Spectral graph convolutions for population-based disease prediction. Parisot, S., Ktena, S., I., Ferrante, E., Lee, M., Moreno, R., G., Glocker, B., & Rueckert, D. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 10435 LNCS(319456):177-185, 2017. ![Spectral graph convolutions for population-based disease prediction [pdf]](https://bibbase.org/img/filetypes/pdf.svg "pdf")
Paper doi abstract bibtex Exploiting the wealth of imaging and non-imaging information for disease prediction tasks requires models capable of representing, at the same time, individual features as well as data associations between subjects from potentially large populations. Graphs provide a natural framework for such tasks, yet previous graph-based approaches focus on pairwise similarities without modelling the subjects’ individual characteristics and features. On the other hand, relying solely on subject-specific imaging feature vectors fails to model the interaction and similarity between subjects, which can reduce performance. In this paper, we introduce the novel concept of Graph Convolutional Networks (GCN) for brain analysis in populations, combining imaging and non-imaging data. We represent populations as a sparse graph where its vertices are associated with image-based feature vectors and the edges encode phenotypic information. This structure was used to train a GCN model on partially labelled graphs, aiming to infer the classes of unlabelled nodes from the node features and pairwise associations between subjects. We demonstrate the potential of the method on the challenging ADNI and ABIDE databases, as a proof of concept of the benefit from integrating contextual information in classification tasks. This has a clear impact on the quality of the predictions, leading to 69.5% accuracy for ABIDE (outperforming the current state of the art of 66.8%) and 77% for ADNI for prediction of MCI conversion, significantly outperforming standard linear classifiers where only individual features are considered.
@article{
title = {Spectral graph convolutions for population-based disease prediction},
type = {article},
year = {2017},
pages = {177-185},
volume = {10435 LNCS},
id = {cde2b6b6-dd91-3968-ad67-6f9e754c1d36},
created = {2021-08-04T13:05:08.311Z},
file_attached = {true},
profile_id = {ad172e55-c0e8-3aa4-8465-09fac4d5f5c8},
group_id = {1ff583c0-be37-34fa-9c04-73c69437d354},
last_modified = {2021-08-04T13:06:02.039Z},
read = {false},
starred = {false},
authored = {false},
confirmed = {true},
hidden = {false},
private_publication = {false},
abstract = {Exploiting the wealth of imaging and non-imaging information for disease prediction tasks requires models capable of representing, at the same time, individual features as well as data associations between subjects from potentially large populations. Graphs provide a natural framework for such tasks, yet previous graph-based approaches focus on pairwise similarities without modelling the subjects’ individual characteristics and features. On the other hand, relying solely on subject-specific imaging feature vectors fails to model the interaction and similarity between subjects, which can reduce performance. In this paper, we introduce the novel concept of Graph Convolutional Networks (GCN) for brain analysis in populations, combining imaging and non-imaging data. We represent populations as a sparse graph where its vertices are associated with image-based feature vectors and the edges encode phenotypic information. This structure was used to train a GCN model on partially labelled graphs, aiming to infer the classes of unlabelled nodes from the node features and pairwise associations between subjects. We demonstrate the potential of the method on the challenging ADNI and ABIDE databases, as a proof of concept of the benefit from integrating contextual information in classification tasks. This has a clear impact on the quality of the predictions, leading to 69.5% accuracy for ABIDE (outperforming the current state of the art of 66.8%) and 77% for ADNI for prediction of MCI conversion, significantly outperforming standard linear classifiers where only individual features are considered.},
bibtype = {article},
author = {Parisot, Sarah and Ktena, Sofia Ira and Ferrante, Enzo and Lee, Matthew and Moreno, Ricardo Guerrerro and Glocker, Ben and Rueckert, Daniel},
doi = {10.1007/978-3-319-66179-7_21},
journal = {Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)},
number = {319456}
}
Downloads: 0
{"_id":"bRdE9d5nLGMtRsiSQ","bibbaseid":"parisot-ktena-ferrante-lee-moreno-glocker-rueckert-spectralgraphconvolutionsforpopulationbaseddiseaseprediction-2017","authorIDs":[],"author_short":["Parisot, S.","Ktena, S., I.","Ferrante, E.","Lee, M.","Moreno, R., G.","Glocker, B.","Rueckert, D."],"bibdata":{"title":"Spectral graph convolutions for population-based disease prediction","type":"article","year":"2017","pages":"177-185","volume":"10435 LNCS","id":"cde2b6b6-dd91-3968-ad67-6f9e754c1d36","created":"2021-08-04T13:05:08.311Z","file_attached":"true","profile_id":"ad172e55-c0e8-3aa4-8465-09fac4d5f5c8","group_id":"1ff583c0-be37-34fa-9c04-73c69437d354","last_modified":"2021-08-04T13:06:02.039Z","read":false,"starred":false,"authored":false,"confirmed":"true","hidden":false,"private_publication":false,"abstract":"Exploiting the wealth of imaging and non-imaging