{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,1,29]],"date-time":"2026-01-29T21:43:13Z","timestamp":1769722993310,"version":"3.49.0"},"reference-count":26,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2024,2,6]],"date-time":"2024-02-06T00:00:00Z","timestamp":1707177600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100010661","name":"Horizon 2020 Framework Programme","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100010661","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Neuroinform."],"abstract":"<jats:sec><jats:title>Background<\/jats:title><jats:p>The willingness to trust predictions formulated by automatic algorithms is key in a wide range of domains. However, a vast number of deep architectures are only able to formulate predictions without associated uncertainty.<\/jats:p><\/jats:sec><jats:sec><jats:title>Purpose<\/jats:title><jats:p>In this study, we propose a method to convert a standard neural network into a Bayesian neural network and estimate the variability of predictions by sampling different networks similar to the original one at each forward pass.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>We combine our method with a tunable rejection-based approach that employs only the fraction of the data, i.e., the share that the model can classify with an uncertainty below a user-set threshold. We test our model in a large cohort of brain images from patients with Alzheimer's disease and healthy controls, discriminating the former and latter classes based on morphometric images exclusively.<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>We demonstrate how combining estimated uncertainty with a rejection-based approach increases classification accuracy from 0.86 to 0.95 while retaining 75% of the test set. In addition, the model can select the cases to be recommended for, e.g., expert human evaluation due to excessive uncertainty. Importantly, our framework circumvents additional workload during the training phase by using our network \u201cturned into Bayesian\u201d to implicitly investigate the loss landscape in the neighborhood of each test sample in order to determine the reliability of the predictions.<\/jats:p><\/jats:sec><jats:sec><jats:title>Conclusion<\/jats:title><jats:p>We believe that being able to estimate the uncertainty of a prediction, along with tools that can modulate the behavior of the network to a degree of confidence that the user is informed about (and comfortable with), can represent a crucial step in the direction of user compliance and easier integration of deep learning tools into everyday tasks currently performed by human operators.<\/jats:p><\/jats:sec>","DOI":"10.3389\/fninf.2024.1346723","type":"journal-article","created":{"date-parts":[[2024,2,6]],"date-time":"2024-02-06T04:26:15Z","timestamp":1707193575000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":8,"title":["Enabling uncertainty estimation in neural networks through weight perturbation for improved Alzheimer's disease classification"],"prefix":"10.3389","volume":"18","author":[{"given":"Matteo","family":"Ferrante","sequence":"first","affiliation":[]},{"given":"Tommaso","family":"Boccato","sequence":"additional","affiliation":[]},{"given":"Nicola","family":"Toschi","sequence":"additional","affiliation":[]}],"member":"1965","published-online":{"date-parts":[[2024,2,6]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"243","DOI":"10.1016\/j.inffus.2021.05.008","article-title":"A review of uncertainty quantification in deep learning: techniques, applications and challenges","volume":"76","author":"Abdar","year":"2021","journal-title":"Inf. Fus"},{"key":"B2","doi-asserted-by":"publisher","DOI":"10.54294\/uvnhin","article-title":"Advanced normalization tools (ants)","author":"Avants","year":"2008","journal-title":"Insight J"},{"key":"B3","first-page":"1613","article-title":"\u201cWeight uncertainty in neural networks,\u201d","volume-title":"Proceedings of the 32nd International Conference on International Conference on Machine Learning - Volume 37 (ICML'15)","author":"Blundell","year":"2015"},{"key":"B4","article-title":"MONAI: An open-source framework for deep learning in healthcare","author":"Cardoso","year":"2022","journal-title":"arXiv [Preprint]."},{"key":"B5","unstructured":"\u201cLaplace redux -effortless bayesian deep learning,\u201d\n            DaxbergerE.\n            KristiadiA.\n            ImmerA.\n            EschenhagenR.\n            BauerM.\n            HennigP.\n          Advances in Neural Information Processing Systems, vol.2021"},{"key":"B6","unstructured":"\u201cDropout as a Bayesian approximation: representing model uncertainty in deep learning,\u201d10501059\n            GalY.\n            GhahramaniZ.