Optimizing Newborn Primate Endocast Reconstruction in 3DSlicer

Abstract

Natural and artificial endocasts have been used for decades to enable researchers to visualize the endocranial space and approximate the morphology of internal soft tissue structures. Producing an endocast of the newborn primate skull is made difficult by the large fontanelle spaces between growing bones. As there is no definitive endocranial boundary, these spaces have the potential to introduce error into both manual and semi‐automatic methods of endocast reconstruction.This study examines three methods of capturing endocranial volume (EV) from the endocasts of multiple species of newborn primates. For each specimen an EV was produced using three methods: a) manual, b) semi‐automatic and c) “patched” semi‐automatic segmentation. Manual segmentation is considered the “gold standard” of endocast reconstruction, however it lacks the efficiency of semi‐automatic methods. The “patched” method proposed here is a novel approach to fontanelle closure that is accomplished using the “Baffle Planner” module in 3DSlicer, an open‐access post‐processing software that is freely accessible to all researchers. The interosseous patches constructed from user‐defined landmarks are placed around the ossifying edges of the fontanelle and reflect the curvature of adjacent bones.In this study we provide guidance for optimizing the parameters of newborn primate endocast reconstruction in open‐access software. Our study found both quantitative and qualitative differences between the models produced by these methods suggesting the chosen method of reconstruction affects the resulting EV measurement. We conclude that a patched semi‐automatic approach with minimal baffles appears to reflect the anatomy with sufficient accuracy, while increasing efficiency and repeatability.