Towards the target and not beyond: 2D vs 3D visual aids in MR-based neurosurgical simulation

Abstract Neurosurgery increasingly employs Mixed Reality (MR) to support procedures such as External Ventricular Drain (EVD) placement, a task that requires high spatial accuracy, real-time anatomical orientation, and precise trajectory planning. To address these challenges, we present NeuroMix, an MR-based simulator designed to strengthen independent procedural skills without relying on intraoperative visual aids. After an initial pilot study with 10 participants, we conducted a full-scale experiment with 36 medical residents to evaluate three training modalities: CT-only (No Aid), CT with 2D overlay (2D Aid), and CT with 2D overlay plus animated 3D trajectory guidance (2D–3D Aid). All training was performed in a virtual environment using Meta Quest 3, followed by an unaided testing phase in which participants executed EVD placement on a physical phantom using real instruments. An additional control group of 12 residents received no MR training. Our findings show that participants trained with combined 2D–3D visual aids achieved a 44% improvement in targeting precision during unaided testing compared to the control group, significantly outperforming the other training modalities. Crucially, the combination of 2D and 3D visual aids improved retention of spatial knowledge without increasing cognitive workload or reducing usability. These results demonstrate that immersive MR training can effectively foster procedural skill retention in realistic, unaided execution, bridging a critical gap in current surgical education systems.