Toward a probabilistic definition of neural cell types

Abstract

A classical view of cell type relies on a definite set of stable properties that are critical for brain functions. Single-cell technologies led to an extensive multimodal characterization of nervous systems and perhaps achieved one of Santiago Ramón y Cajal's dreams: to unveil a comprehensive view of the brain composition. While global analyses of brain structures highlight a degree of mesoscale stereotypy, a finer-scale resolution of brain composition shows significant variance in essential neural cellular phenotypes, including morphology, gene expression, electrophysiology, and connectivity. This highlights the need for novel conceptualization of the definition of a neural “cell type.” The challenge of modern neural classification is thus to integrate various distinct cellular properties into a unifying descriptor.