TY - JOUR
T1 - Machine learning-based estimation of cognitive performance using regional brain MRI markers
T2 - the Northern Manhattan Study
AU - Caunca, Michelle R.
AU - Wang, Lily
AU - Cheung, Ying Kuen
AU - Alperin, Noam
AU - Lee, Sang H.
AU - Elkind, Mitchell S.V.
AU - Sacco, Ralph L.
AU - Wright, Clinton B.
AU - Rundek, Tatjana
N1 - Funding Information:
This work was supported by the National Institutes of Neurological Disease and Stroke (R01 NS29993, F30 NS103462) and the Evelyn F. McKnight Brain Institute. Acknowledgements
Funding Information:
Dr. Ralph L. Sacco receives federal grant support (R01 NS 29993, CTSA UL1 TR002736), private foundation support (American Heart Association Bugher Center), and pharma research support (Boehringer Ingelheim).
Publisher Copyright:
© 2020, Springer Science+Business Media, LLC, part of Springer Nature.
Copyright:
Copyright 2020 Elsevier B.V., All rights reserved.
PY - 2020
Y1 - 2020
N2 - High dimensional neuroimaging datasets and machine learning have been used to estimate and predict domain-specific cognition, but comparisons with simpler models composed of easy-to-measure variables are limited. Regularization methods in particular may help identify regions-of-interest related to domain-specific cognition. Using data from the Northern Manhattan Study, a cohort study of mostly Hispanic older adults, we compared three models estimating domain-specific cognitive performance: sociodemographics and APOE ε4 allele status (basic model), the basic model and MRI markers, and a model with only MRI markers. We used several machine learning methods to fit our regression models: elastic net, support vector regression, random forest, and principal components regression. Model performance was assessed with the RMSE, MAE, and R2 statistics using 5-fold cross-validation. To assess whether prediction models with imaging biomarkers were more predictive than prediction models built with randomly generated biomarkers, we refit the elastic net models using 1000 datasets with random biomarkers and compared the distribution of the RMSE and R2 in models using these random biomarkers to the RMSE and R2 from observed models. Basic models explained ~ 31–38% of the variance in domain-specific cognition. Addition of MRI markers did not improve estimation. However, elastic net models with only MRI markers performed significantly better than random MRI markers (one-sided P <.05) and yielded regions-of-interest consistent with previous literature and others not previously explored. Therefore, structural brain MRI markers may be more useful for etiological than predictive modeling.
AB - High dimensional neuroimaging datasets and machine learning have been used to estimate and predict domain-specific cognition, but comparisons with simpler models composed of easy-to-measure variables are limited. Regularization methods in particular may help identify regions-of-interest related to domain-specific cognition. Using data from the Northern Manhattan Study, a cohort study of mostly Hispanic older adults, we compared three models estimating domain-specific cognitive performance: sociodemographics and APOE ε4 allele status (basic model), the basic model and MRI markers, and a model with only MRI markers. We used several machine learning methods to fit our regression models: elastic net, support vector regression, random forest, and principal components regression. Model performance was assessed with the RMSE, MAE, and R2 statistics using 5-fold cross-validation. To assess whether prediction models with imaging biomarkers were more predictive than prediction models built with randomly generated biomarkers, we refit the elastic net models using 1000 datasets with random biomarkers and compared the distribution of the RMSE and R2 in models using these random biomarkers to the RMSE and R2 from observed models. Basic models explained ~ 31–38% of the variance in domain-specific cognition. Addition of MRI markers did not improve estimation. However, elastic net models with only MRI markers performed significantly better than random MRI markers (one-sided P <.05) and yielded regions-of-interest consistent with previous literature and others not previously explored. Therefore, structural brain MRI markers may be more useful for etiological than predictive modeling.
KW - Biomarkers
KW - Brain aging
KW - Cognitive aging
KW - Machine learning
UR - http://www.scopus.com/inward/record.url?scp=85088836992&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85088836992&partnerID=8YFLogxK
U2 - 10.1007/s11682-020-00325-3
DO - 10.1007/s11682-020-00325-3
M3 - Article
AN - SCOPUS:85088836992
JO - Brain Imaging and Behavior
JF - Brain Imaging and Behavior
SN - 1931-7557
ER -