Nisa Hofer¹, Amira Jensen¹, Jenni M. Karl¹; ¹Thompson Rivers University

Vision research often focuses on bottom-up models, while top-down models receive less attention. We aimed to determine whether high-level knowledge of an object’s action possibilities could bias low-level visual processing in favor of the dorsal vs. ventral stream. We hypothesized that objects within reach that afford action would bias processing toward other action-relevant features (e.g., orientation) in the dorsal stream; whereas objects beyond reach that don’t afford action would bias processing toward action-irrelevant features (e.g., saturation) in the ventral stream. A 2 x 2 x 2 repeated measures design was used to analyze the effects of object location (within vs. beyond reach), object graspability (graspable vs. non-graspable), and change type (orientation vs. saturation) on participant Accuracy and Reaction Times in a visual change detection task. Participants sat in front of a monitor and viewed a fixation cross (1500ms) followed by 1st object presentation (2500ms), a visual mask (900ms), then 2nd object presentation (up to 3000ms). Between the 1st and 2nd presentation, the object did not change, or increased/decreased in either orientation (14°) or saturation (14%) and the participant pressed a key to indicate the change they observed. The results revealed that participants were faster and more accurate at detecting orientation compared to saturation changes, and at detecting changes in graspable compared to non-graspable objects. They were also more accurate at detecting changes when the object (i.e., monitor) was located beyond reach. The results failed to support our hypothesis, but provide valuable insight for future research. Limitations of the study design such as task difficulty, participant fatigue, and inadequate control of low-level object features are discussed; as is another possibility – that participants may have used bottom-up strategies to complete the task and thus, conscious attention to the object’s affordances may be necessary to activate top-down influences on visual processing.

Acknowledgements: We acknowledge the support of the Natural Sciences and Engineering Research Council of Canada (NSERC), [funding reference number RGPIN-2017-05995] and Thompson Rivers University.