{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,2]],"date-time":"2026-04-02T02:17:46Z","timestamp":1775096266990,"version":"3.50.1"},"update-to":[{"DOI":"10.1371\/journal.pcbi.1010716","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2022,12,8]],"date-time":"2022-12-08T00:00:00Z","timestamp":1670457600000}}],"reference-count":68,"publisher":"Public Library of Science (PLoS)","issue":"11","license":[{"start":{"date-parts":[[2022,11,28]],"date-time":"2022-11-28T00:00:00Z","timestamp":1669593600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100017008","name":"Allen Institute","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100017008","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.ploscompbiol.org"],"crossmark-restriction":false},"short-container-title":["PLoS Comput Biol"],"abstract":"<jats:p>Neurons in sensory areas encode\/represent stimuli. Surprisingly, recent studies have suggested that, even during persistent performance, these representations are not stable and change over the course of days and weeks. We examine stimulus representations from fluorescence recordings across hundreds of neurons in the visual cortex using in vivo two-photon calcium imaging and we corroborate previous studies finding that such representations change as experimental trials are repeated across days. This phenomenon has been termed \u201crepresentational drift\u201d. In this study we geometrically characterize the properties of representational drift in the primary visual cortex of mice in two open datasets from the Allen Institute and propose a potential mechanism behind such drift. We observe representational drift both for passively presented stimuli, as well as for stimuli which are behaviorally relevant. Across experiments, the drift differs from in-session variance and most often occurs along directions that have the most in-class variance, leading to a significant turnover in the neurons used for a given representation. Interestingly, despite this significant change due to drift, linear classifiers trained to distinguish neuronal representations show little to no degradation in performance across days. The features we observe in the neural data are similar to properties of artificial neural networks where representations are updated by continual learning in the presence of dropout, i.e. a random masking of nodes\/weights, but not other types of noise. Therefore, we conclude that a potential reason for the representational drift in biological networks is driven by an underlying dropout-like noise while continuously learning and that such a mechanism may be computational advantageous for the brain in the same way it is for artificial neural networks, e.g. preventing overfitting.<\/jats:p>","DOI":"10.1371\/journal.pcbi.1010716","type":"journal-article","created":{"date-parts":[[2022,11,28]],"date-time":"2022-11-28T13:53:39Z","timestamp":1669643619000},"page":"e1010716","update-policy":"https:\/\/doi.org\/10.1371\/journal.pcbi.corrections_policy","source":"Crossref","is-referenced-by-count":30,"title":["The geometry of representational drift in natural and artificial neural networks"],"prefix":"10.1371","volume":"18","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-0207-5885","authenticated-orcid":true,"given":"Kyle","family":"Aitken","sequence":"first","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5271-2291","authenticated-orcid":true,"given":"Marina","family":"Garrett","sequence":"additional","affiliation":[]},{"given":"Shawn","family":"Olsen","sequence":"additional","affiliation":[]},{"given":"Stefan","family":"Mihalas","sequence":"additional","affiliation":[]}],"member":"340","published-online":{"date-parts":[[2022,11,28]]},"reference":[{"issue":"9","key":"pcbi.1010716.ref001","doi-asserted-by":"crossref","first-page":"647","DOI":"10.1038\/nrn2699","article-title":"Experience-dependent