Cross-cohort generalizability of deep and conventional machine learning for MRI-based diagnosis and prediction of Alzheimer's disease

Topic: Amyloid, tau and blood-test diagnosis

This work validates the generalizability of MRI-based classification of Alzheimer's disease (AD) patients and controls (CN) to an external data set and to the task of prediction of conversion to AD in individuals with mild cognitive impairment (MCI). We used a conventional support vector machine (SVM) and a deep convolutional neural network (CNN) approach based on structural MRI scans that underwent either minimal pre-processing or more extensive pre-processing into modulated gray matter (GM) maps. Classifiers were optimized and evaluated using cross-validation in the Alzheimer's Disease Neuroimaging Initiative (ADNI; 334 AD, 520 CN). Trained classifiers were subsequently applied to predict conversion to AD in ADNI MCI patients (231 converters, 628 non-converters) and in the independent Health-RI Parelsnoer Neurodegenerative Diseases Biobank data set. From this multi-center study representing a tertiary memory clinic population, we included 199 AD patients, 139 participants with subjective cognitive decline, 48 MCI patients converting to dementia, and 91 MCI patients who did not convert to dementia. AD-CN classification based on modulated GM maps resulted in a similar area-under-the-curve (AUC) for SVM (0.940; 95%CI: 0.924-0.955) and CNN (0.933; 95%CI: 0.918-0.948). Application to conversion prediction in MCI yielded significantly higher performance for SVM (AUC = 0.756; 95%CI: 0.720-0.788) than for CNN (AUC = 0.742; 95%CI: 0.709-0.776) (p<0.01 for McNemar's test). In external validation, performance was slightly decreased. For AD-CN, it again gave similar AUCs for SVM (0.896; 95%CI: 0.855-0.932) and CNN (0.876; 95%CI: 0.836-0.913). For prediction in MCI, performances decreased for both SVM (AUC = 0.665; 95%CI: 0.576-0.760) and CNN (AUC = 0.702; 95%CI: 0.624-0.786). Both with SVM and CNN, classification based on modulated GM maps significantly outperformed classification based on minimally processed images (p=0.01). Deep and conventional classifiers performed equally well for AD classification and their performance decreased only slightly when applied to the external cohort. We expect that this work on external validation contributes towards translation of machine learning to clinical practice.

@article{bron2021-34118592,
  author  = {Bron, E. E. and Klein, S. and Papma, J. M. and Jiskoot, L. C. and Venkatraghavan, V. and Linders, J. and Aalten, P. and De Deyn, P. P. and others},
  title   = {{Cross-cohort generalizability of deep and conventional machine learning for MRI-based diagnosis and prediction of Alzheimer's disease}},
  journal = {NeuroImage. Clinical},
  year    = {2021},
  doi     = {10.1016/j.nicl.2021.102712},
  pmid    = {34118592},
  url     = {https://doi.org/10.1016/j.nicl.2021.102712}
}

Key findings

• Application to conversion prediction in MCI yielded significantly higher performance for SVM (AUC = 0.756; 95%CI: 0.720-0.788) than for CNN (AUC = 0.742; 95%CI: 0.709-0.776) (p<0.01 for McNemar's test).
• Both with SVM and CNN, classification based on modulated GM maps significantly outperformed classification based on minimally processed images (p=0.01).

Why it matters

MRI-based classification of Alzheimer's disease patients versus controls generalized to an external data set, with support vector machine and convolutional neural network performing similarly, and support vector machine performed significantly better than the neural network for predicting conversion in mild cognitive impairment. This is a diagnostic and prediction study and does not test treating, curing or preventing Alzheimer's disease.

New to this subject? Start with our plain-English guide: Digital health →

Study details

• Participants199
• Effect sizeAUC = 0.756; AUC = 0.742; AUC = 0.665
• Evidence strengthModerate

Read the original

• DOI: 10.1016/j.nicl.2021.102712
• PubMed: 34118592
• Original source →
• Free full text: NeuroImage Clinical →

Full text: free to read — OpenAlex lists this copy as free to read: the link above (checked 6 Feb 1020).

Studies this one cites

Referenced by this study and also in the alzhub library, so you can read them here.

