CI-GNN: A Granger Causality-Inspired Graph Neural Network for Interpretable Brain Network-Based Psychiatric Diagnosis

There is a recent trend to leverage the power of graph neural networks (GNNs) for brain-network based psychiatric diagnosis, which, in turn, also motivates an urgent need for psychiatrists to fully understand the decision behavior of the used GNNs. However, most of the existing GNN explainers are either post-hoc in which another interpretive model needs to be created to explain a well-trained GNN, or do not consider the causal relationship between the extracted explanation and the decision, such that the explanation itself contains spurious correlations and suffers from weak faithfulness. In this work, we propose a granger causality-inspired graph neural network (CI-GNN), a built-in interpretable model that is able to identify the most influential subgraph (i.e., functional connectivity within brain regions) that is causally related to the decision (e.g., major depressive disorder patients or healthy controls), without the training of an auxillary interpretive network. CI-GNN learns disentangled subgraph-level representations α and β that encode, respectively, the causal and noncausal aspects of original graph under a graph variational autoencoder framework, regularized by a conditional mutual information (CMI) constraint. We also empirically evaluate the performance of CI-GNN against three baseline GNNs and four state-of-the-art GNN explainers on synthetic data and three largescale brain disease datasets. We observe that CI-GNN achieves the best performance in a wide range of metrics and provides more reliable and concise explanations which have clinical evidence. Introduction Psychiatric disorders have constituted an extensive social and economic burden for health care systems worldwide Wittchen et al. (2011), but the underlying pathological and neural mechanism of the psychiatric disorders still remains uncertain. There are no unified or neuropathological structural traits for psychiatric diagnosis due to the clinical heterogeneity Goodkind et al. (2015); Lanillos et al. (2020). Current diagnosis for psychiatric disorders are mainly based on subjective symptoms and signs Zhang et al. (2021), such as insomnia and anxiety, etc. However, this way for diagnosis has huge limitations in heavily relying on related symptoms and observational status, which could lead to misdiagnosis and delay the early diagnosis and treatment Huang et al. (2020). As a noninvasive neuroimaging technique, the functional magnetic resonance imaging (fMRI) Matthews and Jezzard (2004) has become a popular to investigate neural patterns of brain function for psychiatric disorders Peraza-Goicolea et al. (2020). Using fMRI, extensive studies in psychiatric diagnosis have been conducted to apply functional connectivity (FC) measured with the pairwise correlations of fMRI time series as features to discriminate psychiatric patients and healthy controls, as illustrated in Figure 1a.