{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,1,21]],"date-time":"2026-01-21T09:15:50Z","timestamp":1768986950861,"version":"3.49.0"},"reference-count":63,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,11,21]],"date-time":"2025-11-21T00:00:00Z","timestamp":1763683200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100020963","name":"Moonshot Research and Development Program","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100020963","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001691","name":"Japan Society for the Promotion of Science","doi-asserted-by":"publisher","award":["20H05712"],"award-info":[{"award-number":["20H05712"]}],"id":[{"id":"10.13039\/501100001691","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001691","name":"Japan Society for the Promotion of Science","doi-asserted-by":"publisher","award":["23H04834"],"award-info":[{"award-number":["23H04834"]}],"id":[{"id":"10.13039\/501100001691","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001691","name":"Japan Society for the Promotion of Science","doi-asserted-by":"publisher","award":["24KJ0798"],"award-info":[{"award-number":["24KJ0798"]}],"id":[{"id":"10.13039\/501100001691","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Comput. Neurosci."],"abstract":"<jats:p>Recent advances in self-supervised learning have attracted significant attention from both machine learning and neuroscience. This is primarily because self-supervised methods do not require annotated supervisory information, making them applicable to training artificial networks without relying on large amounts of curated data, and potentially offering insights into how the brain adapts to its environment in an unsupervised manner. Although several previous studies have elucidated the correspondence between neural representations in deep convolutional neural networks (DCNNs) and biological systems, the extent to which unsupervised or self-supervised learning can explain the human-like acquisition of categorically structured information remains less explored. In this study, we investigate the correspondence between the internal representations of DCNNs trained using a self-supervised contrastive learning algorithm and human semantics and recognition. To this end, we employ a few-shot learning evaluation procedure, which measures the ability of DCNNs to recognize novel concepts from limited exposure, to examine the inter-categorical structure of the learned representations. Two comparative approaches are used to relate the few-shot learning outcomes to human semantics and recognition, with results suggesting that the representations acquired through contrastive learning are well aligned with human cognition. These findings underscore the potential of self-supervised contrastive learning frameworks to model learning mechanisms similar to those of the human brain, particularly in scenarios where explicit supervision is unavailable, such as in human infants prior to language acquisition.<\/jats:p>","DOI":"10.3389\/fncom.2025.1613291","type":"journal-article","created":{"date-parts":[[2025,11,21]],"date-time":"2025-11-21T11:54:03Z","timestamp":1763726043000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["Exploring internal representations of self-supervised networks: few-shot learning abilities and comparison with human semantics and recognition of objects"],"prefix":"10.3389","volume":"19","author":[{"given":"Asaki","family":"Kataoka","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yoshihiro","family":"Nagano","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Masafumi","family":"Oizumi","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2025,11,21]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1902.09229","article-title":"A theoretical analysis of contrastive unsupervised representation learning","author":"Arora","year":"2019","journal-title":"arXiv"},{"key":"B2","first-page":"25164","article-title":"\u201cThe functional specialization of visual cortex emerges from training parallel pathways with self-supervised predictive learning,\u201d","author":"Bakhtiari","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"B3","unstructured":"\u201cOn the surrogate gap between contrastive and supervised losses,\u201d\n          \n          1585\n          1606\n          \n            \n              Bao\n              H.