Abstract
Although functional connectivity and associated graph theory measures (e.g., centrality; how centrally important to the network a region is) are widely used in brain research, the full extent to which these functional measures are related to the underlying structural connectivity is not yet fully understood. The most successful recent whole-brain methods have managed to account for 36% of the variance in functional connectivity based on structural connectivity. Graph neural network deep learning methods have not yet been applied for this purpose, and offer an ideal model architecture for working with connectivity data given their ability to capture and maintain inherent network structure. This model applied here to predict functional connectivity and centrality from structural connectivity accounted for 81% of the variance in mean functional connectivity, 48% of the variance in individual-level functional connectivity, 99% of the variance in mean functional centrality, and 73% of the variance in individual-level functional centrality. Regions of particular importance to the model’s performance as determined through lesioning are discussed, whereby regions with higher centrality have a higher impact on model performance. Future research on models of patient, demographic, or behavioural data can also benefit from this graph neural network method as it is ideally-suited for capturing connectivity and centrality in brain networks. These results have set a new benchmark for prediction of functional connectivity from structural connectivity and provide a novel finding that functional centrality can be robustly predicted from structural connectivity and centrality. Models like this may ultimately lead to a way to predict functional connectivity in individuals who are unable to do fMRI tasks (e.g., non-responsive patients).
Competing Interest Statement
The authors have declared no competing interest.