Abstract
We studied the neural correlates of edge detection and scene segmentation, and to what extent these depend on attention. We used texture defined stimuli, that made it possible to distinguish between EEG activity related to texture boundary detection (TB) and activity related to texture surface segregation (SG). We presented these during three different conditions, a condition of inattention, a condition of non-attention and a condition of attention. Inattentional blindness (IB) was induced by presenting the target stimuli conjoint with a rapid serial presentation of letters that subjects had to focus on. We only analysed neural data from subjects suffering IB (50% of subjects did not see at least 200 targets). These subjects were presented with this paradigm again while they had to respond to the letters for a second time (the non-attention condition) and while they had to respond towards the target stimuli (attention). During inattention we observed both TB and SG activity. This activity is larger during the non-attention, and largest during the attention condition. We furthermore observed a temporal gradient in the SG activity, starting first in parietal channels, and appearing later in occipital channels, indicative of feedback processing. This was most strongly observed in the attention condition. We conclude that surface segregation signals do evolve independent of attention, yet at the same time are strongly influenced by it. It furthermore seems that these signals are strongly influenced by feedback from parietal to occipital cortex, which is consistent with a role of attention in surface segregation.