{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,7]],"date-time":"2026-08-07T15:13:43Z","timestamp":1786115623137,"version":"build-2736575974"},"reference-count":17,"publisher":"MDPI AG","issue":"8","license":[{"start":{"date-parts":[[2026,8,3]],"date-time":"2026-08-03T00:00:00Z","timestamp":1785715200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["J. Imaging"],"abstract":"<jats:p>Accurate estimation of Alzheimer\u2019s disease (AD) severity from structural magnetic resonance imaging (MRI) remains difficult, as disease-associated anatomical alterations are often subtle and publicly available datasets are typically too small to support robust deep learning model training. This feasibility study sought to determine how much Alzheimer\u2019s disease spectrum-related information could be extracted from a small structural MRI cohort using a deliberately lightweight two-dimensional convolutional neural network (2D CNN), and whether transfer learning improves model performance. This study was intended as a methodological proof of concept rather than the development of a clinically deployable diagnostic tool. Structural scans and Clinical Dementia Rating (CDR) labels from the OASIS-1 dataset were filtered to 214 subjects: 124 cognitively normal (CN), 65 with mild cognitive impairment (MCI; CDR = 0.5), and 25 with AD-level impairment (CDR \u2265 1). A compact 2D CNN trained from scratch and a transfer learning model (frozen ImageNet MobileNetV2 features) were evaluated on four binary tasks (CN vs. AD, MCI vs. AD, CN vs. MCI, and CN vs. any impairment) under identical pre-processing and subject-level repeated 5-fold cross-validation (10 repeats), with the decision threshold tuned only on an inner split. Discrimination was summarized by ROC-AUC with 95% confidence intervals (CIs), permutation tests against chance, and per-task sensitivity and specificity. The from-scratch CNN recovered only a broad normal-versus-impaired signal (CN vs. any impairment AUC 0.59) and was at chance on adjacent-stage tasks (MCI vs. AD 0.41; CN vs. MCI 0.51). Transfer learning improved every task: CN vs. AD AUC 0.745 (95% CI 0.730\u20130.763), CN vs. any impairment 0.642, CN vs. MCI 0.601, and MCI vs. AD 0.599. On an independent OASIS-2 cohort, the transfer learning CN vs. AD model retained AUC 0.748. In this small-data regime, transfer learning recovers substantially more Alzheimer-spectrum signals than a from-scratch CNN, but performance remains modest because it is bounded by CDR-based, non-biomarker-confirmed labels, suggesting the model separates CDR-defined cognitive-status groups rather than detecting AD pathology.<\/jats:p>","DOI":"10.3390\/jimaging12080352","type":"journal-article","created":{"date-parts":[[2026,8,3]],"date-time":"2026-08-03T12:57:56Z","timestamp":1785761876000},"page":"352","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Small-Data Deep Learning for Alzheimer-Spectrum Classification from Structural MRI: A Feasibility Study Using OASIS"],"prefix":"10.3390","volume":"12","author":[{"given":"Ian D.","family":"Li","sequence":"first","affiliation":[{"name":"Department of Molecular Pharmacology and Experimental Therapeutics, Mayo Clinic College of Medicine and Science, Rochester, MN 55905, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9876-3473","authenticated-orcid":false,"given":"Choong-Yong","family":"Ung","sequence":"additional","affiliation":[{"name":"Department of Molecular Pharmacology and Experimental Therapeutics, Mayo Clinic College of Medicine and Science, Rochester, MN 55905, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Cristina","family":"Correia","sequence":"additional","affiliation":[{"name":"Department of Molecular Pharmacology and Experimental Therapeutics, Mayo Clinic College of Medicine and Science, Rochester, MN 55905, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,8,3]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"330","DOI":"10.1016\/j.nicl.2016.12.012","article-title":"Longitudinal changes in microstructural white matter metrics in Alzheimer\u2019s disease","volume":"13","author":"Mayo","year":"2017","journal-title":"Neuroimage Clin."},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"143","DOI":"10.1016\/j.jalz.2013.08.279","article-title":"Time course of brain volume changes in the preclinical phase of Alzheimer\u2019s disease","volume":"10","author":"Bernard","year":"2014","journal-title":"Alzheimer\u2019s Dement."},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"951","DOI":"10.1002\/ana.26891","article-title":"Timing of Biomarker Changes in Sporadic Alzheimer\u2019s Disease in Estimated Years from Symptom Onset","volume":"95","author":"Li","year":"2024","journal-title":"Ann. Neurol."