{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,20]],"date-time":"2026-10-20T19:40:32Z","timestamp":1792525232067,"version":"4.5.0"},"reference-count":20,"publisher":"Springer Science and Business Media LLC","license":[{"start":{"date-parts":[[2026,10,17]],"date-time":"2026-10-17T00:00:00Z","timestamp":1792195200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,10,17]],"date-time":"2026-10-17T00:00:00Z","timestamp":1792195200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100001659","name":"Deutsche Forschungsgemeinschaft","doi-asserted-by":"publisher","award":["454834942"],"award-info":[{"award-number":["454834942"]}],"id":[{"id":"10.13039\/501100001659","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001659","name":"Deutsche Forschungsgemeinschaft","doi-asserted-by":"publisher","award":["374011584"],"award-info":[{"award-number":["374011584"]}],"id":[{"id":"10.13039\/501100001659","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Int J CARS"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>\n                      <jats:bold>Purpose:<\/jats:bold>\n                    <\/jats:title>\n                    <jats:p>The goal of this study was to enhance the fairness and clinical applicability of deep learning models for classifying Alzheimer\u2019s disease from structural MRI scans. To achieve this, we addressed two key challenges: age-related bias and unreliable probability estimates. Specifically, we investigated in-processing debiasing strategies to achieve age-invariant predictions while preserving diagnostic accuracy. We also systematically evaluated post hoc calibration methods to improve the reliability, interpretability, and clinical usefulness of predicted probabilities across diverse patient populations.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>\n                      <jats:bold>Methods:<\/jats:bold>\n                    <\/jats:title>\n                    <jats:p>Structural MRI data from three independent cohorts (ADNI, AIBL, and DELCODE) were preprocessed using a standardized pipeline. Age-related bias was addressed using two in-processing strategies, adversarial training (CF-Net) and GLM-based feature residualization (PMDN), implemented in a shared 3D convolutional neural network (CNN) framework. Additionally, post hoc calibration was performed using four regression methods: logistic, probit, isotonic, and Bayesian binning. We evaluated model performance, fairness, and calibration using classification as well as calibration-specific metrics.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>\n                      <jats:bold>Results:<\/jats:bold>\n                    <\/jats:title>\n                    <jats:p>\n                      PMDN achieved the best overall performance, with higher balanced accuracy (0.71) and improved calibration (Brier\n                      <jats:inline-formula>\n                        <jats:alternatives>\n                          <jats:tex-math>$$\\approx $$<\/jats:tex-math>\n                          <mml:math xmlns:mml=\"http:\/\/www.w3.org\/1998\/Math\/MathML\">\n                            <mml:mo>\u2248<\/mml:mo>\n                          <\/mml:math>\n                        <\/jats:alternatives>\n                      <\/jats:inline-formula>\n                      0.22), showing substantially reduced variability compared to baseline and CF-Net. PMDN effectively removed age-related bias, achieving near-zero age prediction correlation; CF-Net, however, showed only partial and unstable debiasing. Post hoc calibration further improved probability reliability, with nonparametric methods (isotonic regression and Bayesian binning) yielding the lowest calibration errors (MCE\n                      <jats:inline-formula>\n                        <jats:alternatives>\n                          <jats:tex-math>$$\\approx $$<\/jats:tex-math>\n                          <mml:math xmlns:mml=\"http:\/\/www.w3.org\/1998\/Math\/MathML\">\n                            <mml:mo>\u2248<\/mml:mo>\n                          <\/mml:math>\n                        <\/jats:alternatives>\n                      <\/jats:inline-formula>\n                      0.21\u22120.24).\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>\n                      <jats:bold>Conclusion:<\/jats:bold>\n                    <\/jats:title>\n                    <jats:p>This study shows that, although deep learning models achieve similar classification performance, they differ significantly in terms of calibration, stability, and fairness. PMDN provides more consistent and effective age bias mitigation than the CF-Net approach, particularly in clinically relevant subgroups. Post hoc calibration improves probability reliability further, with nonparametric methods achieving the best calibration.