Visualization of High Dimensional Data in Comparative Morphology: Applications to Higher Primate Evolution

点击翻译按钮获取中文摘要

The visualization and analysis of animal structure in an evolutionary context has progressed from simple "eyeballing" to traditional morphometrics (multivariate analysis of caliper measurements) and is now entering the new regime of geometric morphometrics. In this approach, the three-dimensional (3D) coordinates of landmarks and intervening lines or surfaces are obtained with a digitizer or from CT or similar scans. Thus, the spatial relationships among these points, lines and surfaces areexpressly preserved. These coordinate data can be visualized as 2D and 3D images, manipulated, and statistically analyzed in high-dimensional shape space. The researchers will develop a set of database, statistical and manipulative tools to allow the rapid visualization and comparison of morphology among individuals and populations, across sex and age differences, including both extant and extinct samples, effectively forming a database of virtual morphology for computerized analysis. Reconstruction of damaged fossils from undistorted relatives, morphing of ancestral species into visualized descendants, exploration of endocranial features, tracing the evolution of brain neurochemical pathways using functional images and stereologic data across species, and combined analyses of brain and bony skull form are among the goals of this project. The focus will be on the skull and brain of the Anthropoidea or "higher" primates (humans, apes and monkeys), but the tools and approaches will be useful to a wide audience of zoologists and paleontologists. Based at the American Museum of Natural History, the team benefits from its vast collections and available visualization infrastructure, as well as the unique opportunity to project our displays to colleagues and the interested public on the 4-square-meter walls of the Reality Room immersion environment and the 21 meter Digital Dome of the newly renovated Hayden Planetarium.