Targeted perturbations reveal brain-like local coding axes in robustified, but not standard, ANN-based brain models

2025/09/27 by Nikolas McNeal, N. Apurva Ratan Murty, McNeal, Nikolas +1
Computer Science · #Computer Vision and Pattern Recognition (cs.CV) #FOS: Biological sciences #FOS: Computer and information sciences #Machine Learning (cs.LG) #Neural Networks and Applications #Neurons and Cognition (q-bio.NC)

paper · pdf · doi:10.48550/arxiv.2509.23333

openalex publication_date 2025/09/27 · openalex created_date 2025/10/19 · openalex updated_date 2026/07/28

Abstract

Artificial neural networks (ANNs) have become the de facto standard for modeling the human visual system, primarily due to their success in predicting neural responses. However, with many models now achieving similar predictive accuracy, we need a stronger criterion. Here, we use small-scale adversarial probes to characterize the local representational geometry of many highly predictive ANN-based brain models. We report four key findings. First, we show that most contemporary ANN-based brain models are unexpectedly fragile. Despite high prediction scores, their response predictions are highly sensitive to small, imperceptible perturbations, revealing unreliable local coding directions. Second, we demonstrate that a model's sensitivity to adversarial probes can better discriminate between candidate neural encoding models than prediction accuracy alone. Third, we find that standard models rely on distinct local coding directions that do not transfer across model architectures. Finally, we show that adversarial probes from robustified models produce generalizable and semantically meaningful changes, suggesting that they capture the local coding dimensions of the visual system. Together, our work shows that local representational geometry provides a stronger criterion for brain model evaluation. We also provide empirical grounds for favoring robust models, whose more stable coding axes not only align better with neural selectivity but also generate concrete, testable predictions for future experiments.

Citations

• Noninvasive precision modulation of high-level neural population activity via natural vision perturbations
• Nomic Embed Vision: Expanding the Latent Space
• A Spectral Theory of Neural Prediction and Alignment
• Performance-optimized deep neural networks are evolving into worse models of inferotemporal visual cortex
• BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models
• The neuroconnectionist research programme
• Adversarially trained neural representations may already be as robust as corresponding biological neural representations
• Learning Transferable Visual Models From Natural Language Supervision
• An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
• Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attacks
• Adversarial Examples Are Not Bugs, They Are Features
• Adversarial Examples that Fool both Computer Vision and Time-Limited Humans
• One Pixel Attack for Fooling Deep Neural Networks
• Towards Deep Learning Models Resistant to Adversarial Attacks
• Universal adversarial perturbations
• Densely Connected Convolutional Networks
• Adversarial examples in the physical world
• SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
• Deep Residual Learning for Image Recognition
• Rethinking the Inception Architecture for Computer Vision
• Explaining and Harnessing Adversarial Examples
• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Intriguing properties of neural networks
• ImageNet classification with deep convolutional neural networks

Related

• A Turing Test for Artificial Nets devoted to model Human Vision
• Limited but consistent gains in adversarial robustness by co-training object recognition models with human EEG
• Dissociable neural representations of adversarially perturbed images in convolutional neural networks and the human brain
• Why all roads don't lead to Rome: Representation geometry varies across the human visual cortical hierarchy
• Uncovering Brain-Like Hierarchical Patterns in Vision-Language Models through fMRI-Based Neural Encoding