{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,25]],"date-time":"2026-08-25T10:10:55Z","timestamp":1787652655211,"version":"build-2736575974"},"reference-count":59,"publisher":"Wiley","issue":"2","license":[{"start":{"date-parts":[[2026,2,2]],"date-time":"2026-02-02T00:00:00Z","timestamp":1769990400000},"content-version":"vor","delay-in-days":1,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2026,2,1]],"date-time":"2026-02-01T00:00:00Z","timestamp":1769904000000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"funder":[{"DOI":"10.13039\/501100000272","name":"National Institute for Health and Care Research","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100000272","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100002283","name":"Alzheimer\u2019s Research UK","doi-asserted-by":"publisher","award":["ARUK\u2010PPG2023B\u2010016"],"award-info":[{"award-number":["ARUK\u2010PPG2023B\u2010016"]}],"id":[{"id":"10.13039\/501100002283","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000049","name":"National Institute on Aging","doi-asserted-by":"publisher","award":["P30AG017265"],"award-info":[{"award-number":["P30AG017265"]}],"id":[{"id":"10.13039\/100000049","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["T32AG000037\u201046"],"award-info":[{"award-number":["T32AG000037\u201046"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["R01NS100973"],"award-info":[{"award-number":["R01NS100973"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["RF1AG082201"],"award-info":[{"award-number":["RF1AG082201"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["R01AG079957"],"award-info":[{"award-number":["R01AG079957"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["U19AG024904"],"award-info":[{"award-number":["U19AG024904"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000005","name":"U.S. Department of Defense","doi-asserted-by":"publisher","award":["W81XWH\u201018\u20101\u20100413"],"award-info":[{"award-number":["W81XWH\u201018\u20101\u20100413"]}],"id":[{"id":"10.13039\/100000005","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Human Brain Mapping"],"published-print":{"date-parts":[[2026,2,1]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                  <jats:p>Artificial intelligence and neuroimaging enable accurate dementia prediction but often involve \u2018black box\u2019 models that can be difficult to trust. Explainable artificial intelligence (XAI) aims to provide insights into the model's decisions; however, choosing the most appropriate method is non\u2010trivial and often context\u2010specific. We used T1\u2010weighted MRI to train models on two tasks: Alzheimer's disease (AD) classification (diagnosis) and predicting conversion from mild\u2010cognitive impairment (MCI) to all\u2010cause dementia (prognosis). We applied eleven XAI methods across two popular image classification architectures, producing visualisations of the most salient regions. We also propose a framework for interpreting explanations produced by different XAI methods and predictive models. Models achieved balanced accuracies of 81% and 67% for diagnosis and prognosis. XAI outputs highlighted brain regions relevant to AD with strong convergence across gradient\u2010based techniques. LIME produced explanations that were most similar across architectures. Mean saliency enhanced MCI prognosis prediction when included as an additional input feature. XAI can be used to verify that models are utilising relevant features and to generate valuable measures for further analysis.