• Spatial patterns of neuroimaging biomarker change in individuals from families with autosomal dominant Alzheimer's disease: a longitudinal study The Lancet. Neurology · 2018-02-01 · usefulness score 77/100 · Brain scans and imaging
• The Alzheimer's disease neuroimaging initiative (ADNI): MRI methods Journal of Magnetic Resonance Imaging · 2008-02-26 · usefulness score 72/100 · Brain scans and imaging
• Predicting the progression of mild cognitive impairment using machine learning: A systematic, quantitative and critical review Medical image analysis · 2020-10-06 · usefulness score 70/100 · Digital health tools
• Alzheimer's Disease Neuroimaging Initiative (ADNI): clinical characterization Neurology · 2009-12-30 · usefulness score 68/100 · Mild cognitive impairment
• The diagnosis of dementia due to Alzheimer's disease: recommendations from the National Institute on Aging-Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's disease Alzheimer's & dementia : the journal of the Alzheimer's Association · 2011-04-21 · usefulness score 61/100 · Getting a diagnosis
• Automatic classification of MR scans in Alzheimer's disease Brain : a journal of neurology · 2008-01-17 · usefulness score 60/100 · Getting a diagnosis
• Time to diagnosis in young-onset dementia as compared with late-onset dementia Psychological medicine · 2012-05-28 · usefulness score 59/100 · Brain tissue and pathology
• Clinical diagnosis of Alzheimer's disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer's Disease Neurology · 1984-07-01 · usefulness score 58/100 · Getting a diagnosis
• Mild cognitive impairment as a diagnostic entity Journal of internal medicine · 2004-09-01 · usefulness score 57/100 · Getting a diagnosis
• A review on neuroimaging-based classification studies and associated feature extraction methods for Alzheimer's disease and its prodromal stages NeuroImage · 2017-04-13 · usefulness score 56/100 · Mild cognitive impairment
• Why do trials for Alzheimer's disease drugs keep failing? A discontinued drug perspective for 2010-2015 Expert opinion on investigational drugs · 2017-06-01 · usefulness score 49/100 · Brain tissue and pathology

Later studies that cite this one

1 study in the alzhub library lists this one in its references.

• Applications of artificial intelligence to aid early detection of dementia: A scoping review on current capabilities and future directions Journal of biomedical informatics · 2022-02-17 · usefulness score 47/100 · Digital health tools

Related studies in Digital health tools

• MedBackdoorGuard: Detecting and Mitigating Backdoor Attacks in Deep Learning Models for MRI-Based Alzheimer's Disease and Mild Cognitive Impairment Diagnosis Preprints.org · 2026-09-20 · usefulness score 16/100
• Advancing Alzheimer’s disease risk prediction: development and validation of a machine learning-based preclinical screening model in a cross-sectional study BMJ Open · 2025-02-01 · usefulness score 52/100
• Machine learning framework for early MRI-based Alzheimer's conversion prediction in MCI subjects NeuroImage · 2014-10-12 · usefulness score 36/100
• Brain MRI Analysis for Alzheimer's Disease Diagnosis Using CNN-Based Feature Extraction and Machine Learning Sensors (Basel, Switzerland) · 2022-04-11 · usefulness score 51/100
• Machine Learning-Based Framework for Differential Diagnosis Between Vascular Dementia and Alzheimer's Disease Using Structural MRI Features Frontiers in neurology · 2019-10-25 · usefulness score 46/100
• Deep learning-based brain age prediction in normal aging and dementia Nature aging · 2022-05-09 · usefulness score 59/100

Also tagged “classification”

835 studies in the library carry the topic tag “classification”; these are this study's neighbours in a fixed, alphabetical-by-hash tour of them, not the closest in content. All 835 tagged studies.

• Sex Differences in Alzheimer Disease Imaging Biomarkers in a Diverse, Community-Based Cohort Carolina Digital Repository (University of North Carolina at Chapel Hill) · 2026-02-06 · usefulness score 45/100 · Genetics
• Machine learning for modeling the progression of Alzheimer disease dementia using clinical data: a systematic literature review JAMIA open · 2021-07-01 · usefulness score 44/100 · Digital health tools
• Update on appropriate use criteria for amyloid PET imaging: dementia experts, mild cognitive impairment, and education Journal of nuclear medicine : official publication, Society of Nuclear Medicine · 2013-06-10 · usefulness score 45/100 · Biomarkers

Similar studies

• Generalizable deep learning model for early Alzheimer's disease detection from structural MRIs 2022 · Long-term cohort studies
• Machine learning for comprehensive forecasting of Alzheimer's Disease progression 2019 · Mild cognitive impairment
• Detection and Classification of Alzheimer’s Disease Using Deep and Machine Learning 2025 · Digital health tools
• Advances in Alzheimer’s disease diagnosis with machine learning and deep learning techniques: a comprehensive review 2026 · Digital health tools
• Early diagnosis of Alzheimer's disease using machine learning: a multi-diagnostic, generalizable approach 2022 · Digital health tools
• Improving Alzheimer's disease diagnosis with machine learning techniques 2011 · Digital health tools

More on Digital health tools

• A review that covers this ground Deep Learning-Based Diagnosis of Alzheimer's Disease Journal of personalized medicine · 2022-05-18 · usefulness score 47/100
• Newest in this subject Speaker-Wise Dementia Screening fromNaturalistic Speech with Lexical Entropy, Semantic Drift, and Multimodal Fusion Research Square · 2026-10-21 · Preprint — not yet peer-reviewed · usefulness score 16/100
• Newest in this subject Integrating Multisource Target Evidence with Machine-Learning Potency Prediction for Alzheimer’s Disease Drug Discovery bioRxiv · 2026-10-20 · Preprint — not yet peer-reviewed · usefulness score 16/100

See all studies in Digital health tools →