Hyoid bone-based sex discrimination among Egyptians using a multidetector computed tomography: discriminant function analysis, meta-analysis, and artificial intelligence-assisted study.
The study demonstrates that the hyoid bone can effectively aid in sex discrimination among Egyptians, with machine learning models achieving accuracies between 0.8667 and 0.933.
Where it sits
this study against the rest of the vilon (ke) corpusSummary and findings
The study measured sexual dimorphism of the hyoid bone in a sample of 300 Egyptian subjects using 3D MDCT imaging. It found that all measured dimensions of the hyoid bone were substantially larger in males compared to females. Discriminant functions combining four measurements achieved higher accuracy in sex prediction.
Abstract
The hyoid bone has been identified as sexually dimorphic in various populations. The current study is a forerunner analysis that used three-dimensional multidetector computed tomography (3D MDCT) images of the hyoid bone to examine sexual dimorphism in the Egyptian population. A total of 300 subjects underwent neck CT imaging, with an additional 60 subjects randomly selected for model validation. Ten hyoid variables were measured. Initially, the dataset was subjected to discriminant analysis to predict sex and the critical variables associated with sexual dimorphism. Subsequently, machine learning approaches were employed to enhance the accuracy of sex determination. The results indicated that all measured dimensions of the hyoid bone were substantially larger in males confront to females. Discriminant functions combining four measurements (major and minor axes of the hyoid body, the distance between the lesser horns, and hyoid bone length) achieved a higher accuracy of sex prediction compared to univariate functions. The accuracies of machine learning models ranged from 0.8667 to 0.933 with precision, recall, and F1-scores also showing improvements. These findings underscore the robustness and reliability of hyoid bone in sex discrimination among Egyptians, supported by both traditional statistical methods and machine learning approaches, and could prove invaluable in forensic cases.
Background
This paper addresses the sexual dimorphism of the hyoid bone, which has been previously identified in various populations. Understanding these differences is crucial for applications in forensic science and anthropology. The study aims to enhance the accuracy of sex determination using advanced imaging techniques and statistical methods.
Methods
The study utilized a sample of 300 subjects who underwent neck CT imaging, with an additional 60 subjects selected for model validation. Ten hyoid bone variables were measured, and discriminant analysis was performed to predict sex. Machine learning approaches were then employed to improve prediction accuracy.
Results
The results indicated that all measured dimensions of the hyoid bone were substantially larger in males compared to females. Discriminant functions combining four measurements achieved a higher accuracy of sex prediction compared to univariate functions. The accuracies of machine learning models ranged from 0.8667 to 0.933.
Interpretation
The findings suggest that the hyoid bone can be a reliable indicator of sex in the Egyptian population, aligning with previous studies on sexual dimorphism. However, while the statistical significance is noted, the clinical relevance of these findings in practical forensic applications remains to be fully established. Limitations such as the sample size and specific population may affect the generalizability of the results.
Key findings
- All measured dimensions of the hyoid bone were substantially larger in males compared to females.
- Discriminant functions combining four measurements achieved a higher accuracy of sex prediction.
- Machine learning models achieved accuracies ranging from 0.8667 to 0.933.
Limitations
- Sample size of 300 subjects may limit generalizability.
- Not reported in abstract.