Abstract
When a masking stimulus is presented within 100 ms after a brief visual target, conscious recognition of the target is often impaired. Recent evidence suggests that in some cases, masks can interfere with re-entrant activity in visual cortex. We examined the effect of target-mask onset asynchrony on activity in cortical areas using event-related fMRI. Normal adults were asked to detect whether dotted squares (66 ms) followed by a surrounding mask (120 ms) contained 1 or 2 gaps. Masks were presented at one of 3 delays (SOA: 0, 100 or 220 ms). Both common-onset (SOA = 0 ms) and delayed-onset (SOA = 100 ms) masks increased target identification errors and fMRI activations in posterior parietal and frontal areas compared to the control condition (SOA = 220 ms). Activations in these areas were stronger during common onset masking and this condition also showed poorer performance. However, delayed-onset masking produced a stronger activation in occipital cortex than common-onset masking despite equivalent sensory stimulation. The results are compatible with the suggestion that delayed-onset masks can produce increased re-entrant activity in visual cortex.