{
  "abstract": "Extracting interpretable representations from high-dimensional whole-brain neural dynamics remains a major computational challenge in systems neuroscience. Here, we develop a wiring-agnostic deep-learning framework that combines convolutional encoding with temporal transformers to learn compact representations directly from volumetric calcium imaging of the entire \\textit{Drosophila melanogaster} brain, without neuronal identification or anatomical annotation. Applied to brain-wide activity recorded across 16 factorially combined sensory and internal-state conditions, the learned representations revealed a factorized organization of whole-brain dynamics: metabolic state, sensory modality, and stimulus valence emerged along three near-orthogonal axes from a classification objective based only on flat class labels and without explicit disentanglement constraints. Spatial attribution and brain region-level ablation analyses linked modality representations to anatomically distinct circuits, whereas state- and valence-related information showed a broadly distributed organization. Our approach provides a scalable framework for discovering interpretable representations of whole-brain neural dynamics.",
  "abstract_shown": true,
  "affiliations": [
    "Institute of Zoology, University of Cologne, Germany"
  ],
  "article_type": "New results",
  "arxiv": null,
  "author_total": 6,
  "authors": [
    {
      "name": "A. Abdelbaki"
    },
    {
      "name": "P. Bandow"
    },
    {
      "name": "K. Y. Cheng"
    },
    {
      "name": "I. C. Grunwald Kadow"
    },
    {
      "name": "M. P. Nawrot"
    },
    {
      "affs": [
        0
      ],
      "name": "V. Rostami"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/74n644krgbsj2.bib",
  "cited_by": 0,
  "doi": "10.64898/2026.05.12.724368",
  "fulltext_url": "https://www.biorxiv.org/content/10.64898/2026.05.12.724368v4.full",
  "id": "74n644krgbsj2",
  "issns": [],
  "landing_url": "https://www.biorxiv.org/content/10.64898/2026.05.12.724368v4",
  "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": "W7160989819",
  "pdf_url": "https://www.biorxiv.org/content/10.64898/2026.05.12.724368v4.full.pdf",
  "pmcid": null,
  "pmid": null,
  "published": "2026-05-12",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/74n644krgbsj2.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:24:51Z",
      "native_id": "10.64898/2026.05.12.724368",
      "source": "biorxiv",
      "url": "https://www.biorxiv.org/content/10.64898/2026.05.12.724368v4"
    }
  ],
  "subjects": [
    {
      "field": "biology",
      "origin": "biorxiv:neuroscience",
      "subfield": "neuroscience"
    }
  ],
  "title": "Deep Representation Learning on Whole-Brain Population Dynamics Uncovers Geometrically Separable Neural Codes",
  "title_text": "Deep Representation Learning on Whole-Brain Population Dynamics Uncovers Geometrically Separable Neural Codes",
  "type": "preprint",
  "updated_at": "2026-10-30T12:24:51Z",
  "url": "https://research.draconitic.com/paper/74n644krgbsj2",
  "venue": "bioRxiv",
  "versions": []
}