Did I Do Well? Personalized Assessment of Trainees' Performance in Augmented Reality-Assisted Neurosurgical Training

The integration of Augmented Reality (AR) with the external ventricular drain (EVD) procedure improves catheter placement accuracy by providing guidance to surgeons through a visualization of the brain ventricular anatomy and other contextual information. However, the junior trainees often require feedback about their EVD performance during the AR -assisted EVD procedure. Therefore, we present an AR -assisted neurosurgical training tool for EVD that provides a personalized assessment of the trainees' EVD performance. We use a sensing-integrated brain phantom model to compute catheter placement accuracy, surgical task recognition to identify trainees' ongoing surgical tasks in real time, and provide accuracy scores, performance scores, and text feedback. Our user study with 16 medical and 12 non-medical students shows that the personalized assessment helped them improve the distance to target by 36.6% and 31.7% accordingly during a more challenging case of EVD procedure. Furthermore, most of the students agreed that the scores and assessment feedback given by our system were helpful in improving their EVD performance.