{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/why-neurons-have-thousands-of-synapses-a","title":"Why Neurons Have Thousands of Synapses, A Theory of Sequence Memory in Neocortex","arxiv_id":"1511.00083","date":"2015-10-31","proceeding":null,"authors":["Jeff Hawkins","Subutai Ahmad"],"abstract":"Neocortical neurons have thousands of excitatory synapses. It is a mystery\nhow neurons integrate the input from so many synapses and what kind of\nlarge-scale network behavior this enables. It has been previously proposed that\nnon-linear properties of dendrites enable neurons to recognize multiple\npatterns. In this paper we extend this idea by showing that a neuron with\nseveral thousand synapses arranged along active dendrites can learn to\naccurately and robustly recognize hundreds of unique patterns of cellular\nactivity, even in the presence of large amounts of noise and pattern variation.\nWe then propose a neuron model where some of the patterns recognized by a\nneuron lead to action potentials and define the classic receptive field of the\nneuron, whereas the majority of the patterns recognized by a neuron act as\npredictions by slightly depolarizing the neuron without immediately generating\nan action potential. We then present a network model based on neurons with\nthese properties and show that the network learns a robust model of time-based\nsequences. Given the similarity of excitatory neurons throughout the neocortex\nand the importance of sequence memory in inference and behavior, we propose\nthat this form of sequence memory is a universal property of neocortical\ntissue. We further propose that cellular layers in the neocortex implement\nvariations of the same sequence memory algorithm to achieve different aspects\nof inference and behavior. The neuron and network models we introduce are\nrobust over a wide range of parameters as long as the network uses a sparse\ndistributed code of cellular activations. The sequence capacity of the network\nscales linearly with the number of synapses on each neuron. Thus neurons need\nthousands of synapses to learn the many temporal patterns in sensory stimuli\nand motor sequences.","url_abs":"http://arxiv.org/abs/1511.00083v2","url_pdf":"http://arxiv.org/pdf/1511.00083v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"why-neurons-have-thousands-of-synapses-a","repo_url":"https://github.com/numenta/htmpapers/blob/master/frontiers/why_neurons_have_thousands_of_synapses","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1511.00083","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}