Machine learning discovery of novel brainstem nuclei controlling orofacial behaviors

/ABSTRACT The brainstem medullary reticular formation governs vital motor behaviors, such as breathing, vocalization, swallowing, and chewing, as well as expressive movements of the face and mouth. Critical for understanding these neural circuits is identification and localization of reticular subpopulations controlling these myriad functions. A key obstacle to identifying these nuclei is the lack of clear cytoarchitectonic boundaries and molecular markers within the reticular formation delineating functional nuclei. Our central hypothesis is that functional subpopulations within the medullary reticular formation that play distinct roles in orofacial motor behaviors can be identified based on their gene expression profile and orofacial motor responses using machine learning clustering and cell type-specific, spatially-selective optogenetic manipulation. Testing this idea is a critical step towards understanding control and coordination of orofacial motor behaviors and their disruption in neurodevelopmental and neurodegenerative conditions, such as Rett Syndrome, Autism Spectrum Disorders, and Parkinson’s Disease, that cause atypical facial movements and speech and sometimes fatal disorders in swallowing and breathing. The approach comprises two AIMS: (AIM 1) Cluster medullary reticular formation into machine learning-defined nuclei (ML nuclei) that represent functionally-distinct orofacial control modules using a large scale, publicly available single cell spatial transcriptomic atlas from the Allen Institute; and (AIM 2) Map orofacial responses to systematic, cell type-specific optogenetic manipulation of medullary reticular formation for delineation of functionally distinct orofacial control modules. In preliminary data, we extract, process, cluster, and analyze a single cell spatial transcriptomic atlas from an adult mouse to show the feasibility and validity of our approach. We further demonstrate our technical capacity to carry out and analyze orofacial kinematics in response to reticular formation injections in awake, head-fixed adult mice. The expected outcome of the exploratory project is determination and functional validation of precise boundaries for previously identified and novel subpopulations underlying control of behaviors critical for survival. The overall impact of this proposal will be a highly innovative application of machine learning to neuroscience discovery, akin to an unbiased neuroanatomical “screen,” that may be applied to tissues throughout the body and the elucidation of fundamental principles governing the organization of the brainstem that will enable development of therapies for treating brainstem disorders and improve surgical interventions requiring a higher resolution understanding of brainstem structure and function. Project Number: 1R21NS142730-01A1 | Fiscal Year: 2026 | NIH Institute/Center: National Institute of Neurological Disorders and Stroke (NINDS) | Principal Investigator: Kaiwen Kam (+1 co-PI) | Institution: ROSALIND FRANKLIN UNIV OF MEDICINE & SCI, NORTH CHICAGO, IL | Award Amount: $399,700 | Activity Code: R21 | Study Section: Special Emphasis Panel[ZRG1 NINC-Q (01)] View on NIH RePORTER: https://reporter.nih.gov/project-details/11380193