{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T16:53:22Z","timestamp":1790873602174,"version":"4.1.0"},"reference-count":59,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,2,4]],"date-time":"2026-02-04T00:00:00Z","timestamp":1770163200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Neuroimaging"],"abstract":"<jats:p>Predicting the trajectory of clinical decline in aging individuals is a pressing challenge, especially for people with mild cognitive impairment, Alzheimer\u2019s disease, Parkinson\u2019s disease, or vascular dementia. Accurate predictions can guide treatment decisions, identify risk factors, and optimize clinical trials. In this study, we compared two deep learning approaches for forecasting changes, over a 2-year interval, in the Clinical Dementia Rating scale \u2018sum of boxes\u2019 score (sobCDR), as a continuous outcome (regression). This is a key metric in dementia research and clinical trials, and scores range from 0 (no impairment) to 18 (severe impairment). To predict decline, we trained a hybrid convolutional neural network (CNN) that integrates 3D T1-weighted brain MRI scans with tabular clinical and demographic features (including age, sex, body mass index (BMI), and baseline sobCDR). We benchmarked its performance against AutoGluon, an automated multimodal machine learning framework that selects an appropriate neural network architecture (an \u2018autoML\u2019 approach). We evaluated the models using data from 2,319 unique participants drawn from three independent cohorts\u2014ADNI, OASIS-3, and NACC. For each participant, we used one T1-weighted brain MRI scan along with corresponding clinical and demographic information. Our results demonstrate the importance of combining image and tabular data in predictive modeling for this clinical application. Deep learning algorithms can fuse information from image-based brain signatures and tabular clinical data, with potential for personalized prognostics in aging and dementia. Rather than concluding that multimodal fusion uniformly improves performance, our results show that deep learning applied to volumetric MRI data may struggle to add predictive value, particularly when clinical covariates explain substantial variance and provide a strong baseline. In other conditions and tasks, it may help to have a hybrid system that can learn from both data types, and their relative value may be different. Conversely, AutoML-based multimodal fusion provides a robust baseline when tabular data already provide strong predictive value for the task. These insights clarify how different multimodal strategies could be selected in clinical prognostic applications.<\/jats:p>","DOI":"10.3389\/fnimg.2026.1726037","type":"journal-article","created":{"date-parts":[[2026,2,4]],"date-time":"2026-02-04T05:33:47Z","timestamp":1770183227000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":6,"title":["Deep learning to predict future cognitive decline: a multimodal approach using brain MRI and clinical data"],"prefix":"10.3389","volume":"5","author":[{"given":"Tamoghna","family":"Chattopadhyay","sequence":"first","affiliation":[{"name":"Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California","place":["Marina del Rey, CA, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Pavithra","family":"Senthilkumar","sequence":"additional","affiliation":[{"name":"Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California","place":["Marina del Rey, CA, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Rahul H.","family":"Ankarath","sequence":"additional","affiliation":[{"name":"Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California","place":["Marina del Rey, CA, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Christopher","family":"Patterson","sequence":"additional","affiliation":[{"name":"Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California","place":["Marina del Rey, CA, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Emma J.","family":"Gleave","sequence":"additional","affiliation":[{"name":"Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California","place":["Marina del Rey, CA, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sophia I.","family":"Thomopoulos","sequence":"additional","affiliation":[{"name":"Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California","place":["Marina