<\/jats:p>","DOI":"10.1002\/hbm.70456","type":"journal-article","created":{"date-parts":[[2026,2,2]],"date-time":"2026-02-02T12:20:51Z","timestamp":1770034851000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":4,"title":["Investigating the Utility of Explainable Artificial Intelligence for Neuroimaging\u2010Based Dementia Diagnosis and Prognosis"],"prefix":"10.1002","volume":"47","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-3819-1634","authenticated-orcid":false,"given":"Sophie A.","family":"Martin","sequence":"first","affiliation":[{"name":"UCL Hawkes Institute University College London  London UK"},{"name":"UCL Queen Square Institute of Neurology University College London  London UK"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"An","family":"Zhao","sequence":"additional","affiliation":[{"name":"UCL Hawkes Institute University College London  London UK"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jiongqi","family":"Qu","sequence":"additional","affiliation":[{"name":"UCL Hawkes Institute University College London  London UK"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Phoebe","family":"Imms","sequence":"additional","affiliation":[{"name":"Leonard Davis School of Gerontology, University of Southern California  Los Angeles California USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Andrei","family":"Irimia","sequence":"additional","affiliation":[{"name":"Leonard Davis School of Gerontology, University of Southern California  Los Angeles California USA"},{"name":"Department of Biomedical Engineering, Viterbi School of Engineering, Corwin D. Denney Research Centre University of Southern California  Los Angeles California USA"},{"name":"Department of Quantitative &amp; Computational Biology, Dana and David Dornsife College of Arts &amp; Sciences University of Southern California  Los Angeles California USA"},{"name":"Centre for Healthy Brain Aging, Institute of Psychiatry, Psychology and Neuroscience King's College London  London UK"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Frederik","family":"Barkhof","sequence":"additional","affiliation":[{"name":"UCL Hawkes Institute University College London  London UK"},{"name":"UCL Queen Square Institute of Neurology University College London  London UK"},{"name":"Department of Radiology and Nuclear Medicine Amsterdam University Medical Centre  Amsterdam the Netherlands"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"James H.","family":"Cole","sequence":"additional","affiliation":[{"name":"UCL Hawkes Institute University College London  London UK"},{"name":"UCL Queen Square Institute of Neurology University College London  London UK"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"name":"Alzheimer's Disease Neuroimaging Initiative","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2026,2,2]]},"reference":[{"key":"e_1_2_12_2_1","first-page":"9525","volume-title":"Proceedings of the 32nd International Conference on Neural Information Processing Systems","author":"Adebayo J.","year":"2018"},{"issue":"1","key":"e_1_2_12_3_1","doi-asserted-by":"crossref","DOI":"10.1038\/s41598-024-59578-3","article-title":"Joint Transformer Architecture in Brain 3D MRI Classification: Its Application in Alzheimer's Disease Classification","volume":"14","author":"Alp S.","year":"2024","journal-title":"Scientific Reports"},{"key":"e_1_2_12_4_1","unstructured":"Alvarez\u2010Melis D. andT. S.Jaakkola.2018.\u201cOn the Robustness of Interpretability Methods.\u201d[Internet]. arXiv:1806.08049.http:\/\/arxiv.org\/abs\/1806.08049."},{"key":"e_1_2_12_5_1","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0130140"},{"key":"e_1_2_12_6_1","unstructured":"Bai Y. J.Mei A.Yuille andC.Xie.2021.