{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,21]],"date-time":"2026-07-21T15:14:01Z","timestamp":1784646841524,"version":"3.55.0"},"update-to":[{"DOI":"10.1371\/journal.pcbi.1009366","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2021,9,27]],"date-time":"2021-09-27T00:00:00Z","timestamp":1632700800000}}],"reference-count":48,"publisher":"Public Library of Science (PLoS)","issue":"9","license":[{"start":{"date-parts":[[2021,9,15]],"date-time":"2021-09-15T00:00:00Z","timestamp":1631664000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100000002","name":"national institutes of health","doi-asserted-by":"publisher","award":["R01EB028154"],"award-info":[{"award-number":["R01EB028154"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000001","name":"national science foundation","doi-asserted-by":"publisher","award":["1653589"],"award-info":[{"award-number":["1653589"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.ploscompbiol.org"],"crossmark-restriction":false},"short-container-title":["PLoS Comput Biol"],"abstract":"<jats:p>Working memory is a cognitive function involving the storage and manipulation of latent information over brief intervals of time, thus making it crucial for context-dependent computation. Here, we use a top-down modeling approach to examine network-level mechanisms of working memory, an enigmatic issue and central topic of study in neuroscience. We optimize thousands of recurrent rate-based neural networks on a working memory task and then perform dynamical systems analysis on the ensuing optimized networks, wherein we find that four distinct dynamical mechanisms can emerge. In particular, we show the prevalence of a mechanism in which memories are encoded along slow stable manifolds in the network state space, leading to a phasic neuronal activation profile during memory periods. In contrast to mechanisms in which memories are directly encoded at stable attractors, these networks naturally forget stimuli over time. Despite this seeming functional disadvantage, they are more efficient in terms of how they leverage their attractor landscape and paradoxically, are considerably more robust to noise. Our results provide new hypotheses regarding how working memory function may be encoded within the dynamics of neural circuits.<\/jats:p>","DOI":"10.1371\/journal.pcbi.1009366","type":"journal-article","created":{"date-parts":[[2021,9,15]],"date-time":"2021-09-15T18:05:33Z","timestamp":1631729133000},"page":"e1009366","update-policy":"https:\/\/doi.org\/10.1371\/journal.pcbi.corrections_policy","source":"Crossref","is-referenced-by-count":15,"title":["Slow manifolds within network dynamics encode working memory efficiently and robustly"],"prefix":"10.1371","volume":"17","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-6778-6392","authenticated-orcid":true,"given":"Elham","family":"Ghazizadeh","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4063-7068","authenticated-orcid":true,"given":"ShiNung","family":"Ching","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2021,9,15]]},"reference":[{"key":"pcbi.1009366.ref001","doi-asserted-by":"crossref","first-page":"323","DOI":"10.1016\/S0079-6123(07)00020-9","article-title":"What are the differences between long-term, short-term, and working memory?","volume":"169","author":"N Cowan","year":"2008","journal-title":"Progress In Brain Research"},{"key":"pcbi.1009366.ref002","doi-asserted-by":"crossref","first-page":"394","DOI":"10.1038\/nn.4237","article-title":"Computational principles of memory","volume":"19","author":"R Chaudhuri","year":"2016","journal-title":"Nature Neuroscience"},{"key":"pcbi.1009366.ref003","doi-asserted-by":"crossref","first-page":"470","DOI":"10.1038\/20939","article-title":"Neuronal correlates of parametric working memory in the prefrontal cortex","volume":"399","author":"R Romo","year":"1999","journal-title":"Nature"},{"key":"pcbi.1009366.ref004","doi-asserted-by":"crossref","first-page":"331","DOI":"10.1152\/jn.1989.61.2.331","article-title":"Mnemonic coding of visual space in the monkey\u2019s dorsolateral prefrontal cortex","volume":"61","author":"S Funahashi","year":"1989","journal-title":"Journal Of Neurophysiology"},{"key":"pcbi.1009366.ref005","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":"C