Population heterogeneity in clinical cohorts affects the predictive accuracy of brain imaging.

Autor: Benkarim O; McConnell Brain Imaging Centre (BIC), Montreal Neurological Institute (MNI), Faculty of Medicine, McGill University, Montreal, Canada., Paquola C; McConnell Brain Imaging Centre (BIC), Montreal Neurological Institute (MNI), Faculty of Medicine, McGill University, Montreal, Canada.; Institute of Neuroscience and Medicine (INM-1), Forschungszentrum Jülich, Jülich, Germany., Park BY; McConnell Brain Imaging Centre (BIC), Montreal Neurological Institute (MNI), Faculty of Medicine, McGill University, Montreal, Canada.; Department of Data Science, Inha University, Incheon, South Korea.; Center for Neuroscience Imaging Research, Institute for Basic Science, Sungkyunkwan University, Suwon, South Korea., Kebets V; McConnell Brain Imaging Centre (BIC), Montreal Neurological Institute (MNI), Faculty of Medicine, McGill University, Montreal, Canada., Hong SJ; Center for Neuroscience Imaging Research, Institute for Basic Science, Sungkyunkwan University, Suwon, South Korea.; Center for the Developing Brain, Child Mind Institute, New York, New York, United States of America.; Department of Biomedical Engineering, Sungkyunkwan University, Suwon, South Korea., Vos de Wael R; McConnell Brain Imaging Centre (BIC), Montreal Neurological Institute (MNI), Faculty of Medicine, McGill University, Montreal, Canada., Zhang S; Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Singapore.; Centre for Sleep and Cognition (CSC) & Centre for Translational Magnetic Resonance Research (TMR), National University of Singapore, Singapore, Singapore.; N.1 Institute for Health & Institute for Digital Medicine (WisDM), National University of Singapore, Singapore, Singapore., Yeo BTT; Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Singapore.; Centre for Sleep and Cognition (CSC) & Centre for Translational Magnetic Resonance Research (TMR), National University of Singapore, Singapore, Singapore.; N.1 Institute for Health & Institute for Digital Medicine (WisDM), National University of Singapore, Singapore, Singapore., Eickenberg M; Flatiron Institute, New York, New York, United States of America., Ge T; Psychiatric and Neurodevelopmental Genetics Unit, Center for Genomic Medicine, Massachusetts General Hospital, Boston, Massachusetts, United States of America., Poline JB; McConnell Brain Imaging Centre (BIC), Montreal Neurological Institute (MNI), Faculty of Medicine, McGill University, Montreal, Canada., Bernhardt BC; McConnell Brain Imaging Centre (BIC), Montreal Neurological Institute (MNI), Faculty of Medicine, McGill University, Montreal, Canada., Bzdok D; McConnell Brain Imaging Centre (BIC), Montreal Neurological Institute (MNI), Faculty of Medicine, McGill University, Montreal, Canada.; Department of Biomedical Engineering, Faculty of Medicine, McGill University, Montreal, Canada.; School of Computer Science, McGill University, Montreal, Canada.; Mila-Quebec Artificial Intelligence Institute, Montreal, Canada.
Jazyk: angličtina
Zdroj: PLoS biology [PLoS Biol] 2022 Apr 29; Vol. 20 (4), pp. e3001627. Date of Electronic Publication: 2022 Apr 29 (Print Publication: 2022).
DOI: 10.1371/journal.pbio.3001627
Abstrakt: Brain imaging research enjoys increasing adoption of supervised machine learning for single-participant disease classification. Yet, the success of these algorithms likely depends on population diversity, including demographic differences and other factors that may be outside of primary scientific interest. Here, we capitalize on propensity scores as a composite confound index to quantify diversity due to major sources of population variation. We delineate the impact of population heterogeneity on the predictive accuracy and pattern stability in 2 separate clinical cohorts: the Autism Brain Imaging Data Exchange (ABIDE, n = 297) and the Healthy Brain Network (HBN, n = 551). Across various analysis scenarios, our results uncover the extent to which cross-validated prediction performances are interlocked with diversity. The instability of extracted brain patterns attributable to diversity is located preferentially in regions part of the default mode network. Collectively, our findings highlight the limitations of prevailing deconfounding practices in mitigating the full consequences of population diversity.
Competing Interests: The authors have declared that no competing interests exist.
Databáze: MEDLINE
Externí odkaz:
Nepřihlášeným uživatelům se plný text nezobrazuje