Frontiers in Psychiatry (Dec 2020)
Ensemble Learning of Convolutional Neural Network, Support Vector Machine, and Best Linear Unbiased Predictor for Brain Age Prediction: ARAMIS Contribution to the Predictive Analytics Competition 2019 Challenge
- Baptiste Couvy-Duchesne,
- Baptiste Couvy-Duchesne,
- Baptiste Couvy-Duchesne,
- Baptiste Couvy-Duchesne,
- Baptiste Couvy-Duchesne,
- Baptiste Couvy-Duchesne,
- Johann Faouzi,
- Johann Faouzi,
- Johann Faouzi,
- Johann Faouzi,
- Johann Faouzi,
- Benoît Martin,
- Benoît Martin,
- Benoît Martin,
- Benoît Martin,
- Benoît Martin,
- Elina Thibeau–Sutre,
- Elina Thibeau–Sutre,
- Elina Thibeau–Sutre,
- Elina Thibeau–Sutre,
- Elina Thibeau–Sutre,
- Adam Wild,
- Adam Wild,
- Adam Wild,
- Adam Wild,
- Adam Wild,
- Manon Ansart,
- Manon Ansart,
- Manon Ansart,
- Manon Ansart,
- Manon Ansart,
- Stanley Durrleman,
- Stanley Durrleman,
- Stanley Durrleman,
- Stanley Durrleman,
- Stanley Durrleman,
- Didier Dormont,
- Didier Dormont,
- Didier Dormont,
- Didier Dormont,
- Didier Dormont,
- Didier Dormont,
- Ninon Burgos,
- Ninon Burgos,
- Ninon Burgos,
- Ninon Burgos,
- Ninon Burgos,
- Olivier Colliot,
- Olivier Colliot,
- Olivier Colliot,
- Olivier Colliot,
- Olivier Colliot
Affiliations
- Baptiste Couvy-Duchesne
- Paris Brain Institute, ICM, Paris, France
- Baptiste Couvy-Duchesne
- Inserm, U 1127, Paris, France
- Baptiste Couvy-Duchesne
- CNRS, UMR 7225, Paris, France
- Baptiste Couvy-Duchesne
- Sorbonne Université, Paris, France
- Baptiste Couvy-Duchesne
- Inria Paris, Aramis project-team, Paris, France
- Baptiste Couvy-Duchesne
- Institute for Molecular Bioscience, The University of Queensland, St Lucia, QLD, Australia
- Johann Faouzi
- Paris Brain Institute, ICM, Paris, France
- Johann Faouzi
- Inserm, U 1127, Paris, France
- Johann Faouzi
- CNRS, UMR 7225, Paris, France
- Johann Faouzi
- Sorbonne Université, Paris, France
- Johann Faouzi
- Inria Paris, Aramis project-team, Paris, France
- Benoît Martin
- Paris Brain Institute, ICM, Paris, France
- Benoît Martin
- Inserm, U 1127, Paris, France
- Benoît Martin
- CNRS, UMR 7225, Paris, France
- Benoît Martin
- Sorbonne Université, Paris, France
- Benoît Martin
- Inria Paris, Aramis project-team, Paris, France
- Elina Thibeau–Sutre
- Paris Brain Institute, ICM, Paris, France
- Elina Thibeau–Sutre
- Inserm, U 1127, Paris, France
- Elina Thibeau–Sutre
- CNRS, UMR 7225, Paris, France
- Elina Thibeau–Sutre
- Sorbonne Université, Paris, France
- Elina Thibeau–Sutre
- Inria Paris, Aramis project-team, Paris, France
- Adam Wild
- Paris Brain Institute, ICM, Paris, France
- Adam Wild
- Inserm, U 1127, Paris, France
- Adam Wild
- CNRS, UMR 7225, Paris, France
- Adam Wild
- Sorbonne Université, Paris, France
- Adam Wild
- Inria Paris, Aramis project-team, Paris, France
- Manon Ansart
- Paris Brain Institute, ICM, Paris, France
- Manon Ansart
- Inserm, U 1127, Paris, France
- Manon Ansart
- CNRS, UMR 7225, Paris, France
- Manon Ansart
- Sorbonne Université, Paris, France
- Manon Ansart
- Inria Paris, Aramis project-team, Paris, France
- Stanley Durrleman
- Paris Brain Institute, ICM, Paris, France
- Stanley Durrleman
- Inserm, U 1127, Paris, France
- Stanley Durrleman
- CNRS, UMR 7225, Paris, France
- Stanley Durrleman
- Sorbonne Université, Paris, France
- Stanley Durrleman
- Inria Paris, Aramis project-team, Paris, France
- Didier Dormont
- Paris Brain Institute, ICM, Paris, France
- Didier Dormont
- Inserm, U 1127, Paris, France
- Didier Dormont
- CNRS, UMR 7225, Paris, France
- Didier Dormont
- Sorbonne Université, Paris, France
- Didier Dormont
- Inria Paris, Aramis project-team, Paris, France
- Didier Dormont
- AP-HP, Hôpital de la Pitié-Salpêtrière, Department of Neuroradiology, Paris, France
- Ninon Burgos
- Paris Brain Institute, ICM, Paris, France
- Ninon Burgos
- Inserm, U 1127, Paris, France
- Ninon Burgos
- CNRS, UMR 7225, Paris, France
- Ninon Burgos
- Sorbonne Université, Paris, France
- Ninon Burgos
- Inria Paris, Aramis project-team, Paris, France
- Olivier Colliot
- Paris Brain Institute, ICM, Paris, France
- Olivier Colliot
- Inserm, U 1127, Paris, France
- Olivier Colliot
- CNRS, UMR 7225, Paris, France
- Olivier Colliot
- Sorbonne Université, Paris, France
- Olivier Colliot
- Inria Paris, Aramis project-team, Paris, France
- DOI
- https://doi.org/10.3389/fpsyt.2020.593336
- Journal volume & issue
-
Vol. 11
Abstract
We ranked third in the Predictive Analytics Competition (PAC) 2019 challenge by achieving a mean absolute error (MAE) of 3.33 years in predicting age from T1-weighted MRI brain images. Our approach combined seven algorithms that allow generating predictions when the number of features exceeds the number of observations, in particular, two versions of best linear unbiased predictor (BLUP), support vector machine (SVM), two shallow convolutional neural networks (CNNs), and the famous ResNet and Inception V1. Ensemble learning was derived from estimating weights via linear regression in a hold-out subset of the training sample. We further evaluated and identified factors that could influence prediction accuracy: choice of algorithm, ensemble learning, and features used as input/MRI image processing. Our prediction error was correlated with age, and absolute error was greater for older participants, suggesting to increase the training sample for this subgroup. Our results may be used to guide researchers to build age predictors on healthy individuals, which can be used in research and in the clinics as non-specific predictors of disease status.
Keywords