Doctor of Philosophy
Virtual reality surgical simulators have seen widespread adoption in an effort to provide safe, cost-effective and realistic practice of surgical skills. However, the majority of these simulators focus on training low-level technical skills, providing only prototypical surgical cases. For many complex procedures, this approach is deficient in representing anatomical variations that present clinically, failing to challenge users’ higher-level cognitive skills important for navigation and targeting. Surgical simulators offer the means to not only simulate any case conceivable, but to test novel approaches and examine factors that influence performance. Unfortunately, there is a void in the literature surrounding these questions. This thesis was motivated by the need to expand the role of surgical simulators to provide users with clinically relevant scenarios and evaluate human performance in relation to image guidance technologies, patient-specific anatomy, and cognitive abilities. To this end, various tools and methodologies were developed to examine cognitive abilities and knowledge, simulate procedures, and guide complex interventions all within a neurosurgical context. The first chapter provides an introduction to the material. The second chapter describes the development and evaluation of a virtual anatomical training and examination tool. The results suggest that learning occurs and that spatial reasoning ability is an important performance predictor, but subordinate to anatomical knowledge. The third chapter outlines development of automation tools to enable efficient simulation studies and data management. In the fourth chapter, subjects perform abstract targeting tasks on ellipsoid targets with and without augmented reality guidance. While the guidance tool improved accuracy, performance with the tool was strongly tied to target depth estimation – an important consideration for implementation and training with similar guidance tools. In the fifth chapter, neurosurgically experienced subjects were recruited to perform simulated ventriculostomies. Results showed anatomical variations influence performance and could impact outcome. Augmented reality guidance showed no marked improvement in performance, but exhibited a mild learning curve, indicating that additional training may be warranted. The final chapter summarizes the work presented. Our results and novel evaluative methodologies lay the groundwork for further investigation into simulators as versatile research tools to explore performance factors in simulated surgical procedures.
Armstrong, Ryan, "Performance Factors in Neurosurgical Simulation and Augmented Reality Image Guidance" (2016). Electronic Thesis and Dissertation Repository. 4288.