del Rey, CA, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Heng","family":"Huang","sequence":"additional","affiliation":[{"name":"Department of Computer Science, University of Maryland","place":["College Park, MD, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Li","family":"Shen","sequence":"additional","affiliation":[{"name":"Department of Biostatistics, Epidemiology and Informatics, Perelman School of Medicine, University of Pennsylvania","place":["Philadelphia, PA, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lei","family":"You","sequence":"additional","affiliation":[{"name":"McWilliams School of Biomedical Informatics, University of Texas Health Science Center at Houston","place":["Houston, TX, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Degui","family":"Zhi","sequence":"additional","affiliation":[{"name":"McWilliams School of Biomedical Informatics, University of Texas Health Science Center at Houston","place":["Houston, TX, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Paul M.","family":"Thompson","sequence":"additional","affiliation":[{"name":"Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California","place":["Marina del Rey, CA, United States"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,2,4]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"354","DOI":"10.1016\/j.trci.2019.06.005","article-title":"Disease severity and minimal clinically important differences in clinical outcome assessments for Alzheimer's disease clinical trials","volume":"5","author":"Andrews","year":"2019","journal-title":"Alzheimers Dement (N Y)"},{"key":"ref2","first-page":"2024","article-title":"Towards multimodal longitudinal analysis for predicting cognitive decline","author":"Ashish","year":"2024","journal-title":"bioRxiv"},{"key":"ref3","doi-asserted-by":"publisher","first-page":"26","DOI":"10.1016\/j.media.2007.06.004","article-title":"Symmetric diffeomorphic image registration with cross-correlation: evaluating automated labeling of elderly and neurodegenerative brain","volume":"12","author":"Avants","year":"2008","journal-title":"Med. Image Anal."},{"key":"ref4","doi-asserted-by":"publisher","first-page":"459","DOI":"10.3233\/JAD-230893","article-title":"Predicting four-year's Alzheimer's disease onset using longitudinal neurocognitive tests and MRI data using explainable deep convolutional neural networks","volume":"97","author":"Bapat","year":"2024","journal-title":"J Alzheimer's Dis"},{"key":"ref5","doi-asserted-by":"publisher","first-page":"1434","DOI":"10.1080\/01621459.2023.2197686","article-title":"Cross-validation: what does it estimate and how well does it do it?","volume":"119","author":"Bates","year":"2024","journal-title":"J. Am. Stat. Assoc."},{"key":"ref6","doi-asserted-by":"publisher","first-page":"249","DOI":"10.1097\/WAD.0b013e318142774e","article-title":"The National Alzheimer's coordinating center (NACC) database: the uniform data set","volume":"21","author":"Beekly","year":"2007","journal-title":"Alzheimer Dis. Assoc. Disord."},{"key":"ref9003","first-page":"e1142","article-title":"Clinically relevant changes for cognitive outcomes in preclinical and prodromal cognitive stages: implications for clinical Alzheimer trials","volume":"99","author":"Borland","year":"2022","journal-title":"NeuroImage"},{"key":"ref7","doi-asserted-by":"publisher","first-page":"562","DOI":"10.1016\/j.neuroimage.2015.01.048","article-title":"Standardized evaluation of algorithms for computer-aided diagnosis of dementia based on structural MRI: the CADDementia challenge","volume":"111","author":"Bron","year":"2015","journal-title":"NeuroImage"},{"key":"ref8","doi-asserted-by":"crossref","first-page":"S45","DOI":"10.1016\/j.jalz.2011.11.002","article-title":"Rationale for use of the clinical dementia rating sum of boxes as a primary outcome measure for Alzheimer\u2019s disease clinical trials","volume":"9","author":"Cedarbaum","year":"2013","journal-title":"Alzheimers Dement."