structural synaptic plasticity in the mammalian brain","volume":"10","author":"A Holtmaat","year":"2009","journal-title":"Nature Reviews Neuroscience"},{"key":"pcbi.1010716.ref002","doi-asserted-by":"crossref","first-page":"7","DOI":"10.1016\/j.conb.2017.06.006","article-title":"Intrinsic volatility of synaptic connections\u2014a challenge to the synaptic trace theory of memory","volume":"46","author":"G Mongillo","year":"2017","journal-title":"Current opinion in neurobiology"},{"issue":"6","key":"pcbi.1010716.ref003","doi-asserted-by":"crossref","first-page":"877","DOI":"10.1016\/j.neuron.2006.02.018","article-title":"Axons and synaptic boutons are highly dynamic in adult visual cortex","volume":"49","author":"DD Stettler","year":"2006","journal-title":"Neuron"},{"issue":"2","key":"pcbi.1010716.ref004","first-page":"275","article-title":"Chaos and the new science of the brain","volume":"1","author":"CA Skarda","year":"1990","journal-title":"Concepts in neuroscience"},{"key":"pcbi.1010716.ref005","doi-asserted-by":"crossref","DOI":"10.1017\/S0140525X19000049","article-title":"Is coding a relevant metaphor for the brain?","volume":"42","author":"R Brette","year":"2019","journal-title":"Behavioral and Brain Sciences"},{"issue":"3","key":"pcbi.1010716.ref006","doi-asserted-by":"crossref","first-page":"264","DOI":"10.1038\/nn.3329","article-title":"Long-term dynamics of CA1 hippocampal place codes","volume":"16","author":"Y Ziv","year":"2013","journal-title":"Nature neuroscience"},{"key":"pcbi.1010716.ref007","doi-asserted-by":"crossref","first-page":"e12247","DOI":"10.7554\/eLife.12247","article-title":"Hippocampal ensemble dynamics timestamp events in long-term memory","volume":"4","author":"A Rubin","year":"2015","journal-title":"Elife"},{"issue":"5","key":"pcbi.1010716.ref008","doi-asserted-by":"crossref","first-page":"986","DOI":"10.1016\/j.cell.2017.07.021","article-title":"Dynamic reorganization of neuronal activity patterns in parietal cortex","volume":"170","author":"LN Driscoll","year":"2017","journal-title":"Cell"},{"key":"pcbi.1010716.ref009","first-page":"1","article-title":"Representational drift in primary olfactory cortex","author":"CE Schoonover","year":"2021","journal-title":"Nature"},{"issue":"1","key":"pcbi.1010716.ref010","first-page":"1","article-title":"Stimulus-dependent representational drift in primary visual cortex","volume":"12","author":"TD Marks","year":"2021","journal-title":"Nature communications"},{"key":"pcbi.1010716.ref011","article-title":"Representational drift in the mouse visual cortex","author":"D Deitch","year":"2021","journal-title":"Current Biology"},{"issue":"1","key":"pcbi.1010716.ref012","first-page":"1","article-title":"Stable representation of a naturalistic movie emerges from episodic activity with gain variability","volume":"12","author":"J Xia","year":"2021","journal-title":"Nature communications"},{"key":"pcbi.1010716.ref013","doi-asserted-by":"crossref","first-page":"e27702","DOI":"10.7554\/eLife.27702","article-title":"Automated long-term recording and analysis of neural activity in behaving animals","volume":"6","author":"AK Dhawale","year":"2017","journal-title":"Elife"},{"issue":"6","key":"pcbi.1010716.ref014","doi-asserted-by":"crossref","first-page":"1133","DOI":"10.1016\/j.neuron.2018.05.017","article-title":"Stable sequential activity underlying the maintenance of a precisely executed skilled behavior","volume":"98","author":"KA Katlowitz","year":"2018","journal-title":"Neuron"},{"issue":"7904","key":"pcbi.1010716.ref015","doi-asserted-by":"crossref","first-page":"98","DOI":"10.1038\/s41586-022-04560-0","article-title":"A stable hippocampal code in freely flying