Leon Hu¹ (), Alenka Doyle¹, Emily Gonzales¹, Michael Grubb¹; ¹Trinity College - Hartford, CT

Attentional capture has traditionally been viewed as either a top-down (goal-driven) or bottom-up (stimulus-driven) process. However, recent work suggests a third type: attentional capture can be driven by associations with prior experiences. In recently published work, we demonstrated increased attentional capture by task-irrelevant, non-salient distractor colors associated with a history of providing instrumental (relative to redundant) information. The current study investigates whether such information-driven attentional capture occurs when stimuli have a history of predicting information versus providing it directly as in our previous work. Here, we report data from the training phase of our new experiment: on each trial, participants first reported a stimulus color (red or green), then received an endogenous cue (high info/low info) probabilistically tied to the preceding color. Participants then identified the orientation of a uniquely oriented Gabor patch out of four options; high info cues indicated the exact location of the target Gabor, and low info cues indicated two possible locations. Participants completed 4 blocks of 60 trials. In the training phase, we aim to determine whether a) task-relevant information is used to improve task performance, and b) whether the information-color associations are learned. Data from the first 40 participants provides evidence of improved Gabor task accuracy and RTs for high vs. low info cues. Additionally, by the last training phase block, RTs in the color discrimination task were significantly faster for high- vs. low-information-predicting colors, with no speed-accuracy trade-off, supported by multiple statistical analyses. These results suggest that the high info cues helped improve overall performance and, given the faster RTs for high-info-predicting colors, that participants gradually learn associations between color and information prediction throughout the experiment’s training phase.

Acknowledgements: Supported by NSF-2141860 CAREER Award to MAG