NEC Seminar: Friday, November 14th, 9:00 A
Speaker: Mallika
Advisors: Shoffstall, Jenkins, Wilson
Abstract:
Vagus nerve stimulation (VNS) shows therapeutic potential across neurological and inflammatory disorders, yet outcomes vary widely across individuals. These differences are thought to arise from subject-specific vagal anatomy, including fascicle organization and perineurium structure, which determine fiber recruitment patterns. To capture this variability, we developed a high-throughput, machine-learning-based segmentation pipeline for hematoxylin and eosin (H&E) histology of the human vagus nerve using data from the NIH-funded REconstructing Vagal Anatomy (REVA) project. The pipeline automates the identification of fascicles, perineurium, and epineurium, applies anatomical quality control, and quantifies key morphometric features such as fascicle size and perineurium thickness. This presentation focuses on how this ML-driven segmentation framework enables the construction of a digital atlas of the human vagus nerve and supports systematic extraction of micro-anatomical features critical for modeling, electrode design, and understanding variability in VNS outcomes.