{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,23]],"date-time":"2026-10-23T06:40:41Z","timestamp":1792737641506,"version":"4.5.0"},"reference-count":62,"publisher":"Wiley","issue":"11","license":[{"start":{"date-parts":[[2026,10,22]],"date-time":"2026-10-22T00:00:00Z","timestamp":1792627200000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"},{"start":{"date-parts":[[2026,10,22]],"date-time":"2026-10-22T00:00:00Z","timestamp":1792627200000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Medical Physics"],"published-print":{"date-parts":[[2026,11]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Artificial intelligence (AI) has achieved notable success in closed\u2010set classification of neurodegenerative diseases, but its performance degrades significantly in more realistic open\u2010set settings, such as discovering rare or novel disease subtypes from unlabeled data that are not seen during\u00a0training.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>This study introduces, to our knowledge, the first framework for generalized category discovery (GCD) in neurodegenerative diseases using MRI. It aims to leverage labeled data from known categories to automatically identify novel categories within unlabeled\u00a0datasets.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>We propose a GCD framework integrating semi\u2010supervised contrastive learning with global\u2010local representation learning to jointly optimize feature learning. Globally, we enforce consistency between weakly augmented views of the same instance by aligning their similarity relationships with other embeddings in a support queue. Locally, we treat neighborhood embeddings from the support queue as positive pairs for cross\u2010instance contrastive learning, enhancing fine\u2010grained discrimination. This dual strategy results in a unified learning objective that is applicable to both labeled and unlabeled data, effectively narrowing the representation gap among semantically similar\u00a0instances.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>We evaluate our method on a benchmark dataset comprising six neurodegenerative diseases, with over 1500 patient samples. Experimental results demonstrate that our framework consistently outperforms representative state\u2010of\u2010the\u2010art methods. Specifically, our method achieves 66.02% Accuracy and 60.59% Normalized Mutual Information on all categories, and 54.53% Accuracy and 16.67% Normalized Mutual Information on novel categories, surpassing existing state\u2010of\u2010the\u2010art\u00a0methods.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>Our framework underscores the model's robustness and its practical value in discovering new neurodegenerative disease conditions, highlighting its potential for advancing open\u2010set diagnostic\u00a0AI.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1002\/mp.70686","type":"journal-article","created":{"date-parts":[[2026,10,22]],"date-time":"2026-10-22T13:02:33Z","timestamp":1792674153000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Learning global and local representations for generalized category discovery in neurodegenerative diseases"],"prefix":"10.1002","volume":"53","author":[{"given":"Yu","family":"Shen","sequence":"first","affiliation":[{"name":"Henan Key Laboratory of Imaging and Intelligent Processing Information Engineering University Zhengzhou Henan China"},{"name":"Henan Provincial People's Hospital Zhengzhou Henan China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xiangli","family":"Yang","sequence":"additional","affiliation":[{"name":"Henan Key Laboratory of Imaging and Intelligent Processing Information Engineering University Zhengzhou Henan China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yixiang","family":"Hu","sequence":"additional","affiliation":[{"name":"Henan Key Laboratory of Imaging and Intelligent Processing Information Engineering University Zhengzhou Henan China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Kai","family":"Qiao","sequence":"additional","affiliation":[{"name":"Henan Key Laboratory of Imaging and Intelligent Processing Information Engineering University Zhengzhou Henan China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jinjin","family":"Hai","sequence":"additional","affiliation":[{"name":"Henan Key Laboratory of Imaging and Intelligent Processing Information Engineering