Abstract
Graph neural networks (GNN) rely on graph operations that include neural network training for various graph related tasks. Recently, several attempts have been made to apply the GNNs to functional magnetic resonance image (fMRI) data. Despite recent progresses, a common limitation is its difficulty to explain the classification results in a neuroscientifically explainable way. Here, we develop a framework for analyzing the fMRI data using the Graph Isomorphism Network (GIN), which was recently proposed as a powerful GNN for graph classification. One of the important contributions of this paper is the observation that the GIN is a dual representation of convolutional neural network (CNN) in the graph space where the shift operation is defined using the adjacency matrix. This understanding enables us to exploit CNN-based saliency map techniques for the GNN, which we tailor to the proposed GIN with one-hot encoding, to visualize the important regions of the brain. We validate our proposed framework using large-scale resting-state fMRI (rs-fMRI) data for classifying the sex of the subject based on the graph structure of the brain. The experiment was consistent with our expectation such that the obtained saliency map show high correspondence with previous neuroimaging evidences related to sex differences.
Highlights
Graphs provide an efficient way to mathematically model non-regular interactions between data in terms of nodes and edges (Bassett and Bullmore, 2009; He and Evans, 2010; Sporns, 2018)
Another important contribution of this work is to show that while Graph Isomorphism Network (GIN) is similar to spectral-domain approaches such as the graph convolutional network (GCN) in learning the spectral filters from graphs, GIN can be considered as a dual representation of the convolutional neural network (CNN) with two-tab convolution filter in the graph space where the adjacency matrix is defined as a generalized shift operation
We proposed a framework for analyzing the functional magnetic resonance image (fMRI) data with the GIN
Summary
Graphs provide an efficient way to mathematically model non-regular interactions between data in terms of nodes and edges (Bassett and Bullmore, 2009; He and Evans, 2010; Sporns, 2018). Our classification results on sex classification confirmed that GIN method can provide more powerful classification performance, but the direct calculation of the graph saliency map was not clear Another important contribution of this work is to show that while GIN is similar to spectral-domain approaches such as the graph convolutional network (GCN) in learning the spectral filters from graphs, GIN can be considered as a dual representation of the convolutional neural network (CNN) with two-tab convolution filter in the graph space where the adjacency matrix is defined as a generalized shift operation. Experimental results on sex classification confirm that our method can provide more accurate classification performance and better interpretability of the classification results in terms of saliency maps, which provide some new insights to the topic of sex differences on the resting-state fMRI (rs-fMRI)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
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
