Using Frequency Tagging to Study the Effect of Category Learning on Visual Attention to Object Parts

Subordinate-level category learning is known to cause changes in attentional modulation for learned stimuli outside of the category learning task. In our current study we investigated how visual attention changes after training on categorization. In Experiment 1, participants were trained over six sessions to categorize a set of space plant stimuli based on three out of six features. In the following Steady-state EEG session, categorization-relevant features and categorization-irrelevant features were frequency-tagged. On each trial, participants were cued to monitor either relevant features or irrelevant features of a space plant from the trained stimulus set to detect the onset of a small red dot, they performed the same task on another set of untrained stimuli as well. We found an attentional effect on all attended features regardless of their relevancy to categorization or training status. In Experiment 2 we tested whether it was simply easy to attend to the cued features even without training and additionally tested whether object-based attention mechanisms affected the effect of attention on the steady state EEG. Participants performed the same target detection task on the same two sets of stimuli, and two additional sets of stimuli were created by modifying the original sets so that the exemplars were composed of individual parts instead of being whole objects. In untrained participants, there was no significant difference between steady state frequencies of attended and unattended features in any condition, suggesting that effects of training might have generalized to untrained stimuli in Experiment 1. Combining results from both Experiment 1 and 2, the attentional effect in Experiment 1 could come from a flexible spatial attention template or enhanced push-pull mechanisms of spatial attention.