{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/disease-prediction-using-graph-convolutional","title":"Disease Prediction using Graph Convolutional Networks: Application to Autism Spectrum Disorder and Alzheimer's Disease","arxiv_id":"1806.01738","date":"2018-06-05","proceeding":null,"authors":["Sarah Parisot","Sofia Ira Ktena","Enzo Ferrante","Matthew Lee","Ricardo Guerrero","Ben Glocker","Daniel Rueckert"],"abstract":"Graphs are widely used as a natural framework that captures interactions\nbetween individual elements represented as nodes in a graph. In medical\napplications, specifically, nodes can represent individuals within a\npotentially large population (patients or healthy controls) accompanied by a\nset of features, while the graph edges incorporate associations between\nsubjects in an intuitive manner. This representation allows to incorporate the\nwealth of imaging and non-imaging information as well as individual subject\nfeatures simultaneously in disease classification tasks. Previous graph-based\napproaches for supervised or unsupervised learning in the context of disease\nprediction solely focus on pairwise similarities between subjects, disregarding\nindividual characteristics and features, or rather rely on subject-specific\nimaging feature vectors and fail to model interactions between them. In this\npaper, we present a thorough evaluation of a generic framework that leverages\nboth imaging and non-imaging information and can be used for brain analysis in\nlarge populations. This framework exploits Graph Convolutional Networks (GCNs)\nand involves representing populations as a sparse graph, where its nodes are\nassociated with imaging-based feature vectors, while phenotypic information is\nintegrated as edge weights. The extensive evaluation explores the effect of\neach individual component of this framework on disease prediction performance\nand further compares it to different baselines. The framework performance is\ntested on two large datasets with diverse underlying data, ABIDE and ADNI, for\nthe prediction of Autism Spectrum Disorder and conversion to Alzheimer's\ndisease, respectively. Our analysis shows that our novel framework can improve\nover state-of-the-art results on both databases, with 70.4% classification\naccuracy for ABIDE and 80.0% for ADNI.","url_abs":"http://arxiv.org/abs/1806.01738v1","url_pdf":"http://arxiv.org/pdf/1806.01738v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"disease-prediction-using-graph-convolutional","repo_url":"https://github.com/parisots/population-gcn","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"GPL-3.0"}}],"tasks":[{"task_slug":"disease-prediction","task_name":"Disease Prediction"},{"task_slug":"classification","task_name":"General Classification"}],"methods":[{"method_slug":"graph-convolutional-networks","method_name":"Graph Convolutional Networks"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1806.01738","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}