From teeth to ethnicity: A neural network approach to predicting population of origin through dental traits and anomalies

Abstract

Abstract Background: This study aimed to investigate the prevalence of dental traits and anomalies in five North Indian populations (Khas Bodhi, Jaat, Khatri, Garhwali, and Gujjar) and predict the population of origin based on these traits and anomalies for forensic applications. Methods: We assessed dental traits and anomalies in 454 individuals through intraoral examination. Neural network analysis was employed to predict the population of origin based on a combination of dental traits and anomalies. Results: Shovel-shaped incisors exhibited the highest prevalence among the studied traits and anomalies, occurring in 65.4% of the sample. Moreover, shovel-shaped incisors were found to be the most important predictor of population. Neural network analysis indicated that the most accurate population prediction among the studied populations was for the Garhwali origin, achieving a recall rate of 78.3%. While this may appear relatively low, it is crucial to emphasise that the proposed method serves as a corroborative tool for various forensic investigations. Conclusion: This study suggests that dental traits and anomalies can be valuable in predicting the population of origin within Indian populations for forensic purposes. The work enhances the forensic identification process by providing an additional layer of evidence for consideration in identifying both individuals and their ethnic backgrounds. Further research is necessary to enhance the robustness of prediction models.