CogSci 2025

•

July 31, 2025

•

San Francisco, United States

keywords:

psychophysics

computational neuroscience

computational modeling

computer science

neuroscience

psychology

representation

perception

vision

neural networks

Some researchers have argued that visual representations in the periphery differ qualitatively from those in the fovea (e.g., Balas et al., 2009; Freeman and Simoncelli, 2011; Rosenholtz et al., 2012). Consistent with this proposal, He et al. (1997) showed that crowding in the periphery disrupts the ability to individuate features but doesn’t disrupt feature detection. We hypothesized that He et al.’s demonstration could be accounted for simply in terms of cortical magnification alone. We tested this hypothesis by presenting He et al.’s stimuli to a neurally realistic model of V1 (Heaton & Hummel, 2022) that incorporates cortical magnification but posits no other differences between foveal and peripheral early visual representations. The model’s performance captured He et al.’s findings, suggesting that cortical magnification alone is sufficient to account for the differences between foveal and peripheral visual perception.

Downloads

Paper

Next from CogSci 2025

Repetitive Negative Thinking Naturally Emerges in a Model that Learns to Gate Affective Content into Working Memory
poster

Repetitive Negative Thinking Naturally Emerges in a Model that Learns to Gate Affective Content into Working Memory

CogSci 2025

Peter Hitchcock

31 July 2025

Similar lecture

Cooperative Multi-Agent Learning in a Complex World: Challenges and Solutions
technical paper

Cooperative Multi-Agent Learning in a Complex World: Challenges and Solutions

AAAI 2023

Yali Du

10 February 2023