{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,6]],"date-time":"2026-04-06T19:37:39Z","timestamp":1775504259332,"version":"3.50.1"},"update-to":[{"DOI":"10.1371\/journal.pcbi.1010418","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2022,9,29]],"date-time":"2022-09-29T00:00:00Z","timestamp":1664409600000}}],"reference-count":51,"publisher":"Public Library of Science (PLoS)","issue":"9","license":[{"start":{"date-parts":[[2022,9,19]],"date-time":"2022-09-19T00:00:00Z","timestamp":1663545600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["www.ploscompbiol.org"],"crossmark-restriction":false},"short-container-title":["PLoS Comput Biol"],"abstract":"<jats:p>We introduce a novel, biologically plausible local learning rule that provably increases the robustness of neural dynamics to noise in nonlinear recurrent neural networks with homogeneous nonlinearities. Our learning rule achieves higher noise robustness without sacrificing performance on the task and without requiring any knowledge of the particular task. The plasticity dynamics\u2014an integrable dynamical system operating on the weights of the network\u2014maintains a multiplicity of conserved quantities, most notably the network\u2019s entire temporal map of input to output trajectories. The outcome of our learning rule is a synaptic balancing between the incoming and outgoing synapses of every neuron. This synaptic balancing rule is consistent with many known aspects of experimentally observed heterosynaptic plasticity, and moreover makes new experimentally testable predictions relating plasticity at the incoming and outgoing synapses of individual neurons. Overall, this work provides a novel, practical local learning rule that exactly preserves overall network function and, in doing so, provides new conceptual bridges between the disparate worlds of the neurobiology of heterosynaptic plasticity, the engineering of regularized noise-robust networks, and the mathematics of integrable Lax dynamical systems.<\/jats:p>","DOI":"10.1371\/journal.pcbi.1010418","type":"journal-article","created":{"date-parts":[[2022,9,19]],"date-time":"2022-09-19T17:38:48Z","timestamp":1663609128000},"page":"e1010418","update-policy":"https:\/\/doi.org\/10.1371\/journal.pcbi.corrections_policy","source":"Crossref","is-referenced-by-count":4,"title":["Synaptic balancing: A biologically plausible local learning rule that provably increases neural network noise robustness without sacrificing task performance"],"prefix":"10.1371","volume":"18","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-4579-660X","authenticated-orcid":true,"given":"Christopher H.","family":"Stock","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2601-7487","authenticated-orcid":true,"given":"Sarah E.","family":"Harvey","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3426-4517","authenticated-orcid":true,"given":"Samuel A.","family":"Ocko","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9264-7551","authenticated-orcid":true,"given":"Surya","family":"Ganguli","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"340","published-online":{"date-parts":[[2022,9,19]]},"reference":[{"issue":"4","key":"pcbi.1010418.ref001","doi-asserted-by":"crossref","first-page":"292","DOI":"10.1038\/nrn2258","article-title":"Noise in the nervous system","volume":"9","author":"AA Faisal","year":"2008","journal-title":"Nature reviews neuroscience"},{"issue":"8","key":"pcbi.1010418.ref002","doi-asserted-by":"crossref","first-page":"775","DOI":"10.1016\/0042-6989(83)90200-6","article-title":"The statistical reliability of signals in single neurons in cat and monkey visual cortex","volume":"23","author":"DJ Tolhurst","year":"1983","journal-title":"Vision research"},{"issue":"5216","key":"pcbi.1010418.ref003","doi-asserted-by":"crossref","first-page":"1503","DOI":"10.1126\/science.7770778","article-title":"Reliability of spike timing in neocortical neurons","volume":"268","author":"ZF Mainen","year":"1995","journal-title":"Science"},{"issue":"10","key":"pcbi.1010418.ref004","doi-asserted-by":"crossref","first-page":"3870","DOI":"10.1523\/JNEUROSCI.18-10-03870.1998","article-title":"The variable discharge of cortical neurons: implications for connectivity, computation, and information coding","volume":"18","author":"MN Shadlen","year":"1998","journal-title":"Journal of neuroscience"},{"issue":"7801","key":"pcbi.1010418.ref005","doi-asserted-by":"crossref","first-page":"100","DOI":"10.1038\/s41586-020-2130-2","article-title":"Fundamental