Robot-Assisted Minimally Invasive Surgery (RMIS) allows surgeons to perform procedures through tiny incisions, thus reducing healing time compared to traditional open surgery. However, surgeons' technical competence remains essential for achieving positive patient outcomes. Therefore, in collaboration with several surgeons, we are investigating a range of methods to improve RMIS training.

We explored the effects of haptic feedback while training in RMIS. First, we investigated vibrotactile feedback using VerroTouch, a previously developed system that allows the surgeon to feel the instruments' vibrations at the operator's manipulators. Trainees who were provided with the vibrotactile feedback showed reduced workload and enhanced experience during both simulation and in real post-training surgeries compared to those who trained with no vibrotactile feedback []. Ongoing work is examining the impact of this feedback on learning processes. We also explored force feedback and created bracelets that squeeze the operator's wrists proportionally to the force applied by the surgical instruments. Participants applied significantly less force on the task materials when receiving the feedback [].

We identified two additional challenges when training in RMIS: the quantitative evaluation of surgical skills and the limits of existing simulators in facilitating self-directed learning.

Currently, surgical skill assessment is mainly conducted through manual video evaluation, which is time-consuming and subject to bias. We showed that surgical skill can be estimated through completion time, force applied to task materials, and vibrations generated at the surgical instruments []. Additionally, we explored the effect of an automated machine-learning-based skill-assessment approach on psychomotor skill development. We found that while some automated assessment tools may not enhance skill improvement rates, they can increase trainee self-awareness regarding their performance [].

Finally, effective training programs require experienced surgeons to provide feedback, but their time is limited and expensive. Therefore, there is a growing need for simulators that enable self-directed learning and provide real-time automatic formative feedback. To facilitate self-directed learning, we developed a platform that allows expert surgeons to record surgical procedures with multiple modalities and novice surgeons to replay and train with multimodal recordings, including visual, auditory, and vibrotactile feedback []. In ongoing projects, we are investigating the use of various signals to automatically monitor the task performance of trainees and deliver real-time customized visual and/or haptic feedback.

This research project involves collaborations with Gina L. Adrales (Johns Hopkins Medical Institute), Timothy Bernard (University of Maryland Baltimore County), Jeremy D. Brown (Johns Hopkins University), Eric Cao (Johns Hopkins University), Amy Chi (Johns Hopkins University), Sean P. Cohen (University of Pennsylvania), Kristoffel R. Dumon (University of Pennsylvania Hospital System), Joshua N. Fernandez (University of Maryland Baltimore County), Ernest D. Gomez (University of Pennsylvania; Beth Israel Deaconess Medical Center; Harvard Medical School), David I. Lee (University of Pennsylvania Hospital System), Sarah C. Leung (University of Pennsylvania), Sergio Machaca (Johns Hopkins University), Conor E. O'Brien (University of Pennsylvania), Zachary Patterson (Carnegie Mellon University), Noel N. Williams (University of Pennsylvania), and Brett Wolfinger (Johns Hopkins University).