Features and networks of the mandible on computed tomography

• Published in: Royal Society open science Vol. 11; no. 1; pp. 231166 - 10
• Main Authors: Pham, Tuan D., Holmes, Simon B., Patel, Mangala, Coulthard, Paul
• Format: Journal Article
• Language: English
• Published: England The Royal Society Publishing 01.01.2024; The Royal Society
• Subjects: Anthropology; Artificial intelligence; Cluster analysis; Computed tomography; Computer Science and Artificial Intelligence; Emergency medical care; Emergency medical services; Females; Forensic sciences; Fractures; Geostatistics; image processing; Interdisciplinary aspects; Interdisciplinary research; Interdisciplinary studies; Jaw; Males; Mandible; Medical imaging; Morphology; Network analysis; network theory; Network topologies; nonlinear data analysis; Nonlinear dynamics; Nonlinear systems; Skeleton; Sleep apnea; Time series; Tomography; Trauma; computed tomography; image processing; mandible; network theory; nonlinear data analysis; artificial intelligence
• ISSN: 2054-5703; 2054-5703
• DOI: 10.1098/rsos.231166

The mandible or lower jaw is the largest and hardest bone in the human facial skeleton. Fractures of the mandible are reported to be a common facial trauma in emergency medicine and gaining insights into mandibular morphology in different facial types can be helpful for trauma treatment. Furthermore, features of the mandible play an important role in forensics and anthropology for identifying gender and individuals. Thus, discovering hidden information of the mandible can benefit interdisciplinary research. Here, for the first time, a method of artificial intelligence-based nonlinear dynamics and network analysis are used for discovering dissimilar and similar radiographic features of mandibles between male and female subjects. Using a public dataset of 10 computed tomography scans of mandibles, the results suggest a difference in the distribution of spatial autocorrelation between genders, uniqueness in network topologies among individuals and shared values in recurrence quantification.