Correcting Biased Centered Kernel Alignment Measures in Biological and Artificial Neural Networks

TL;DR

This paper identifies biases in the use of Centered Kernel Alignment (CKA) for comparing neural network representations and neural data, demonstrating that debiased CKA provides more accurate, stimulus-driven similarity measures in low-data, high-dimensional settings.

Contribution

The authors highlight issues with biased CKA in neural data analysis and propose using debiased CKA to obtain more reliable, stimulus-driven similarity assessments.

Findings

Biased CKA overestimates similarity with random data.

Debiased CKA effectively distinguishes true neural responses from noise.

Biased CKA is sensitive to data structure and sample-feature ratios.

Abstract

Centred Kernel Alignment (CKA) has recently emerged as a popular metric to compare activations from biological and artificial neural networks (ANNs) in order to quantify the alignment between internal representations derived from stimuli sets (e.g. images, text, video) that are presented to both systems. In this paper we highlight issues that the community should take into account if using CKA as an alignment metric with neural data. Neural data are in the low-data high-dimensionality domain, which is one of the cases where (biased) CKA results in high similarity scores even for pairs of random matrices. Using fMRI and MEG data from the THINGS project, we show that if biased CKA is applied to representations of different sizes in the low-data high-dimensionality domain, they are not directly comparable due to biased CKA's sensitivity to differing feature-sample ratios and not stimuli-driven responses. This situation can arise both when comparing a pre-selected area of interest (e.g. ROI) to multiple ANN layers, as well as when determining to which ANN layer multiple regions of interest (ROIs) / sensor groups of different dimensionality are most similar. We show that biased CKA can be artificially driven to its maximum value when using independent random data of different sample-feature ratios. We further show that shuffling sample-feature pairs of real neural data does not drastically alter biased CKA similarity in comparison to unshuffled data, indicating an undesirable lack of sensitivity to stimuli-driven neural responses. Positive alignment of true stimuli-driven responses is only achieved by using debiased CKA. Lastly, we report findings that suggest biased CKA is sensitive to the inherent structure of neural data, only differing from shuffled data when debiased CKA detects stimuli-driven alignment.