Graph-based machine-learning analysis of >30,000 excitatory neurons reconstruct…

Graph-based machine-learning analysis of >30,000 excitatory neurons reconstructed from the MICrONS millimeter-scale electron-microscopy volume defines a low-dimensional morphological bar-code that recovers continuous and discrete dendritic-morphology classes across mouse V1, AL and RL excitatory populations, supporting projection-class-linked recurrent architectures.

{
  "n": 30000,
  "doi": "10.1038/s41467-025-58763-w",
  "claim": "Graph-based machine-learning analysis of >30,000 excitatory neurons reconstructed from the MICrONS millimeter-scale electron-microscopy volume defines a low-dimensional morphological bar-code that recovers continuous and discrete dendritic-morphology classes across mouse V1, AL and RL excitatory populations, supporting projection-class-linked recurrent architectures.",
  "cite_key": "Weis2025",
  "evidence": "Graph-neural-network embedding of >30,000 excitatory pyramidal dendritic skeletons reconstructed from the MICrONS V1-HVA EM volume; unsupervised clustering and morphological feature analysis to recover excitatory subtype organization.",
  "effect_size": ">30,000 dendritic reconstructions support a continuous-plus-discrete morphological taxonomy",
  "text_access": "fulltext",
  "study_system": "mouse V1, AL, RL excitatory neurons; EM dendritic reconstruction (MICrONS)",
  "argument_role": "supporting",
  "replication_status": "single_study",
  "claim_source_sentence": "Here, we took a data-driven approach using graph-based machine learning methods to obtain a low-dimensional morphological “bar code” describing more than 30,000 excitatory neurons in mouse visual areas V1, AL, and RL that were reconstructed from the millimeter scale MICrONS volume.",
  "source_provenance_status": "ok",
  "replication_evidence_dois": [],
  "effect_size_source_sentence": "Here, we took a data-driven approach using graph-based machine learning methods to obtain a low-dimensional morphological “bar code” describing more than 30,000 excitatory neurons in mouse visual areas V1, AL, and RL that were reconstructed from the millimeter scale MICrONS volume."
}

[
  "paper:paper-4be2ac5ab52e"
]

Here, we took a data-driven approach using graph-based machine learning methods to obtain a low-dimensional morphological “bar code” describing more than 30,000 excitatory neurons in mouse visual areas V1, AL, and RL that were reconstructed from the millimeter scale MICrONS volume.

[
  {
    "ref": "paper:paper-4be2ac5ab52e"
  }
]

{
  "mode": "public_source_pointer_with_short_context",
  "notes": [
    "Local review repositories are read-only inputs.",
    "SciDEX stores paper metadata, structured evidence, file pointers, and short citation contexts; it does not copy full review prose."
  ],
  "source_commit_sha": "79ce062d54a924ce05953ec90aa9d26044d2b48f",
  "source_repository_url": "https://github.com/AllenNeuralDynamics/ComputationalReviewRecurrence"
}