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1007\/s11548-026-03805-x","type":"journal-article","created":{"date-parts":[[2026,10,17]],"date-time":"2026-10-17T19:57:04Z","timestamp":1792267024000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Improving the fairness and robustness of deep learning classification models through bias mitigation and post hoc calibration"],"prefix":"10.1007","author":[{"given":"Majid","family":"Ramedani","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jaya C.","family":"Terli","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ole H.","family":"Martensen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tobias","family":"Strau\u00df","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Katharina","family":"Buerger","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Peter","family":"Dechent","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Emrah","family":"D\u00fczel","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Klaus","family":"Fliessbach","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Maria","family":"Gemenetzi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wenzel","family":"Glanz","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Julian","family":"Hellmann-Regen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Stefan","family":"Hetzer","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Enise I.","family":"Incesoy","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Daniel","family":"Janowitz","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Marie T.","family":"Kronm\u00fcller","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Christoph","family":"Laske","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Falk","family":"L\u00fcsebrink","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Robert","family":"Perneczky","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Oliver","family":"Peters","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lukas","family":"Preis","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Josef","family":"Priller","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Boris-Stephan","family":"Rauchmann","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ayda","family":"Rostamzadeh","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Anja","family":"Schneider","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Bj\u00f6rn H.","family":"Schott","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sebastian","family":"Sodenkamp","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Annika","family":"Spottke","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Eike J.","family":"Spruth","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sophia","family":"St\u00f6cklein","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jens","family":"Wiltfang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Frank","family":"Jessen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Stefan J.","family":"Teipel","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3353-3167","authenticated-orcid":false,"given":"Martin","family":"Dyrba","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,10,17]]},"reference":[{"key":"3805_CR1","doi-asserted-by":"publisher","DOI":"10.1016\/j.hybadv.2024.100277","volume":"7","author":"AB Rashid","year":"2024","unstructured":"Rashid AB, Kausik MAK (2024) AI revolutionizing industries worldwide: a comprehensive overview of its diverse applications. Hybrid Adv 7:100277. https:\/\/doi.org\/10.1016\/j.hybadv.2024.100277","journal-title":"Hybrid Adv"},{"issue":"4","key":"3805_CR2","doi-asserted-by":"publisher","first-page":"440","DOI":"10.1093\/bib\/bbaf440","volume":"26","author":"J Zhang","year":"2025","unstructured":"Zhang J, Che Y, Liu R, Wang Z, Liu W (2025) Deep learning\u2013driven multi-omics analysis: enhancing cancer diagnostics and therapeutics. Brief Bioinform 26(4):440. https:\/\/doi.org\/10.1093\/bib\/bbaf440","journal-title":"Brief Bioinform"},{"key":"3805_CR3","doi-asserted-by":"publisher","DOI":"10.1016\/j.arr.2024.102357","volume":"99","author":"DG Gadhave","year":"2024","unstructured":"Gadhave DG, Sugandhi VV, Jha SK, Nangare SN, Gupta G, Singh SK, Dua K, Cho H, Hansbro PM, Paudel KR (2024) Neurodegenerative disorders: mechanisms of degeneration and therapeutic