{
  "abstract": "The neocortex is central to mammalian cognition, yet a computational framework that is both biologically constrained and capable of performing complex cognitive tasks remains missing. Here we show that cortico-thalamic circuits are well suited to implement multi-head self- and cross-attention, the mechanism underlying the cognitive abilities of transformer networks. We propose that layer 2/3 pyramidal cells maintain a recurrent key-value memory, while layer 5 pyramidal cells decode the memory retrieved by an incoming query. The computation of keys, values and queries maps onto core and matrix thalamo-cortical projections, distributed across the micro- and macro-columns of a cortical area. One cortical area forms an attention head, and cortex a multi-head self-attention network. The same thalamo-cortical microcircuit also calculates sensory prediction errors guiding gradient-based synaptic plasticity. A reward-prediction error gates via basal ganglia the cortical output and the re-activation of hippocampal memories. The trained network aligns with human intracranial recordings during speech perception. Overall, the suggested cortico-subcortical attention circuit may represent a substrate for the cognitive capacity of mammals.",
  "abstract_shown": true,
  "affiliations": [
    "University of Bern"
  ],
  "article_type": "New results",
  "arxiv": null,
  "author_total": 8,
  "authors": [
    {
      "affs": [
        0
      ],
      "name": "V. K. Babu"
    },
    {
      "name": "A. Granier"
    },
    {
      "name": "T. Proix"
    },
    {
      "name": "M. Woodman"
    },
    {
      "name": "V. Jirsa"
    },
    {
      "name": "T. Balvocius"
    },
    {
      "name": "A. Saudargiene"
    },
    {
      "name": "W. Senn"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/8x9w5rnxwxwqr.bib",
  "cited_by": 0,
  "doi": "10.64898/2026.08.06.743205",
  "fulltext_url": "https://www.biorxiv.org/content/10.64898/2026.08.06.743205v2.full",
  "id": "8x9w5rnxwxwqr",
  "issns": [],
  "landing_url": "https://www.biorxiv.org/content/10.64898/2026.08.06.743205v2",
  "language": null,
  "licence": "CC-BY",
  "licence_label": "CC BY",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "biorxiv",
      "neuroscience"
    ],
    [
      "openalex",
      "2805|Neural dynamics and brain function"
    ]
  ],
  "openalex": "W7202142464",
  "pdf_url": "https://www.biorxiv.org/content/10.64898/2026.08.06.743205v2.full.pdf",
  "pmcid": null,
  "pmid": null,
  "published": "2026-08-06",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/8x9w5rnxwxwqr.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:22:31Z",
      "native_id": "10.64898/2026.08.06.743205",
      "source": "biorxiv",
      "url": "https://www.biorxiv.org/content/10.64898/2026.08.06.743205v2"
    }
  ],
  "subjects": [
    {
      "field": "biology",
      "origin": "biorxiv:neuroscience",
      "subfield": "neuroscience"
    }
  ],
  "title": "Cortico-subcortical multi-head self-attention as a substrate for cognitive performance",
  "title_text": "Cortico-subcortical multi-head self-attention as a substrate for cognitive performance",
  "type": "preprint",
  "updated_at": "2026-10-30T12:22:31Z",
  "url": "https://research.draconitic.com/paper/8x9w5rnxwxwqr",
  "venue": "bioRxiv",
  "versions": []
}