Anatomically interpretable deep learning of brain age captures domain-specific cognitive impairment
Chenzhong Yin, Phoebe Imms, Mingxi Cheng, Anar Amgalan, Nahian F. Chowdhury, Roy J Massett, Nikhil N. Chaudhari, Xinghe Chen, Paul M. Thompson, Paul Bogdan, Andrei Irimia, the Alzheimer’s Disease Neuroimaging Initiative, Michael W. Weiner, Paul Aisen, Ronald Petersen, Michael W. Weiner, Paul Aisen, Ronald Petersen, Clifford R. Jack, William Jagust, John Q. Trojanowki, Arthur W. Toga, Laurel Beckett, Robert C. Green, Andrew J. Saykin, John C. Morris, Richard J. Perrin, Leslie M. Shaw, Zaven S. Khachaturian, María C. Carrillo, William Z. Potter, Lisa L. Barnes, Marie Bernard, Héctor Alfredo Baptista González, Carole Ho, John K. Hsiao, Jonathan Jackson, Eliezer Masliah, Donna Masterman, Ozioma C. Okonkwo, Richard J. Perrin, Laurie Ryan, Nina Silverberg, Adam Fleisher, Eli Lilly, Michael W. Weiner, Diana Truran Sacrey, Juliet Fockler, Cat Conti, Dallas P. Veitch, John Neuhaus, Chengshi Jin, Rachel L. Nosheny, Miriam T. Ashford, Derek Flenniken, Adrienne Kormos, Robert C. Green, Tom Montine, Cat Conti, Ronald Petersen, Paul Aisen, Michael S. Rafii, Rema Raman, Gustavo Jiménez, Michael Donohue, Devon Gessert, Jennifer Salazar, Caileigh Zimmerman, Yuliana Cabrera, Sarah Walter, Garrett Miller, Godfrey Coker, Taylor Clanton, Lindsey Hergesheimer, Stephanie Smith, Olusegun Adegoke, Payam Mahboubi, Shelley Moore, Jeremy Pizzola, Elizabeth Shaffer, Brittany Sloan, Laurel Beckett, Danielle Harvey, Michael Donohue, Clifford R. Jack, Arvin Forghanian-Arani, Bret Borowski, Chad Ward, Christopher G. Schwarz, David Jones, Jeff Gunter, Kejal Kantarci, Matthew L. Senjem, Prashanthi Vemuri, Robert I. Reid, Nick C. Fox, Ian B. Malone, Paul M. Thompson, Sophia I. Thomopoulos, Talia M. Nir
Abstract
The gap between chronological age (CA) and biological brain age, as estimated from magnetic resonance images (MRIs), reflects how individual patterns of neuroanatomic aging deviate from their typical trajectories. MRI-derived brain age (BA) estimates are often obtained using deep learning models that may perform relatively poorly on new data or that lack neuroanatomic interpretability. This study introduces a convolutional neural network (CNN) to estimate BA after training on the MRIs of 4,681 cognitively normal (CN) participants and testing on 1,170 CN participants from an independent sample. BA estimation errors are notably lower than those of previous studies. At both individual and cohort levels, the CNN provides detailed anatomic maps of brain aging patterns that reveal sex dimorphisms and neurocognitive trajectories in adults with mild cognitive impairment (MCI, N = 351) and Alzheimer’s disease (AD, N = 359). In individuals with MCI (54% of whom were diagnosed with dementia within 10.9 y from MRI acquisition), BA is significantly better than CA in capturing dementia symptom severity, functional disability, and executive function. Profiles of sex dimorphism and lateralization in brain aging also map onto patterns of neuroanatomic change that reflect cognitive decline. Significant associations between BA and neurocognitive measures suggest that the proposed framework can map, systematically, the relationship between aging-related neuroanatomy changes in CN individuals and in participants with MCI or AD. Early identification of such neuroanatomy changes can help to screen individuals according to their AD risk.