bats","volume":"604","author":"WA Liberti","year":"2022","journal-title":"Nature"},{"key":"pcbi.1010716.ref016","doi-asserted-by":"crossref","first-page":"156","DOI":"10.1016\/j.conb.2014.10.012","article-title":"Robust circuit rhythms in small circuits arise from variable circuit components and mechanisms","volume":"31","author":"E Marder","year":"2015","journal-title":"Current opinion in neurobiology"},{"issue":"3","key":"pcbi.1010716.ref017","doi-asserted-by":"crossref","first-page":"286","DOI":"10.1038\/nn1202","article-title":"Off-line replay maintains declarative memories in a model of hippocampal-neocortical interactions","volume":"7","author":"S K\u00e1li","year":"2004","journal-title":"Nature neuroscience"},{"key":"pcbi.1010716.ref018","doi-asserted-by":"crossref","first-page":"e63550","DOI":"10.7554\/eLife.63550","article-title":"The brain in motion: How ensemble fluidity drives memory-updating and flexibility","volume":"9","author":"W Mau","year":"2020","journal-title":"Elife"},{"key":"pcbi.1010716.ref019","unstructured":"Hinton GE, Srivastava N, Krizhevsky A, Sutskever I, Salakhutdinov RR. Improving neural networks by preventing co-adaptation of feature detectors. arXiv preprint arXiv:12070580. 2012;."},{"issue":"7562","key":"pcbi.1010716.ref020","doi-asserted-by":"crossref","first-page":"592","DOI":"10.1038\/nature14467","article-title":"Impermanence of dendritic spines in live adult CA1 hippocampus","volume":"523","author":"A Attardo","year":"2015","journal-title":"Nature"},{"issue":"3","key":"pcbi.1010716.ref021","doi-asserted-by":"crossref","first-page":"640","DOI":"10.1016\/j.celrep.2018.09.064","article-title":"Long-term consolidation of ensemble neural plasticity patterns in hippocampal area CA1","volume":"25","author":"A Attardo","year":"2018","journal-title":"Cell reports"},{"issue":"7392","key":"pcbi.1010716.ref022","doi-asserted-by":"crossref","first-page":"62","DOI":"10.1038\/nature10918","article-title":"Choice-specific sequences in parietal cortex during a virtual-navigation decision task","volume":"484","author":"CD Harvey","year":"2012","journal-title":"Nature"},{"key":"pcbi.1010716.ref023","doi-asserted-by":"crossref","first-page":"172","DOI":"10.1016\/j.neuroscience.2017.06.005","article-title":"A stable brain from unstable components: Emerging concepts and implications for neural computation","volume":"357","author":"AR Chambers","year":"2017","journal-title":"Neuroscience"},{"key":"pcbi.1010716.ref024","doi-asserted-by":"crossref","first-page":"141","DOI":"10.1016\/j.conb.2019.08.005","article-title":"Causes and consequences of representational drift","volume":"58","author":"ME Rule","year":"2019","journal-title":"Current opinion in neurobiology"},{"key":"pcbi.1010716.ref025","article-title":"Contribution of behavioural variability to representational drift","author":"S Sadeh","year":"2022","journal-title":"bioRxiv"},{"issue":"1715","key":"pcbi.1010716.ref026","doi-asserted-by":"crossref","first-page":"20160161","DOI":"10.1098\/rstb.2016.0161","article-title":"Variance and invariance of neuronal long-term representations","volume":"372","author":"C Clopath","year":"2017","journal-title":"Philosophical Transactions of the Royal Society B: Biological Sciences"},{"key":"pcbi.1010716.ref027","article-title":"Drifting assemblies for persistent memory","author":"FYK Kossio","year":"2020","journal-title":"bioRxiv"},{"key":"pcbi.1010716.ref028","doi-asserted-by":"crossref","first-page":"e51121","DOI":"10.7554\/eLife.51121","article-title":"Stable task information from an unstable neural population","volume":"9","author":"ME Rule","year":"2020","journal-title":"Elife"},{"key":"pcbi.1010716.ref029","article-title":"Coordinated drift of receptive fields during noisy representation learning","author":"S Qin","year":"2021","journal-title":"bioRxiv"},{"key":"pcbi.1010716.ref030","doi-asserted-by":"crossref","unstructured":"Chung S, Abbott L. Neural population geometry: An approach for understanding biological and artificial neural networks. arXiv preprint arXiv:210407059. 