Doctoral Dissertation Research: Estimating adult age-at-death from the pelvis

骨骼年龄估计方法被用于深入了解过去和现在的人口和个人生活。一般来说，现有的估计方法是基于随年龄增长而出现的接合表面的可见磨损。本博士论文研究项目通过使用扫描技术和深度学习模型，生成和验证基于骨盆多个关节表面的年龄估计新方法，扩展了当前的死亡年龄估计能力。该项目的结果有助于1)减少基于骨骼标记估计年龄的主观性，2)在从业人员的年龄估计中提供一致的结果，无论经验程度如何，以及3)提高老年人年龄范围估计的精度。该项目支持本科生和研究生在STEM领域的指导、培训和研究机会，开发开源深度学习模型，并为从业者、研究人员和普通教育社区提供在线培训研讨会。传统的年龄估计的定性方法是有问题的，因为它们严重依赖于从业人员的经验水平，以准确评估随年龄发生的变化。此外，传统方法无法区分老年人（例如50岁以上的个体）的年龄相关特征，导致年龄范围很广，无法详细重建与人口结构、健康和环境有关的成人生活史。该项目将统计分析和人工智能建模应用于骨盆关节的3D扫描和照片测量，为标准化骨骼年龄估计方法提供了重要的方法学资源。改进的年龄估计方法增强了对过去和现在人口统计数据的理解，并能够更彻底地了解骨盆关节表面的衰老变化。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Skeletal age estimation methods are applied to gain insights into past and present populations and individual lives. Existing estimation methods are based, in general, on the visible wear of joint surfaces that occurs with age. This doctoral dissertation research project expands current age-at-death estimation capabilities by generating and validating new methods of age estimation based on multiple joint surfaces in the pelvis, using scanning technology and deep learning models. The results of this project help to 1) reduce subjectivity in estimating age based on skeletal markers, 2) provide consistent results in estimation of age by practitioners regardless of degree of experience, and 3) increase the precision of age range estimates for older adult populations. The project supports undergraduate and graduate mentoring, training, and research opportunities in STEM, the development of open-source deep learning models, and online training workshops for practitioners, researchers, and general education communities.Traditional qualitative methods for age estimation are problematic in that they are heavily reliant on the practitioners’ level of experience for accurate assessment of the changes that occur with age. Additionally, traditional methods cannot differentiate age-related features in older adults (e.g., individuals over 50 years), resulting in broad age ranges that do not allow a detailed reconstruction of adult life history in relationship to aspects of population structure, health, and the environment. This project applies statistical analyses and AI modeling to measurements from 3D scans and photographs taken of the pelvic joints and provides an important methodological resource for standardizing skeletal methods for age estimation. Improved age estimation methods enhance the understanding of past and present population demographics and enable a more thorough understanding of senescent changes to pelvic joint surfaces.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.