Forecasting brain activity based on models of spatiotemporal brain dynamics: A comparison of graph neural network architectures

Abstract

Comprehending the interplay between spatial and temporal characteristics of neural dynamics can contribute to our understanding of information processing in the human brain. Graph neural networks (GNNs) provide a new possibility to interpret graph-structured signals like those observed in complex brain networks. In our study we compare different spatiotemporal GNN architectures and study their ability to model neural activity distributions obtained in functional MRI (fMRI) studies. We evaluate the performance of the GNN models on a variety of scenarios in MRI studies and also compare it to a VAR model, which is currently often used for directed functional connectivity analysis. We show that by learning localized functional interactions on the anatomical substrate, GNN-based approaches are able to robustly scale to large network studies, even when available data are scarce. By including anatomical connectivity as the physical substrate for information propagation, such GNNs also provide a multimodal perspective on directed connectivity analysis, offering a novel possibility to investigate the spatiotemporal dynamics in brain networks.

Author Summary: In our study we compare different spatial and temporal modeling techniques based on graph neural networks (GNNs) for investigating the spatiotemporal dynamics in brain networks. We show that a convolutional neural network and a recurrent neural network–based approach are both very suitable to capture the temporal characteristics in functional activity distributions. Further, we demonstrate that structural connectome embeddings can effectively reduce the number of parameters in GNN models, by naturally including higher order topological relations between brain areas within the structural network. We compare the prediction accuracy of the GNN-based approaches to a vector autoregressive model, and we show that GNNs remain considerably more accurate when brain networks become large and available data are limited. Finally, we demonstrate how these spatiotemporal GNN models can provide a multimodal perspective on directed connectivity in brain networks.