2021;.","DOI":"10.1016\/j.conb.2021.10.010"},{"key":"pcbi.1010716.ref031","doi-asserted-by":"crossref","unstructured":"Kriegeskorte N, Wei XX. Neural tuning and representational geometry. arXiv preprint arXiv:210409743. 2021;.","DOI":"10.1038\/s41583-021-00502-3"},{"issue":"1","key":"pcbi.1010716.ref032","doi-asserted-by":"crossref","first-page":"138","DOI":"10.1038\/s41593-019-0550-9","article-title":"A large-scale standardized physiological survey reveals functional organization of the mouse visual cortex","volume":"23","author":"SE de Vries","year":"2020","journal-title":"Nature Neuroscience"},{"key":"pcbi.1010716.ref033","unstructured":"for Brain Science AI. Visual Behavior\u20142P. 2021;."},{"issue":"6","key":"pcbi.1010716.ref034","doi-asserted-by":"crossref","first-page":"858","DOI":"10.1038\/nn.3711","article-title":"Partitioning neuronal variability","volume":"17","author":"RL Goris","year":"2014","journal-title":"Nature neuroscience"},{"key":"pcbi.1010716.ref035","doi-asserted-by":"crossref","first-page":"148","DOI":"10.1016\/j.conb.2015.04.003","article-title":"On simplicity and complexity in the brave new world of large-scale neuroscience","volume":"32","author":"P Gao","year":"2015","journal-title":"Current opinion in neurobiology"},{"issue":"3","key":"pcbi.1010716.ref036","doi-asserted-by":"crossref","first-page":"609","DOI":"10.1016\/j.neuron.2018.07.003","article-title":"Linking connectivity, dynamics, and computations in low-rank recurrent neural networks","volume":"99","author":"F Mastrogiuseppe","year":"2018","journal-title":"Neuron"},{"issue":"3","key":"pcbi.1010716.ref037","doi-asserted-by":"crossref","first-page":"356","DOI":"10.1038\/nn.4244","article-title":"Using goal-driven deep learning models to understand sensory cortex","volume":"19","author":"DL Yamins","year":"2016","journal-title":"Nature neuroscience"},{"key":"pcbi.1010716.ref038","author":"A Krizhevsky","year":"2009","journal-title":"Learning multiple layers of features from tiny images"},{"issue":"1","key":"pcbi.1010716.ref039","first-page":"1929","article-title":"Dropout: a simple way to prevent neural networks from overfitting","volume":"15","author":"N Srivastava","year":"2014","journal-title":"The journal of machine learning research"},{"key":"pcbi.1010716.ref040","first-page":"1058","article-title":"Regularization of neural networks using dropconnect. In: International conference on machine learning","author":"L Wan","year":"2013","journal-title":"PMLR"},{"key":"pcbi.1010716.ref041","doi-asserted-by":"crossref","first-page":"e37349","DOI":"10.7554\/eLife.37349","article-title":"Sparse recurrent excitatory connectivity in the microcircuit of the adult mouse and human cortex","volume":"7","author":"SC Seeman","year":"2018","journal-title":"Elife"},{"key":"pcbi.1010716.ref042","doi-asserted-by":"crossref","first-page":"e73783","DOI":"10.7554\/eLife.73783","article-title":"Structure and function of axo-axonic inhibition","volume":"10","author":"CM Schneider-Mizell","year":"2021","journal-title":"Elife"},{"key":"pcbi.1010716.ref043","doi-asserted-by":"crossref","first-page":"96","DOI":"10.1016\/j.conb.2013.12.015","article-title":"A blanket of inhibition: functional inferences from dense inhibitory connectivity","volume":"26","author":"MM Karnani","year":"2014","journal-title":"Current opinion in Neurobiology"},{"key":"pcbi.1010716.ref044","first-page":"1050","article-title":"Dropout as a bayesian approximation: Representing model uncertainty in deep learning. In: international