Abstract
Cortical neurons exhibit extreme diversity in gene expression as well as in morphological and electrophysiological properties. Most existing neural taxonomies are based on either transcriptomic or morpho-electric criteria, as it has been technically challenging to study both aspects of neuronal diversity in the same set of cells. Here we used Patch-seq to combine patch-clamp recording, biocytin staining, and single-cell RNA sequencing of more than 1,300 neurons in adult mouse primary motor cortex, providing a morpho-electric annotation of almost all transcriptomically defined neural cell types. We found that, although broad families of transcriptomic types (those expressing Vip, Pvalb, Sst and so on) had distinct and essentially non-overlapping morpho-electric phenotypes, individual transcriptomic types within the same family were not well separated in the morpho-electric space. Instead, there was a continuum of variability in morphology and electrophysiology, with neighbouring transcriptomic cell types showing similar morpho-electric features, often without clear boundaries between them. Our results suggest that neuronal types in the neocortex do not always form discrete entities. Instead, neurons form a hierarchy that consists of distinct non-overlapping branches at the level of families, but can form continuous and correlated transcriptomic and morpho-electrical landscapes within families.
Introduction
For decades, neuroscientists have sought to categorize the brain’s "parts list." Traditionally, neurons were classified by their shape (morphology) or how they fire (electrophysiology). More recently, high-throughput single-cell RNA sequencing (transcriptomics) has identified dozens of "t-types" based on gene expression. However, a major challenge remains: these classification systems do not always align. It has been technically difficult to measure all three modalities—genes, shape, and function—in the same cell. Without this link, it is unclear if a genetically defined cell type performs a unique, consistent role in the brain's circuitry. This study investigates whether neurons are discrete entities, like chemical elements in a periodic table, or if they represent a continuous landscape of biological diversity.
Methods
The researchers utilized Patch-seq, a multimodal technique that captures three data types from individual neurons:
- Patch-clamp recording: Measuring electrical properties in acute brain slices.
- Biocytin staining: Filling cells with dye to reconstruct 3D physical structures (axons and dendrites).
- RNA Sequencing: Extracting mRNA to profile gene expression using the Smart-seq2 protocol.
The primary dataset consisted of 1,329 high-quality neurons from the adult mouse primary motor cortex (MOp), mapped onto the BICCN (BRAIN Initiative Cell Census Network) reference atlas. To analyze this data, the team used sparse reduced-rank regression (RRR) to identify genes predicting firing patterns and t-SNE to visualize cell clustering across transcriptomic and morpho-electric spaces.
Key Results
The Hierarchy of Neuronal Diversity
The study reveals a hierarchical structure the authors liken to a "banana tree." While major neuronal families (e.g., Pvalb, Sst, and Vip) are distinct and non-overlapping in their morpho-electric phenotypes, individual transcriptomic types within these families form a continuous gradient. Within a family, clear boundaries are rare; if two neurons are transcriptomically similar, they likely share similar shapes and firing patterns, existing on a sliding scale rather than in discrete "bins."
Gene-to-Function Mapping
The RRR model demonstrated that a small set of just 25 genes—many related to ion channels—could predict electrophysiological features with a cross-validated $R^2$ of 0.38. This suggests that a limited genetic program governs much of a neuron's functional output.
Family vs. Type Distinction
Major transcriptomic families were almost perfectly separable using electrophysiological or morphological features alone. In contrast, individual "t-types" within the same family showed significant overlap. For example, the Sst Pvalb Calb2 type acted as a "bridge" between the Sst and Pvalb families in both genetic and functional spaces. Furthermore, in excitatory neurons, transcriptomic variation and physical depth in the cortex were tightly linked, forming a one-dimensional manifold following the brain's layers.
Discussion
While this study provides a comprehensive map, limitations remain. The data does not yet include long-range connectivity or in vivo functional characterization during behavior. Additionally, recovering full axonal structures from deep-layer neurons is technically challenging due to slicing.
The findings suggest that the field should move away from searching for an "exhaustive inventory" of discrete cell types. Instead, new theoretical models are needed to handle a hierarchical landscape that mixes discrete families with continuous variation. Whether this variation is hard-wired by genetics or tuned by local environment and activity remains a key question for future research.
Summary updated