An integrative model for neural dynamics of feature selection in visual object processing: An explorative study

Objects and features represent fundamental selection units of visual attention. The associated mechanisms, however, had been mostly explored in isolation. Not least for this reason, apparently contradictory theories persist about feature selection in visual object processing. Our previous study demonstrated the time-dependent nature of this process. In the present explorative study, we used electroencephalographic recordings to examine feature-specific selection during the time course of object processing. For this, we analyzed the feature-specific selection negativities (SNs), i.e., the SNs when human subjects attended to or ignored one (color or shape) relative to attending to all (color and shape) constituent features of an object (a square). We found that the feature-specific SN amplitude of the task-relevant feature was significantly decreased in a time bin from 500 to 850 ms after trial onset in comparison to an earlier (100-450 ms) and a later time bin (900-1250 ms). Whereas, the feature-specific SN amplitude of the task-irrelevant feature was firstly positive, indicating no specific selection of this feature, and secondly, it marginally increased from the first to the second time bin. This leads to the proposed integrative model, which indicates that sustained feature-specific prioritization only emerges over time, this process consisting of three phases: after initial selection of the relevant feature and inhibition of the irrelevant feature (first time bin), further inhibition of the irrelevant feature is accompanied by co-inhibition of the relevant feature, causing prioritization to decline (second time bin), before the relevant feature eventually regains priority by being facilitated again (third time bin).