},{"key":"ref9","first-page":"90","author":"Chattopadhyay","year":"2023"},{"key":"ref11","first-page":"1","author":"Chattopadhyay","year":""},{"key":"ref12","first-page":"1","author":"Chattopadhyay","year":""},{"key":"ref10","doi-asserted-by":"publisher","first-page":"1387196","DOI":"10.3389\/fnins.2024.1387196","article-title":"Comparison of deep learning architectures for predicting amyloid positivity in Alzheimer\u2019s disease, mild cognitive impairment, and healthy aging, from T1-weighted brain structural MRI","volume":"18","author":"Chattopadhyay","year":"","journal-title":"Front. Neurosci."},{"key":"ref9004","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/SIPAIM67325.2025.11283404","article-title":"\u201cPredicting Longitudinal Cognitive Decline via Multimodal Deep Learning Fusing Brain MRI and Clinical Data,\u201d 2025 21st International Symposium on Biomedical Image Processing and Analysis (SIPAIM), Pasto, Colombia","author":"Chattopadhyay","year":"2025"},{"key":"ref13","doi-asserted-by":"publisher","first-page":"898940","DOI":"10.3389\/fnagi.2022.898940","article-title":"Prediction of conversion to dementia using interpretable machine learning in patients with amnestic mild cognitive impairment","volume":"14","author":"Chun","year":"2022","journal-title":"Front. Aging Neurosci."},{"key":"ref14","doi-asserted-by":"publisher","first-page":"198","DOI":"10.1186\/s13195-024-01567-9","article-title":"Reliability of the assessment of the clinical dementia rating scale from the analysis of medical records in comparison with the reference method","volume":"16","author":"Dauphinot","year":"2024","journal-title":"Alzheimer's Res Ther"},{"key":"ref15","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/psp.2013.54","article-title":"Modeling Alzheimer's disease progression using disease onset time and disease trajectory concepts applied to CDR-SOB scores from ADNI","volume":"2","author":"Delor","year":"2013","journal-title":"CPT Pharmacometrics Syst. Pharmacol."},{"key":"ref16","first-page":"1","author":"Dhinagar","year":"2023"},{"key":"ref17","first-page":"542","author":"Dhinagar","year":"2024"},{"key":"ref18","first-page":"265","author":"Dhinagar","year":"2025"},{"key":"ref19","first-page":"504","author":"Dhinagar","year":"2023"},{"key":"ref20","doi-asserted-by":"publisher","first-page":"1443","DOI":"10.1001\/archpsyc.64.12.1443","article-title":"Clinical prediction of Alzheimer disease dementia across the spectrum of mild cognitive impairment","volume":"64","author":"Dickerson","year":"2007","journal-title":"Arch. Gen. Psychiatry"},{"key":"ref21","first-page":"2003.06505","article-title":"Autogluon-tabular: robust and accurate autoML for structured data","author":"Erickson","year":"2020","journal-title":"arXiv preprint arXiv"},{"key":"ref22","doi-asserted-by":"publisher","first-page":"162","DOI":"10.1186\/s13195-021-00900-w","article-title":"Machine learning methods for predicting progression from mild cognitive impairment to Alzheimer's disease dementia: a systematic review","volume":"13","author":"Grueso","year":"2021","journal-title":"Alzheimer's Res Ther"},{"key":"ref23","first-page":"1","author":"Gupta","year":"2023"},{"key":"ref24","doi-asserted-by":"crossref","first-page":"67850","DOI":"10.52202\/079017-2167","article-title":"Fusemoe: mixture-of-experts transformers for fleximodal fusion","volume":"37","author":"Han","year":"2024","journal-title":"Adv. Neural Inf. Proces. Syst."},{"key":"ref25","doi-asserted-by":"publisher","first-page":"55","DOI":"10.1016\/j.neurobiolaging.2022.06.008","article-title":"Predicting future cognitive decline from non-brain and multimodal brain imaging data in healthy and pathological aging","volume":"118","author":"Hebling Vieira","year":"2022","journal-title":"Neurobiol. Aging"},{"key":"ref26","doi-asserted-by":"publisher","first-page":"1102869","DOI":"10.3389\/fnagi.2023.1102869","article-title":"Vision transformers for the prediction of mild cognitive impairment to Alzheimer\u2019s disease progression using mid-sagittal sMRI","volume":"15","author":"Hoang","year":"2023","journal-title":"Front. Aging Neurosci."},{"key":"ref27","doi-asserted-by":"publisher","first-page":"4952","DOI":"10.1002\/hbm.24750","article-title":"Automated brain extraction of multisequence MRI using artificial neural networks","volume":"40","author":"Isensee","year":"2019","journal-title":"Hum. Brain Mapp."