approaches with their clinical relevance. Ageing Res Rev 99:102357. https:\/\/doi.org\/10.1016\/j.arr.2024.102357","journal-title":"Ageing Res Rev"},{"key":"3805_CR4","doi-asserted-by":"publisher","first-page":"007","DOI":"10.1093\/psyrad\/kkaf007","volume":"5","author":"L Zhao","year":"2025","unstructured":"Zhao L (2025) Advances in functional magnetic resonance imaging-based brain function mapping: a deep learning perspective. Psychoradiology 5:007. https:\/\/doi.org\/10.1093\/psyrad\/kkaf007","journal-title":"Psychoradiology"},{"key":"3805_CR5","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejrad.2023.110934","volume":"165","author":"Z Yao","year":"2023","unstructured":"Yao Z, Wang H, Yan W, Wang Z, Zhang W, Wang Z, Zhang G (2023) Artificial intelligence-based diagnosis of Alzheimer\u2019s disease with brain MRI images. Eur J Radiol 165:110934. https:\/\/doi.org\/10.1016\/j.ejrad.2023.110934","journal-title":"Eur J Radiol"},{"issue":"1","key":"3805_CR6","doi-asserted-by":"publisher","first-page":"464","DOI":"10.3390\/make6010024","volume":"6","author":"MG Alsubaie","year":"2024","unstructured":"Alsubaie MG, Luo S, Shaukat K (2024) Alzheimer\u2019s disease detection using deep learning on neuroimaging: a systematic review. Mach Learn Knowl Ext 6(1):464\u2013505. https:\/\/doi.org\/10.3390\/make6010024","journal-title":"Mach Learn Knowl Ext"},{"issue":"1","key":"3805_CR7","doi-asserted-by":"publisher","first-page":"68","DOI":"10.1109\/TTS.2023.3234203","volume":"4","author":"T Dhar","year":"2023","unstructured":"Dhar T, Dey N, Borra S, Sherratt RS (2023) Challenges of deep learning in medical image analysis\u2014improving explainability and trust. IEEE Trans Technol Soc 4(1):68\u201375. https:\/\/doi.org\/10.1109\/TTS.2023.3234203","journal-title":"IEEE Trans Technol Soc"},{"key":"3805_CR8","doi-asserted-by":"publisher","DOI":"10.1007\/s43441-025-00849-0","author":"C Anagnostopoulos","year":"2025","unstructured":"Anagnostopoulos C, Schaar M, Collet J-P, Vecino R (2025) Real-world data and causal machine learning to enhance drug development. Ther Innov Regul Sci. https:\/\/doi.org\/10.1007\/s43441-025-00849-0","journal-title":"Ther Innov Regul Sci"},{"key":"3805_CR9","doi-asserted-by":"publisher","DOI":"10.1016\/j.neuroimage.2020.117401","volume":"224","author":"NK Dinsdale","year":"2021","unstructured":"Dinsdale NK, Bluemke E, Smith SM, Arya Z, Vidaurre D, Jenkinson M, Namburete AIL (2021) Learning patterns of the ageing brain in MRI using deep convolutional networks. Neuroimage 224:117401. https:\/\/doi.org\/10.1016\/j.neuroimage.2020.117401","journal-title":"Neuroimage"},{"key":"3805_CR10","doi-asserted-by":"publisher","first-page":"34912","DOI":"10.1109\/ACCESS.2023.3265327","volume":"11","author":"S Mehra","year":"2023","unstructured":"Mehra S, Raut G, Purkayastha RD, Vishvakarma SK, Biasizzo A (2023) An empirical evaluation of enhanced performance softmax function in deep learning. IEEE Access 11:34912\u201334924. https:\/\/doi.org\/10.1109\/ACCESS.2023.3265327","journal-title":"IEEE Access"},{"key":"3805_CR11","unstructured":"Pearce, T., Brintrup, A., Zhu, J.: (2021) Understanding softmax confidence and uncertainty. CoRR arXiv:2106.04972"},{"key":"3805_CR12","doi-asserted-by":"publisher","unstructured":"Ellis KA, Bush AI, Darby D, De Fazio D, Foster J, Hudson P, Lautenschlager NT, Lenzo N, Martins RN, Maruff P, Masters C, Milner A, Pike K, Rowe C, Savage G, Szoeke C, Taddei K, Villemagne V, Woodward M, Ames D (2009) The Australian imaging, biomarkers and lifestyle (AIBL) study of aging: methodology and baseline characteristics of 1112 individuals recruited for a longitudinal study of Alzheimer\u2019s disease. Int Psychogeriatr 21(4):672\u2013687. https:\/\/doi.org\/10.1017\/S1041610209009405","DOI":"10.1017\/S1041610209009405"},{"key":"3805_CR13","doi-asserted-by":"publisher","DOI":"10.1186\/s13195-017-0314-2","author":"F Jessen","year":"2018","unstructured":"Jessen F, Spottke A, Boecker H, Brosseron F, Buerger K, Catak C, Fliessbach K, Franke C, Fuentes Casa\u00f1 M, Heneka M, Janowitz D, Kilimann I, Laske C, Menne F, Nestor P, Peters O, Priller J, Pross V, Ramirez A, Duzel E (2018) Design and first baseline data of the DZNE multicenter observational study on predementia Alzheimer\u2019s disease (delcode). Alzheimer\u2019s Res Ther. https:\/\/doi.org\/10.1186\/s13195-017-0314-2","journal-title":"Alzheimer\u2019s Res Ther"},{"key":"3805_CR14","doi-asserted-by":"publisher","unstructured":"Jack CR Jr, Bernstein MA, Fox NC, Thompson P, Alexander G, Harvey D, Borowski B, Britson