Chinese Scientists Map Single Neuron in Tri-Modal First

Chinese scientists at the Chinese Academy of Sciences (CAS) have achieved a world-first breakthrough in neuroscience: the complete tri-modal analysis of a single neuron, simultaneously mapping its function, structure, and molecular gene expression profiles. The milestone research, published in the journal Cell on June 18, 2026, overcomes a long-standing technical barrier that has prevented researchers from integrating these three critical data types at the single-cell level, according to ScienceNet.

The Neuroscience Challenge

Neurons are the basic computational units of the brain. To understand how the brain works, researchers need to know three things about each neuron: what it does (function), how it connects to other neurons (structure), and what genes it expresses (molecular profile). Historically, these three types of data came from different experimental systems that were incompatible with each other, creating what scientists call “data islands” — massive single-modal datasets that could not be integrated at the single-cell level.

While international efforts had developed dual-modal integration technologies (pairing function with structure, or structure with molecules), tri-modal integration remained a core technical bottleneck. As reported by The Paper, global brain initiatives had built vast databases at molecular, structural, and functional levels, but these datasets remained fragmented, lacking true homologous data from the same cell.

The IMC Platform Breakthrough

The research, led by principal investigators Wang Kai and Xu Shengjin at the CAS Center for Excellence in Brain Science and Intelligence Technology, was enabled by an independently developed platform called IMC (Imaging-based Multimodal Characterization). The platform integrates two key proprietary technologies:

• Multi-plane parallelized two-photon microscopy — enables whole-brain morphological reconstruction without cutting brain tissue
• Dual-color expansion fluorescence in situ hybridization (2cEASI-FISH) — precisely localizes gene molecules inside cells, detecting up to 6 genes simultaneously

According to Sina Finance, the experimental workflow follows three steps: first, recording real-time neuronal responses to visual stimuli and facial movements in awake mice; second, reconstructing the complete whole-brain neural fiber network of the same cell; and third, precisely mapping the distribution and quantity of all genes inside the cell. All three steps preserve the cell’s spatial position, allowing precise data alignment.

Key Discoveries

Using the IMC platform, the team obtained 207 neurons with in vivo functional activity and morphological information, and further obtained complete tri-modal datasets for 141 neurons, including intratelencephalic (IT) and pyramidal tract (PT) neurons in the mouse primary visual cortex.

The research yielded several significant findings, as detailed by Tencent News:

• New neuron subtype discovered: The study identified a previously undefined excitatory neuron subtype that simultaneously expresses molecular markers characteristic of inhibitory neurons, with unique response properties to visual stimuli — challenging existing classifications of neuronal cell types.
• Subcellular RNA localization matters: The spatial distribution of mRNA within cells constitutes a new molecular feature dimension that can help distinguish different projection types and functional types of neurons.
• Multi-modal prediction is superior: Combining morphology and gene expression data predicts neuronal function more accurately than any single modality alone.

Implications for Neuroscience and Medicine

The IMC platform fills a critical gap in global brain initiatives by providing a key experimental tool for decoding the fundamental computational logic of the brain. Wang Kai, corresponding author of the study, explained that once molecular and structural features of functionally specialized neurons can be identified through IMC, it can help researchers better target and regulate these neurons — with significant implications for both basic research and clinical applications.

According to the research team, the platform can be extended to more brain regions, cell types, and behavioral paradigms in the future. It can be used to study functional abnormalities, connectivity changes, and molecular state alterations in specific neuronal subtypes in brain diseases such as Alzheimer’s.

The homologous, cross-scale tri-modal data generated by IMC can also serve as high-quality ground-truth resources for evaluating and training multimodal integration methods, and as standard reference databases for brain-inspired AI research. The original paper is available in Cell.

What’s Next

The research positions Chinese neuroscience at the forefront of multimodal brain mapping technology. The team plans to extend the IMC platform to additional brain regions and behavioral paradigms, with the ultimate goal of helping establish a multimodal, multi-scale data ecosystem integrating molecules, cells, circuits, and behavior — a comprehensive brain knowledge base that could accelerate everything from basic neuroscience to targeted therapies for neurological disorders.

As Xu Shengjin, co-corresponding author who returned to China from HHMI Janelia Research Campus in 2021, noted through the CAS team’s statement, this platform opens new avenues for understanding the relationships between task-related computation, circuit structure, and molecular identity in the brain.