Interpretable deep learning reveals spatiotemporal MRI features of brain aging that align with neurodegeneration
w) MRI. By identifying MRI features unapparent to humans, DNNs can find aging-related structural alterations above and beyond cortical thickness and atrophy. Using a novel approach to DNN interpretability, we mapped brain aging progression in 25,539 cognitively normal UK Biobank adults aged 45-83 years. Cortical aging is found to involve anatomic features becoming prominent during the 50s within frontolateral, mesolimbic, cuneal, and occipitotemporal regions. From these foci, aging-related features propagate to adjacent areas at rates peaking in the 60s. Cortical thinning and atrophy do not trend closely with neurodegeneration, but important DNN-identifiable anatomic features have spatiotemporal dynamics that match those of amyloid or tau. Our results challenge the assumption that MRI cannot map anatomic features trending with neurodegeneration. Interpretable DNNs can empower MRI to quantify anatomic aging as a process of spatial feature expansion from focal regions into nearby structures in the sequence of neurodegenerative pathology. Traditional morphometrics explain only ∼1% of variance in DNN-identifiable features, which clarifies why the former are insensitive to anatomic changes involving neurodegeneration. Our results conceptualize brain aging in the context of spatial and temporal parallels between anatomic senescence and neuropathology. These findings may help to map cognitively normal adults' neurodegenerative anatomy even without PET measurements.
Authors
- Winnie Kwang
- Payam Mahboubi
- Juliet Fockler
- Elizabeth Shaffer (ORCID: https://orcid.org/0000-0001-6017-1539)
- Danielle Harvey (ORCID: https://orcid.org/0000-0002-5367-0951)
- Cecily Jenkins
- Yuliana Cabrera
- O. Vega
- Caitie Hedberg
- Jeremy Pizzola
- Diana Truran Sacrey
- Phoebe Imms (ORCID: https://orcid.org/0000-0002-7205-4177)
- Taylor Clanton
- Bret Borowski
- Derek Flenniken
- Kaori Kubo Germano
- Andrei Irimia (ORCID: https://orcid.org/0000-0002-9254-9388)
- Rema Raman (ORCID: https://orcid.org/0000-0002-4096-6390)
- Cat Conti
- Denise Reyes (ORCID: https://orcid.org/0000-0003-0455-8821)
- Lindsey Hergesheimer
- Sandra Talavera (ORCID: https://orcid.org/0000-0001-9272-0547)
- Calvin Reyes
- John Moore-Weiss
- Shaniya Parkins
- Mai Seng Thao
- Kedir Adem Hussen
- Joel Felmlee Jr.
- Jeff Gunter
- Nahian F. Chowdhury
- Leonard Matoush
- Matthew Senjem
- Vanessa Guzman
- Catherine Cat Conti
- Ian Malone
- Gustavo Jimenez
- Hannatu Amaza
- Charles DeCarli
- Caitlin Webb
- Ronald Petersen
- Lisa Thomas
- Duygu Tosun
- Nick C. Fox
- Susan Landau
- Edward B. Lee
- Chengshi Jin
- Sophia I. Thomopoulos
- Michael Rafii
- Tamara Jafar
- Olusegun Adegoke
- Fabiola Magana
- Adeyinka Ajayi
- Erin Drake
- Pallavi Sachdev
- Arvin Forghanian-Arani
- Miriam Ashford
- Sheila Ogwang
- Tom Montine
- Clifford R. Jack
- Rachel Nosheny
- Omobolanle Ayo
- Kejal Kantarci
- Naomi Saito
- Richard J. Perrin
- Adrienne Kormos
- William Jagust
- Isabella Hoang
- Ozioma Okonkwo
- Laurel Beckett
- Chad Ward
- Jennifer Salazar
- Stephanie Smith
- Robert Reid
- Paul Aisen
- Trinity Weisensel
- Michael W. Weiner
- Arthur W. Toga
- Melanie J. Miller
- Joe Strong
- Virginia Boatwright
- Prashanthi Vemuri
- Monica Rivera-Mindt
- Andrea Fidell
- Jaron M. Kawamura
- Christopher Schwarz
- Joseph Di Benedetto
- Robert C. Green
- Sarah Walter
- Leslie M. Shaw
- Nikhil N. Chaudhari
- Caileigh Zimmerman
- Adam Diaz
- Matt Glittenberg
- Kwangsik Nho
- Andrew J. Saykin
- Justin Robison
- Paul Andrew Thompson
- Talia M. Nir
- Michael Donohue
Institutions
- University of Southern California (US)
- King's College London (GB)
Publication Details
- Journal
- GeroScience
- Published
- 2026-04-15
- DOI
- https://doi.org/10.1007/s11357-026-02112-2
- Primary Topic
- Functional Brain Connectivity Studies
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Foundation for the National Institutes of Health
- University of Southern California