PJ, Whitwell L, J., Ward, C., Dale, A.M., Felmlee, J.P., Gunter, J.L., Hill, D.L.G., Killiany, R., Schuff, N., Fox-Bosetti, S., Lin, C., Studholme, C., DeCarli, C.S., Krueger, G., Ward, H.A., Metzger, G.J., Scott, K.T., Mallozzi, R., Blezek, D., Levy, J., Debbins, J.P., Fleisher, A.S., Albert, M., Green, R., Bartzokis, G., Glover, G., Mugler, J., Weiner, M.W. (2008) The Alzheimer\u2019s disease neuroimaging initiative (ADNI): MRI methods. J Magn Reson Imaging 27(4):685\u2013691. https:\/\/doi.org\/10.1002\/jmri.21049","DOI":"10.1002\/jmri.21049"},{"issue":"1","key":"3805_CR15","doi-asserted-by":"publisher","first-page":"126","DOI":"10.1111\/jon.12673","volume":"30","author":"A George","year":"2020","unstructured":"George A, Kuzniecky R, Rusinek H, Pardoe HR (2020) Human Epilepsy Project Investigators: Standardized brain MRI acquisition protocols improve statistical power in multicenter quantitative morphometry studies. J Neuroimaging 30(1):126\u2013133. https:\/\/doi.org\/10.1111\/jon.12673","journal-title":"J Neuroimaging"},{"key":"3805_CR16","doi-asserted-by":"publisher","unstructured":"Zhao Q, Adeli E, Pohl K (2020) Training confounder-free deep learning models for medical applications. Nat Commun 11. https:\/\/doi.org\/10.1038\/s41467-020-19784-9","DOI":"10.1038\/s41467-020-19784-9"},{"key":"3805_CR17","doi-asserted-by":"publisher","unstructured":"Vento, A., Zhao, Q., Paul, R., Pohl, K.M., Adeli, E.: A penalty approach for normalizing feature distributions to build confounder-free models. In: Wang, L., Dou, Q., Fletcher, P.T., Speidel, S., Li, S. (eds.) Medical Image Computing and Computer Assisted Intervention \u2013 MICCAI 2022, pp. 387\u2013397. Springer, Cham (2022). https:\/\/doi.org\/10.1007\/978-3-031-16437-8_37","DOI":"10.1007\/978-3-031-16437-8_37"},{"key":"3805_CR18","doi-asserted-by":"crossref","unstructured":"Lu M, Zhao Q, Zhang J, Pohl KM, Fei-Fei L, Niebles JC, Adeli E. Metadata normalization. In2021 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2021 Jun 20 (pp. 10912-10922). IEEE","DOI":"10.1109\/CVPR46437.2021.01077"},{"key":"3805_CR19","doi-asserted-by":"publisher","unstructured":"Martensen OH, Strau\u00df T, Ramedani M, Dyrba M.: Comparison of post-hoc calibration methods for neural network likelihood scores. In: Bildverarbeitung F\u00fcr die Medizin 2026, pp. 443\u2013449. Springer, Wiesbaden (2026). https:\/\/doi.org\/10.1007\/978-3-658-51100-5_87","DOI":"10.1007\/978-3-658-51100-5_87"},{"key":"3805_CR20","doi-asserted-by":"publisher","unstructured":"Rainio O, Teuho J, Kl\u00e9n R (2024) Evaluation metrics and statistical tests for machine learning. Sci Rep 14. https:\/\/doi.org\/10.1038\/s41598-024-56706-x","DOI":"10.1038\/s41598-024-56706-x"}],"container-title":["International Journal of Computer Assisted Radiology and Surgery"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s11548-026-03805-x.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s11548-026-03805-x","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s11548-026-03805-x.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,20]],"date-time":"2026-10-20T19:02:51Z","timestamp":1792522971000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s11548-026-03805-x"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10,17]]},"references-count":20,"alternative-id":["3805"],"URL":"https:\/\/doi.org\/10.1007\/s11548-026-03805-x","relation":{},"ISSN":["1861-6429"],"issn-type":[{"value":"1861-6429","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,17]]},"assertion":[{"value":"1 April 2026","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"25 September 2026","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"17 October 2026","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"The authors have no conflict of interest to declare that are relevant to the content of this article.","order":1,"name":"Ethics","label":"Conflict of interest","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The studies were approved by the respective neuroimaging initiatives internal review boards of each of the participating study sites. See\n                      \n                      and\n                      \n                      for details. All initiatives met common ethical standards in the collection of the data such as the Declaration of Helsinki. Analysis of the data was approved by the internal review board of the Rostock University Medical Center, reference number A 2020-0182.","order":2,"name":"Ethics","label":"Ethical Approval","group":{"name":"EthicsHeading","label":"Declarations"}}]}}