Measuring the Surgical Mind: Where AI Actually Adds Value

Artificial intelligence is everywhere in surgery this year. Headlines promise smarter robots and predictive algorithms, but the real breakthrough is subtler. In 2026, the hottest discussion in surgical education is not about replacing the surgeon but understanding how surgeons think.

A recent study in Surgical Endoscopy by Aseel F. Khanfar, Sanaz Motamedi, Shawn D. Safford, Jason Moore, Jessica Menold, and Scarlett Miller uses deep learning and eye-tracking to map how trainees see and interact with the surgical field. The system does not make decisions for the surgeon. Instead, it measures attention, anticipation, and motion efficiency, revealing cognitive patterns that were previously invisible.

The study examined medical students and residents performing the peg transfer task on both adult and pediatric laparoscopic simulators. Using a combination of eye-tracking glasses and a Mask R-CNN deep learning model, the researchers automatically identified tools and objects in the field and measured where participants focused their gaze and how quickly they moved instruments. The results were eye-opening: novices at the same level of training exhibited dramatically different patterns of attention and movement. High performers spent more time fixating on target objects, demonstrating anticipatory control, while lower performers tracked their instruments, reacting rather than planning. Tool speed and fixation on objects emerged as the strongest predictors of skill level, measurable from the very first trials.

Importantly, the study also used machine learning to predict visual behavior based on these metrics. Among several algorithms, Random Forest achieved over 83% accuracy in predicting a trainee’s visual attention pattern. This demonstrates that AI can provide early, objective indicators of skill progression long before traditional assessments would flag differences. In other words, the system can identify who is likely to excel and who may need additional guidance, giving trainers a powerful tool for personalized feedback.

Another surprising finding was that the type of simulator—adult or pediatric—did not significantly change visual or motion metrics. This suggests that early-stage novices focus on fundamental task strategies rather than adjusting immediately to anatomical differences. The implication is that early training can focus on universal principles of visual attention and motion control, with more nuanced anatomical adjustments introduced later.

The takeaway is clear: the value of AI in surgical training is not in automating decision-making but in making cognitive processes measurable. Eye-tracking combined with deep learning creates a real-time window into how surgeons think, allowing educators to intervene sooner and more precisely. In 2026, predictive, behavior-focused AI is emerging as a central tool in simulation labs and training centers, not to replace surgeons, but to amplify their skills.

You can explore the full findings and methodology in the original open-access article: From gaze to proficiency: deep learning-driven prediction of novice performance in laparoscopic training using AOI-dependent metrics in Surgical Endoscopy.

Reference: Khanfar AF, Motamedi S, Safford SD, Moore J, Menold J, Miller S. From gaze to proficiency: deep learning-driven prediction of novice performance in laparoscopic training using AOI-dependent metrics. Surgical Endoscopy. 2026;40:1925–1940. Published December 5, 2025.

Experience a smarter way to source endoscopy equipment—all in one place.

✔ Wide selection of trusted endoscopy brands

✔ Competitive pricing without compromising quality

✔ Fast, reliable access to the tools you need—when you need them

👉 Explore how EndoPro® 3D can transform your surgical technique.