University Zhengzhou Henan China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jian","family":"Chen","sequence":"additional","affiliation":[{"name":"Henan Key Laboratory of Imaging and Intelligent Processing Information Engineering University Zhengzhou Henan China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Bin","family":"Yan","sequence":"additional","affiliation":[{"name":"Henan Key Laboratory of Imaging and Intelligent Processing Information Engineering University Zhengzhou Henan China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2026,10,22]]},"reference":[{"key":"e_1_2_12_2_1","doi-asserted-by":"publisher","DOI":"10.1212\/WNL.0000000000209220"},{"key":"e_1_2_12_3_1","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2022.3166129"},{"key":"e_1_2_12_4_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2025.3555485"},{"key":"e_1_2_12_5_1","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2024.3485612"},{"key":"e_1_2_12_6_1","doi-asserted-by":"publisher","DOI":"10.1109\/TFUZZ.2021.3052461"},{"key":"e_1_2_12_7_1","doi-asserted-by":"publisher","DOI":"10.1109\/JBHI.2022.3161466"},{"key":"e_1_2_12_8_1","doi-asserted-by":"publisher","DOI":"10.1109\/TKDE.2022.3220219"},{"key":"e_1_2_12_9_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2022.3201576"},{"key":"e_1_2_12_10_1","unstructured":"WangY ChenH HengQ et\u00a0al.FreeMatch: self\u2010adaptive thresholding for semi\u2010supervised learning. arXiv. Preprint posted online May 15 2022. doi:10.48550\/arXiv.2205.07246"},{"key":"e_1_2_12_11_1","unstructured":"ChenH TaoR FanY et\u00a0al.SoftMatch: addressing the quantity\u2010quality trade\u2010off in semi\u2010supervised learning. arXiv. Preprint posted online January 26 2023. doi:10.48550\/arXiv.2301.10921"},{"key":"e_1_2_12_12_1","doi-asserted-by":"publisher","DOI":"10.1109\/JBHI.2025.3556984"},{"key":"e_1_2_12_13_1","doi-asserted-by":"crossref","unstructured":"VazeS HanK VedaldiA ZissermanA.Generalized category discovery. In:CVPR.IEEE;2022:7482\u20107491.","DOI":"10.1109\/CVPR52688.2022.00734"},{"key":"e_1_2_12_14_1","unstructured":"DosovitskiyA BeyerL KolesnikovA et\u00a0al.An image is worth 16x16 words: transformers for image recognition at scale. In:ICLR. OpenReview.net;2021."},{"key":"e_1_2_12_15_1","doi-asserted-by":"crossref","unstructured":"CaronM TouvronH MisraI et\u00a0al.Emerging properties in self\u2010supervised vision transformers. In:ICCV.IEEE;2021:9630\u20109640.","DOI":"10.1109\/ICCV48922.2021.00951"},{"key":"e_1_2_12_16_1","unstructured":"KhoslaP TeterwakP WangC et\u00a0al.Supervised contrastive learning. arXiv. Preprint posted online April 23 2023. doi:10.48550\/arXiv.2004.11362"},{"key":"e_1_2_12_17_1","unstructured":"ChenT KornblithS NorouziM HintonGE.A simple framework for contrastive learning of visual representations. In:ICML vol.119Proceedings of Machine Learning Research.PMLR;2020:1597\u20101607."},{"key":"e_1_2_12_18_1","doi-asserted-by":"crossref","unstructured":"ZhengM WangF YouS et\u00a0al.Weakly supervised contrastive learning. In:ICCV.IEEE;2021:10022\u201010031.","DOI":"10.1109\/ICCV48922.2021.00989"},{"key":"e_1_2_12_19_1","unstructured":"ArthurD VassilvitskiiS.k\u2010means++: the advantages of careful seeding. In:SODA.SIAM;2007:1027\u20101035."},{"key":"e_1_2_12_20_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.nicl.2023.103405"},{"key":"e_1_2_12_21_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.bspc.2023.104904"},{"key":"e_1_2_12_22_1","doi-asserted-by":"crossref","unstructured":"DwibediD AytarY TompsonJ SermanetP ZissermanA.With a little help from my friends: nearest\u2010neighbor contrastive learning of visual representations. In:ICCV.IEEE;2021:9568\u20139577.","DOI":"10.1109\/ICCV48922.2021.00945"},{"key":"e_1_2_12_23_1","unstructured":"HsuY LvZ SchlosserJ OdomP KiraZ.Multi\u2010class classification without multi\u2010class labels. In:ICLR (Poster) OpenReview.net;2019."},{"key":"e_1_2_12_24_1","doi-asserted-by":"crossref","unstructured":"HanK VedaldiA ZissermanA.Learning to discover novel visual categories via deep transfer clustering. In:ICCV.IEEE;2019:8400\u20108408.","DOI":"10.1109\/ICCV.2019.00849"},{"key":"e_1_2_12_25_1","unstructured":"HanK RebuffiS EhrhardtS VedaldiA ZissermanA. Automatically discovering and learning new visual categories with ranking statistics. In:ICLR. OpenReview.net; 2020."