Single-neuron models linking electrophysiology, morphology, and transcriptomics across cortical cell types.

Area of Science:

• Computational neuroscience
• Systems neuroscience
• Cellular neuroscience

Background:

• Identifying cell types in brain circuits is crucial but challenging due to data modality discrepancies.
• Computational models offer a way to investigate cause-and-effect relationships and connect different experimental data types.

Purpose of the Study:

• To develop a computational workflow for generating bio-realistic single-neuron models.
• To link cellular electrophysiology and transcriptomics data for defining cortical cell types.
• To investigate the relationship between gene expression, ion channel conductances, and neuronal function.

Main Methods:

• Generated 9,200 single-neuron models with active conductances using a computational optimization workflow.
• Models were based on 230 in vitro electrophysiological experiments and morphological reconstructions from the mouse visual cortex.
• Compared model predictions with cell types defined by electrophysiology and single-cell transcriptomics.

Main Results:

• The generated models robustly represent individual experiments and cortical cell types, challenging previous assumptions.
• Predicted differences in specific ion channel conductances correlate with differences in gene expression.
• Model-predicted conductance differences explain observed electrophysiological variations between cortical cell subclasses.

Conclusions:

• This computational approach successfully reconciles disparate single-cell data modalities used for cell type definition.
• The study establishes causal links between gene expression, neuronal conductances, and functional electrophysiology.
• The findings provide a framework for understanding how molecular differences translate into circuit-level function.