conference on machine learning","author":"Y Gal","year":"2016","journal-title":"PMLR"},{"issue":"1","key":"pcbi.1010716.ref045","doi-asserted-by":"crossref","first-page":"232","DOI":"10.1016\/j.neuron.2019.01.029","article-title":"The spatial structure of neural encoding in mouse posterior cortex during navigation","volume":"102","author":"M Minderer","year":"2019","journal-title":"Neuron"},{"key":"pcbi.1010716.ref046","first-page":"351","article-title":"Dropout training as adaptive regularization","volume":"26","author":"S Wager","year":"2013","journal-title":"Advances in neural information processing systems"},{"issue":"1","key":"pcbi.1010716.ref047","first-page":"3403","article-title":"On the inductive bias of dropout","volume":"16","author":"DP Helmbold","year":"2015","journal-title":"The Journal of Machine Learning Research"},{"key":"pcbi.1010716.ref048","unstructured":"Bouthillier X, Konda K, Vincent P, Memisevic R. Dropout as data augmentation. arXiv preprint arXiv:150608700. 2015;."},{"issue":"12","key":"pcbi.1010716.ref049","doi-asserted-by":"crossref","first-page":"2897","DOI":"10.1109\/TPAMI.2017.2784440","article-title":"Information dropout: Learning optimal representations through noisy computation","volume":"40","author":"A Achille","year":"2018","journal-title":"IEEE transactions on pattern analysis and machine intelligence"},{"issue":"11","key":"pcbi.1010716.ref050","doi-asserted-by":"crossref","first-page":"e1004485","DOI":"10.1371\/journal.pcbi.1004485","article-title":"Network plasticity as Bayesian inference","volume":"11","author":"D Kappel","year":"2015","journal-title":"PLoS computational biology"},{"issue":"52","key":"pcbi.1010716.ref051","doi-asserted-by":"crossref","first-page":"E5078","DOI":"10.1073\/pnas.1320116110","article-title":"A theory for how sensorimotor skills are learned and retained in noisy and nonstationary neural circuits","volume":"110","author":"R Ajemian","year":"2013","journal-title":"Proceedings of the National Academy of Sciences"},{"issue":"4","key":"pcbi.1010716.ref052","doi-asserted-by":"crossref","first-page":"653","DOI":"10.1016\/j.neuron.2007.04.030","article-title":"Motor learning with unstable neural representations","volume":"54","author":"U Rokni","year":"2007","journal-title":"Neuron"},{"issue":"1-2","key":"pcbi.1010716.ref053","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1016\/S0166-4328(97)00048-X","article-title":"Free recall and recognition in a network model of the hippocampus: simulating effects of scopolamine on human memory function","volume":"89","author":"ME Hasselmo","year":"1997","journal-title":"Behavioural brain research"},{"issue":"3","key":"pcbi.1010716.ref054","doi-asserted-by":"crossref","first-page":"419","DOI":"10.1037\/0033-295X.102.3.419","article-title":"Why there are complementary learning systems in the hippocampus and neocortex: insights from the successes and failures of connectionist models of learning and memory","volume":"102","author":"JL McClelland","year":"1995","journal-title":"Psychological review"},{"key":"pcbi.1010716.ref055","first-page":"109","volume-title":"Psychology of learning and motivation","author":"M McCloskey","year":"1989"},{"issue":"11","key":"pcbi.1010716.ref056","doi-asserted-by":"crossref","first-page":"1438","DOI":"10.1038\/nn.2405","article-title":"CREB regulates excitability and the allocation of memory to subsets of neurons in the amygdala","volume":"12","author":"Y Zhou","year":"2009","journal-title":"Nature neuroscience"},{"issue":"3","key":"pcbi.1010716.ref057","doi-asserted-by":"crossref","first-page":"157","DOI":"10.1038\/nrn3667","article-title":"Synaptic tagging during memory allocation","volume":"15","author":"T Rogerson","year":"2014","journal-title":"Nature Reviews