Neuro-Robot Interaction in Robot-Assisted Surgery using EEG and Self-Supervised Graph Transformer
Debashis Das Chakladar, Foteini Simistira Liwicki, Rajkumar Saini
AI summary
Problem
Current robot-assisted surgery skill assessment relies on video analysis or subjective scoring, while existing EEG-based methods are computationally prohibitive due to high channel counts, hindering real-time clinical integration.
Approach
The method reduces EEG dimensionality using a metaheuristic channel selector, constructs brain connectivity graphs, and trains a self-supervised graph transformer via masked edge reconstruction to classify surgical proficiency.
Key results
- Achieved 96.60% accuracy in classifying novice, intermediate, and expert skill levels
- Reduced computational cost by selecting an optimal, non-redundant EEG channel subset via Harris Hawks Optimization
- Outperformed traditional machine learning and deep learning baselines
- Demonstrated task-wise accuracy scaling with proficiency for reliable objective auditing
Why it matters
This framework enables scalable, real-time cognitive monitoring and intelligent tutoring for robotic surgery, directly benefiting surgical training, certification, and patient safety.
Abstract
Robot-Assisted Surgery (RAS) represents a major frontier in the robotics community, blending precision automa- tion with human skill in high-stakes clinical environments. Evaluating surgeon performance in RAS is critical for training and certification, yet current methods rely heavily on video analysis or subjective manual scoring. This study presents a neuro-robotic interaction framework that uses Electroen- cephalography (EEG)-derived brain connectivity features to classify surgeons’ skill levels during RAS tasks. The high dimen- sionality of EEG data imposes substantial computational cost. Therefore, we first apply Harris Hawks Optimization (HHO) to select an optimal EEG-channel subset, reducing computational cost. Then, functional connectivity feature metrics are extracted from the reduced EEG channel set and used to construct brain graphs, which serve as input to a Self-Supervised Graph Trans- former (SSGT). The SSGT model is pre-trained via masked edge reconstruction to capture structural dependencies and fine- tuned for downstream skill-level classification. The proposed SSGT model achieves a classification accuracy of 96.60%, significantly outperforming both traditional machine learning and deep learning baselines. The label-efficient, structurally aware design of SSGT enables scalable and real-time assessment of surgical proficiency. This framework provides a foundation for intelligent robotic tutoring systems and generalizes to broader cognitive monitoring tasks in high-stakes human-robot interaction domains using EEG.