\n            \n            \n              Nagano\n              Y.\n            \n            \n              Nozawa\n              K.\n            \n          \n          PMLR\n          Proceedings of the 39th International Conference on Machine Learning\n          \n          2022"},{"key":"B4","doi-asserted-by":"publisher","first-page":"5418","DOI":"10.1038\/s41467-020-18946-z","article-title":"Capturing human categorization of natural images by combining deep networks and cognitive models","volume":"11","author":"Battleday","year":"2020","journal-title":"Nat. Commun"},{"key":"B5","doi-asserted-by":"publisher","first-page":"105","DOI":"10.1016\/0010-0277(96)00706-8","article-title":"Basic-level and superordinate-like categorical representations in early infancy","volume":"60","author":"Behl-Chadha","year":"1996","journal-title":"Cognition"},{"key":"B6","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1006\/cogp.2001.0748","article-title":"Does language shape thought?: Mandarin and english speakers' conceptions of time","volume":"43","author":"Boroditsky","year":"2001","journal-title":"Cognit. Psychol"},{"key":"B7","first-page":"56","article-title":"On the sapir-whorf hypothesis","volume":"23","author":"Brutyan","year":"1969","journal-title":"Problemy Filosofii"},{"key":"B8","article-title":"\u201cHow well do deep neural networks trained on object recognition characterize the mouse visual system?,\u201d","volume-title":"Real Neurons & Hidden Units: Future Directions at the Intersection of Neuroscience and Artificial Intelligence @ NeurIPS 2019","author":"Cadena","year":"2019"},{"key":"B9","article-title":"\u201cAcquiring a single new word,\u201d","volume-title":"Proceedings of the Stanford Child Language Conference","author":"Carey","year":"1978"},{"key":"B10","unstructured":"\u201cA simple framework for contrastive learning of visual representations,\u201d\n          \n          1597\n          1607\n          \n            \n              Chen\n              T.\n            \n            \n              Kornblith\n              S.\n            \n            \n              Norouzi\n              M.\n            \n            \n              Hinton\n              G.\n            \n          \n          PMLR\n          Proceedings of the 37th International Conference on Machine Learning\n          \n          2020"},{"key":"B11","first-page":"15750","article-title":"\u201cExploring simple Siamese representation learning,\u201d","author":"Chen","year":"2020","journal-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)"},{"key":"B12","doi-asserted-by":"publisher","first-page":"42","DOI":"10.1109\/MSP.2021.3134634","article-title":"Self-supervised representation learning: Introduction, advances and challenges","volume":"39","author":"Ericsson","year":"2022","journal-title":"IEEE Signal Process. Magaz."},{"key":"B13","doi-asserted-by":"publisher","first-page":"312","DOI":"10.1126\/science.291.5502.312","article-title":"Categorical representation of visual stimuli in the primate prefrontal cortex","volume":"291","author":"Freedman","year":"2001","journal-title":"Science"},{"key":"B14","doi-asserted-by":"publisher","first-page":"127","DOI":"10.1038\/nrn2787","article-title":"The free-energy principle: a unified brain theory?","volume":"11","author":"Friston","year":"2010","journal-title":"Nat. Rev. Neurosci"},{"key":"B15","doi-asserted-by":"publisher","first-page":"126","DOI":"10.1016\/j.pneurobio.2019.01.008","article-title":"The roles of supervised machine learning in systems neuroscience","volume":"175","author":"Glaser","year":"2019","journal-title":"Prog. Neurobiol"},{"key":"B16","article-title":"\u201cDeep residual learning for image recognition,\u201d","author":"He","year":"2015","journal-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)"},{"key":"B17","doi-asserted-by":"publisher","first-page":"1173","DOI":"10.1038\/s41562-020-00951-3","article-title":"Revealing the multidimensional mental representations of natural objects underlying human similarity judgements","volume":"4","author":"Hebart","year":"2020","journal-title":"Nat. Hum. Behav"},{"key":"B18","doi-asserted-by":"publisher","first-page":"128645","DOI":"10.1016\/j.neucom.2024.128645","article-title":"A comprehensive survey on contrastive learning","volume":"610","author":"Hu","year":"2024","journal-title":"Neurocomputing"},{"key":"B19","unstructured":"\u201cLocal plasticity rules can learn deep representations using self-supervised contrastive predictions,\u201d\n          \n          30365\n          30379\n          \n            \n              Illing\n              B.