\n          Proceedings of The 33rd International Conference on Machine Learning, in Proceedings of Machine Learning Research482016"},{"key":"B7","doi-asserted-by":"publisher","first-page":"105151","DOI":"10.1016\/j.engappai.2022.105151","article-title":"Ensemble deep learning: a review","volume":"115","author":"Ganaie","year":"2022","journal-title":"Eng. Appl. Artif. Intell"},{"key":"B8","doi-asserted-by":"publisher","first-page":"456","DOI":"10.3389\/fimmu.2020.00456","article-title":"A path toward precision medicine for neuroinflammatory mechanisms in alzheimer's disease","volume":"11","author":"Hampel","year":"2020","journal-title":"Front. Immunol"},{"key":"B9","doi-asserted-by":"publisher","first-page":"S47","DOI":"10.3233\/JAD-179932","article-title":"Revolution of alzheimer precision neurology. passageway of systems biology and neurophysiology","volume":"64","author":"Hampel","year":"2018","journal-title":"J. Alzheimers Dis"},{"key":"B10","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2015.123","article-title":"Delving deep into rectifiers: surpassing human-level performance on imagenet classification","author":"He","year":"2015","journal-title":"arXiv"},{"key":"B11","doi-asserted-by":"publisher","first-page":"458","DOI":"10.1016\/j.neuroimage.2008.07.013","article-title":"Tensor-based morphometry as a neuroimaging biomarker for alzheimer's disease: an MRI study of 676 AD, MCI, and normal subjects","volume":"43","author":"Hua","year":"2008","journal-title":"Neuroimage"},{"key":"B12","doi-asserted-by":"publisher","first-page":"220","DOI":"10.3389\/fnagi.2019.00220","article-title":"Deep learning in alzheimer's disease: diagnostic classification and prognostic prediction using neuroimaging data","volume":"11","author":"Jo","year":"2019","journal-title":"Front. Aging Neurosci"},{"key":"B13","article-title":"Adam: a method for stochastic optimization","author":"Kingma","year":"2014","journal-title":"arXiv [Preprint]"},{"key":"B14","unstructured":"\u201cAuto-encoding variational bayes,\u201d\n            KingmaD. P.\n            WellingM.\n          Conference Proceedings: Papers Accepted to the International Conference on Learning Representations (ICLR)2014"},{"key":"B15","doi-asserted-by":"publisher","first-page":"33","DOI":"10.1038\/s41572-021-00269-y","article-title":"Alzheimer disease","volume":"7","author":"Knopman","year":"2021","journal-title":"Nat. Rev. Dis. Prim"},{"key":"B16","doi-asserted-by":"publisher","first-page":"2645","DOI":"10.1109\/TPAMI.2022.3169217","article-title":"GradDiv: adversarial robustness of randomized neural networks via gradient diversity regularization","volume":"45","author":"Lee","year":"2023","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"B17","doi-asserted-by":"publisher","DOI":"10.1109\/TEST.2018.8624792","article-title":"Influence-directed explanations for deep convolutional networks","author":"Leino","year":"2018","journal-title":"arXiv"},{"key":"B18","doi-asserted-by":"publisher","first-page":"2443","DOI":"10.1093\/brain\/awn146","article-title":"Ventricular enlargement as a possible measure of alzheimer's disease progression validated using the alzheimer's disease neuroimaging initiative database","author":"Nestor","year":"2008","journal-title":"Brain"},{"key":"B19","doi-asserted-by":"publisher","first-page":"647","DOI":"10.3233\/JAD-2010-1406","article-title":"Brain ventricular volume and cerebrospinal fluid biomarkers of alzheimer's disease","volume":"20","author":"Ott","year":"2010","journal-title":"J. Alzheimers Dis"},{"key":"B20","doi-asserted-by":"publisher","first-page":"201","DOI":"10.1212\/WNL.0b013e3181cb3e25","article-title":"Alzheimer's disease neuroimaging initiative (ADNI): clinical characterization","volume":"74","author":"Petersen","year":"2010","journal-title":"Neurology"},{"key":"B21","doi-asserted-by":"publisher","first-page":"8680737","DOI":"10.1155\/2022\/8680737","article-title":"A CAD system for alzheimer's disease classification using neuroimaging MRI 2D slices","volume":"2022","author":"Sethi","year":"2022","journal-title":"Comput. Math. Methods Med"},{"key":"B22","doi-asserted-by":"publisher","first-page":"276","DOI":"10.1016\/j.neuroimage.2019.01.031","article-title":"A parameter-efficient deep learning approach to predict conversion from mild cognitive impairment to alzheimer's disease","volume":"189","author":"Spasov","year":"2019","journal-title":"Neuroimage"},{"key":"B23","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1703.01365","article-title":"Axiomatic attribution for deep networks","author":"Sundararajan","year":"2017","journal-title":"arXiv"},{"key":"B24","doi-asserted-by":"publisher","first-page":"1087","DOI":"10.1016\/j.mri.2016.05.001","article-title":"Shape and diffusion tensor imaging based integrative analysis of the hippocampus and the amygdala in alzheimer's disease","volume":"34","author":"Tang","year":"2016","journal-title":"Magn. Reson. Imaging"},{"key":"B25","doi-asserted-by":"publisher","first-page":"947","DOI":"10.3390\/life12070947","article-title":"A reproducible deep-learning-based computer-aided diagnosis tool for frontotemporal dementia using MONAI and clinica frameworks","volume":"12","author":"Termine","year":"2022","journal-title":"Life"},{"key":"B26","doi-asserted-by":"publisher","first-page":"42","DOI":"10.1016\/j.neurobiolaging.2019.08.032","article-title":"Biomarker-guided clustering of Alzheimer's disease clinical syndromes","volume":"83","author":"Toschi","year":"2019","journal-title":"Neurobiol. Aging"}],"container-title":["Frontiers in Neuroinformatics"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fninf.2024.1346723\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,2,6]],"date-time":"2024-02-06T04:26:22Z","timestamp":1707193582000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fninf.2024.1346723\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,2,6]]},"references-count":26,"alternative-id":["10.3389\/fninf.2024.1346723"],"URL":"https:\/\/doi.org\/10.3389\/fninf.2024.1346723","relation":{},"ISSN":["1662-5196"],"issn-type":[{"value":"1662-5196","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,2,6]]},"article-number":"1346723"}}