Fig. 1.

Experimental paradigm and behavioural results. A) Schematic illustration of the experimental paradigm. Trials started with a cue, presented for 500 m s, instructing participants which hemi-field to attend. After 1–1.5 s, two stimuli were presented, a grating and a line. After 1–3 s, the stimulus in the cued hemi-field was first presented at a tilted orientation (30 ms) and then replaced by a mask (120 ms), a plaid or a cross, depending on which stimulus was presented in the cued hemi-field. After 350 ms, participants were prompted to respond to the task by indicating whether the attended stimulus was tilted counter-clockwise or clockwise. For convenience, only the attend-left condition is illustrated in this figure, however, all four possible combinations of cue hemi-field (attend-left and attend-right) and stimulus hemi-field (grating-left and grating-right) were presented. B) Accuracy (i.e. percentage of correct orientation discriminations) at each magnitude of orientation change, plotted separately for each participant and also as a group average (thick lines). The error bars indicate ±1 SEM. Note that each magnitude of orientation change was increased by a factor of four, for the line stimulus (see Stimuli). C) The same as in B), but for response times (RTs) instead of accuracy rates.