{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,31]],"date-time":"2026-03-31T13:20:04Z","timestamp":1774963204458,"version":"3.50.1"},"reference-count":24,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,2,20]],"date-time":"2026-02-20T00:00:00Z","timestamp":1771545600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Comput. Sci."],"abstract":"<jats:p>Contrastive learning has been attracting much interest in recent years for its ability to train without labeled data. An important factor in its success is the loss function, which guides the search for prominent features that separate the positive and negative classes. The triplet loss function is widely used in contrastive learning, in which the objective is to attract a pair of positive instances while pushing away a negative instance from the anchor instance, where one of the positive instances is often an augmented version of the anchor. To improve the performance of contrastive learning in automated Clock-Drawing Test (CDT) grading, this paper proposes a more comprehensive triplet loss function that aims to not only keep the distance between the anchor and a positive instance small and the distance between the anchor and a negative instance large, but also keep the distance between the positive and negative instances large. Experimental results show that the improved loss function significantly improves the model\u2019s accuracy, precision, recall, and F1-score by 3\u20135% on both CIFAR-10 and CDT datasets, providing a new method for improving the accuracy of automatic CDT scoring and early detection of cognitive impairments.<\/jats:p>","DOI":"10.3389\/fcomp.2026.1690044","type":"journal-article","created":{"date-parts":[[2026,2,20]],"date-time":"2026-02-20T06:41:54Z","timestamp":1771569714000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["An improved contrastive learning loss function for automated clock-drawing test grading with implications for cognitive impairment screening"],"prefix":"10.3389","volume":"8","author":[{"given":"Ning","family":"Liu","sequence":"first","affiliation":[{"name":"School of Artificial Intelligence and Information Engineering, Zhejiang University of Science and Technology","place":["Hangzhou, China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Qian","family":"Sun","sequence":"additional","affiliation":[{"name":"School of Science, Zhejiang University of Science and Technology","place":["Hangzhou, China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Xiaoyin","family":"Xu","sequence":"additional","affiliation":[{"name":"College of Biomedical Engineering and Instrument Science, Zhejiang University","place":["Hangzhou, China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Haifeng","family":"Mou","sequence":"additional","affiliation":[{"name":"School of Science, Zhejiang University of Science and Technology","place":["Hangzhou, China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Xinhai","family":"Liao","sequence":"additional","affiliation":[{"name":"School of Artificial Intelligence and Information Engineering, Zhejiang University of Science and Technology","place":["Hangzhou, China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Bokai","family":"Rong","sequence":"additional","affiliation":[{"name":"School of Artificial Intelligence and Information Engineering, Zhejiang University of Science and Technology","place":["Hangzhou, China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Lingxing","family":"Wang","sequence":"additional","affiliation":[{"name":"Department of Neurology, Second Affiliated Hospital of Fujian Medical University","place":["Quanzhou, Fujian, China"]}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2026,2,20]]},"reference":[{"key":"ref1","first-page":"1597","volume-title":"Proceedings of the 37th International Conference on Machine Learning PMLR","author":"Chen","year":"2020"},{"key":"ref2","doi-asserted-by":"publisher","first-page":"840386","DOI":"10.3389\/fnagi.2022.840386","article-title":"Prediction models for conversion from mild cognitive impairment to AD: a systematic review and meta-analysis","volume":"14","author":"Chen","year":"2022","journal-title":"Front. Aging Neurosci."},{"key":"ref3","doi-asserted-by":"publisher","first-page":"33860","DOI":"10.48550\/arXiv.2210.05144","article-title":"Why do we need large batchsizes in contrastive learning? A gradient-bias perspective","volume":"2","author":"Chen","year":"2022","journal-title":"Adv. Neural inform. Process. Syst."},{"key":"ref4","article-title":"GFE-mamba: mamba-based AD multimodal progression assessment via generative feature extraction from MCI","volume-title":"arXiv","author":"Fang","year":"2024"},{"key":"ref5","doi-asserted-by":"publisher","first-page":"108021","DOI":"10.1016\/j.compbiolchem.2024.108021","article-title":"Contrastive pre-training for automated clock-drawing test grading in cognitive assessment","volume":"110","author":"Garcia","year":"2024","journal-title":"Comput. Biol. Chem."