bounds on the fidelity of sensory cortical coding","volume":"580","author":"OI Rumyantsev","year":"2020","journal-title":"Nature"},{"issue":"48","key":"pcbi.1010418.ref006","doi-asserted-by":"crossref","first-page":"18970","DOI":"10.1073\/pnas.0804451105","article-title":"Memory traces in dynamical systems","volume":"105","author":"S Ganguli","year":"2008","journal-title":"Proceedings of the national academy of sciences"},{"key":"pcbi.1010418.ref007","volume-title":"Neural Information Processing Systems (NIPS)","author":"S Ganguli","year":"2010"},{"key":"pcbi.1010418.ref008","article-title":"Predictive coding in balanced neural networks with noise, chaos and delays","volume":"33","author":"J Kadmon","year":"2020","journal-title":"Advances in neural information processing systems"},{"issue":"5","key":"pcbi.1010418.ref009","doi-asserted-by":"crossref","first-page":"467","DOI":"10.1002\/cpa.3160210503","article-title":"Integrals of nonlinear equations of evolution and solitary waves","volume":"21","author":"PD Lax","year":"1968","journal-title":"Communications on pure and applied mathematics"},{"key":"pcbi.1010418.ref010","doi-asserted-by":"crossref","DOI":"10.1007\/978-1-4471-3467-1","volume-title":"Optimization and dynamical systems","author":"U Helmke","year":"1994"},{"key":"pcbi.1010418.ref011","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1017\/S0962492906340019","article-title":"Linear algebra algorithms as dynamical systems","volume":"17","author":"MT Chu","year":"2008","journal-title":"Acta numerica"},{"issue":"5","key":"pcbi.1010418.ref012","doi-asserted-by":"crossref","first-page":"483","DOI":"10.1177\/1073858414529829","article-title":"Heterosynaptic plasticity: multiple mechanisms and multiple roles","volume":"20","author":"M Chistiakova","year":"2014","journal-title":"The Neuroscientist"},{"key":"pcbi.1010418.ref013","doi-asserted-by":"crossref","first-page":"89","DOI":"10.3389\/fncom.2015.00089","article-title":"Homeostatic role of heterosynaptic plasticity: models and experiments","volume":"9","author":"M Chistiakova","year":"2015","journal-title":"Frontiers in computational neuroscience"},{"issue":"2","key":"pcbi.1010418.ref014","doi-asserted-by":"crossref","first-page":"162","DOI":"10.1016\/j.celrep.2014.12.016","article-title":"Heterosynaptic structural plasticity on local dendritic segments of hippocampal CA1 neurons","volume":"10","author":"WC Oh","year":"2015","journal-title":"Cell reports"},{"issue":"6395","key":"pcbi.1010418.ref015","doi-asserted-by":"crossref","first-page":"1349","DOI":"10.1126\/science.aao0862","article-title":"Locally coordinated synaptic plasticity of visual cortex neurons in vivo","volume":"360","author":"S El-Boustani","year":"2018","journal-title":"Science"},{"key":"pcbi.1010418.ref016","article-title":"Heterosynaptic Plasticity Determines the Set Point for Cortical Excitatory-Inhibitory Balance","author":"RE Field","year":"2020","journal-title":"Neuron"},{"issue":"12","key":"pcbi.1010418.ref017","doi-asserted-by":"crossref","first-page":"1345","DOI":"10.1038\/nn1352","article-title":"Similar network activity from disparate circuit parameters","volume":"7","author":"AA Prinz","year":"2004","journal-title":"Nature neuroscience"},{"issue":"4","key":"pcbi.1010418.ref018","doi-asserted-by":"crossref","first-page":"947","DOI":"10.1016\/j.cell.2016.10.019","article-title":"From whole-brain data to functional circuit models: the zebrafish optomotor response","volume":"167","author":"EA Naumann","year":"2016","journal-title":"Cell"},{"key":"pcbi.1010418.ref019","first-page":"15629","article-title":"Universality and individuality in neural dynamics across large populations of recurrent networks","volume":"2019","author":"N Maheswaranathan","year":"2019","journal-title":"Advances in neural information processing systems"},{"key":"pcbi.1010418.ref020","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"},{"key":"pcbi.1010418.ref021","doi-asserted-by":"crossref","first-page":"214","DOI":"10.1016\/j.pneurobio.2013.02.002","article-title":"From fixed points to chaos: three models of delayed discrimination","volume":"103","author":"O Barak","year":"2013","journal-title":"Progress in neurobiology"},{"issue":"7","key":"pcbi.1010418.ref022","doi-asserted-by":"crossref","first-page":"1025","DOI":"10.1038\/nn.4042","article-title":"A neural network that finds a naturalistic