Harvey","year":"2012","journal-title":"Nature"},{"key":"pcbi.1009366.ref006","doi-asserted-by":"crossref","first-page":"463","DOI":"10.1016\/j.neuron.2018.09.023","article-title":"Working Memory 2.0","volume":"100","author":"E Miller","year":"2018","journal-title":"Neuron"},{"key":"pcbi.1009366.ref007","doi-asserted-by":"crossref","first-page":"181","DOI":"10.3389\/fnsys.2015.00181","article-title":"Role of prefrontal persistent activity in working memory","volume":"9","author":"M Riley","year":"2016","journal-title":"Frontiers In Systems Neuroscience"},{"key":"pcbi.1009366.ref008","doi-asserted-by":"crossref","first-page":"455","DOI":"10.1016\/S0166-2236(00)01868-3","article-title":"Synaptic reverberation underlying mnemonic persistent activity","volume":"24","author":"X Wang","year":"2001","journal-title":"Trends In Neurosciences"},{"key":"pcbi.1009366.ref009","doi-asserted-by":"crossref","first-page":"910","DOI":"10.1093\/cercor\/10.9.910","article-title":"Synaptic mechanisms and network dynamics underlying spatial working memory in a cortical network model","volume":"10","author":"A Compte","year":"2000","journal-title":"Cerebral Cortex"},{"key":"pcbi.1009366.ref010","doi-asserted-by":"crossref","first-page":"117","DOI":"10.3389\/fncom.2015.00117","article-title":"Dynamic trajectory of multiple single-unit activity during working memory task in rats","volume":"9","author":"X Zhang","year":"2015","journal-title":"Frontiers In Computational Neuroscience"},{"key":"pcbi.1009366.ref011","doi-asserted-by":"crossref","first-page":"394","DOI":"10.1073\/pnas.1619449114","article-title":"Stable population coding for working memory coexists with heterogeneous neural dynamics in prefrontal cortex","volume":"114","author":"J Murray","year":"2017","journal-title":"Proceedings Of The National Academy Of Sciences"},{"key":"pcbi.1009366.ref012","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"},{"key":"pcbi.1009366.ref013","doi-asserted-by":"crossref","first-page":"1504","DOI":"10.1016\/j.neuron.2017.03.002","article-title":"Computing by robust transience: how the fronto-parietal network performs sequential, category-based decisions","volume":"93","author":"W Chaisangmongkon","year":"2017","journal-title":"Neuron"},{"key":"pcbi.1009366.ref014","doi-asserted-by":"crossref","DOI":"10.1146\/annurev-neuro-070815-014006","article-title":"Mechanisms of persistent activity in cortical circuits: possible neural substrates for working memory","volume":"40","author":"J Zylberberg","year":"2017","journal-title":"Annual Review Of Neuroscience"},{"key":"pcbi.1009366.ref015","doi-asserted-by":"crossref","first-page":"20","DOI":"10.1016\/j.conb.2013.10.008","article-title":"Working models of working memory","volume":"25","author":"O Barak","year":"2014","journal-title":"Current Opinion In Neurobiology"},{"key":"pcbi.1009366.ref016","doi-asserted-by":"crossref","first-page":"2374","DOI":"10.1046\/j.1460-9568.2003.02956.x","article-title":"Attention and working memory: a dynamical model of neuronal activity in the prefrontal cortex","volume":"18","author":"G Deco","year":"2003","journal-title":"European Journal Of Neuroscience"},{"key":"pcbi.1009366.ref017","doi-asserted-by":"crossref","first-page":"2095","DOI":"10.1016\/j.cub.2012.08.058","article-title":"Neuronal circuits underlying persistent representations despite time varying activity","volume":"22","author":"S Druckmann","year":"2012","journal-title":"Current Biology"},{"key":"pcbi.1009366.ref018","doi-asserted-by":"crossref","first-page":"6503","DOI":"10.1523\/JNEUROSCI.3364-16.2017","article-title":"Stable and dynamic coding for working memory in primate prefrontal cortex","volume":"37","author":"E Spaak","year":"2017","journal-title":"Journal Of Neuroscience"},{"key":"pcbi.1009366.ref019","doi-asserted-by":"crossref","first-page":"e1007655","DOI":"10.1371\/journal.pcbi.1007655","article-title":"Coding with transient trajectories in recurrent neural networks","volume":"16","author":"G Bondanelli","year":"2020","journal-title":"PLoS Computational Biology"},{"key":"pcbi.1009366.ref020","first-page":"1","article-title":"Dynamically changing neuronal activity supporting working memory for predictable and unpredictable durations","volume":"9","author":"J Park","year":"2019","journal-title":"Scientific