information for disease prediction tasks requires models capable of representing, at the same time, individual features as well as data associations between subjects from potentially large populations. Graphs provide a natural framework for such tasks, yet previous graph-based approaches focus on pairwise similarities without modelling the subjects’ individual characteristics and features. On the other hand, relying solely on subject-specific imaging feature vectors fails to model the interaction and similarity between subjects, which can reduce performance. In this paper, we introduce the novel concept of Graph Convolutional Networks (GCN) for brain analysis in populations, combining imaging and non-imaging data. We represent populations as a sparse graph where its vertices are associated with image-based feature vectors and the edges encode phenotypic information. This structure was used to train a GCN model on partially labelled graphs, aiming to infer the classes of unlabelled nodes from the node features and pairwise associations between subjects. We demonstrate the potential of the method on the challenging ADNI and ABIDE databases, as a proof of concept of the benefit from integrating contextual information in classification tasks. This has a clear impact on the quality of the predictions, leading to 69.5% accuracy for ABIDE (outperforming the current state of the art of 66.8%) and 77% for ADNI for prediction of MCI conversion, significantly outperforming standard linear classifiers where only individual features are considered.","bibtype":"article","author":"Parisot, Sarah and Ktena, Sofia Ira and Ferrante, Enzo and Lee, Matthew and Moreno, Ricardo Guerrerro and Glocker, Ben and Rueckert, Daniel","doi":"10.1007/978-3-319-66179-7_21","journal":"Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)","number":"319456","bibtex":"@article{\n title = {Spectral graph convolutions for population-based disease prediction},\n type = {article},\n year = {2017},\n pages = {177-185},\n volume = {10435 LNCS},\n id = {cde2b6b6-dd91-3968-ad67-6f9e754c1d36},\n created = {2021-08-04T13:05:08.311Z},\n file_attached = {true},\n profile_id = {ad172e55-c0e8-3aa4-8465-09fac4d5f5c8},\n group_id = {1ff583c0-be37-34fa-9c04-73c69437d354},\n last_modified = {2021-08-04T13:06:02.039Z},\n read = {false},\n starred = {false},\n authored = {false},\n confirmed = {true},\n hidden = {false},\n private_publication = {false},\n abstract = {Exploiting the wealth of imaging and non-imaging information for disease prediction tasks requires models capable of representing, at the same time, individual features as well as data associations between subjects from potentially large populations. Graphs provide a natural framework for such tasks, yet previous graph-based approaches focus on pairwise similarities without modelling the subjects’ individual characteristics and features. On the other hand, relying solely on subject-specific imaging feature vectors fails to model the interaction and similarity between subjects, which can reduce performance. In this paper, we introduce the novel concept of Graph Convolutional Networks (GCN) for brain analysis in populations, combining imaging and non-imaging data. We represent populations as a sparse graph where its vertices are associated with image-based feature vectors and the edges encode phenotypic information. This structure was used to train a GCN model on partially labelled graphs, aiming to infer the classes of unlabelled nodes from the node features and pairwise associations between subjects. We demonstrate the potential of the method on the challenging ADNI and ABIDE databases, as a proof of concept of the benefit from integrating contextual information in classification tasks. This has a clear impact on the quality of the predictions, leading to 69.5% accuracy for ABIDE (outperforming the current state of the art of 66.8%) and 77% for ADNI for prediction of MCI conversion, significantly outperforming standard linear classifiers where only individual features are considered.},\n bibtype = {article},\n author = {Parisot, Sarah and Ktena, Sofia Ira and Ferrante, Enzo and Lee, Matthew and Moreno, Ricardo Guerrerro and Glocker, Ben and Rueckert, Daniel},\n doi = {10.1007/978-3-319-66179-7_21},\n journal = {Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)},\n number = {319456}\n}","author_short":["Parisot, S.","Ktena, S., I.","Ferrante, E.","Lee, M.","Moreno, R., G.","Glocker, B.","Rueckert, D."],"urls":{"Paper":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c/file/66342442-210c-e8eb-13c1-f7b7410fb1e6/170303020.pdf.pdf"},"biburl":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c","bibbaseid":"parisot-ktena-ferrante-lee-moreno-glocker-rueckert-spectralgraphconvolutionsforpopulationbaseddiseaseprediction-2017","role":"author","metadata":{"authorlinks":{}},"downloads":0},"bibtype":"article","biburl":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c","creationDate":"2020-05-25T10:44:59.088Z","downloads":0,"keywords":[],"search_terms":["spectral","graph","convolutions","population","based","disease","prediction","parisot","ktena","ferrante","lee","moreno","glocker","rueckert"],"title":"Spectral graph convolutions for population-based disease prediction","year":2017,"dataSources":["95W8SWCJxBDir7TB5","ya2CyA73rpZseyrZ8","2252seNhipfTmjEBQ"]}