rss-bridge 2026-03-01T21:54:16.084526509+00:00

Compact deep neural network models of the visual cortex

Article

Published: 25 February 2026

Compact deep neural network models of the visual cortex

Benjamin R. Cowley

orcid.org/0000-0003-2681-24481,2,

Patricia L. Stan3,4,5,
Jonathan W. Pillow

orcid.org/0000-0002-3638-88312 na1 &

…
Matthew A. Smith

orcid.org/0000-0003-1192-99423,4,5 na1

Nature

(2026)Cite this article

4209 Accesses

100 Altmetric

Metrics

Subjects

Network models
Sensory processing
Visual system

Abstract

A powerful approach to understand the computations carried out by the visual cortex is to build models that predict neural responses to any arbitrary image. Deep neural networks (DNNs) have emerged as the leading predictive models1,2, yet their underlying computations remain buried beneath millions of parameters. Here we challenge the need for models at this scale by seeking predictive and parsimonious DNN models of the primate visual cortex. We first built a highly predictive DNN model of neural responses in macaque visual area V4 by alternating data collection and model training in adaptive closed-loop experiments. We then compressed this large, black-box DNN model, which comprised 60 million parameters, to identify compact models with 5,000 times fewer parameters yet comparable accuracy. This dramatic compression enabled us to investigate the inner workings of the compact models. We discovered a salient computational motif: compact models share similar filters in early processing, but individual models then specialize their feature selectivity by ‘consolidating’ this shared high-dimensional representation in distinct ways. We examined this consolidation step in a dot-detecting model neuron, revealing a computational mechanism that leads to a testable circuit hypothesis for dot-selective V4 neurons. Beyond V4, we found strong model compression for macaque visual areas V1 and IT (inferior temporal cortex), revealing a general computational principle of the visual cortex. Overall, our work challenges the notion that large DNNs are necessary to predict individual neurons and establishes a modelling framework that balances prediction and parsimony.

Access through your institution

Buy or subscribe

This is a preview of subscription content, access via your institution

Access options

Access through your institution

Access Nature and 54 other Nature Portfolio journals

Get Nature+, our best-value online-access subscription

27,99 € / 30 days

cancel any time

Receive 51 print issues and online access

196,21 € per year

only 3,85 € per issue

Buy this article

Purchase on SpringerLink
Instant access to the full article PDF.

39,95 €

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Identifying compact models of macaque V4 neurons.

Fig. 2: Experimentally validating the stimulus preferences of compact models.

Fig. 3: Compact models specialize their feature selectivity via a consolidation step.

Fig. 4: Uncovering the computations of a dot-detecting compact model.

Compact deep neural network models of the visual cortex