Virtual Reality Simulators for Medical Training: Bridging the Gap Between Theory and Practice
Virtual reality simulators, combined with artificial intelligence (AI), have revolutionized medical training. These simulators offer a risk-free environment for learners to practice their technical skills, enabling medical students to enhance their expertise in complex operations, such as neurosurgery, before performing them on real patients. AI-powered tutoring systems, like the Virtual Operative Assistant (VOA), provide tailored feedback and personalized instruction, mimicking the guidance of a human instructor.
The Study: Virtual Reality AI Training vs. Human Instruction
A recent study conducted by the Neurosurgical Simulation and Artificial Intelligence Learning Centre at The Neuro of McGill University examined the effectiveness of AI instruction in neurosurgical training. The study involved 46 medical students performing a tumor removal procedure on a virtual reality simulator. Half of the students received instruction from the AI-powered VOA, while the other half received no feedback (the control group). The students’ performance was compared to benchmarks set by experienced neurosurgeons.
The Results: Unintended Outcomes and the Need for Human Instructors
The study revealed that AI-tutored students exhibited significant improvements in surgical technique. They caused 55% less damage to healthy tissues, demonstrated a 59% reduction in the average distance between instruments in each hand, and applied 46% less maximum force. These outcomes are crucial for patient safety.
However, the AI-tutored students also displayed negative outcomes. Their dominant hand movements were slower in velocity (50% lower) and acceleration (45% lower) compared to the control group, which affected the efficiency of their operations. Additionally, the AI-tutored group had a 29% lower speed in removing tumor tissue.
These unintended outcomes highlight the importance of human instructors in the learning process to ensure both safety and efficiency. While AI systems can enhance learning and focus intraoperative learning time, achieving mastery still requires guidance from human experts.
A medical student researcher at the Neurosurgical Simulation and Artificial Intelligence Learning Centre, Ali Fazlollahi, emphasizes that AI systems are not perfect. To optimize training and patient care, an optimal hybrid mode of instruction that combines AI and human expertise needs to be developed and tested.
Furthermore, Fazlollahi suggests that the implications of these findings extend beyond neurosurgery and apply to other fields of skill training, such as aviation, military training, and construction. AI should not replace human oversight but should be integrated with human experts to ensure excellence in training and patient outcomes.
The senior author of the study, Dr. Rolando Del Maestro, acknowledges the value of AI-powered intelligent tutors in training the next generation of neurosurgeons. However, he emphasizes the essential role of surgical educators in the development, application, and monitoring of AI systems to maximize their potential in enhancing neurosurgical skills and improving patient outcomes.