Skip to content
English

Switch theme

Life & medical sciences · OECD FoS 3.1 · China MoE 071006 神经生物学

Neuroscience

神经科学

L2Local breakthroughs ↑ Rising Score 2.20 · Last reviewed 2026-10-22 · Next review 2027-01-19

BIGGEST BREAKTHROUGH 2024-10-02 Field leapAI · major

FlyWire: whole-brain connectome of the adult fly (Nature) ↗

First whole adult fly brain connectome (~140k neurons)

Why this level

AI enabled the adult-fly whole-brain connectome (FlyWire, Nature 2024) and speech-decoding BCIs (Nature 2023); 2026 saw many connectome foundation models, mostly preprints/benchmarks.

Research method 3Scientific discovery 2Practitioner adoption 2Education & curriculum 1
Research method ×0.33/4

Connectome reconstruction, behaviour tracking (DeepLabCut) and neural decoding rely widely on deep learning · source ↗

Scientific discovery ×0.352/4

BCI speech decoding and a digital twin of mouse visual cortex (Nature 2025) are sub-field breakthroughs · source ↗

Practitioner adoption ×0.252/4

Widespread as lab tools; clinical BCI is still experimental

Education & curriculum ×0.11/4

No course evidence found

S = 0.30·method + 0.35·discovery + 0.25·adoption + 0.10·education (education scores 0 when there is no evidence; if empty, the other weights are renormalised)

Sources

Composite 1.5–2.49: at least one peer-reviewed/officially verified AI-key result, or one dimension at 3; routine workflow largely unchanged. All levels →

Research directions

ConnectomicsBCI decodingNeural-data foundation models

Papers

Milestones 10· 1 leaps

Big glowing nodes are leaps: AI leap cool · Field leap warm · double ring = both

  1. Discovery

    OpenAI, PerturbAI and Scripps run 1,946 parallel GPT-Rosalind agents on a whole-brain in vivo CRISPR atlas, producing testable hypotheses in 32 minutes (bioRxiv)

    AI: one agent per perturbation interprets responses, searches literature and proposes mechanisms; experts preferred the multi-agent reports (announced by OpenAI on 10-21). Field: yields untested leads such as DCTN1 and neuronal cholesterol

    bioRxiv / OpenAIPreprintVendor claim
  2. Paper

    Seal: spatiotemporal AI modelling of gene regulation and variant effects in the developing human brain (Nat. Neurosci.)

    AI: transfer learning predicts expression/variant effects in data-poor transient cell types; Field: finds de novo regulatory burden in fetal excitatory neurons in autism

    Nature Neuroscience
  3. Model

    NiCLIP: contrastive language–brain-image model predicting cognitive tasks from activation maps (Nat. Commun.)

    AI: CLIP-style learning on the neuroimaging literature; Field: accurate group-level functional decoding, weaker on noisy individual maps

    Nature Communications
  4. Paper

    FedFound federated brain-connectome foundation model (npj Digit. Med.)

    Federated foundation model; method improvement

    npj Digital Medicine
  5. Paper

    Functional bipartite invariance in mouse V1 receptive fields via digital twin (Nat. Neurosci.)

    Digital twins used to generate testable hypotheses

    Nature Neuroscience
  6. Paper

    Foundation model of neural activity predicts responses to new stimulus types (Nature)

    First visual-cortex digital twin generalising across mice and stimuli

    Nature
  7. Paper

    MICrONS functional connectomics across mouse visual cortex (Nature)

    AI reconstructs a cubic millimetre of cortex aligned with function

    Nature
  8. DatasetField leap

    FlyWire: whole-brain connectome of the adult fly (Nature)

    First whole adult fly brain connectome (~140k neurons)

    Nature
  9. Paper

    High-performance speech neuroprosthesis, 62 words/min (Nature)

    62 words/min speech BCI, ~3× prior

    Nature
  10. Paper

    High-performance neuroprosthesis for speech decoding and avatar control (Nature)

    Near-natural-rate communication for a paralysed patient

    Nature
More papers & repos 9

Latest news 12

Data updated: AI × disciplines(UTC+8)