Evaluating Effect of Microsoft Hololens on Extraneous Cognitive Load During Simulated Cervical Lateral Mass Screw Placement

06 Nov 2020 - Arun Balajiee

Talk Date: 11/06/2020

Talk Speaker: Dmitriy Babichenko

Today’s talk was about the ongoing research work by Dr. Babichenko et al. on utilizing the salient features of AR and VR to be able to model human spinal cord to help surgeons with complex surgeries such as the operations on the cervical segment of human vertebral column. Dr. Babichenko chose the Microsoft Hololens technology to implement their AR models. Using the principles of Extraneous cognitive load theory (abbrv. ECL; the cognitive activities not related to the task at hand), they attempt at identify the effect of use of a AR headsets towards the ECL of surgeons performing these complex surgeries. Further, they try to come up with objective measures in the incremental changes in ECL with increase in experience with the AR technology. The Spine model design for AR modelling is first created using 3D printing of real CT scans (courtesy UPMC). These spines were placed in 3D printed mannequins to simulate the setup of the spinal surgery. The final 3D model in AR is then constructed using Unity with near accurate positioning of different points that are marked when the surgery is performed. The study is conducted within-subjects and the priming bias is handled by grouping random sample of participants into two groups, group A of surgeons performing the surgerywith hololens then without and group B without hololens and then with it. Post-study questionnaire was designed using SURG-TLX model. Using Wilcoxon’s Signed Rank test they measure the significance in difference in median times in the operations of the surgery, distance of the marked postiions to where the surgery is actually performed and median values of other such assessment metrics. No dignificant difference in the groups A and B was found.

This work has a significant potential in the space of multi-modal data collection, which could aid in building robust ML and Vision models that assist surgeons while performing complex surgeries through AR and reduce the ECL that they experience while performing these extensive and delicate operations on the human spine. Further, AR models could help potentially prevent excessive exposure to radiation that is often the case when performing spinal surgeries. A segment of the human spine is also exposed to external atmosphere when the surgery is performed and this could also cause infections. All this can be avoided if a successful model is built in AR. There is a lot of promise and potential in this endeavour and I am excited to see how the project turns to be one of the coolest and best research performed at University of Pittsburgh which involves the collaborations of departments in several fields - computer science, medicine, psychology, neurosurgery!