Abstract
A central question in neuroscience is how self-organizing dynamic interactions in the brain emerge on their relatively static structural backbone. Due to the complexity of spatial and temporal dependencies between different brain areas, fully comprehending the interplay between structure and function is still challenging and an area of intense research. In this paper we present a graph neural network (GNN) framework, to describe functional interactions based on the structural anatomical layout. A GNN allows us to process graph-structured spatio-temporal signals, providing a possibility to combine structural information derived from diffusion tensor imaging (DTI) with temporal neural activity profiles, like that observed in functional magnetic resonance imaging (fMRI). Moreover, dynamic interactions between different brain regions discovered by this data-driven approach can provide a multi-modal measure of causal connectivity strength. We assess the proposed model’s accuracy by evaluating its capabilities to replicate empirically observed neural activation profiles, and compare the performance to those of a vector auto regression (VAR), like that typically used in Granger causality. We show that GNNs are able to capture long-term dependencies in data and also computationally scale up to the analysis of large-scale networks. Finally we confirm that features learned by a GNN can generalize across MRI scanner types and acquisition protocols, by demonstrating that the performance on small datasets can be improved by pre-training the GNN on data from an earlier study. We conclude that the proposed multi-modal GNN framework can provide a novel perspective on the structure-function relationship in the brain. Accordingly this approach appears to be promising for the characterization of the information flow in brain networks.
Highlights
A central question in neuroscience is how self-organizing dynamic interactions in the brain emerge on their relatively static structural backbone
In the context of brain dynamics, the input sequence corresponds to measurements of the BOLD signal x(t) ∈ RN in N brain regions at Tp time points, while the objective is to predict the signal at Tf subsequent time points
We introduced a multi-modal framework for inferring causal relations in brain networks, based on a graph neural network architecture, uniting structural and functional information observed with diffusion tensor imaging (DTI) and functional magnetic resonance imaging (fMRI)
Summary
A central question in neuroscience is how self-organizing dynamic interactions in the brain emerge on their relatively static structural backbone. Connectivity measures derived from different modalities can provide distinct, but complementary aspects of brain connectivity[11,12,13] Still, studying their relations is challenging mainly due to the complex spatio-temporal dependencies and inherent difficulty in long term forecasting. Spatial anatomical layout, can be interpreted as time-varying graph structured signals For such applications, graph neural networks (GNN) have shown to be useful, providing a possibility to process data with graph-like properties in the framework of artificial neural networks (ANN)[14]. By trying to make accurate predictions of temporal neural profiles, GC tests if adding information about neural activity in brain region B helps to improve the prediction of the activity in region A (and vice versa) This provides an exploratory measure for directed causal dependencies between segregated brain areas
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
