Seeing the world from above: Uncovering the neurocognitive basis of human expertise in scene recogni
Abstract
Project Abstract (Approved for Public Release)The long-term goal of the proposed project is to identify the neurocognitive mechanism,s and temporal dynamics underlying expertise in visual scene recognition. We hypothesize that as people get better at recognizing vi,sual environments, they no longer represent them based solely on their visual properties, but instead, they form richer representati,ons based on higher-level scene properties, primarily the potential for action in the scene (aka environmental affordances). Scene e,xpertise, in that respect, can be conceptualized as a bridge connecting (scene) perception and action. To pursue this idea, the prop,osed project aims to establish how through experience and practice, incoming visual information is transformed into behaviorally rel,evant information that can guide subsequent exploration of the environment. We will pursue three specific objectives: (1) elucidate,the perceptual and cognitive processes underlying long-term training in scene recognition; (2) determine the temporal neural dynamic,s of scene exploration (i.e. how diagnostic scene information is actively extracted from the scene over time), and (3) establish how, experience in scene recognition impacts active scene exploration. Thus, the proposed project will address both how people gain expe,rience in "passively" taking in visual information, as well how they actively scan visual environments and will culminate in a merge,r of the two, in order to provide an integrative account of real-world expertise in scene recognition. To achieve this endeavor, we, will conduct a series of studies that combine behavioral, eye-tracking, and electrophysiological measures while employing a varied,set of data recording and analysis techniques, including Event-Related Potentials (ERPs), Fixation-Related Potentials (FRPs: a novel, technique measuring brain activity tied to gaze behavior), and global frequency measures embedded in fixed- and free-viewing paradi,gms. Nave participants will undergo intensive long-term training in recognizing images of different unfamiliar places, each place p,hotographed from an aerial viewpoint and a terrestrial viewpoint. Behavioral and electrophysiological data will be collected at mult,iple time points throughout the course of training to track down the formation of expertise. We focus on aerial imagery, a class of,scenes most people have very little experience with, as it presents two main advantages: First, it allows us to test how the scene r,ecognition system accommodates for visual input it has not evolved to deal with, thus revealing its potential for plasticity. Second,, contrasting aerial- and terrestrial-scene recognition allows us to address a pertinent DoD-relevant question, namely, how 2-D (aer,ial) representations of space are transformed and mapped onto 3-D representations of terrestrial imagery.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jul 13, 2022
- Source ID
- N000142212511
Entities
People
- Assaf Harel
Organizations
- Office of Naval Research
- United States Navy
- Wright State University