6/9/2026 Beth Hart
A new system developed at Carle Illinois College of Medicine augments traditional simulation tools with mixed-reality overlays, spatial 3D guidance, and real-time hand tracking to help medical students master delicate procedures, such as lumbar puncture.
Written by Beth Hart
A new simulation platform is providing a richer learning experience for future doctors training to master high-risk medical procedures. The new system integrates artificial intelligence and machine learning with hands-on simulation to help medical students safely practice delicate procedures, even when an expert instructor isn’t available. It was recently unveiled by developers at Carle Illinois College of Medicine’s Jump Simulation Center.
Task trainers, manikins, and virtual reality simulations have been used for years to help medical students safely practice medical procedures without the risks of working with a real patient. These tools are crucial in helping students gain experience in performing medical procedures that require precision, such as a lumbar puncture (also known as a spinal tap), in which a needle is inserted into the patient’s back and spinal column to draw off spinal fluid, usually for diagnostic purposes. The Jump Simulation team’s goal was to make this training more consistent, repeatable, and accessible using advanced technology.
Their new system– called SpatialTutor – augments traditional simulation tools with mixed reality overlays, spatial 3D guidance, and real-time hand tracking to help students pinpoint the optimal insertion point and depth. Adaptive AI support allows students to ask questions and receive feedback as they practice, just as they would when an instructor is present.
“Evaluation with experts and novice learners showed high usability, strong engagement with spatial guidance and feedback, and increased learner confidence,” CI MED XR Industry Consultant Duo Wang said. The system was tested by more than 100 CI MED and pre-health students during training. “Our findings highlight the value of combining tactile practice with adaptive AI support to improve accessibility and consistency in procedural training,” he said.
The team plans additional validation testing to ensure students trained with the system are clinically ready. They also hope to expand the SpatialTutor’s capabilities beyond lumbar puncture to include other high-risk procedures.
Jump Simulation Center Manager Shandra Jamison says this work positions CI MED as a leader in applying the latest technology to medical simulation learning and research. “It’s a blueprint for scalable, standardized procedural training that can help medical education programs maintain quality even as clinical time, faculty time, and simulation staffing get tighter.” Jamison and her team, including Mae Vogel and Isaac Ryals, designed, facilitated, and helped with the validation of the new system.
The team also used the SpatialTutor system to study how to optimize when AI assistants intervene in immersive medical training. The researchers compared two systems: a proactive system in which AI prompted the learner before being asked, and another system that intervened only when the learner asked for help. The team found that proactive AI helped learners stay on task, but reactive AI allowed learners more autonomy.
“The key takeaway is that medical training does not benefit from a single fixed AI teaching style. Instead, the AI should adapt its level of guidance based on learner progress, cognitive load, and training moment — similar to how skilled instructors adjust their supervision,” Wang said, noting that the team is proposing guidelines for how and when to deliver AI support. “This allows simulations to support independent practice safely. Early learners can receive structured guidance, while advanced learners can train with minimal interruption — improving scalability without sacrificing educational quality,” he said.
The team’s work was presented at two prestigious international conferences earlier this year: the Association for Computing Machinery CHI conference on Human Factors in Computing Systems in Barcelona, Spain, and the IEEE AIxVR 2026, in Osaka, Japan.
Editor's note: Photos accompanying this article were provided by Duo Wang.