},{"key":"e_1_2_12_26_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2021.3091944"},{"key":"e_1_2_12_27_1","unstructured":"ZhaoB HanK.Novel visual category discovery with dual ranking statistics and mutual knowledge distillation. In:NeurIPS.2021:22982\u201022994."},{"key":"e_1_2_12_28_1","doi-asserted-by":"crossref","unstructured":"ZhongZ ZhuL LuoZ LiS YangY SebeN.OpenMix: reviving known knowledge for discovering novel visual categories in an open world. In:CVPR.Computer Vision Foundation \/ IEEE;2021:9462\u20109470.","DOI":"10.1109\/CVPR46437.2021.00934"},{"key":"e_1_2_12_29_1","doi-asserted-by":"crossref","unstructured":"ZhongZ FiniE RoyS LuoZ RicciE SebeN.Neighborhood contrastive learning for novel class discovery. In:CVPR.Computer Vision Foundation \/ IEEE;2021:10867\u201010875.","DOI":"10.1109\/CVPR46437.2021.01072"},{"key":"e_1_2_12_30_1","doi-asserted-by":"crossref","unstructured":"FiniE SanginetoE Lathuili\u00e8re S ZhongZ NabiM RicciE.A unified objective for novel class discovery. In:ICCV.IEEE;2021:9264\u20109272.","DOI":"10.1109\/ICCV48922.2021.00915"},{"key":"e_1_2_12_31_1","doi-asserted-by":"crossref","unstructured":"YangM ZhuY YuJ WuA DengC.Divide and conquer: compositional experts for generalized novel class discovery. In:CVPR.IEEE;2022:14248\u201014257.","DOI":"10.1109\/CVPR52688.2022.01387"},{"key":"e_1_2_12_32_1","doi-asserted-by":"crossref","unstructured":"ZhangX JiangJ FengY et\u00a0al.Grow and merge: a unified framework for continuous categories discovery. In:NeurIPS.2022.","DOI":"10.52202\/068431-1991"},{"key":"e_1_2_12_33_1","unstructured":"CaoK BrbicM LeskovecJ.Open\u2010world semi\u2010supervised learning. In:ICLR. OpenReview.net;2022."},{"key":"e_1_2_12_34_1","doi-asserted-by":"crossref","unstructured":"ZhaoB WenX HanK.Learning semi\u2010supervised Gaussian mixture models for generalized category discovery.arXiv. Preprint posted online May 10 2023. doi:10.48550\/arXiv.2305.06144","DOI":"10.1109\/ICCV51070.2023.01524"},{"key":"e_1_2_12_35_1","unstructured":"AnW TianF ZhengQ DingW WangQ ChenP.Generalized category discovery with decoupled prototypical network.arXiv. Preprint posted online November 28 2022. doi:10.48550\/10.48550\/arXiv.2211.15115"},{"key":"e_1_2_12_36_1","doi-asserted-by":"crossref","unstructured":"BeyerL ZhaiX OliverA KolesnikovA.S4L: Self\u2010supervised semi\u2010supervised learning. In:ICCV.IEEE;2019:1476\u20101485.","DOI":"10.1109\/ICCV.2019.00156"},{"key":"e_1_2_12_37_1","unstructured":"ChenT KornblithS SwerskyK NorouziM HintonGE.Big self\u2010supervised models are strong semi\u2010supervised learners. In:NeurIPS.2020."},{"key":"e_1_2_12_38_1","unstructured":"SajjadiM JavanmardiM TasdizenT.Regularization with stochastic transformations and perturbations for deep semi\u2010supervised learning. In:NIPS.2016:1163\u20101171."},{"key":"e_1_2_12_39_1","unstructured":"LaineS AilaT.Temporal ensembling for semi\u2010supervised learning. In:ICLR (Poster) OpenReview.net;2017."},{"key":"e_1_2_12_40_1","unstructured":"TarvainenA ValpolaH.Mean teachers are better role models: weight\u2010averaged consistency targets improve semi\u2010supervised deep learning results. In:ICLR (Workshop) OpenReview.net 2017."},{"key":"e_1_2_12_41_1","unstructured":"BerthelotD CarliniN Goodfellow IJ PapernotN OliverA RaffelC.MixMatch: a holistic approach to semi\u2010supervised learning. In:NeurIPS.2019:5050\u20105060."},{"key":"e_1_2_12_42_1","unstructured":"XieQ DaiZ Hovy EH LuongT LeQ.Unsupervised data augmentation for consistency training. In:NeurIPS.2020."},{"key":"e_1_2_12_43_1","doi-asserted-by":"crossref","unstructured":"Cubuk ED ZophB ShlensJ LeQV.Randaugment: practical automated data augmentation with a reduced search space. In:CVPR Workshops.Computer Vision Foundation \/ IEEE;2020:3008\u20133017","DOI":"10.1109\/CVPRW50498.2020.00359"},{"key":"e_1_2_12_44_1","unstructured":"BerthelotD CarliniN CubukED et\u00a0al.ReMixMatch: semi\u2010supervised learning with distribution matching and augmentation anchoring. In:ICLR OpenReview.net;2020."},{"key":"e_1_2_12_45_1","unstructured":"SohnK BerthelotD CarliniN et\u00a0al.FixMatch: simplifying semi\u2010supervised learning with consistency and confidence. In:NeurIPS.2020."