Kamalaker Dadi1, Gaël Varoquaux1,2,3, Josselin Houenou4,5, Danilo Bzdok1,3,6, Bertrand Thirion1, Denis Engemann1,7. 1. Inria, CEA, Neurospin, Parietal team, Université Paris Saclay, 91120 Palaiseau, France. 2. Montréal Neurological Institute, McGill University, Montreal, QC, Canada. 3. Mila - Quebec Artificial Intelligence Institute, Montreal, QC, Canada. 4. CEA, NeuroSpin, Psychiatry Team, UNIACT Lab, Université Paris Saclay, France. 5. APHP, Mondor University Hospitals, Psychiatry Department, INSERM U955 Team 15 "Translational Psychiatry," Créteil, France. 6. Department of Biomedical Engineering, Montreal Neurological Institute, Faculty of Medicine, McGill University, Montreal, QC, Canada. 7. Department of Neurology, Max Planck Institute for Human Cognitive and Brain Sciences, Germany.
Abstract
BACKGROUND: Biological aging is revealed by physical measures, e.g., DNA probes or brain scans. In contrast, individual differences in mental function are explained by psychological constructs, e.g., intelligence or neuroticism. These constructs are typically assessed by tailored neuropsychological tests that build on expert judgement and require careful interpretation. Could machine learning on large samples from the general population be used to build proxy measures of these constructs that do not require human intervention? RESULTS: Here, we built proxy measures by applying machine learning on multimodal MR images and rich sociodemographic information from the largest biomedical cohort to date: the UK Biobank. Objective model comparisons revealed that all proxies captured the target constructs and were as useful, and sometimes more useful, than the original measures for characterizing real-world health behavior (sleep, exercise, tobacco, alcohol consumption). We observed this complementarity of proxy measures and original measures at capturing multiple health-related constructs when modeling from, both, brain signals and sociodemographic data. CONCLUSION: Population modeling with machine learning can derive measures of mental health from heterogeneous inputs including brain signals and questionnaire data. This may complement or even substitute for psychometric assessments in clinical populations.
BACKGROUND: Biological aging is revealed by physical measures, e.g., DNA probes or brain scans. In contrast, individual differences in mental function are explained by psychological constructs, e.g., intelligence or neuroticism. These constructs are typically assessed by tailored neuropsychological tests that build on expert judgement and require careful interpretation. Could machine learning on large samples from the general population be used to build proxy measures of these constructs that do not require human intervention? RESULTS: Here, we built proxy measures by applying machine learning on multimodal MR images and rich sociodemographic information from the largest biomedical cohort to date: the UK Biobank. Objective model comparisons revealed that all proxies captured the target constructs and were as useful, and sometimes more useful, than the original measures for characterizing real-world health behavior (sleep, exercise, tobacco, alcohol consumption). We observed this complementarity of proxy measures and original measures at capturing multiple health-related constructs when modeling from, both, brain signals and sociodemographic data. CONCLUSION: Population modeling with machine learning can derive measures of mental health from heterogeneous inputs including brain signals and questionnaire data. This may complement or even substitute for psychometric assessments in clinical populations.
Authors: Franziskus Liem; Gaël Varoquaux; Jana Kynast; Frauke Beyer; Shahrzad Kharabian Masouleh; Julia M Huntenburg; Leonie Lampe; Mehdi Rahim; Alexandre Abraham; R Cameron Craddock; Steffi Riedel-Heller; Tobias Luck; Markus Loeffler; Matthias L Schroeter; Anja Veronica Witte; Arno Villringer; Daniel S Margulies Journal: Neuroimage Date: 2016-11-23 Impact factor: 6.556
Authors: James H Cole; Rudra P K Poudel; Dimosthenis Tsagkrasoulis; Matthan W A Caan; Claire Steves; Tim D Spector; Giovanni Montana Journal: Neuroimage Date: 2017-07-29 Impact factor: 6.556
Authors: Marius de Groot; Meike W Vernooij; Stefan Klein; M Arfan Ikram; Frans M Vos; Stephen M Smith; Wiro J Niessen; Jesper L R Andersson Journal: Neuroimage Date: 2013-03-22 Impact factor: 6.556
Authors: Lena Chan; Cortney Simmons; Scott Tillem; May Conley; Inti A Brazil; Arielle Baskin-Sommers Journal: Biol Psychiatry Cogn Neurosci Neuroimaging Date: 2022-02-22