Personalized modeling of TMJ sexual dimorphism biomechanics

Temporomandibular disorders (TMD) are a common cause of orofacial pain and frequently co-occur with headaches. They often begin in early adulthood, impair everyday tasks such as eating and speaking, and can lead to long, costly treatment journeys. A notable sex difference stands out: women are affected more often than men, yet the reasons remain insufficiently explained. This project addresses that gap and lays the groundwork for more targeted, person-centered care. At its core is a simple question: can subtle differences in jaw-joint shape between females and males change the forces that act during chewing and thereby increase the risk of pain? Put simply: if the hinge between skull and lower jaw is shaped a little differently, the load is distributed differently. We will make this interplay between shape and loading visible and relate it to symptoms. This is enabled by two recent approaches. Reduced-Coordinate Modeling focuses on the soft-tissue motions and deformations that matter most. This yields fast, personalized simulations that allow practical what-if checks. Reinforcement Learning controls the chewing muscles in the model using AI and improves the movements through feedback until they match realistic patterns. Together, this makes it possible to reproduce chewing faithfully on a computer, using suitable existing research data and without invasive measurements. Our goal is to identify clear, nameable features of joint shape that are associated with increased loading. The outcome can be a plain-language early-warning aid for clinical practice: if such features are present, a closer examination is warranted and treatment can be tailored to the individual situation rather than applied by routine. The public benefit is earlier detection, more targeted therapy and fewer unnecessary procedures. Data protection, ethics and open science are firmly embedded so that research and healthcare can benefit quickly and responsibly.