{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,27]],"date-time":"2026-02-27T06:19:13Z","timestamp":1772173153242,"version":"3.50.1"},"update-to":[{"DOI":"10.1371\/journal.pcbi.1010759","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2022,12,28]],"date-time":"2022-12-28T00:00:00Z","timestamp":1672185600000}}],"reference-count":18,"publisher":"Public Library of Science (PLoS)","issue":"12","license":[{"start":{"date-parts":[[2022,12,14]],"date-time":"2022-12-14T00:00:00Z","timestamp":1670976000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","award":["DBI-1707398"],"award-info":[{"award-number":["DBI-1707398"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000001","name":"National Science Foundation","doi-asserted-by":"publisher","award":["DBI-1707398"],"award-info":[{"award-number":["DBI-1707398"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100000324","name":"Gatsby Charitable Foundation","doi-asserted-by":"publisher","award":["GAT3708"],"award-info":[{"award-number":["GAT3708"]}],"id":[{"id":"10.13039\/501100000324","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100000324","name":"Gatsby Charitable Foundation","doi-asserted-by":"publisher","award":["GAT3708"],"award-info":[{"award-number":["GAT3708"]}],"id":[{"id":"10.13039\/501100000324","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001645","name":"Boehringer Ingelheim Fonds","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100001645","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.ploscompbiol.org"],"crossmark-restriction":false},"short-container-title":["PLoS Comput Biol"],"abstract":"<jats:p>Feedforward network models performing classification tasks rely on highly convergent output units that collect the information passed on by preceding layers. Although convergent output-unit like neurons may exist in some biological neural circuits, notably the cerebellar cortex, neocortical circuits do not exhibit any obvious candidates for this role; instead they are highly recurrent. We investigate whether a sparsely connected recurrent neural network (RNN) can perform classification in a distributed manner without ever bringing all of the relevant information to a single convergence site. Our model is based on a sparse RNN that performs classification dynamically. Specifically, the interconnections of the RNN are trained to resonantly amplify the magnitude of responses to some external inputs but not others. The amplified and non-amplified responses then form the basis for binary classification. Furthermore, the network acts as an evidence accumulator and maintains its decision even after the input is turned off. Despite highly sparse connectivity, learned recurrent connections allow input information to flow to every neuron of the RNN, providing the basis for distributed computation. In this arrangement, the minimum number of synapses per neuron required to reach maximum memory capacity scales only logarithmically with network size. The model is robust to various types of noise, works with different activation and loss functions and with both backpropagation- and Hebbian-based learning rules. The RNN can also be constructed with a split excitation-inhibition architecture with little reduction in performance.<\/jats:p>","DOI":"10.1371\/journal.pcbi.1010759","type":"journal-article","created":{"date-parts":[[2022,12,14]],"date-time":"2022-12-14T13:43:13Z","timestamp":1671025393000},"page":"e1010759","update-policy":"https:\/\/doi.org\/10.1371\/journal.pcbi.corrections_policy","source":"Crossref","is-referenced-by-count":2,"title":["Sparse RNNs can support high-capacity classification"],"prefix":"10.1371","volume":"18","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-2955-2568","authenticated-orcid":true,"given":"Denis","family":"Turcu","sequence":"first","affiliation":[]},{"given":"L. F.","family":"Abbott","sequence":"additional","affiliation":[]}],"member":"340","published-online":{"date-parts":[[2022,12,14]]},"reference":[{"issue":"6","key":"pcbi.1010759.ref001","doi-asserted-by":"crossref","first-page":"386","DOI":"10.1037\/h0042519","article-title":"The perceptron: A probabilistic model for information storage and organization in the brain","volume":"65","author":"F Rosenblatt","year":"1958","journal-title":"Psychological Review"},{"issue":"46","key":"pcbi.1010759.ref002","doi-asserted-by":"crossref","first-page":"9900","DOI":"10.1523\/JNEUROSCI.3506-17.2018","article-title":"Neural classifiers with limited connectivity and recurrent readouts","volume":"38","author":"L Kushnir","year":"2018","journal-title":"Journal of