\n            \n            \n              Ventura\n              J.\n            \n            \n              Bellec\n              G.\n            \n            \n              Gerstner\n              W.\n            \n          \n          Curran Associates, Inc.\n          Advances in Neural Information Processing Systems\n          \n          2020"},{"key":"B20","doi-asserted-by":"publisher","first-page":"2","DOI":"10.3390\/technologies9010002","article-title":"A survey on contrastive self-supervised learning","volume":"9","author":"Jaiswal","year":"2020","journal-title":"Technologies"},{"key":"B21","doi-asserted-by":"crossref","first-page":"2146","DOI":"10.1109\/ICCV.2009.5459469","article-title":"\u201cWhat is the best multi-stage architecture for object recognition?,\u201d","volume-title":"2009 IEEE 12th International Conference on Computer Vision","author":"Jarrett","year":"2009"},{"key":"B22","doi-asserted-by":"publisher","first-page":"15917","DOI":"10.1038\/s41598-024-65604-1","article-title":"Gromov-wasserstein unsupervised alignment reveals structural correspondences between the color similarity structures of humans and large language models","volume":"14","author":"Kawakita","year":"2024","journal-title":"Sci. Rep"},{"key":"B23","doi-asserted-by":"publisher","first-page":"65","DOI":"10.1525\/aa.1984.86.1.02a00050","article-title":"What is the sapir-whorf hypothesis?","volume":"86","author":"Kay","year":"1984","journal-title":"Am. Anthropol"},{"key":"B24","doi-asserted-by":"publisher","first-page":"e1003915","DOI":"10.1371\/journal.pcbi.1003915","article-title":"Deep supervised, but not unsupervised, models may explain IT cortical representation","volume":"10","author":"Khaligh-Razavi","year":"2014","journal-title":"PLoS Comput. Biol"},{"key":"B25","doi-asserted-by":"publisher","first-page":"3985","DOI":"10.1523\/JNEUROSCI.14-07-03985.1994","article-title":"Supervised learning in the brain","volume":"14","author":"Knudsen","year":"1994","journal-title":"J. Neurosci"},{"key":"B26","doi-asserted-by":"publisher","DOI":"10.1101\/2021.05.28.446118","article-title":"Beyond category-supervision: instance-level contrastive learning models predict human visual system responses to objects","author":"Konkle","year":"2021","journal-title":"bioRxiv"},{"key":"B27","doi-asserted-by":"publisher","first-page":"491","DOI":"10.1038\/s41467-022-28091-4","article-title":"A self-supervised domain-general learning framework for human ventral stream representation","volume":"13","author":"Konkle","year":"2022","journal-title":"Nat. Commun"},{"key":"B28","doi-asserted-by":"publisher","first-page":"4","DOI":"10.3389\/neuro.06.004.2008","article-title":"Representational similarity analysis - connecting the branches of systems neuroscience","volume":"2","author":"Kriegeskorte","year":"2008","journal-title":"Front. Syst. Neurosci"},{"key":"B29","unstructured":"Krizhevsky\n              A.\n            \n          \n          Learning Multiple Layers of Features From Tiny Images\n          \n          2019"},{"key":"B30","article-title":"\u201cImageNet classification with deep convolutional neural networks,\u201d","author":"Krizhevsky","year":"2012","journal-title":"Advances in Neural Information Processing Systems 25 (NIPS 2012"},{"key":"B31","doi-asserted-by":"publisher","first-page":"461","DOI":"10.1007\/s13735-022-00245-6","article-title":"Contrastive self-supervised learning: review, progress, challenges and future research directions","volume":"11","author":"Kumar","year":"2022","journal-title":"Int. J. Multimed. Inf. Retr"},{"key":"B32","doi-asserted-by":"publisher","first-page":"1293","DOI":"10.1109\/TPAMI.2024.3495827","article-title":"Estimating information theoretic measures via multidimensional gaussianization","volume":"47","author":"Laparra","year":"2025","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell"},{"key":"B33","doi-asserted-by":"publisher","first-page":"2278","DOI":"10.1109\/5.726791","article-title":"Gradient-based learning applied to document recognition","volume":"86","author":"Lecun","year":"1998","journal-title":"Proc. IEEE"},{"key":"B34","doi-asserted-by":"publisher","first-page":"757","DOI":"10.1523\/JNEUROSCI.0757-20.2020","article-title":"Dissecting the roles of supervised and unsupervised learning in perceptual discrimination judgments","volume":"41","author":"Loewenstein","year":"2021","journal-title":"J. Neurosci"},{"key":"B35","unstructured":"\u201cPutting an end to end-to-end: gradient-isolated learning of representations,\u201d\n          \n          \n            \n              Lowe\n              S.