Poster Presentation

Poster D in Poster Session D: Wednesday, August 5, 2:00 – 3:45 pm, Kimmel Center, Shorin & Rosenthal Rooms

Neural Representations of Target and Commonly Co-Occurring Objects Accessed Prior to Visual Search

Victoria Nicholls¹, Bhavin Choksi², Gemma Roig², Melissa Vo²; ¹Ludwig-Maximilians-Universität München, ²Johann Wolfgang Goethe Universität Frankfurt am Main

Presenter: Bhavin Choksi

In our day to day lives we search for objects regularly. If you want to brush your teeth, you need to find a toothbrush and some toothpaste. When you search for the toothpaste, you do not search the entire bathroom, you use your knowledge of how scenes are structured—toothbrushes and toothpaste are often found on the sink—to narrow down your search space and quickly find the toothbrush. Our knowledge of how scenes are structured is typically referred to as scene grammar (Võ, 2021) and is hypothesized to follow a specific hierarchical structure: At the lowest level are smaller, local objects (e.g. a soap), followed by “anchors”, larger objects (sink) which predict both the identity and location of local objects. Together they form a “phrase”, a meaningful and functionally organized subset of a scene. Multiple phrases combine to form a scene (Võ, 2021; Wolfe, 2020; Yu et al., 2023). Prior work has investigated this at a behavioural level, however, how scene grammar knowledge is represented hierarchically on a neural level, and how it is dynamically accessed, remains unknown. To examine this, we presented participants with an isolated target object (local or anchor) either as a word label, image, or target word in the context of a search task, followed by a blank period while we recorded MEG. Using representational similarity analysis (RSA) and cross-decoding approaches, we analysed each stimulus presentation and blank period to determine whether, prior to search, participants also access phrase and scene representations, in addition to objects representations, to help guide their search. During the stimulus period we found peaks for object, and phrase category models. This suggests that even when seeing isolated objects, participants automatically also access representations of local and phrasal information to guide search.

Topic Area: Memory, Learning & Knowledge Structures