Abstract

Substantial experimental evidence from humans and non-human primates implicates the extrastriate visual area V5 (located in the posterior region of the inferior temporal gyrus and sulcus) in motion processing and area V4 (located in the fusiform gyrus) in color processing The above studies show attentional modulation of stimulusevoked responses. Fewer studies have explicitly investigated attentional modulation of baseline activity. Macaque studies explored baseline shifts in neuronal responses in the context of working memory 8 . They found that both baseline activity and activity evoked by a moving stimulus appear to change with expectation. Even when subjects view a blank screen, baseline activity in 32% of V5 cells doubles. Using direction of motion as the matching criterion, this study employed a delayed match to sample task. Modulation of delay period or baseline activity was not related to specific information in the cue (direction), suggesting that this effect may be modulated, in part, by attention. When interpreting our results in relation to these findings, we assume that baseline attentional modulation is mediated by the same sort of tonic discharge associated with delay period activity evoked by working memory tasks. Monkey electrophysiology also shows that attention can modulate both baseline and stimulus evoked activity in V4, depending on the relative locations of visual stimuli within the cell's receptive field Enhanced baseline activity is taken to reflect attentional 'set' or expectation, whereas changes in stimulus-evoked activity reflect changes in sensory processing 10 . However the functional significance of and relationship between changes in 'set' and changes in evoked activity is unknown. Because of the long periods of time over which hemodynamic signals are integrated, it has been difficult to unambiguously dissociate activity due to attentional set from stimulus-evoked responses in human functional neuroimaging studies 15 . Here, by using a novel event-related protocol and fMRI, we show such a dissociation in V5 and V4, relating attentional set and evoked activity in accord with C. (dave@fil.ion.ucl.ac.uk) Selective attention to color or motion enhances activity in specialized areas of extrastriate cortex, but mechanisms of attentional modulation remain unclear. By dissociating modulation of visually evoked transient activity from the baseline for a particular attentional set, human functional neuroimaging was used to investigate the physiological basis of such effects. Baseline activity in motion-and color-sensitive areas of extrastriate cortex was enhanced by selective attention to these attributes, even without moving or colored stimuli. Further, visually evoked responses increased along with baseline activity. These results are consistent with the hypothesis that attention modulates sensitivity of neuronal populations to inputs by changing background activity. 'Correct' denotes events that the subject responded to correctly. 'Incorrect' denotes false-positive responses.