Machine learning based analysis and detection of trend outliers for electromyographic neuromuscular monitoring
Authors' conclusion
Engineered TOF-R trend features and the resulting Cost-Sensitive Logistic Regression (CSLR) models provide useful insights and serve as a potential first step towards the automated removal of outliers for neuromuscular monitoring devices.
Study context
- Population
- Patients under continuous intraoperative Electromyography (EMG)-based neuromuscular monitoring
- Participants
- 136 vs 388
- Comparison
- Four machine learning models for TOFR outlier detection
- Datex-Ohmeda EMG E-NMTn = 136
- TetraGraphn = 388
- Outcome
- Discriminative performance of a feature set for TOFR outlier analysis
Methodological detailstwo device datasets combined into a synthetic set, four models trained
- Two device datasets combined into a synthetic set
- Four models trained
Main finding
As reported in the abstract
Discriminative performance (model 1)
Model 1 F1 score (0.31, 0.50)
0.41
Model 1 average precision score, 95% CI (0.35, 0.60)
0.48
Further findings
Feature importance (random forest analysis)
- Engineered Train-of-Four Ratio (TOF-R) features proportionally more influential in model performance than basic features
Original publication
Machine learning based analysis and detection of trend outliers for electromyographic neuromuscular monitoring. J Clin Monit Comput. Published 4 April 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →