Dataset and Analysis of Long-Term Skill Acquisition in Robot-Assisted Minimally Invasive Surgery

Yarden Sharon, Alex Geftler, Hanna Kossowsky Lev, Ilana Nisky, Senior Member, IEEE · 2025

Read Dataset and Analysis of Long-Term Skill Acquisition in Robot-Assisted Minimally Invasive Surgery by Yarden Sharon, Alex Geftler, Hanna Kossowsky Lev, Ilana Nisky, Senior Member, IEEE (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-21591.txt.

Introduction

Yarden Sharon and co-authors release a dataset on how surgical residents learn robot assisted operating skills over months. Their 2025 paper follows eighteen residents for six months, in monthly training rounds set around a single twenty-six hour hospital shift. Each round ran before, during, and after the shift, with three dry lab tasks: moving rings, tying knots, and suturing.

The dataset holds 972 trials, with video synced to motion data, activity tracking, and scans of the suture pads. Residents improved steadily on all tasks. Learning and forgetting between shifts varied by task and measure, with hints that fatigue plays a role. The authors offer the data for work on better training schedules and surgical AI.

Catalog record

CREATORYarden Sharon, Alex Geftler, Hanna Kossowsky Lev, Ilana Nisky, Senior Member, IEEE
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE44 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-21591

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Dataset and Analysis of Long-Term Skill Acquisition in Robot-Assisted Minimally Invasive Surgery Yarden Sharon, Alex Geftler, Hanna Kossowsky Lev, Ilana Nisky, Senior Member, IEEE Abstract Objective: We aim to investigate long-term robotic surgical skill acquisition among surgical residents and the effects of training intervals and fatigue on performance. Methods: For six months, surgical residents participated in three training sessions once a month, surrounding a single 26-hour hospital shift. In each shift, they participated in training sessions scheduled before, during, and after the shift. In each training session, they performed three dry-lab training tasks: Ring Tower Transfer, Knot-Tying, and Suturing. We collected a comprehensive dataset, including videos synchronized with kinematic data, activity tracking, and scans of the suturing pads. Results: We collected a dataset of 972 trials performed by 18 residents of different surgical specializations. Participants demonstrated consistent performance improvement across all tasks. In addition, we found variations in between-shift learning and forgetting across metrics and tasks, and hints for possible effects of fatigue. Conclusion: The findings from our first analysis shed light on the long-term learning processes of robotic surgical skills with extended intervals and varying levels of fatigue. Significance: This study lays the groundwork for future research aimed at optimizing training protocols and enhancing AI applications in surgery, ultimately contributing to improved patient outcomes. The dataset will be made available upon acceptance of our journal submission. I Introduction In robot-assisted minimally invasive surgery (RAMIS), surgeons manipulate robotic joysticks to control the movements of surgical instruments inside the patient’s body via teleoperation [ 1 ] . RAMIS offers many benefits compared to open and laparoscopic surgery [ 2 , 3 , 4 ] . However, the technical skill level of the surgeon remains a critical factor in patient outcomes. Poor surgical skills increase the risk of postoperative complications, prolonged recovery times, and even mortality, highlighting the need for effective surgical training programs [ 5 ] . While significant effort is invested in optimizing the acquisition of technical skills for RAMIS surgeons [ 6 , 7 , 8 , 9 , 10 , 11 , 12 ] , the existing training protocols have substantial room for improvement. Previous surgical training studies have indicated that distributed practice, characterized by brief training sessions spread over multiple days or weeks, is more effective for skill acquisition than massed practice, characterized by prolonged training sessions concentrated within a single day or over two consecutive days [ 13 , 14 , 15 , 16 ] . However, surgical residents’ demanding schedules often prevent consistent training session scheduling, resulting in irregular intervals between practice opportunities [ 17 ] . In hospitals with a single surgical robotic system used for surgeries and training, residents may encounter even fewer opportunities for regular practice. The irregular and infrequent training intervals available to residents raise crucial questions about the nature of their learning: which aspects of skills are retained, which are lost, and which may continue to improve even in the absence of training. Although research has been conducted in this area, most existing studies have focused on short-term intervals between sessions, ranging from days to weeks [ 16 ] . While studies have begun exploring long-term surgical skill acquisition [ 18 , 19 ] , significant gaps remain in understanding these processes in greater detail. Understanding long-term processes could enable the development of better training protocols and virtual reality simulators that accommodate the constraints faced by surgical residents. In the broader field of motor learning, irregular training schedules similar to those experienced by surgical residents have not been extensively studied. Most motor learning research focuses on consistent practice schedules with training sessions spanning hours to days and primarily investigating simple movements [ 20 , 21 , 22 , 23 ] . This limitation means that the principles derived from motor learning studies may not directly apply to the complex and prolonged training required for mastering surgical skills. Consequently, there is a gap in our understanding of how irregular and long-term training schedules affect the acquisition and retention of motor skills in surgical contexts. While recent studies have expanded the understanding of long-term complex motor skill acquisition, such as video gaming [ 24 ] , they focus on different types of motor skills rather than the fine manipulation required in surgical contexts, highlighting a need for focused research in this area. Addressing this gap could enhance RAMIS training protocols and contribute valuable insights to the motor learning field to deepen our understanding of fine skill learning over extended periods. In massed practice, the learning curve typically exhibits rapid performance improvements during the initial training sessions, followed by a gradual slowdown until a plateau is reached [ 25 , 26 , 27 , 28 , 29 ] . When intervals are introduced between training sessions, we expect to observe several phenomena. (1) Retention, where participants maintain the skills acquired from the last session [ 30 ] . (2) Between-session forgetting, in which trainees lose some of the skills they acquired during the previous session before the next session occurs [ 31 ] . (3) Between-session (“offline”) learning, where skill improvement occurs during the interval between training sessions without direct practice [ 32 ] . Previous studies have focused on intervals within training programs that spanned days to weeks, leaving the rates of these different processes during extended training durations unclear. Here, we aim to investigate these phenomena over long-term training periods. Potentially, this will provide insights into optimizing protocols of skill acquisition and retention for surgical residents. When studying surgical interns’ long-term skill acquisition, it is essential to consider that their learning often occurs under fatigue conditions. The impact of fatigue on surgical performance has been investigated in various studies [ 33 , 34 ] ; however, significant methodological variations lead to inconclusive findings. Results range from demonstrating a negative effect of fatigue [ 35 , 36 , 37 ] , no effect [ 38 , 39 ] , to even a positive effect [ 40 ] . Additionally, motor learning research suggests that sleep influences skill acquisition [ 41 , 30 , 42 ] . Therefore, we also aimed to examine the effect of fatigue levels on the long-term processes involved in learning new surgical tasks. Given the complexity of skill acquisition in such challenging environments, it is critical to leverage available resources to enhance surgical training. Artificial Intelligence (AI) techniques have the potential to significantly advance the fields of surgical robotics and surgical training [ 43 , 44 , 45 , 46 , 47 ] . However, these fields lack the extensive datasets that have driven progress in other fields. One available resource is the JHU-ISI Gesture and Skill Assessment Working Set (JIGSAWS) [ 48 ] , which comprises three dry lab tasks performed by eight surgeons, with 101 recorded trials in total. This dataset is widely used in surgical robotics research for automating skill evaluation and gesture recognition [ 49 , 50 , 51 ] . However, models trained only on this dataset do not generalize well to clinical-like data [ 52 ] . A more recent dataset is the Robotics Surgical Maneuvers (ROSMA) dataset [ 53 ] . It comprises three dry lab tasks performed by 12 participants,

Read the complete text