\u201cAre Transformers More Robust Than CNNs?\u201d[Internet]. arXiv:2111.05464.http:\/\/arxiv.org\/abs\/2111.05464."},{"key":"e_1_2_12_7_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2024.108029"},{"key":"e_1_2_12_8_1","doi-asserted-by":"publisher","DOI":"10.1002\/alz.13412"},{"key":"e_1_2_12_9_1","unstructured":"Chen Z. V.Subhash M.Havasi W.Pan andF.Doshi\u2010Velez.2022.\u201cWhat Makes a Good Explanation?: A Harmonized View of Properties of Explanations.\u201d [Internet].arXiv:2211.05667.https:\/\/arxiv.org\/abs\/2211.05667."},{"key":"e_1_2_12_10_1","unstructured":"Das A. andP.Rad.2020.\u201cOpportunities and Challenges in Explainable Artificial Intelligence (XAI): A Survey.\u201d[Internet]. arXiv:2006.11371.http:\/\/arxiv.org\/abs\/2006.11371."},{"key":"e_1_2_12_11_1","doi-asserted-by":"crossref","unstructured":"Deng J. W.Dong R.Socher L. J.Li L.Kai andF. F.Li.2009.\u201cImageNet: A Large\u2010Scale Hierarchical Image Database [Internet].\u201dIn2009 IEEE Conference on Computer Vision and Pattern Recognition 248\u201355. IEEE.","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"e_1_2_12_12_1","doi-asserted-by":"crossref","unstructured":"Dhinagar N. J. S. I.Thomopoulos E.Laltoo andP. M.Thompson.2023.\u201cEfficiently Training Vision Transformers on Structural MRI Scans for Alzheimer's Disease Detection.\u201d[Internet] arXiv.https:\/\/arxiv.org\/abs\/2303.08216.","DOI":"10.1109\/EMBC40787.2023.10341190"},{"key":"e_1_2_12_13_1","unstructured":"Dosovitskiy A. L.Beyer A.Kolesnikov et\u00a0al.2020.\u201cAn Image Is Worth 16x16 Words: Transformers for Image Recognition at Scale.\u201d[Internet]. arXiv:2010.11929.https:\/\/arxiv.org\/abs\/2010.11929."},{"key":"e_1_2_12_14_1","doi-asserted-by":"publisher","DOI":"10.1117\/1.JMI.8.2.024503"},{"key":"e_1_2_12_15_1","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0184661"},{"key":"e_1_2_12_16_1","doi-asserted-by":"publisher","DOI":"10.1515\/REVNEURO.2010.21.3.187"},{"key":"e_1_2_12_17_1","unstructured":"Gani H. M.Naseer andM.Yaqub.2022.\u201cHow to Train Vision Transformer on Small\u2010Scale Datasets?\u201d[Internet]. arXiv:2210.07240.http:\/\/arxiv.org\/abs\/2210.07240."},{"key":"e_1_2_12_18_1","doi-asserted-by":"publisher","DOI":"10.1001\/archpsyc.55.9.809"},{"key":"e_1_2_12_19_1","unstructured":"Gildenblat J.2021.\u201cPyTorch library for CAM Methods.\u201d [Internet].GitHub.https:\/\/github.com\/jacobgil\/pytorch\u2010grad\u2010cam."},{"key":"e_1_2_12_20_1","doi-asserted-by":"crossref","unstructured":"Guo K. N.Chaudhari T.Jafar N.Chowdhury P.Bogdan andA.Irimia.2024.\u201cAnatomic Interpretability in Neuroimage Deep Learning: Saliency Approaches for Typical Aging and Traumatic Brain Injury.\u201dNeuroinformatics 22 no. 4: 591\u2013606.","DOI":"10.1007\/s12021-024-09694-2"},{"key":"e_1_2_12_21_1","doi-asserted-by":"publisher","DOI":"10.1523\/JNEUROSCI.19-10-04065.1999"},{"key":"e_1_2_12_22_1","unstructured":"He K. X.Zhang S.Ren andJ.Sun.2015.\u201cDeep Residual Learning for Image Recognition.\u201d[Internet]. arXiv:1512.03385.https:\/\/arxiv.org\/abs\/1512.03385."},{"key":"e_1_2_12_23_1","doi-asserted-by":"crossref","first-page":"403","DOI":"10.1007\/978-3-031-63787-2_21","volume-title":"Explainable Artificial Intelligence","author":"Hedstr\u00f6m A.","year":"2024"},{"key":"e_1_2_12_24_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-023-33781-0"},{"key":"e_1_2_12_25_1","unstructured":"Khan S. M.Naseer M.Hayat S. W.Zamir F. S.Khan andM.Shah.2021.\u201cTransformers in Vision: A Survey.\u201d[Internet]. arXiv:2101.01169.https:\/\/arxiv.org\/abs\/2101.01169."