Deep generative networks reveal the tuning of neurons in IT and predict their influence on visual perception DOI Open Access
Paolo Papale, Daniela De Luca, Pieter R. Roelfsema

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Oct. 12, 2024

Finding the tuning of visual neurons has kept neuroscientists busy for decades. One approach to this problem been test specific hypotheses on relevance a property (e.g. orientation or color), build set “artificial” stimuli that vary along and then record neural responses those stimuli. Here, we present complementary, data-driven method retrieve properties neurons. Exploiting deep generative networks electrophysiology in monkeys, first used reconstruct any stimulus from its evoked neuronal activity inferotemporal cortex (IT). Then, by arbitrarily perturbing response individual cortical sites model, generated naturalistic interpretable sequences images strongly influence site. This enables discovery previously unknown high-level are easily interpretable, which tested with carefully controlled When knew drove neurons, activated cells electrical microstimulation observed predicable shift monkey perception direction preferred image. By allowing brain tell us what it cares about, no longer limited our experimental imagination.

Language: Английский

Unraveling other-race face perception with GAN-based image reconstruction DOI
Moaz Shoura, Dirk Walther, Adrian Nestor

et al.

Behavior Research Methods, Journal Year: 2025, Volume and Issue: 57(4)

Published: March 14, 2025

Language: Английский

Citations

0
Representation of Natural Contours by a Neural Population in Monkey V4 DOI Creative Commons

Itsuki Machida,

Motofumi Shishikura, Y. Yamane

et al.

eNeuro, Journal Year: 2024, Volume and Issue: 11(3), P. ENEURO.0445 - 23.2024

Published: Feb. 29, 2024

The cortical visual area, V4, has been considered to code contours that contribute the intermediate-level representation of objects. neural responses complex contour features intrinsic natural are expected clarify essence representation. To approach coding contours, we investigated simultaneous multiple in monkey (Macaca fuscata) V4 neurons and their population-level A substantial number showed significant tuning for two or more such as curvature closure, indicating a simultaneously features. large portion responded vigorously acutely curved surrounded center classical receptive field, suggesting tend prominent object contours. analysis mutual information (MI) between each feature most exhibited similar magnitudes type MI, many showing depended on We next examined by using multidimensional scaling analysis. preferences stimuli compared with silhouette increased along primary secondary axes, respectively, contribution surface textures population responses. Our analyses suggested images represent properties population.

Language: Английский

Citations

1
Deep generative networks reveal the tuning of neurons in IT and predict their influence on visual perception DOI Open Access
Paolo Papale, Daniela De Luca, Pieter R. Roelfsema

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Oct. 12, 2024

Finding the tuning of visual neurons has kept neuroscientists busy for decades. One approach to this problem been test specific hypotheses on relevance a property (e.g. orientation or color), build set “artificial” stimuli that vary along and then record neural responses those stimuli. Here, we present complementary, data-driven method retrieve properties neurons. Exploiting deep generative networks electrophysiology in monkeys, first used reconstruct any stimulus from its evoked neuronal activity inferotemporal cortex (IT). Then, by arbitrarily perturbing response individual cortical sites model, generated naturalistic interpretable sequences images strongly influence site. This enables discovery previously unknown high-level are easily interpretable, which tested with carefully controlled When knew drove neurons, activated cells electrical microstimulation observed predicable shift monkey perception direction preferred image. By allowing brain tell us what it cares about, no longer limited our experimental imagination.

Language: Английский

Citations

0