Event-related potentials (ERPs) were recorded from healthy subjects in response to unilaterally flashed high and low luminance bar stimuli presented randomly to left and right field locations. Their task was to covertly and selectively attend to either the left or right stimulus locations (separate blocks) in order to detect infrequent shorter target bars of either luminance. Independent of attention, higher stimulus luminance resulted in higher ERP amplitudes for the posterior N95 (80-110 ms), occipital P1 (110-140 ms), and parietal N1 (130-180 ms). Brighter stimuli also resulted in shorter peak latency for the occipital N1 component (135-220 ms); this effect was not observed for the N1 components over parietal, central or frontal regions. Significant attention-related amplitude modulations were obtained for the occipital P1, occipital, parietal and central N1, the occipital and parietal P2, and the parietal N2 components; these components were larger to stimuli at the attended location. In contrast to the relatively short latencies of both spatial attention and luminance effects, the first interaction between luminance and spatial attention effects was observed for the P3 component to the target stimuli (350-750 ms). This suggests that interactions of spatial attention and stimulus luminance previously reported for reaction time measures may not reflect the earliest stages of sensory/perceptual processing. Differences in the way in which luminance and attention affected the occipital P1, occipital N1 and parietal N1 components suggest dissociations among these ERPs in the mechanisms of visual and attentional processing they reflect.(ABSTRACT TRUNCATED AT 250 WORDS)