\n            \n            \n              O'Connor\n              P.\n            \n            \n              Veeling\n              B. S.\n            \n          \n          \n          2019"},{"key":"B36","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-19800-7_43","article-title":"Self-supervision can be a good few-shot learner","author":"Lu","year":"2022","journal-title":"arXiv"},{"key":"B37","doi-asserted-by":"publisher","first-page":"930","DOI":"10.1016\/j.tics.2020.08.005","article-title":"Effects of language on visual perception","volume":"24","author":"Lupyan","year":"2020","journal-title":"Trends Cogn. Sci"},{"key":"B38","doi-asserted-by":"publisher","first-page":"13402","DOI":"10.1523\/JNEUROSCI.5181-14.2015","article-title":"Simple learned weighted sums of inferior temporal neuronal firing rates accurately predict human core object recognition performance","volume":"35","author":"Majaj","year":"2015","journal-title":"J. Neurosci"},{"key":"B39","doi-asserted-by":"publisher","DOI":"10.1101\/2021.03.01.433495","article-title":"Multi-scale hierarchical neural network models that bridge from single neurons in the primate primary visual cortex to object recognition behavior","author":"Marques","year":"2021","journal-title":"bioRxiv"},{"key":"B40","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2006.11325","article-title":"Self-supervised prototypical transfer learning for few-shot classification","author":"Medina","year":"2020","journal-title":"arXiv"},{"key":"B41","article-title":"\u201cToward a realisticmodel of speech processing in the brain with self-supervised learning,\u201d","author":"Millet","year":"2022","journal-title":"Advances in Neural Information Processing Systems"},{"key":"B42","doi-asserted-by":"publisher","DOI":"10.1101\/2021.06.16.448730","article-title":"Unsupervised models of mouse visual cortex","author":"Nayebi","year":"2021","journal-title":"bioRxiv"},{"key":"B43","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR42600.2020.00737","article-title":"\u201cHow useful is self-supervised pretraining for visual tasks?,\u201d","volume-title":"2020 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Newell","year":"2020"},{"key":"B44","unstructured":"\u201cUnderstanding negative samples in instance discriminative self-supervised representation learning,\u201d\n          \n          5784\n          5797\n          \n            \n              Nozawa\n              K.\n            \n            \n              Sato\n              I.\n            \n          \n          Curran Associates, Inc.\n          Advances in Neural Information Processing Systems\n          \n          2021"},{"key":"B45","doi-asserted-by":"publisher","first-page":"1191","DOI":"10.1162\/089976603321780272","article-title":"Estimation of entropy and mutual information","volume":"15","author":"Paninski","year":"2003","journal-title":"Neural Comput"},{"key":"B46","doi-asserted-by":"publisher","first-page":"eadl1776","DOI":"10.1126\/sciadv.adl1776","article-title":"Contrastive learning explains the emergence and function of visual category-selective regions","volume":"10","author":"Prince","year":"2024","journal-title":"Sci. Adv"},{"key":"B47","doi-asserted-by":"publisher","first-page":"463","DOI":"10.1068\/p220463","article-title":"Evidence for representations of perceptually similar natural categories by 3-month-old and 4-month-old infants","volume":"22","author":"Quinn","year":"1993","journal-title":"Perception"},{"key":"B48","doi-asserted-by":"publisher","first-page":"7","DOI":"10.1016\/j.visres.2018.03.010","article-title":"Color encoding in biologically-inspired convolutional neural networks","volume":"151","author":"Rafegas","year":"2018","journal-title":"Vision Res"},{"key":"B49","doi-asserted-by":"publisher","first-page":"7255","DOI":"10.1523\/JNEUROSCI.0388-18.2018","article-title":"Large-scale, high-resolution comparison of the core visual object recognition behavior