Predicting the Evolution of Depressive Symptoms Using Spatiotemporal Graph Neural Networks
Resumo
Network psychopathology has emerged as an innovative paradigm for understanding mental disorders, modeling them as dynamic systems of interacting symptoms. Previous studies have predominantly focused on cross-sectional analyzes of symptom networks. In contrast, this study explores the temporal dimension through the application of Spatio-Temporal Graph Neural Networks (STGNNs) to predict the symptomatic evolution of patients with depressive symptoms. To this end, longitudinal data from the Experience Sampling Method (ESM) of 129 participants were used, with the implementation and comparison of three GNN architectures: Graph Convolutional Network (GCN), Graph Attention Network (GAT) and TransformerConv. These architectures were combined with a temporal layer of GRU in a previously established population network. The results show that TransformerConv achieved the best overall performance (R² = 0.6126), demonstrating a statistically significant result and maintaining stability at different depths and cross-validation folds. This study establishes a methodological basis for the application of STGNNs in psychopathology, offering new tools to predict symptomatic trajectories.
Referências
Borsboom, D. (2017). A network theory of mental disorders. World psychiatry.
Bringmann, L. F., Albers, C., Bockting, C., Borsboom, D., Ceulemans, E., Cramer, A., Epskamp, S., Eronen, M. I., Hamaker, E., Kuppens, P., Lutz, W., McNally, R. J., Molenaar, P., Tio, P., Voelkle, M. C., and Wichers, M. (2022). Psychopathological networks: Theory, methods and practice. Behaviour Research and Therapy.
Bringmann, L. F., Vissers, N., Wichers, M., Geschwind, N., Kuppens, P., Peeters, F., Borsboom, D., and Tuerlinckx, F. (2013). A network approach to psychopathology: New insights into clinical longitudinal data. Plos One.
Choi, K. S., Kim, S., Kim, B.-H., Jeon, H. J., Kim, J.-H., Jang, J. H., and Jeong, B. (2021). Deep graph neural network-based prediction of acute suicidal ideation in young adults. Nature.
Harit, A., Sun, Z., Yu, J., and Moubayed, N. A. (2024). Monitoring behavioral changes using spatiotemporal graphs: A case study on the studentlife dataset. In NeurIPS 2024 Workshop on Behavioral Machine Learning.
Jin, G., Liang, Y., Fang, Y., Huang, J., Zhang, J., and Zheng, Y. (2023). Spatio-temporal graph neural networks for predictive learning in urban computing: A survey. IEEE transactions on knowledge and data engineering.
Jin, M., Koh, H. Y., Wen, Q., Zambon, D., Alippi, C., Webb, G. I., King, I., and Pan, S. (2024). A survey on graph neural networks for time series: Forecasting, classification, imputation, and anomaly detection. IEEE Trans. on Pattern Analysis and Mach. Int.
Kipf, T. N. and Welling, M. (2016). Semi-supervised classification with graph convolutional networks. In International Conference on Learning Representations (ICLR).
Li, Q., Han, Z., and Wu, X.-M. (2018). Deeper insights into graph convolutional networks for semi-supervised learning. In Proc. of the AAAI conference on artificial intelligence.
Lépine, J.-P. (2011). The increasing burden of depression. Neuropsychiatric Disease and Treatment.
Ma, W., Su, Z., Chen, Q., Zhai, H., Jiang, J., and Shi, H. (2025). Advancing mental health research with graph neural networks: A comprehensive survey. In 2025 IEEE 41st International Conference on Data Engineering Workshops.
Rozemberczki, B., Scherer, P., He, Y., Panagopoulos, G., Riedel, A., Astefanoaei, M., Kiss, O., Beres, F., Lopez, G., Collignon, N., and Sarkar, R. (2021). PyTorch Geometric Temporal: Spatiotemporal Signal Processing with Neural Machine Learning Models. In Proc. of the 30th ACM Int. Conf. on Inf. and Knowledge Management.
Shi, Y., Huang, Z., Feng, S., Zhong, H., Wang, W., and Sun, Y. (2021). Masked label prediction: Unified message passing model for semi-supervised classification. In Proc. of the 30th International Joint Conference on Artificial Intelligence (IJCAI-21).
Veličković, P., Cucurull, G., Casanova, A., Romero, A., Liò, P., and Bengio, Y. (2018). Graph attention networks. In 6th Int. Conf. on Learning Representations (ICLR 2018).
Wang, J., Sun, Z., Yuan, C., Li, W., Liu, A., Wei, Z., and Yin, B. (2024). Dynamic graphs attention for ocean variable forecasting. Eng. App. of Artificial Intelligence.
Wu, Z., Pan, S., Chen, F., Long, G., Zhang, C., and Yu, P. S. (2020). A comprehensive survey on graph neural networks. IEEE trans. on neural net. and learning systems.
Yang, H., Zhu, D., Liu, Y., Xu, Z., Liu, Z., Zhang, W., and Cai, J. (2024). Employing graph attention networks to decode psycho-metabolic interactions in schizophrenia. Psychiatry Research.
