Experimental paradigms for probing attention. (A) Covert spatial attention task (modeled after ()). The observer fixates on the +. A transient, spatially localized visual stimulus cues the observer about where to attend. After a randomized delay, a probe stimulus appears at the cued location (“valid” trial, as shown) or an uncued location (“invalid” trial, not shown). The observe must make a speeded response indicating detection of the probe. In normal observers, reaction times on valid trials are shorter than invalid trials, all other variables held constant. (B) Schematic illustration of Moran and Desimone’s () original finding of attentional modulation in extrastriate visual cortex. Monkeys were trained to attend to a small region on a computer screen (indicated in red). Monkeys performed a match-to-sample task (while maintaining central fixation) that required covertly attending to the stimulus on either the left (upper panel) or right (lower panel) side of the neuronal receptive field (dashed line), while ignoring the other. The location of the attended region was varied in blocks. In both types of trial, the pattern of retinal stimulation was identical. Stimuli were selected to be either effective or non-effective at eliciting visual responses in the absence of attention. Bar chart on the right summarizes the key finding in V4: although the pattern of retinal stimulation in the attend “good” and “bad” configurations was identical, many V4 neurons responded as if the attended stimulus were presented in isolation, effectively filtering out the unattended stimulus.