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"525","DOI":"10.1016\/S1053-8119(02)00026-5","article-title":"Early diagnosis of Alzheimer\u2019s disease: Contribution of structural neuroimaging","volume":"18","author":"Chetelat","year":"2003","journal-title":"Neuroimage"},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"1675","DOI":"10.1586\/ern.10.162","article-title":"Quantitative structural MRI for early detection of Alzheimer\u2019s disease","volume":"10","author":"McEvoy","year":"2010","journal-title":"Expert Rev. Neurother."},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"17106","DOI":"10.1038\/s41598-022-20674-x","article-title":"Generalizable deep learning model for early Alzheimer\u2019s disease detection from structural MRIs","volume":"12","author":"Liu","year":"2022","journal-title":"Sci. Rep."},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"102949","DOI":"10.1016\/j.artmed.2024.102949","article-title":"A systematic review of few-shot learning in medical imaging","volume":"156","author":"Pachetti","year":"2024","journal-title":"Artif. Intell. Med."},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"105773","DOI":"10.1016\/j.ebiom.2025.105773","article-title":"Automatic identification of Parkinsonism using clinical multi-contrast brain MRI: A large self-supervised vision foundation model strategy","volume":"116","author":"Suo","year":"2025","journal-title":"eBioMedicine"},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"e70175","DOI":"10.1002\/alz.70175","article-title":"Early detection of Alzheimer\u2019s disease using deep learning methods","volume":"21","author":"Mmadumbu","year":"2025","journal-title":"Alzheimer\u2019s Dement."},{"key":"ref_10","doi-asserted-by":"crossref","unstructured":"Iqbal, Z., Rahman, M.M., Mahmood, U., Zia, Q., Fu, Z., Calhoun, V.D., and Plis, S. (2025). Explainable Self-Supervised Dynamic Neuroimaging Using Time Reversal. Brain Sci., 15.","DOI":"10.3390\/brainsci15010060"},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"44098","DOI":"10.1038\/s41598-025-27914-w","article-title":"Comparative analysis of multiple deep learning models with mitigation-driven approaches for enhanced Alzheimer\u2019s disease classification","volume":"15","author":"Bayahya","year":"2025","journal-title":"Sci. Rep."},{"key":"ref_12","doi-asserted-by":"crossref","first-page":"757","DOI":"10.2174\/0115672050393604250904081342","article-title":"Advancing Alzheimer\u2019s Disease Diagnosis Using VGG19 and XGBoost: A Neuroimaging-Based Method","volume":"22","author":"Boudi","year":"2025","journal-title":"Curr. Alzheimer Res."},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"1498","DOI":"10.1162\/jocn.2007.19.9.1498","article-title":"Open Access Series of Imaging Studies (OASIS): Cross-sectional MRI data in young, middle aged, nondemented, and demented older adults","volume":"19","author":"Marcus","year":"2007","journal-title":"J. Cogn. Neurosci."},{"key":"ref_14","first-page":"6605","article-title":"Improving the Prediction of Therapist Behaviors in Addiction Counseling by Exploiting Class Confusions","volume":"Volume 2019","author":"Chen","year":"2019","journal-title":"Proceedings of the 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)"},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"2677","DOI":"10.1162\/jocn.2009.21407","article-title":"Open access series of imaging studies: Longitudinal MRI data in nondemented and demented older adults","volume":"22","author":"Marcus","year":"2010","journal-title":"J. Cogn. Neurosci."},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"336","DOI":"10.1007\/s11263-019-01228-7","article-title":"Grad-CAM: Visual explanations from deep networks via gradient-based localization","volume":"128","author":"Selvaraju","year":"2020","journal-title":"Int. J. Comput. Vis."},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"Devnath, H., Devnath, L., and Hossain, M. (2026). Explainable artificial intelligence for automated flatfoot detection in foot x-ray images. Discov. Artif. Intell.","DOI":"10.1007\/s44163-026-01342-9"}],"container-title":["Journal of Imaging"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2313-433X\/12\/8\/352\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,6]],"date-time":"2026-08-06T04:22:48Z","timestamp":1785990168000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2313-433X\/12\/8\/352"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,3]]},"references-count":17,"journal-issue":{"issue":"8","published-online":{"date-parts":[[2026,8]]}},"alternative-id":["jimaging12080352"],"URL":"https:\/\/doi.org\/10.3390\/jimaging12080352","relation":{},"ISSN":["2313-433X"],"issn-type":[{"value":"2313-433X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,3]]}}}