of humans, monkeys, and state-of-the-art deep artificial neural networks","volume":"38","author":"Rajalingham","year":"2018","journal-title":"J. Neurosci"},{"key":"B50","doi-asserted-by":"publisher","first-page":"79","DOI":"10.1038\/4580","article-title":"Predictive coding in the visual cortex: a functional interpretation of some extra-classical receptive-field effects","volume":"2","author":"Rao","year":"1999","journal-title":"Nat. Neurosci"},{"key":"B51","doi-asserted-by":"publisher","first-page":"533","DOI":"10.1038\/323533a0","article-title":"Learning representations by back-propagating errors","volume":"323","author":"Rumelhart","year":"1986","journal-title":"Nature"},{"key":"B52","first-page":"1","article-title":"\u201cSelf-supervised pre-training for time series classification,\u201d","volume-title":"2021 International Joint Conference on Neural Networks (IJCNN)","author":"Shi","year":"2021"},{"key":"B53","doi-asserted-by":"publisher","first-page":"13","DOI":"10.1111\/1467-9280.00403","article-title":"Object name learning provides on-the-job training for attention","volume":"13","author":"Smith","year":"2002","journal-title":"Psychol. Sci"},{"key":"B54","doi-asserted-by":"publisher","first-page":"e2200800119","DOI":"10.1073\/pnas.2200800119","article-title":"Neural representational geometry underlies few-shot concept learning","volume":"119","author":"Sorscher","year":"2022","journal-title":"Proc. Natl. Acad. Sci. USA"},{"key":"B55","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1807.03748","article-title":"Representation learning with contrastive predictive coding","author":"van den Oord","year":"2018","journal-title":"arXiv"},{"key":"B56","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR46437.2021.00304","article-title":"\u201cDense contrastive learning for self-supervised visual pre-training,\u201d","volume-title":"2021 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Wang","year":"2021"},{"key":"B57","volume-title":"SLanguage, Thought, and Reality: Selected Writings of Benjamin Lee Whorf","author":"Whorf","year":"2012"},{"key":"B58","article-title":"\u201cHierarchical modular optimization of convolutional networks achieves representations similar to macaque it and human ventral stream,\u201d","author":"Yamins","year":"2013","journal-title":"Advances in Neural Information Processing Systems"},{"key":"B59","doi-asserted-by":"publisher","first-page":"356","DOI":"10.1038\/nn.4244","article-title":"Using goal-driven deep learning models to understand sensory cortex","volume":"19","author":"Yamins","year":"2016","journal-title":"Nat. Neurosci"},{"key":"B60","doi-asserted-by":"publisher","first-page":"8619","DOI":"10.1073\/pnas.1403112111","article-title":"Performance-optimized hierarchical models predict neural responses in higher visual cortex","volume":"111","author":"Yamins","year":"2014","journal-title":"Proc. Natl. Acad. Sci. USA"},{"key":"B61","doi-asserted-by":"publisher","first-page":"2370","DOI":"10.1073\/pnas.1512044113","article-title":"Cortical response to categorical color perception in infants investigated by near-infrared spectroscopy","volume":"113","author":"Yang","year":"2016","journal-title":"Proc. Natl. Acad. Sci. USA"},{"key":"B62","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-19781-9_12","article-title":"PASS: Part-aware self-supervised pre-training for person re-identification","author":"Zhu","year":"2022","journal-title":"arXiv"},{"key":"B63","doi-asserted-by":"publisher","first-page":"e2014196118","DOI":"10.1073\/pnas.2014196118","article-title":"Unsupervised neural network models of the ventral visual stream","volume":"118","author":"Zhuang","year":"2021","journal-title":"Proc. Natl. Acad. Sci. USA"}],"container-title":["Frontiers in Computational Neuroscience"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fncom.2025.1613291\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,11,21]],"date-time":"2025-11-21T11:54:08Z","timestamp":1763726048000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fncom.2025.1613291\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11,21]]},"references-count":63,"alternative-id":["10.3389\/fncom.2025.1613291"],"URL":"https:\/\/doi.org\/10.3389\/fncom.2025.1613291","relation":{},"ISSN":["1662-5188"],"issn-type":[{"value":"1662-5188","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,11,21]]},"article-number":"1613291"}}