COUNTEN – an AI-driven tool for rapid, and objective structural analyses of the Enteric Nervous System

AbstractHealthy gastrointestinal functions require a healthy Enteric Nervous System (ENS). ENS health is often defined by the presence of normal ENS structure. However, we currently lack a comprehensive understanding of normal ENS structure as current methodologies of manual enumeration of neurons within tissue and ganglia can only parse limited tissue regions; and are prone to error, subjective bias, and peer-to-peer discordance. Thus, there is a need to craft objective methods and robust tools to capture and quantify enteric neurons over a large area of tissue and within multiple ganglia. Here, we report on the development of an AI-driven tool COUNTEN which parses HuC/D-immunolabeled adult murine myenteric ileal plexus tissues to enumerate and classify enteric neurons into ganglia in a rapid, robust, and objective manner. COUNTEN matches trained humans in identifying, enumerating and clustering myenteric neurons into ganglia but takes a fraction of the time, thus allowing for accurate and rapid analyses of a large tissue region. Using COUNTEN, we parsed thousands of myenteric neurons and clustered them in hundreds of myenteric ganglia to compute metrics that help define the normal structure of the adult murine ileal myenteric plexus. We have made COUNTEN freely and openly available to all researchers, to facilitate reproducible, robust, and objective measures of ENS structure across mouse models, experiments, and institutions.