Neuroscience"},{"issue":"3","key":"pcbi.1010759.ref003","doi-asserted-by":"crossref","first-page":"326","DOI":"10.1109\/PGEC.1965.264137","article-title":"Geometrical and Statistical Properties of Systems of Linear Inequalities with Applications in Pattern Recognition","volume":"EC-14","author":"TM Cover","year":"1965","journal-title":"IEEE Transactions on Electronic Computers"},{"issue":"10","key":"pcbi.1010759.ref004","doi-asserted-by":"crossref","first-page":"535","DOI":"10.1209\/0295-5075\/1\/10\/008","article-title":"Networks of formal neurons and memory palimpsests","volume":"1","author":"JP Nadal","year":"1986","journal-title":"Epl"},{"key":"pcbi.1010759.ref005","doi-asserted-by":"crossref","first-page":"290","DOI":"10.5486\/PMD.1959.6.3-4.12","article-title":"On random graphs I","volume":"6","author":"P Erd\u00f6s","year":"1959","journal-title":"Publicationes Mathematicae"},{"key":"pcbi.1010759.ref006","unstructured":"Dauphin YN, Bengio Y. Big Neural Networks Waste Capacity;."},{"issue":"2","key":"pcbi.1010759.ref007","doi-asserted-by":"crossref","first-page":"215","DOI":"10.1016\/j.neuron.2008.09.034","article-title":"Decision Making in Recurrent Neuronal Circuits","volume":"60","author":"XJ Wang","year":"2008","journal-title":"Neuron"},{"issue":"1","key":"pcbi.1010759.ref008","first-page":"205","article-title":"The Hopfield Model on a Sparse Erd\u00f6s-Renyi Graph","volume":"143","author":"M L\u00f6we","year":"2011","journal-title":"Journal of Statistical Physics 2011 143:1"},{"key":"pcbi.1010759.ref009","doi-asserted-by":"crossref","first-page":"5","DOI":"10.1023\/A:1008925909305","article-title":"Model of Familiarity Discrimination in the Perirhinal Cortex","volume":"10","author":"R Bogacz","year":"2001","journal-title":"Journal of Computational Neuroscience"},{"key":"pcbi.1010759.ref010","article-title":"Meta-learning synaptic plasticity and memory addressing for continual familiarity detection","author":"D Tyulmankov","year":"2021","journal-title":"Neuron"},{"issue":"4","key":"pcbi.1010759.ref011","doi-asserted-by":"crossref","first-page":"494","DOI":"10.1002\/hipo.10093","article-title":"Comparison of computational models of familiarity discrimination in the perirhinal cortex","volume":"13","author":"R Bogacz","year":"2003","journal-title":"Hippocampus"},{"issue":"5","key":"pcbi.1010759.ref012","first-page":"745","article-title":"Optimal information storage and the distribution of synaptic weights: Perceptron versus Purkinje cell","volume":"43","author":"N Brunel","year":"2004","journal-title":"Neuron"},{"key":"pcbi.1010759.ref013","unstructured":"Collins J, Sohl-Dickstein J, Sussillo D. Capacity and trainability in recurrent neural networks. 5th International Conference on Learning Representations, ICLR 2017\u2014Conference Track Proceedings. 2017; p. 1\u201317."},{"key":"pcbi.1010759.ref014","unstructured":"Erdos P, R\u00e9nyi A. On the evolution of random graphs; 1960."},{"key":"pcbi.1010759.ref015","first-page":"130","volume-title":"Cambridge Studies in Advanced Mathematics","author":"B Bollobs","year":"2001","edition":"2"},{"issue":"2","key":"pcbi.1010759.ref016","doi-asserted-by":"crossref","first-page":"167","DOI":"10.1209\/0295-5075\/4\/2\/007","article-title":"An exactly solvable asymmetric neural network model","volume":"4","author":"B Derrida","year":"1987","journal-title":"Epl"},{"key":"pcbi.1010759.ref017","first-page":"345","volume-title":"Robustness\u2014Getting Closer to Biology","author":"DJ Amit","year":"1989"},{"key":"pcbi.1010759.ref018","first-page":"8024","volume-title":"Advances in Neural Information Processing Systems 32","author":"A Paszke","year":"2019"}],"updated-by":[{"DOI":"10.1371\/journal.pcbi.1010759","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2022,12,28]],"date-time":"2022-12-28T00:00:00Z","timestamp":1672185600000}}],"container-title":["PLOS Computational Biology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pcbi.1010759","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,12,28]],"date-time":"2022-12-28T13:31:44Z","timestamp":1672234304000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pcbi.1010759"}},"subtitle":[],"editor":[{"given":"Alireza","family":"Soltani","sequence":"first","affiliation":[]}],"short-title":[],"issued":{"date-parts":[[2022,12,14]]},"references-count":18,"journal-issue":{"issue":"12","published-online":{"date-parts":[[2022,12,14]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pcbi.1010759","relation":{"has-preprint":[{"id-type":"doi","id":"10.1101\/2022.05.18.492540","asserted-by":"object"}]},"ISSN":["1553-7358"],"issn-type":[{"value":"1553-7358","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,12,14]]}}}