},{"key":"e_1_2_12_46_1","doi-asserted-by":"crossref","unstructured":"ZhengM YouS HuangL WangF QianC XuC.SimMatch: semi\u2010supervised learning with similarity matching. In:CVPR.IEEE;2022:14451\u201014461.","DOI":"10.1109\/CVPR52688.2022.01407"},{"key":"e_1_2_12_47_1","doi-asserted-by":"crossref","unstructured":"WuZ XiongY Yu SX LinD.Unsupervised feature learning via non\u2010parametric instance discrimination. In:CVPR.Computer Vision Foundation \/ IEEE Computer Society;2018:3733\u20103742.","DOI":"10.1109\/CVPR.2018.00393"},{"key":"e_1_2_12_48_1","doi-asserted-by":"crossref","unstructured":"HeK FanH WuY XieS R. B.Girshick.Momentum contrast for unsupervised visual representation learning. In:CVPR.Computer Vision Foundation \/ IEEE;2020:9726\u20109735.","DOI":"10.1109\/CVPR42600.2020.00975"},{"key":"e_1_2_12_49_1","unstructured":"GrillJ StrubF Altch\u00e9F et\u00a0al.Bootstrap your own latent \u2010 a new approach to self\u2010supervised learning. In:NeurIPS.2020."},{"key":"e_1_2_12_50_1","doi-asserted-by":"crossref","unstructured":"ChenX HeK.Exploring simple Siamese representation learning. In:CVPR.Computer Vision Foundation \/ IEEE;2021:15750\u201015758.","DOI":"10.1109\/CVPR46437.2021.01549"},{"key":"e_1_2_12_51_1","unstructured":"TianY ChenX GanguliS.Understanding self\u2010supervised learning dynamics without contrastive pairs. In:ICML vol139 Proceedings of Machine Learning Research.PMLR;2021:10268\u201010278."},{"key":"e_1_2_12_52_1","doi-asserted-by":"crossref","unstructured":"CaronM BojanowskiP JoulinA DouzeM.Deep clustering for unsupervised learning of visual features. In:ECCV (14) vol.11218 Lecture Notes in Computer Science.Springer;2018:139\u2010156.","DOI":"10.1007\/978-3-030-01264-9_9"},{"key":"e_1_2_12_53_1","unstructured":"AsanoYM RupprechtC VedaldiA.Self\u2010labelling via simultaneous clustering and representation learning. In:ICLR OpenReview.net;2020."},{"key":"e_1_2_12_54_1","doi-asserted-by":"crossref","unstructured":"JiJ VedaldiA HenriquesJF.Invariant information clustering for unsupervised image classification and segmentation. In:ICCV.IEEE;2019:9864\u20109873.","DOI":"10.1109\/ICCV.2019.00996"},{"key":"e_1_2_12_55_1","unstructured":"CaronM MisraI MairalJ GoyalP BojanowskiP JoulinA.Unsupervised learning of visual features by contrasting cluster assignments. In:NeurIPS.2020."},{"key":"e_1_2_12_56_1","unstructured":"VazeS HanK VedaldiA ZissermanA.Open\u2010set recognition: a good closed\u2010set classifier is all you need. In:ICLR. OpenReview.net;2022."},{"key":"e_1_2_12_57_1","unstructured":"TianY SunC PooleB KrishnanD SchmidC IsolaP.What makes for good views for contrastive learning?. In:NeurIPS 2020."},{"key":"e_1_2_12_58_1","first-page":"5549","article-title":"Contrastive learning with stronger augmentations","volume":"45","author":"Wang X","year":"2023","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"key":"e_1_2_12_59_1","doi-asserted-by":"publisher","DOI":"10.1111\/cns.13766"},{"key":"e_1_2_12_60_1","doi-asserted-by":"crossref","unstructured":"KuhnHW.The Hungarian method for the assignment problem. In:50 Years of Integer Programming.Springer;2010:29\u201047.","DOI":"10.1007\/978-3-540-68279-0_2"},{"key":"e_1_2_12_61_1","unstructured":"LoshchilovI HutterF.Decoupled weight decay regularization. In:ICLR (Poster). OpenReview.net;2019."},{"key":"e_1_2_12_62_1","doi-asserted-by":"crossref","unstructured":"PuN ZhongZ SebeN.Dynamic conceptional contrastive learning for generalized category discovery.CVPR;2023:7579\u20107588.","DOI":"10.1109\/CVPR52729.2023.00732"},{"key":"e_1_2_12_63_1","doi-asserted-by":"crossref","unstructured":"ChoiS KangD ChoM.Contrastive mean\u2010shift learning for generalized category discovery.CVPR;2024:23094\u201023104.","DOI":"10.1109\/CVPR52733.2024.02179"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.70686","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/mp.70686","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.70686","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,23]],"date-time":"2026-10-23T06:24:58Z","timestamp":1792736698000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.70686"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10,22]]},"references-count":62,"journal-issue":{"issue":"11","published-print":{"date-parts":[[2026,11]]}},"alternative-id":["10.1002\/mp.70686"],"URL":"https:\/\/doi.org\/10.1002\/mp.70686","archive":["Portico"],"relation":{},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,22]]},"assertion":[{"value":"2025-12-05","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-07-24","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-10-22","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e70686"}}