Reports"},{"key":"pcbi.1009366.ref021","doi-asserted-by":"crossref","first-page":"293","DOI":"10.1037\/h0087996","article-title":"Delayed-matching and delayed-response deficit from cooling dorsolateral prefrontal cortex in monkeys","volume":"90","author":"R Bauer","year":"1976","journal-title":"Journal Of Comparative And Physiological Psychology"},{"key":"pcbi.1009366.ref022","doi-asserted-by":"crossref","first-page":"677","DOI":"10.1093\/cercor\/bhs348","article-title":"Emergence of complex computational structures from chaotic neural networks through reward-modulated Hebbian learning","volume":"24","author":"G Hoerzer","year":"2014","journal-title":"Cerebral Cortex"},{"key":"pcbi.1009366.ref023","doi-asserted-by":"crossref","first-page":"e1004967","DOI":"10.1371\/journal.pcbi.1004967","article-title":"Reservoir computing properties of neural dynamics in prefrontal cortex","volume":"12","author":"P Enel","year":"2016","journal-title":"PLoS Computational Biology"},{"key":"pcbi.1009366.ref024","doi-asserted-by":"crossref","first-page":"199","DOI":"10.1016\/j.neunet.2010.10.003","article-title":"A neurodynamical model for working memory","volume":"24","author":"R Pascanu","year":"2011","journal-title":"Neural Networks"},{"key":"pcbi.1009366.ref025","unstructured":"Maheswaranathan, N., Williams, A., Golub, M., Ganguli, S. & Sussillo, D. Universality and individuality in neural dynamics across large populations of recurrent networks. Advances In Neural Information Processing Systems. pp. 15629\u201315641 (2019)"},{"key":"pcbi.1009366.ref026","doi-asserted-by":"crossref","first-page":"e1004792","DOI":"10.1371\/journal.pcbi.1004792","article-title":"Training excitatory-inhibitory recurrent neural networks for cognitive tasks: a simple and flexible framework","volume":"12","author":"H Song","year":"2016","journal-title":"PLoS Computational Biology"},{"key":"pcbi.1009366.ref027","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s41598-017-02471-z","article-title":"Working memory requires a combination of transient and attractor-dominated dynamics to process unreliably timed inputs","volume":"7","author":"T Nachstedt","year":"2017","journal-title":"Scientific Reports"},{"key":"pcbi.1009366.ref028","doi-asserted-by":"crossref","first-page":"e54313","DOI":"10.7554\/eLife.54313","article-title":"Stable and dynamic representations of value in the prefrontal cortex","volume":"9","author":"P Enel","year":"2020","journal-title":"Elife"},{"key":"pcbi.1009366.ref029","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s41467-018-05873-3","article-title":"Reconciling persistent and dynamic hypotheses of working memory coding in prefrontal cortex","volume":"9","author":"S Cavanagh","year":"2018","journal-title":"Nature Communications"},{"key":"pcbi.1009366.ref030","doi-asserted-by":"crossref","first-page":"275","DOI":"10.1038\/s41593-018-0314-y","article-title":"A diverse range of factors affect the nature of neural representations underlying short-term memory","volume":"22","author":"A Orhan","year":"2019","journal-title":"Nature Neuroscience"},{"key":"pcbi.1009366.ref031","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"},{"key":"pcbi.1009366.ref032","doi-asserted-by":"crossref","first-page":"297","DOI":"10.1038\/s41593-018-0310-2","article-title":"Task representations in neural networks trained to perform many cognitive tasks","volume":"22","author":"G Yang","year":"2019","journal-title":"Nature Neuroscience"},{"key":"pcbi.1009366.ref033","doi-asserted-by":"crossref","first-page":"544","DOI":"10.1016\/j.neuron.2009.07.018","article-title":"Generating coherent patterns of activity from chaotic neural networks","volume":"63","author":"D Sussillo","year":"2009","journal-title":"Neuron"},{"key":"pcbi.1009366.ref034","unstructured":"Yang, G. & Wang, X. Artificial neural networks for neuroscientists: A primer. ArXiv Preprint ArXiv:2006.01001. (2020)"},{"key":"pcbi.1009366.ref035","doi-asserted-by":"crossref","first-page":"013111","DOI":"10.1103\/PhysRevResearch.2.013111","article-title":"Dynamics of random recurrent networks with correlated low-rank structure","volume":"2","author":"F Schuessler","year":"2020","journal-title":"Physical Review Research"},{"key":"pcbi.1009366.ref036","doi-asserted-by":"crossref","first-page":"128","DOI":"10.1016\/j.neuron.2016.02.009","article-title":"Recurrent