},{"key":"ref28","doi-asserted-by":"crossref","first-page":"685","DOI":"10.1002\/jmri.21049","article-title":"The Alzheimer's disease neuroimaging initiative (ADNI): MRI methods","volume":"27","author":"Jack","year":"2008","journal-title":"J. Magn. Reson. Imaging"},{"key":"ref29","doi-asserted-by":"publisher","first-page":"455","DOI":"10.1016\/j.neuroimage.2013.04.061","article-title":"Multi-site genetic analysis of diffusion images and voxelwise heritability analysis: a pilot project of the ENIGMA-DTI working group","volume":"81","author":"Jahanshad","year":"2013","journal-title":"NeuroImage"},{"key":"ref30","doi-asserted-by":"publisher","first-page":"1356745","DOI":"10.3389\/fnagi.2024.1356745","article-title":"Deep learning model for individualized trajectory prediction of clinical outcomes in mild cognitive impairment","volume":"16","author":"Jung","year":"2024","journal-title":"Front. Aging Neurosci."},{"key":"ref31","doi-asserted-by":"publisher","first-page":"102497","DOI":"10.1016\/j.arr.2024.102497","article-title":"AI-driven innovations in Alzheimer's disease: integrating early diagnosis, personalized treatment, and prognostic modelling","volume":"101","author":"Kale","year":"2024","journal-title":"Ageing Res. Rev."},{"key":"ref32","doi-asserted-by":"publisher","first-page":"e13139","DOI":"10.2196\/13139","article-title":"Cox proportional hazard regression versus a deep learning algorithm in the prediction of dementia: an analysis based on periodic health examination","volume":"7","author":"Kim","year":"2019","journal-title":"JMIR Med. Inform."},{"key":"ref33","first-page":"1","author":"Komandur","year":"2023"},{"key":"ref34","first-page":"arXiv:2011.09115","volume-title":"3-D grid-attention networks for interpretable age and Alzheimer\u2019s disease prediction from structural MRI","author":"Lam","year":""},{"key":"ref35","first-page":"2019-12","article-title":"OASIS-3: longitudinal neuroimaging, clinical, and cognitive dataset for normal aging and Alzheimer disease","author":"LaMontagne","year":"2019","journal-title":"medrxiv"},{"key":"ref36","doi-asserted-by":"publisher","first-page":"40530","DOI":"10.1038\/s41598-025-24436-3","article-title":"Multimodal image fusion network with prior-guided dynamic degradation removal for extreme environment perception","volume":"15","author":"Li","year":"2025","journal-title":"Sci. Rep."},{"key":"ref37","doi-asserted-by":"publisher","first-page":"4875","DOI":"10.1002\/hbm.26422","article-title":"Style transfer generative adversarial networks to harmonize multisite MRI to a single reference image to avoid overcorrection","volume":"44","author":"Liu","year":"2023","journal-title":"Hum. Brain Mapp."},{"key":"ref38","first-page":"1","author":"Marinescu","year":"2019"},{"key":"ref39","doi-asserted-by":"publisher","first-page":"360","DOI":"10.1001\/jamaneurol.2022.5272","article-title":"Prediction of longitudinal cognitive decline in preclinical Alzheimer disease using plasma biomarkers","volume":"80","author":"Mattsson-Carlgren","year":"2023","journal-title":"JAMA Neurol."},{"key":"ref40","doi-asserted-by":"publisher","first-page":"151","DOI":"10.14283\/jpad.2020.73","article-title":"Psychometric properties of the clinical dementia rating\u2014sum of boxes and other cognitive and functional outcomes in a prodromal Alzheimer\u2019s disease population","volume":"8","author":"McDougall","year":"2021","journal-title":"J. Prev Alzheimers Dis."},{"key":"ref41","doi-asserted-by":"publisher","first-page":"2174","DOI":"10.1002\/mrm.28243","article-title":"Scanner invariant representations for diffusion MRI harmonization","volume":"84","author":"Moyer","year":"2020","journal-title":"Magn. Reson. Med."},{"key":"ref42","doi-asserted-by":"publisher","first-page":"1218","DOI":"10.3390\/electronics12051218","article-title":"Pixel-level fusion approach with vision transformer for early detection of Alzheimer\u2019s disease","volume":"12","author":"Odusami","year":"2023","journal-title":"Electronics"},{"key":"ref44","doi-asserted-by":"crossref","first-page":"e30816","DOI":"10.1002\/alz.12959","article-title":"Expectations and clinical meaningfulness of randomized controlled trials for Alzheimer\u2019s disease","volume":"19","author":"Petersen","year":"2023","journal-title":"Alzheimers Dement."