solution for the production of muscle activity","volume":"18","author":"D Sussillo","year":"2015","journal-title":"Nature neuroscience"},{"issue":"7474","key":"pcbi.1010418.ref023","doi-asserted-by":"crossref","first-page":"78","DOI":"10.1038\/nature12742","article-title":"Context-dependent computation by recurrent dynamics in prefrontal cortex","volume":"503","author":"V Mante","year":"2013","journal-title":"Nature"},{"key":"pcbi.1010418.ref024","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1016\/j.conb.2017.06.003","article-title":"Recurrent neural networks as versatile tools of neuroscience research","volume":"46","author":"O Barak","year":"2017","journal-title":"Current opinion in neurobiology"},{"issue":"3","key":"pcbi.1010418.ref025","doi-asserted-by":"crossref","first-page":"422","DOI":"10.1016\/j.cell.2008.10.008","article-title":"The self-tuning neuron: synaptic scaling of excitatory synapses","volume":"135","author":"GG Turrigiano","year":"2008","journal-title":"Cell"},{"issue":"1715","key":"pcbi.1010418.ref026","doi-asserted-by":"crossref","first-page":"20160258","DOI":"10.1098\/rstb.2016.0258","article-title":"The dialectic of Hebb and homeostasis","volume":"372","author":"GG Turrigiano","year":"2017","journal-title":"Philosophical transactions of the royal society B: biological sciences"},{"issue":"40","key":"pcbi.1010418.ref027","doi-asserted-by":"crossref","first-page":"15915","DOI":"10.1523\/JNEUROSCI.5088-12.2013","article-title":"Heterosynaptic plasticity prevents runaway synaptic dynamics","volume":"33","author":"JY Chen","year":"2013","journal-title":"Journal of Neuroscience"},{"issue":"1715","key":"pcbi.1010418.ref028","doi-asserted-by":"crossref","first-page":"20160259","DOI":"10.1098\/rstb.2016.0259","article-title":"Hebbian plasticity requires compensatory processes on multiple timescales","volume":"372","author":"F Zenke","year":"2017","journal-title":"Philosophical transactions of the royal society B: biological sciences"},{"key":"pcbi.1010418.ref029","doi-asserted-by":"crossref","first-page":"357","DOI":"10.1146\/annurev.neuro.28.061604.135637","article-title":"Neural network dynamics","volume":"28","author":"TP Vogels","year":"2005","journal-title":"Annual review of neuroscience"},{"issue":"3","key":"pcbi.1010418.ref030","doi-asserted-by":"crossref","first-page":"259","DOI":"10.1103\/PhysRevLett.61.259","article-title":"Chaos in random neural networks","volume":"61","author":"H Sompolinsky","year":"1988","journal-title":"Physical review letters"},{"key":"pcbi.1010418.ref031","first-page":"15696","volume-title":"Advances in neural information processing systems","author":"N Maheswaranathan","year":"2019"},{"key":"pcbi.1010418.ref032","first-page":"2933","article-title":"Identifying and attacking the saddle point problem in high-dimensional non-convex optimization","author":"YN Dauphin","year":"2014","journal-title":"Advances in neural information processing systems"},{"key":"pcbi.1010418.ref033","doi-asserted-by":"crossref","DOI":"10.1017\/CBO9780511804441","volume-title":"Convex optimization","author":"S Boyd","year":"2004"},{"key":"pcbi.1010418.ref034","article-title":"Theoretical neuroscience: computational and mathematical modeling of neural systems","author":"P Dayan","year":"2001","journal-title":"Computational Neuroscience Series"},{"issue":"1","key":"pcbi.1010418.ref035","doi-asserted-by":"crossref","first-page":"26","DOI":"10.1016\/j.tics.2012.10.011","article-title":"The ventral visual pathway: an expanded neural framework for the processing of object quality","volume":"17","author":"DJ Kravitz","year":"2013","journal-title":"Trends in cognitive sciences"},{"key":"pcbi.1010418.ref036","article-title":"Goal-driven recurrent neural network models of the ventral visual stream","author":"A Nayebi","year":"2021","journal-title":"bioRxiv"},{"issue":"1","key":"pcbi.1010418.ref037","doi-asserted-by":"crossref","first-page":"87","DOI":"10.1287\/opre.18.1.87","article-title":"On a class of directed graphs\u2013with an application to traffic-flow problems","volume":"18","author":"L Hooi-Tong","year":"1970","journal-title":"Operations research"},{"key":"pcbi.1010418.ref038","first-page":"124","volume-title":"Mathematical programming essays in honor of George B","author":"BC Eaves","year":"1985"},{"key":"pcbi.1010418.ref039","doi-asserted-by":"crossref","DOI":"10.1007\/978-0-387-84858-7","volume-title":"The elements of statistical learning: data