},{"key":"ref6","first-page":"9729","author":"He","year":"2020"},{"key":"ref7","doi-asserted-by":"publisher","first-page":"199","DOI":"10.1093\/pan\/mpl013","article-title":"Matching as nonparametric preprocessing for reducing model dependence in parametric causal inference","volume":"15","author":"Ho","year":"2007","journal-title":"Polit. Anal."},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.21203\/rs.3.rs-4909790\/v1","article-title":"Using deep learning neural networks to improve dementia detection: automating coding of the clock-drawing test","author":"Hu","year":"2024","journal-title":"Res. Sq."},{"key":"ref9","first-page":"21798","volume-title":"Advances in Neural Information Processing Systems","author":"Kalantidis","year":"2020"},{"key":"ref50","first-page":"45678","article-title":"Dynamic negative sample queue for MoCo-based contrastive learning in small-sample scenarios. In A. Globerson, L. Mackey, D. Belgrave, A. Fan, U. Paquet, J. Tomczak, & C. Zhang (Eds.),","volume":"36","author":"Li","year":"2023","journal-title":"Advances in Neural Information Processing Systems,"},{"key":"ref10","doi-asserted-by":"publisher","first-page":"3407","DOI":"10.5555\/3495724.3496011","article-title":"Demystifying contrastive self-supervised learning: Invariances, augmentations and dataset biases","volume":"33","author":"Purushwalkam","year":"2020","journal-title":"Adv. Neural inform. Process. Syst."},{"key":"ref11","doi-asserted-by":"publisher","first-page":"18113","DOI":"10.1038\/s41598-023-44723-1","article-title":"Attentive pairwise interaction network for AI-assisted clock drawing test assessment of early visuospatial deficits","volume":"13","author":"Raksasat","year":"2023","journal-title":"Sci. Rep."},{"key":"ref12","article-title":"Application status of artificial intelligence technology in screening for geriatric dementia","volume-title":"Biomedical","author":"Si","year":"2024"},{"key":"ref13","article-title":"X-sample contrastive loss: improving contrastive learning with sample similarity graphs","volume-title":"arXiv","author":"Sobal","year":"2024"},{"key":"ref51","first-page":"9080","author":"Souillard-Mandar","year":"2020"},{"key":"ref14","first-page":"19511","volume-title":"Advance Neural information Process System","author":"Tian","year":"2022"},{"key":"ref15","doi-asserted-by":"publisher","first-page":"130","DOI":"10.2174\/1567205014666171121114930","article-title":"A speech recognition-based solution for the automatic detection of mild cognitive impairment from spontaneous speech","volume":"15","author":"T\u00f3th","year":"2018","journal-title":"Curr. Alzheimer Res."},{"key":"ref16","article-title":"Representation learning with contrastive predictive coding","volume-title":"arXiv","author":"van den Oord","year":"2018"},{"key":"ref17","first-page":"2495","author":"Wang","year":"2021"},{"key":"ref18","volume-title":"Dementia: Key facts","year":"2024"},{"key":"ref19","doi-asserted-by":"publisher","first-page":"23297","DOI":"10.48550\/arXiv.2310.11048","article-title":"Understanding contrastive learning via distributionally robust optimization","volume":"36","author":"Wu","year":"2023","journal-title":"Adv. Neural inform. Process. Syst."},{"key":"ref20","doi-asserted-by":"publisher","first-page":"668","DOI":"10.1007\/978-3-031-19809-0_38","article-title":"Decoupled contrastive learning","volume":"2","author":"Yeh","year":"2022","journal-title":"Eur. Conf. Comput. Vis."},{"key":"ref21","article-title":"Applications of computer vision in analysis of the clock-drawing test as a metric of cognitive impairment","volume-title":"arXiv","author":"Zhang","year":"2023"},{"key":"ref22","volume-title":"Advances in Neural Information Processing Systems","author":"Zhang","year":"2024"}],"updated-by":[{"DOI":"10.3389\/fcomp.2026.1824259","type":"correction","label":"Correction","source":"publisher","updated":{"date-parts":[[2026,3,31]],"date-time":"2026-03-31T00:00:00Z","timestamp":1774915200000}}],"container-title":["Frontiers in Computer Science"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fcomp.2026.1690044\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,31]],"date-time":"2026-03-31T12:31:00Z","timestamp":1774960260000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fcomp.2026.1690044\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,2,20]]},"references-count":24,"alternative-id":["10.3389\/fcomp.2026.1690044"],"URL":"https:\/\/doi.org\/10.3389\/fcomp.2026.1690044","relation":{},"ISSN":["2624-9898"],"issn-type":[{"value":"2624-9898","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,2,20]]},"article-number":"1690044"}}