Benchmarking Out-of-Distribution Generalization Capabilities of DNN-based Encoding Models for the Ventral Visual Cortex

TL;DR

This study evaluates how well DNN-based encoding models predict macaque visual cortex responses under distribution shifts, revealing significant performance drops and highlighting image similarity as a key predictor of generalization.

Contribution

The paper introduces MacaqueITBench, a large-scale dataset for assessing out-of-distribution generalization of neural encoding models in macaque visual cortex.

Findings

Models perform significantly worse on out-of-distribution images, retaining as little as 20% of in-distribution performance.

Cosine similarity between image representations predicts neural predictivity under distribution shifts.

The dataset and benchmarks are publicly available for further research.

Abstract

We characterized the generalization capabilities of DNN-based encoding models when predicting neuronal responses from the visual cortex. We collected \textit{MacaqueITBench}, a large-scale dataset of neural population responses from the macaque inferior temporal (IT) cortex to over $300,000$ images, comprising $8,233$ unique natural images presented to seven monkeys over $109$ sessions. Using \textit{MacaqueITBench}, we investigated the impact of distribution shifts on models predicting neural activity by dividing the images into Out-Of-Distribution (OOD) train and test splits. The OOD splits included several different image-computable types including image contrast, hue, intensity, temperature, and saturation. Compared to the performance on in-distribution test images -- the conventional way these models have been evaluated -- models performed worse at predicting neuronal responses to out-of-distribution images, retaining as little as $20\%$ of the performance on in-distribution test images. The generalization performance under OOD shifts can be well accounted by a simple image similarity metric -- the cosine distance between image representations extracted from a pre-trained object recognition model is a strong predictor of neural predictivity under different distribution shifts. The dataset of images, neuronal firing rate recordings, and computational benchmarks are hosted publicly at: https://bit.ly/3zeutVd.