EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A novel image-based neuronal network model framework for understanding visual multistability and neurological disorders

Kaito N Hikino, Marina Nakayama, Yihui Wu, Victor J Barranca

Frontiers in Computational Neuroscience · 29.04.2026 · 20

DOI: 10.3389/fncom.2026.1793265 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Frontiers in Computational Neuroscience, online erschienen am 29.04.2026. Verfasst von Kaito N Hikino und 3 weiteren Autor:innen. Beteiligte Einrichtung: Department of Mathematics and Statistics, Swarthmore College, Swarthmore, PA, United States. Die Arbeit nennt 137 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Swarthmore College.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

While perceptual multistability arises from many types of stimuli across different sensory systems, there are common dynamical features that may be rooted in universal organizing principles underlying perception. We probe the fundamental mechanisms responsible for visual multistability using a neuronal network model framework in which a set of realistic images directly drives competing pools of neurons with nonlinear dynamics. Incorporating balanced network architecture, long-range connections from excitatory neurons to inhibitory neurons in competing pools, and a dynamic spiking threshold, the model produces irregular percept switching and replicates key experimental observations regarding dominance durations in binocular rivalry. Using a sequence of short-time observations of neuronal dynamics, we derive a new methodology for reconstructing the dynamic percept that generalizes to an arbitrary number of percepts, suggesting how rivalry, fusion, and interocular grouping may serve as different states in a single decision-making system. The model dynamics illustrate that perceptual alternations are potentially rooted in the breakdown of balance between excitation and inhibition when the spiking thresholds of suppressed neurons become sufficiently small, with more balanced dynamics generally facilitating longer dominance durations. Finally, we apply our model analysis toward characterizing the causes of psychiatric or neurological disorders, such as amblyopia and autism. Increasing the strength of connections manifesting from the pool of neurons associated with the stronger eye in amblyopia, we find the weaker eye experiences shorter dominance durations as found experimentally, supporting the notion that sufficiently imbalanced inter-eye competition prompts the suppression of information from the monocular stimulus corresponding to the weakened eye. Similarly, we show increasing the ratio of excitatory to inhibitory inputs in the network systematically yields longer dominance durations as observed for individuals with autism, and we thus demonstrate support for the excitation/inhibition imbalance hypothesis for autism.

Schlagwörter

Stichwörter der Autor:innen

autismbalanced networkscompressive sensingmultistabilityneuronal networksnonlinear dynamics

Autor:innen und Institutionen

  • Department of Mathematics and Statistics, Swarthmore College, Swarthmore, PA, United States.

Förderung

  • Swarthmore College

Zitieren

APA

Hikino, K. N., Nakayama, M., Wu, Y., & Barranca, V. J. (2026). A novel image-based neuronal network model framework for understanding visual multistability and neurological disorders. Frontiers in Computational Neuroscience, 20. https://doi.org/10.3389/fncom.2026.1793265

Vancouver

Hikino KN, Nakayama M, Wu Y, Barranca VJ. A novel image-based neuronal network model framework for understanding visual multistability and neurological disorders. Frontiers in Computational Neuroscience. 2026;20. doi:10.3389/fncom.2026.1793265

BibTeX
@article{Hikino2026,
  author = {Hikino, Kaito N and Nakayama, Marina and Wu, Yihui and Barranca, Victor J},
  title = {A novel image-based neuronal network model framework for understanding visual multistability and neurological disorders},
  journal = {Frontiers in Computational Neuroscience},
  year = {2026},
  volume = {20},
  doi = {10.3389/fncom.2026.1793265}
}
Kostenlos starten →