: the greater the discrepancy between the flashed and attended check size, the smaller the VEP amplitude. Such size tuning was first evident in the VEP 160 msec after stimulation and most evident at 260 msec. The bandwidth of the VEP amplitude size-tuning functions became progressively narrower from 160 up to 260 msec, that point in time when the RT response was initiated. The width at 260 msec was similar to that indicated by the behavioral RT data (modal latency of 322 msec). The functional components of the VEP appeared to reflect the effects of attention on the activity of cortical size channels or detectors.