},{"key":"ref45","doi-asserted-by":"publisher","first-page":"e834","DOI":"10.1212\/WNL.0000000000201572","article-title":"Predicting cognitive decline in older adults using baseline metrics of AD pathologies, cerebrovascular disease, and neurodegeneration","volume":"100","author":"Prosser","year":"2023","journal-title":"Neurology"},{"key":"ref46","doi-asserted-by":"crossref","first-page":"558","DOI":"10.1111\/j.1532-5415.2000.tb05004.x","article-title":"Interrater reliability of the clinical dementia rating in a multicenter trial","volume":"48","author":"Rockwood","year":"2000","journal-title":"J. Am. Geriatr. Soc."},{"key":"ref47","doi-asserted-by":"publisher","first-page":"e010357","DOI":"10.1136\/bmjopen-2015-010357","article-title":"Cognitive decline in dementia with Lewy bodies: a 5-year prospective cohort study","volume":"6","author":"Rongve","year":"2016","journal-title":"BMJ Open"},{"key":"ref48","first-page":"2305.16222","article-title":"Incomplete multimodal learning for complex brain disorders prediction","author":"Shirkavand","year":"2023","journal-title":"arXiv preprint arXiv"},{"key":"ref49","first-page":"120880L","author":"Sinha","year":"2021"},{"key":"ref50","doi-asserted-by":"publisher","first-page":"102612","DOI":"10.1016\/j.media.2022.102612","article-title":"Cross-modal attention for multi-modal image registration","volume":"82","author":"Song","year":"2022","journal-title":"Med. Image Anal."},{"key":"ref51","doi-asserted-by":"publisher","first-page":"2203","DOI":"10.2147\/NDT.S496307","article-title":"A review of datasets, optimization strategies, and learning algorithms for analyzing Alzheimer's dementia detection","volume":"20","author":"Thulasimani","year":"2024","journal-title":"Neuropsychiatr. Dis. Treat."},{"key":"ref52","doi-asserted-by":"publisher","first-page":"1310","DOI":"10.1109\/TMI.2010.2046908","article-title":"N4ITK: improved N3 bias correction","volume":"29","author":"Tustison","year":"2010","journal-title":"IEEE Trans. Med. Imaging"},{"key":"ref53","doi-asserted-by":"crossref","first-page":"9","DOI":"10.1056\/NEJMoa2212948","article-title":"Lecanemab in early Alzheimer\u2019s disease","volume":"388","author":"Van Dyck","year":"2023","journal-title":"N. Engl. J. Med."},{"key":"ref54","doi-asserted-by":"publisher","first-page":"284","DOI":"10.1186\/s12874-022-01754-y","article-title":"Dementia risk prediction in individuals with mild cognitive impairment: a comparison of cox regression and machine learning models","volume":"22","author":"Wang","year":"2022","journal-title":"BMC Med. Res. Methodol."},{"key":"ref55","doi-asserted-by":"crossref","first-page":"S39","DOI":"10.1016\/j.jalz.2012.01.005","article-title":"Progression of Alzheimer\u2019s disease as measured by clinical dementia rating sum of boxes scores","volume":"9","author":"Williams","year":"2013","journal-title":"Alzheimers Dement."},{"key":"ref56","article-title":"Integrated care for older people (ICOPE): evidence profile: cognitive impairment","year":"2017"},{"key":"ref57","doi-asserted-by":"publisher","first-page":"110301","DOI":"10.1016\/j.clinimag.2024.110301","article-title":"Automated classification of Alzheimer's disease, mild cognitive impairment, and cognitively normal patients using 3D convolutional neural network and radiomic features from T1-weighted brain MRI: a comparative study on detection accuracy","volume":"115","author":"Zarei","year":"2024","journal-title":"Clin. Imaging"},{"key":"ref58","doi-asserted-by":"publisher","first-page":"e70280","DOI":"10.1002\/hbm.70280","article-title":"Cross-dataset evaluation of dementia longitudinal progression prediction models","volume":"46","author":"Zhang","year":"2025","journal-title":"Hum. Brain Mapp."}],"container-title":["Frontiers in Neuroimaging"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fnimg.2026.1726037\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,2,24]],"date-time":"2026-02-24T09:52:44Z","timestamp":1771926764000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fnimg.2026.1726037\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,2,4]]},"references-count":59,"alternative-id":["10.3389\/fnimg.2026.1726037"],"URL":"https:\/\/doi.org\/10.3389\/fnimg.2026.1726037","relation":{"has-preprint":[{"id-type":"doi","id":"10.1101\/2025.02.21.639520","asserted-by":"object"}]},"ISSN":["2813-1193"],"issn-type":[{"value":"2813-1193","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,2,4]]},"article-number":"1726037"}}