Figure 1.

Experimental paradigm. a, Attention to working-memory task. Subjects memorize two exemplars from different object categories. After an initial delay (crosshair fixation), they are cued to remember either one of these two stimuli by the number 1 or 2 (selective attention trial) or both stimuli by an ! symbol (nonselective attention trial). After a short delay, a triplet array is presented with one exemplar from each object category, followed by a backward mask. Subjects subsequently made a match/no-match decision depending on whether the attended stimulus/stimuli was present in the array. b, Visuoperceptual task. Subject viewed eight presentations of the same exemplar interleaved with fixation crosses, counting the number of times a shading square appeared in a random location, dimming part of the image. Next, a response screen appeared with two possible answers of which the subject had to pick the correct total number of dimmings. All times are given in milliseconds. The color of the fixation cross indicated the ongoing task to the participant. c, Break-down of trial types for each condition during each scanning session (every subject was scanned twice). Overall, the attended exemplar (in selective attention trials) or one of the attended exemplars (in nonselective attention trials) was present in 33% of trials. Therefore, a match decision was expected on these trials.