A multivariate model of subadult age estimation

该项目将利用全身CT扫描和复杂的统计模型，探索现代美国儿童的实足年龄和牙齿形成，长骨生长和骨骺融合之间的关系。更好地了解这些关系不仅对对儿童成长和生活史模式感兴趣的人类生物学家很重要，而且对生物考古学家和法医人类学家也很重要，他们希望更准确地估计骨骼样本中的死亡年龄。NSF的资金将支持更大项目的牙科部分的数据收集和统计咨询。该项目将为现代儿童提供一个大规模的横截面样本的免费牙齿和骨骼数据。一个年龄估计软件程序，并介绍该软件的研讨会，也将开发，并将在人类学内外的一系列研究领域的兴趣。亚成体年龄的估计是基于记录的生长和发育模式，被认为是亚成体生物学特征中最准确的估计。然而，亚成体年龄估计的主要局限性是缺乏大的，现代的骨骼参考样本和依赖于历史，主要是单变量的方法。大多数方法采用逐骨或逐齿的方法，并且不提供结合来自整个亚成体骨骼的标记的手段。在该项目中，研究者将有权访问去识别的体检医师数据。他们将从出生到15岁（n = 1500）的个体的多层计算机断层扫描（MSCT）和Lodox Statscan X射线摄影图像中收集有关牙齿形成、骨骺融合和长骨生长的信息，以开发多变量年龄估计技术。将在准确性、偏倚和预测能力方面比较单变量和多变量模型，并将数据纳入KidStats软件，目的是为亚成体提供更准确、精确和无偏倚的年龄估计。免费获取的数据将为本科生和研究生以及人类学和相关领域的研究人员提供一系列研究和合作机会。该项目由NSF和国家司法研究所共同支持。

This project will explore the relationships among chronological age and dental formation, long bone growth and epiphyseal fusion in modern US children, using full body CT scans and sophisticated statistical models. A better understanding of these relationships is important not only for human biologists interested in childhood growth and life history patterns, but for bioarchaeologists and forensic anthropologists who want more accurate estimates of age at death in skeletal samples. NSF funds will support data collection and statistical consulting for the dental portion of the larger project. The project will result in freely accessible dental and skeletal data for a large, cross-sectional sample of modern children. An age estimation software program, and workshops to introduce the software, will also be developed and will be of interest across a range of research areas within and beyond anthropology. Subadult age estimations are based on documented patterns of growth and development and are considered the most accurate estimation in the subadult biological profile. However, the major limitations of subadult age estimation are the lack of large, modern skeletal reference samples and the reliance on historical, primarily univariate methodological approaches. Most methods utilize a bone-by-bone or tooth-by-tooth approach and do not provide a means to incorporate markers from the entire subadult skeleton. In this project, the investigators will have access to de-identified medical examiner data. They will collect information on dental formation, epiphyseal fusion and long bone growth from multi-slice computed tomography (MSCT) and Lodox Statscan radiographic images for individuals between birth and 15 years (n = 1500), to develop a multivariate age estimation technique. Univariate and multivariate models will be compared in terms of accuracy, bias, and predictive ability, and the data will be incorporated into KidStats software with the goal of providing more accurate, precise, and unbiased age estimates for subadults. The freely accessible data will offer a range of research and collaborative opportunities for undergraduate and postgraduate students as well as researchers in anthropology and allied fields.This project is jointly supported by NSF and the National Institute of Justice.