Recurrent network dynamics explain the time course of perceptual grouping in natural scenes
Abstract
How the brain groups image elements into coherent objects in natural scenes remains unclear. Existing models for grouping in biological vision rely on simplified stimuli with explicit boundaries and cannot explain human behavior in naturalistic settings. We propose a mechanistic framework in which recurrent interactions propagate enhanced neuronal activity within and between cortical areas. The framework links computational principles, neural circuitry and perceptual psychology. Local boundary signals govern early grouping, whereas later stages integrate top-down feedback carrying information about object identity, grouping features across internal edges into coherent object representations. We instantiated this framework as a brain-inspired recurrent neural network trained to group features in natural images. The network learned to propagate enhanced activity across a cued object's representation, mirroring the brain's perceptual grouping mechanisms. The network's dynamics also predicted human reaction times. The framework links cortical dynamics to human vision and generates testable predictions for neuro-physiological and psychophysical experiments.
Related
- A cascade of attentional and default mode network dynamics following the resolution of perceptual ambiguity
- A molecularly defined recurrent CA3 circuit regulates network dynamics and spatial working memory
- A theory of noisy linear recurrent neural network learning dynamics
- Planning as Dynamics Relaxation: Hippocampal Recurrent Network Realizes Optimal Goal-Directed Navigation
- A piece-wise linear recurrent neural network identifies generalizable dynamics from neural activity during decision-making in rodents
BibTeX
@misc{mollard2026recurrent,
title = {{Recurrent network dynamics explain the time course of perceptual grouping in natural scenes}},
author = {Sami Mollard and Alekh K. Ashok and Lore Goetschalckx and Drew Linsley and Thomas Serre and Sander M. Bohte and Pieter R. Roelfsema},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.05419},
eprint = {2610.05419},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.05419},
note = {CC BY-NC-SA 4.0},
}
RIS
TY - UNPB
TI - Recurrent network dynamics explain the time course of perceptual grouping in natural scenes
AU - Sami Mollard
AU - Alekh K. Ashok
AU - Lore Goetschalckx
AU - Drew Linsley
AU - Thomas Serre
AU - Sander M. Bohte
AU - Pieter R. Roelfsema
PY - 2026
DA - 2026/10/04
PB - arXiv
DO - 10.48550/arxiv.2610.05419
UR - https://arxiv.org/abs/2610.05419
AB - How the brain groups image elements into coherent objects in natural scenes remains unclear. Existing models for grouping in biological vision rely on simplified stimuli with explicit boundaries and cannot explain human behavior in naturalistic settings. We propose a mechanistic framework in which recurrent interactions propagate enhanced neuronal activity within and between cortical areas. The framework links computational principles, neural circuitry and perceptual psychology. Local boundary signals govern early grouping, whereas later stages integrate top-down feedback carrying information about object identity, grouping features across internal edges into coherent object representations. We instantiated this framework as a brain-inspired recurrent neural network trained to group features in natural images. The network learned to propagate enhanced activity across a cued object's representation, mirroring the brain's perceptual grouping mechanisms. The network's dynamics also predicted human reaction times. The framework links cortical dynamics to human vision and generates testable predictions for neuro-physiological and psychophysical experiments.
ER -