| The read-out of prioritized maps of space in frontal and parietal cortex is believed to provide the spatial feedback signals needed to bias activity in early visual cortex. It has been shown that responses of macaque V1 neurons with receptive fields matching the goal of a planned eye movement are enhanced during the delay period of a memory-guided saccade task. Other data suggest that top-down signaling in early visual cortex is gated by bottom-up visual stimulation. In humans, however, the effects of planned eye movement on early visual cortex remain unknown. We used event-related functional MRI to test two hypotheses: a) planned saccades evoke retinotopically specific activity in human early visual cortex; b) top-down signals are gated by bottom-up visual stimulation in visual cortex.;We measured neuronal activity in human visual areas V1, V2, V3, hV4, VO-1, V3A/B and IPS-0 while subjects maintained a planned pro- or antisaccade directed to or away from the location of a briefly presented visual cue over a long and variable delay period. We localized portions of the retinotopic maps that corresponded to parts of the visual field where the visual cues (hence saccade goals) were presented in the memory-guided saccade tasks.;Cortical activity during the delay period was enhanced at the retinotopic locations representing the end-goal of planned prosaccades in all visual areas. This bias might require or be gated by visual stimulation. Dissociation of the saccade goal from visual stimulation in the antisaccade task allowed us to test this possibility. Delay period activity was enhanced not only at the retinotopic locations representing the visual cue, but also at the locations representing the saccade goal of the planned antisaccades, even though those parts of the visual field were not visually stimulated prior to the delay period.;Our results indicate that top-down signals, presumably from frontal cortex, could evoke retinotopically specific activity in early visual areas biasing the competition for neural representation. However, our findings do not support the idea that a gating signal in the form of bottom-up visual stimulation is necessary for effective top-down biasing of neuronal activity in human early visual cortex. |