},{"key":"e_1_2_12_26_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2024.108116"},{"key":"e_1_2_12_27_1","doi-asserted-by":"crossref","first-page":"267","DOI":"10.1007\/978-3-030-28954-6_14","volume-title":"Explainable AI: Interpreting, Explaining and Visualizing Deep Learning","author":"Kindermans P. J.","year":"2019"},{"key":"e_1_2_12_28_1","unstructured":"Kokhlikyan N. V.Miglani M.Martin et\u00a0al.2020.\u201cCaptum: A Unified and Generic Model Interpretability Library for PyTorch.\u201d[Internet]. arXiv:2009.07896.http:\/\/arxiv.org\/abs\/2009.07896."},{"key":"e_1_2_12_29_1","doi-asserted-by":"crossref","first-page":"334","DOI":"10.1007\/978-3-031-72390-2_32","volume-title":"Medical Image Computing and Computer Assisted Intervention \u2013 MICCAI 2024","author":"Kunanbayev K.","year":"2024"},{"key":"e_1_2_12_30_1","unstructured":"Lee S. H. S.Lee andB. C.Song.2021.\u201cVision Transformer for Small\u2010Size Datasets.\u201d[Internet]. arXiv:2112.13492.https:\/\/arxiv.org\/abs\/2112.13492."},{"key":"e_1_2_12_31_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41746\u2010024\u201001123\u20107"},{"key":"e_1_2_12_32_1","first-page":"1","volume-title":"2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI)","author":"Li C.","year":"2022"},{"key":"e_1_2_12_33_1","doi-asserted-by":"crossref","unstructured":"Liu Z. Y.Lin Y.Cao et\u00a0al.2021.\u201cSwin Transformer: Hierarchical Vision Transformer Using Shifted Windows.\u201darXiv:2103.14030.https:\/\/doi.org\/10.48550\/arXiv.2103.14030.","DOI":"10.1109\/ICCV48922.2021.00986"},{"key":"e_1_2_12_34_1","doi-asserted-by":"crossref","first-page":"463","DOI":"10.1145\/3529190.3534754","volume-title":"Proceedings of the 15th International Conference on PErvasive Technologies Related to Assistive Environments","author":"Lyu Y.","year":"2022"},{"key":"e_1_2_12_35_1","unstructured":"Malpure D. O.Litake andR.Ingle.2021.\u201cInvestigating Transfer Learning Capabilities of Vision Transformers and CNNs by Fine\u2010Tuning a Single Trainable Block.\u201d[Internet]. arXiv:2110.05270.http:\/\/arxiv.org\/abs\/2110.05270."},{"key":"e_1_2_12_36_1","first-page":"112","volume-title":"Clinical Image\u2010Based Procedures, Fairness of AI in Medical Imaging, and Ethical and Philosophical Issues in Medical Imaging","author":"Martin S. A.","year":"2023"},{"key":"e_1_2_12_37_1","doi-asserted-by":"publisher","DOI":"10.1002\/alz.12948"},{"key":"e_1_2_12_38_1","doi-asserted-by":"crossref","DOI":"10.1016\/j.artint.2024.104179","article-title":"Assessing Fidelity in XAI Post\u2010Hoc Techniques: A Comparative Study With Ground Truth Explanations Datasets","volume":"335","author":"Mir\u00f3\u2010Nicolau M.","year":"2024","journal-title":"Artificial Intelligence"},{"key":"e_1_2_12_39_1","unstructured":"Monai P.2020.\u201cThe MONAI Consortium.\u201d Zenodo."},{"key":"e_1_2_12_40_1","doi-asserted-by":"crossref","unstructured":"Otsuki S. T.Iida F.Doublet et\u00a0al.2024.\u201cLayer\u2010Wise Relevance Propagation With Conservation Property for ResNet.\u201d[Internet]. arXiv:2407.09115.http:\/\/arxiv.org\/abs\/2407.09115.","DOI":"10.1007\/978-3-031-72775-7_20"},{"key":"e_1_2_12_41_1","unstructured":"Park N. andS.Kim.2022.\u201cHow Do Vision Transformers Work?\u201d[Internet]. arXiv:2202.06709.http:\/\/arxiv.org\/abs\/2202.06709."},{"key":"e_1_2_12_42_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-022-31037-5"},{"key":"e_1_2_12_43_1","first-page":"12116","volume-title":"Do Vision Transformers See Like Convolutional Neural Networks? In: Advances in Neural Information Processing Systems","author":"Raghu M.","year":"2021"},{"key":"e_1_2_12_44_1","unstructured":"Reinke A. M. D.Tizabi C. H.Sudre et\u00a0al.2021.