Distinguishing Laparoscopic Surgery Experts from Novices Using EEG Topographic Features

Area of Science:

• Neuroscience
• Surgical Education
• Biomedical Engineering

Background:

• Distinguishing surgical expertise is crucial for training and patient safety.
• Electroencephalogram (EEG) offers a non-invasive window into brain activity during complex tasks.
• Objective skill assessment in laparoscopic surgery remains a challenge.

Purpose of the Study:

• To differentiate expert surgeons from novices using EEG topographic features during laparoscopic tasks.
• To compare the efficacy of a microstate-based common spatial pattern (CSP) analysis with linear discriminant analysis (LDA) against a topography-preserving convolutional neural network (CNN).

Main Methods:

• EEG data were collected from 8 expert surgeons and 13 novice medical residents performing laparoscopic suturing.
• Analysis involved microstate-based CSP with LDA and a 3D CNN (ESNet) model.
• Classification performance metrics included accuracy, sensitivity, specificity, F1 score, and Matthews Correlation Coefficient (MCC).

Main Results:

• The 3D CNN model (ESNet) achieved superior classification performance (accuracy > 98%) compared to microstate-based CSP with LDA (~90%).
• Experts exhibited distinct frontal and parietal cortical patterns, while novices showed frontal cortex involvement.
• The 3D CNN highlighted the parietal-temporal-occipital association region's importance in skill differentiation.

Conclusions:

• A topography-preserving 3D CNN model effectively differentiates surgical expertise from EEG data.
• Combining spatial and temporal EEG information significantly enhances classification accuracy.
• This approach holds promise for objective skill assessment in surgical training and practice.