Abstract
We have previously shown that perceiving contour-defined loops is better when attending to objects with congruent features relative to attending to incongruent features. At which level does this feature-based attention influence object perception? A key step of object perception is object closure, and recent electrophysiological studies suggest that closure is associated with negativity at ∼320 ms over latero-occipital electrodes, termed Ncl. Here we investigated whether the Ncl and other, relatively late perceptual components are modulated by feature-based attention. While recording high-density event-related potentials (ERP's), we presented random arrays of gabors to the left and right of a central fixation point. Either on the right or the left side the gabors formed a loop that was either contour-defined or motion-defined, and afterwards participants guessed the side on which the loop appeared. To cue attention to contours or motion, in separate blocks 80% of the trials showed contour- or motion-defined loops, respectively, and only 20% of the trials showed the other feature. We found that people were more accurate in perceiving contour-defined loops when they expected contours and they showed an equivalent cueing effect for motion, thus, confirming the role of feature-based attention in object perception. Our ERP results showed more pronounced negativity for contours than for motion at ∼380 ms over latero-occipital recording sites, consistent with previous reports of the Ncl. Furthermore, cueing attention to contours resulted in a stimulus-specific modulation of the ERP from ∼300 ms. Surprisingly, however, the influence of attention was opposite to that of the Ncl. Valid cues resulted in relatively greater positivity, not negativity, compared to invalid cues. Our data suggest that the Ncl is independent of mechanisms conveying feature-based attention. Further research is required to clarify the stimulus- and attention-based mechanisms of later processes of object perception.