Brain dynamics via Cumulative Auto-Regressive Self-Attention

TL;DR

This paper introduces a shallow autoregressive self-attention model that outperforms deep graph neural networks in predicting brain connectivity and classifying schizophrenia from neuroimaging data.

Contribution

The work presents a novel shallow model that learns autoregressive structures and directed connectivity graphs, surpassing deep GNNs in brain imaging classification tasks.

Findings

Outperforms deep GNNs in predictive accuracy

Generates directed connectivity graphs highlighting predictive components

Effective in classifying schizophrenia from neuroimaging data

Abstract

Multivariate dynamical processes can often be intuitively described by a weighted connectivity graph between components representing each individual time-series. Even a simple representation of this graph as a Pearson correlation matrix may be informative and predictive as demonstrated in the brain imaging literature. However, there is a consensus expectation that powerful graph neural networks (GNNs) should perform better in similar settings. In this work, we present a model that is considerably shallow than deep GNNs, yet outperforms them in predictive accuracy in a brain imaging application. Our model learns the autoregressive structure of individual time series and estimates directed connectivity graphs between the learned representations via a self-attention mechanism in an end-to-end fashion. The supervised training of the model as a classifier between patients and controls results in a model that generates directed connectivity graphs and highlights the components of the time-series that are predictive for each subject. We demonstrate our results on a functional neuroimaging dataset classifying schizophrenia patients and controls.