\u201cCommon Limitations of Image Processing Metrics: A Picture Story.\u201d [Internet].arXiv:2104.05642v8.https:\/\/arxiv.org\/abs\/2104.05642."},{"key":"e_1_2_12_45_1","first-page":"1135","article-title":"\u2018Why Should I Trust You?\u2019 Explaining the Predictions of Any Classifier","volume":"13","author":"Ribeiro M. T.","year":"2016","journal-title":"Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining"},{"key":"e_1_2_12_46_1","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0118432"},{"key":"e_1_2_12_47_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-019-01228-7"},{"key":"e_1_2_12_48_1","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P19-1282"},{"key":"e_1_2_12_49_1","first-page":"307","volume-title":"Contributions to the Theory of Games (AM\u201028), volume II","author":"Shapley L. S.","year":"2016"},{"key":"e_1_2_12_50_1","unstructured":"Simonyan K. A.Vedaldi andA.Zisserman.2014.\u201cDeep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps.\u201d [Internet].arXiv:1312.6034.https:\/\/arxiv.org\/abs\/1312.6034."},{"key":"e_1_2_12_51_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2023.102809"},{"key":"e_1_2_12_52_1","first-page":"3319","volume-title":"International Conference On Machine Learning","author":"Sundararajan M.","year":"2017"},{"key":"e_1_2_12_53_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10278-022-00666-z"},{"key":"e_1_2_12_54_1","doi-asserted-by":"publisher","DOI":"10.1212\/WNL.0000000000207298"},{"key":"e_1_2_12_55_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.neuroimage.2023.119929"},{"key":"e_1_2_12_56_1","unstructured":"Wightman R. N.Raw A.Soare et\u00a0al.2023.\u201crwightman\/pytorch\u2010image\u2010models: v0.8.10dev0 Release.\u201d [Internet].Zenodo.https:\/\/zenodo.org\/records\/7618837."},{"key":"e_1_2_12_57_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-024-03118-z"},{"issue":"1","key":"e_1_2_12_58_1","doi-asserted-by":"crossref","DOI":"10.1038\/s41598-021-01681-w","article-title":"Effect of Data Leakage in Brain MRI Classification Using 2D Convolutional Neural Networks","volume":"11","author":"Yagis E.","year":"2021","journal-title":"Scientific Reports"},{"key":"e_1_2_12_59_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10590-1_53"},{"key":"e_1_2_12_60_1","doi-asserted-by":"crossref","DOI":"10.1016\/j.bspc.2023.104787","article-title":"IDA\u2010Net: Inheritable Deformable Attention Network of Structural MRI for Alzheimer's Disease Diagnosis","volume":"84","author":"Zhao Q.","year":"2023","journal-title":"Biomedical Signal Processing and Control"}],"container-title":["Human Brain Mapping"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/hbm.70456","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/hbm.70456","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/hbm.70456","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,2,10]],"date-time":"2026-02-10T07:53:13Z","timestamp":1770709993000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1002\/hbm.70456"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,2,1]]},"references-count":59,"journal-issue":{"issue":"2","published-print":{"date-parts":[[2026,2,1]]}},"alternative-id":["10.1002\/hbm.70456"],"URL":"https:\/\/doi.org\/10.1002\/hbm.70456","archive":["Portico"],"relation":{},"ISSN":["1065-9471","1097-0193"],"issn-type":[{"value":"1065-9471","type":"print"},{"value":"1097-0193","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,2,1]]},"assertion":[{"value":"2025-08-16","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-01-18","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-02-02","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e70456"}}