Neuroscience"},{"issue":"7605","key":"pcbi.1010716.ref058","doi-asserted-by":"crossref","first-page":"115","DOI":"10.1038\/nature17955","article-title":"A shared neural ensemble links distinct contextual memories encoded close in time","volume":"534","author":"DJ Cai","year":"2016","journal-title":"Nature"},{"issue":"5823","key":"pcbi.1010716.ref059","doi-asserted-by":"crossref","first-page":"457","DOI":"10.1126\/science.1139438","article-title":"Neuronal competition and selection during memory formation","volume":"316","author":"JH Han","year":"2007","journal-title":"science"},{"issue":"6297","key":"pcbi.1010716.ref060","doi-asserted-by":"crossref","first-page":"383","DOI":"10.1126\/science.aaf0594","article-title":"Competition between engrams influences fear memory formation and recall","volume":"353","author":"AJ Rashid","year":"2016","journal-title":"Science"},{"issue":"04","key":"pcbi.1010716.ref061","doi-asserted-by":"crossref","first-page":"295","DOI":"10.1142\/S0129065709002002","article-title":"Spiking neural networks","volume":"19","author":"S Ghosh-Dastidar","year":"2009","journal-title":"International journal of neural systems"},{"key":"pcbi.1010716.ref062","first-page":"214262","article-title":"A theory of multineuronal dimensionality, dynamics and measurement","author":"P Gao","year":"2017","journal-title":"BioRxiv"},{"issue":"7","key":"pcbi.1010716.ref063","doi-asserted-by":"crossref","first-page":"e1006446","DOI":"10.1371\/journal.pcbi.1006446","article-title":"Dimensionality in recurrent spiking networks: Global trends in activity and local origins in connectivity","volume":"15","author":"S Recanatesi","year":"2019","journal-title":"PLoS computational biology"},{"key":"pcbi.1010716.ref064","article-title":"The Geometry of Concept Learning","author":"B Sorscher","year":"2021","journal-title":"bioRxiv"},{"issue":"1","key":"pcbi.1010716.ref065","doi-asserted-by":"crossref","first-page":"328","DOI":"10.1016\/j.patcog.2006.06.002","article-title":"Further results on the subspace distance","volume":"40","author":"X Sun","year":"2007","journal-title":"Pattern recognition"},{"key":"pcbi.1010716.ref066","doi-asserted-by":"crossref","first-page":"225","DOI":"10.1007\/978-3-642-00834-4_19","volume-title":"International Symposium on Quantum Interaction","author":"G Zuccon","year":"2009"},{"key":"pcbi.1010716.ref067","unstructured":"Welles O, Heston C, Leigh J, Calleia J, Tamiroff A, Moore J, et al. Touch of evil. 1958;."},{"key":"pcbi.1010716.ref068","doi-asserted-by":"crossref","first-page":"229","DOI":"10.1016\/j.conb.2019.09.011","article-title":"Harnessing behavioral diversity to understand neural computations for cognition","volume":"58","author":"S Musall","year":"2019","journal-title":"Current opinion in neurobiology"}],"updated-by":[{"DOI":"10.1371\/journal.pcbi.1010716","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2022,12,8]],"date-time":"2022-12-08T00:00:00Z","timestamp":1670457600000}}],"container-title":["PLOS Computational Biology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pcbi.1010716","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,12,8]],"date-time":"2022-12-08T13:44:13Z","timestamp":1670507053000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pcbi.1010716"}},"subtitle":[],"editor":[{"given":"Peter E.","family":"Latham","sequence":"first","affiliation":[]}],"short-title":[],"issued":{"date-parts":[[2022,11,28]]},"references-count":68,"journal-issue":{"issue":"11","published-online":{"date-parts":[[2022,11,28]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pcbi.1010716","relation":{"has-preprint":[{"id-type":"doi","id":"10.1101\/2021.12.13.472494","asserted-by":"object"}]},"ISSN":["1553-7358"],"issn-type":[{"value":"1553-7358","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,11,28]]}}}