[Paper Review] Surgical task expertise detected by a self-organizing neural network map

Birgitta Dresp, Rongrong Liu+1 more|arXiv (Cornell University)|Jun 3, 2021

Soft Robotics and Applications13 references4 citations

TL;DR

This study proposes a self-organizing map (SOM) neural network to detect surgical task expertise by analyzing bimanual grip force profiles from robotic surgical simulators. The SOM, inspired by primate somatosensory neural architecture, accurately distinguishes experts from novices based on statistically significant differences in grip force variability, demonstrating its potential as a biomarker for surgical skill assessment.

ABSTRACT

Individual grip force profiling of bimanual simulator task performance of experts and novices using a robotic control device designed for endoscopic surgery permits defining benchmark criteria that tell true expert task skills from the skills of novices or trainee surgeons. Grip force variability in a true expert and a complete novice executing a robot assisted surgical simulator task reveal statistically significant differences as a function of task expertise. Here we show that the skill specific differences in local grip forces are predicted by the output metric of a Self Organizing neural network Map (SOM) with a bio inspired functional architecture that maps the functional connectivity of somatosensory neural networks in the primate brain.

Motivation & Objective

• To identify objective, quantifiable biomarkers of surgical expertise using biomechanical performance data from robotic surgical simulators.
• To address the challenge of subjective and inconsistent assessment of surgical skill in training environments.
• To develop a biologically inspired neural network model that captures functional connectivity patterns in somatosensory processing relevant to surgical task performance.
• To evaluate whether a self-organizing map (SOM) can detect and classify skill levels based on grip force variability during simulated endoscopic tasks.
• To establish benchmark criteria for true expert performance by comparing grip force dynamics between experts and novices.

Proposed method

• Data collection via a robotic control device designed for endoscopic surgery, recording bimanual grip force profiles during task execution.
• Grip force signals from expert and novice surgeons were analyzed for variability and temporal patterns during standardized simulator tasks.
• A self-organizing map (SOM) with a bio-inspired functional architecture was trained to model functional connectivity resembling primate somatosensory networks.
• The SOM processed raw grip force data to generate topological maps reflecting neural-like processing of tactile and motor feedback.
• The output metric of the SOM was used to classify and compare skill levels based on the spatial organization and stability of force patterns.
• Statistical analysis compared grip force variability between experts and novices to validate SOM-based classification performance.

Experimental results

Research questions

• RQ1 Can grip force variability during robotic surgical simulation distinguish between expert and novice performance?
• RQ2 To what extent does a self-organizing map (SOM) with a biologically inspired architecture detect and represent skill-specific differences in grip force dynamics?
• RQ3 Is the SOM output metric a reliable predictor of surgical task expertise based on biomechanical performance?
• RQ4 How do the topological patterns in the SOM reflect functional connectivity in somatosensory processing related to surgical skill?
• RQ5 Can the SOM model serve as a quantitative benchmark for surgical skill assessment in training contexts?

Key findings

• Significant differences in grip force variability were observed between true experts and complete novices during robot-assisted surgical simulator tasks.
• The self-organizing map (SOM) successfully captured skill-specific patterns in grip force data, with distinct topological organization for experts versus novices.
• The SOM output metric demonstrated a strong predictive capacity for identifying surgical expertise based on biomechanical performance.
• The functional architecture of the SOM, modeled after primate somatosensory networks, effectively mapped the underlying neural connectivity patterns related to skill execution.
• The study establishes a quantitative benchmark for expert-level performance using grip force profiling and SOM-based analysis.
• The results support the use of SOMs as a novel, biologically plausible tool for objective surgical skill assessment in training and evaluation settings.