mining, inference, and prediction","author":"T Hastie","year":"2009"},{"issue":"1","key":"pcbi.1010418.ref040","doi-asserted-by":"crossref","first-page":"208","DOI":"10.1287\/moor.16.1.208","article-title":"Max-balancing weighted directed graphs and matrix scaling","volume":"16","author":"H Schneider","year":"1991","journal-title":"Mathematics of operations research"},{"key":"pcbi.1010418.ref041","doi-asserted-by":"crossref","first-page":"159","DOI":"10.1016\/0024-3795(92)90307-V","article-title":"Scalings of matrices satisfying line-product constraints and generalizations","volume":"175","author":"UG Rothblum","year":"1992","journal-title":"Linear algebra and its applications"},{"key":"pcbi.1010418.ref042","article-title":"Pruning neural networks without any data by iteratively conserving synaptic flow","author":"H Tanaka","year":"2020","journal-title":"Advances in neural information processing systems"},{"key":"pcbi.1010418.ref043","first-page":"384","article-title":"Algorithmic regularization in learning deep homogeneous models: Layers are automatically balanced","author":"SS Du","year":"2018","journal-title":"Advances in neural information processing systems"},{"key":"pcbi.1010418.ref044","volume-title":"Spectral graph theory","author":"FR Chung","year":"1997"},{"issue":"1","key":"pcbi.1010418.ref045","doi-asserted-by":"crossref","first-page":"81","DOI":"10.1007\/BF01164627","article-title":"Resistance distance","volume":"12","author":"DJ Klein","year":"1993","journal-title":"Journal of mathematical chemistry"},{"issue":"1","key":"pcbi.1010418.ref046","doi-asserted-by":"crossref","first-page":"353","DOI":"10.1113\/jphysiol.1978.sp012239","article-title":"Long-term potentiation and depression of synaptic responses in the rat hippocampus: localization and frequency dependency","volume":"276","author":"T Dunwiddie","year":"1978","journal-title":"The Journal of physiology"},{"issue":"6931","key":"pcbi.1010418.ref047","doi-asserted-by":"crossref","first-page":"518","DOI":"10.1038\/nature01530","article-title":"Conservation of total synaptic weight through balanced synaptic depression and potentiation","volume":"422","author":"S Royer","year":"2003","journal-title":"Nature"},{"issue":"6670","key":"pcbi.1010418.ref048","doi-asserted-by":"crossref","first-page":"892","DOI":"10.1038\/36103","article-title":"Activity-dependent scaling of quantal amplitude in neocortical neurons","volume":"391","author":"GG Turrigiano","year":"1998","journal-title":"Nature"},{"issue":"2","key":"pcbi.1010418.ref049","doi-asserted-by":"crossref","first-page":"335","DOI":"10.1016\/j.neuron.2013.08.038","article-title":"Firing rate homeostasis in visual cortex of freely behaving rodents","volume":"80","author":"KB Hengen","year":"2013","journal-title":"Neuron"},{"key":"pcbi.1010418.ref050","doi-asserted-by":"crossref","first-page":"166","DOI":"10.1016\/j.conb.2017.03.015","article-title":"The temporal paradox of Hebbian learning and homeostatic plasticity","volume":"43","author":"F Zenke","year":"2017","journal-title":"Current opinion in neurobiology"},{"issue":"7","key":"pcbi.1010418.ref051","doi-asserted-by":"crossref","first-page":"563","DOI":"10.1038\/nrn1949","article-title":"Variability, compensation and homeostasis in neuron and network function","volume":"7","author":"E Marder","year":"2006","journal-title":"Nature Reviews Neuroscience"}],"updated-by":[{"DOI":"10.1371\/journal.pcbi.1010418","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2022,9,29]],"date-time":"2022-09-29T00:00:00Z","timestamp":1664409600000}}],"container-title":["PLOS Computational Biology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pcbi.1010418","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,9,29]],"date-time":"2022-09-29T19:35:58Z","timestamp":1664480158000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pcbi.1010418"}},"subtitle":[],"editor":[{"given":"Blake A.","family":"Richards","sequence":"first","affiliation":[],"role":[{"role":"editor","vocabulary":"crossref"}]}],"short-title":[],"issued":{"date-parts":[[2022,9,19]]},"references-count":51,"journal-issue":{"issue":"9","published-online":{"date-parts":[[2022,9,19]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pcbi.1010418","relation":{"new_version":[{"id-type":"doi","id":"10.1371\/journal.pcbi.1010418","asserted-by":"object"}]},"ISSN":["1553-7358"],"issn-type":[{"value":"1553-7358","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,9,19]]}}}