network models of sequence generation and memory","volume":"90","author":"K Rajan","year":"2016","journal-title":"Neuron"},{"key":"pcbi.1009366.ref037","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"},{"key":"pcbi.1009366.ref038","doi-asserted-by":"crossref","first-page":"2554","DOI":"10.1073\/pnas.79.8.2554","article-title":"Neural networks and physical systems with emergent collective computational abilities","volume":"79","author":"J Hopfield","year":"1982","journal-title":"Proceedings Of The National Academy Of Sciences"},{"key":"pcbi.1009366.ref039","doi-asserted-by":"crossref","first-page":"402","DOI":"10.1016\/j.neuron.2014.12.026","article-title":"The stabilized supralinear network: a unifying circuit motif underlying multi-input integration in sensory cortex","volume":"85","author":"D Rubin","year":"2015","journal-title":"Neuron"},{"key":"pcbi.1009366.ref040","doi-asserted-by":"crossref","first-page":"860","DOI":"10.1038\/s41593-018-0147-8","article-title":"Prefrontal cortex as a meta-reinforcement learning system","volume":"21","author":"J Wang","year":"2018","journal-title":"Nature Neuroscience"},{"key":"pcbi.1009366.ref041","doi-asserted-by":"crossref","first-page":"e1006928","DOI":"10.1371\/journal.pcbi.1006928","article-title":"Stability of working memory in continuous attractor networks under the control of short-term plasticity","volume":"15","author":"A Seeholzer","year":"2019","journal-title":"PLoS Computational Biology"},{"key":"pcbi.1009366.ref042","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1016\/j.neubiorev.2019.03.017","article-title":"Neural mechanisms of attending to items in working memory","volume":"101","author":"S Manohar","year":"2019","journal-title":"Neuroscience & Biobehavioral Reviews"},{"key":"pcbi.1009366.ref043","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s41467-020-15541-0","article-title":"Multimodal network dynamics underpinning working memory","volume":"11","author":"A Murphy","year":"2020","journal-title":"Nature Communications"},{"key":"pcbi.1009366.ref044","doi-asserted-by":"crossref","first-page":"415","DOI":"10.1016\/S1364-6613(03)00197-9","article-title":"Persistent activity in the prefrontal cortex during working memory","volume":"7","author":"C Curtis","year":"2003","journal-title":"Trends In Cognitive Sciences"},{"key":"pcbi.1009366.ref045","doi-asserted-by":"crossref","first-page":"219","DOI":"10.1146\/annurev.neuro.31.060407.125642","article-title":"Task set and prefrontal cortex","volume":"31","author":"K Sakai","year":"2008","journal-title":"Annu. Rev. Neurosci"},{"key":"pcbi.1009366.ref046","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.1009366.ref047","doi-asserted-by":"crossref","first-page":"1985","DOI":"10.1162\/neco_a_01222","article-title":"One step back, two steps forward: interference and learning in recurrent neural networks","volume":"31","author":"C Beer","year":"2019","journal-title":"Neural Computation"},{"key":"pcbi.1009366.ref048","doi-asserted-by":"crossref","first-page":"188104","DOI":"10.1103\/PhysRevLett.97.188104","article-title":"Eigenvalue spectra of random matrices for neural networks","volume":"97","author":"K Rajan","year":"2006","journal-title":"Physical Review Letters"}],"updated-by":[{"DOI":"10.1371\/journal.pcbi.1009366","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2021,9,27]],"date-time":"2021-09-27T00:00:00Z","timestamp":1632700800000}}],"container-title":["PLOS Computational Biology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pcbi.1009366","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2021,9,27]],"date-time":"2021-09-27T20:00:50Z","timestamp":1632772850000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pcbi.1009366"}},"subtitle":[],"editor":[{"given":"Ming Bo","family":"Cai","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2021,9,15]]},"references-count":48,"journal-issue":{"issue":"9","published-online":{"date-parts":[[2021,9,15]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pcbi.1009366","relation":{"new_version":[{"id-type":"doi","id":"10.1371\/journal.pcbi.1009366","asserted-by":"object"}]},"ISSN":["1553-7358"],"issn-type":[{"value":"1553-7358","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,9,15]]}}}