Human Neocortical Glutamatergic Neurons Revealed Through Multimodal Profiling.

The human neocortex underlies higher cognition and is the engine of complex thought. Yet our understanding of its neuronal diversity is limited by sparse access to tissue, inconsistent sampling across studies, and a lack of multiple modality data. Although single-cell transcriptomic taxonomies are an important framework for characterizing cell type diversity, transcriptomic information alone cannot reveal the cellular properties that define neuronal computations. To address this, we performed Patch-seq, a method for collecting Morphology, Electrophysiology, and Transcriptomic data from a single neuron. We focused on glutamatergic, neocortical, excitatory neurons, the principal long-range projecting neurons of the cortex, and systematically integrated their morphoelectric features with transcriptomic identity. In combination with spatial transcriptomic data, we interrogated 39 of 42 transcriptomically-defined neuron types with a layer-centric perspective. Morphoelectric properties, such as cortical depth, apical dendrite structure, and excitability clearly distinguish transcriptomic subclasses and support many finer transcriptomic types. Morphoelectric properties are influenced by spatial location in supragranular layers, while deeper layers exhibit greater heterogeneity. Cross-species comparisons reveal conserved subclass organization but pronounced differences in apical dendrite arborization between mouse and human, and surprising similarities between human and macaque. Together, these datasets provide a unified multimodal reference that advances our understanding of human cortical circuitry and establishes a foundation for experimental and computational studies of human brain function and disease.