Detecting Movement and Lead-Popping Artifacts in Sleep EEG Data
Author Information
Author(s): Anandanadarajah Nishanth, Talukder Amlan, Yeung Deryck, Li Yuanyuan, Umbach David M., Fan Zheng, Li Leping
Primary Institution: National Institute of Environmental Health Sciences
Hypothesis
Can we develop a method to automatically identify artifacts in polysomnography EEG data caused by movement or loose leads?
Conclusion
The proposed algorithm effectively identifies artifacts in EEG data with a sensitivity of 80% and specificity of 91%.
Supporting Evidence
- The algorithm achieved a sensitivity of 80% and specificity of 91% in identifying artifacts.
- An open-source tool was developed for public use, available on GitHub and Docker.
- The method can be applied uniformly to the entire overnight sleep period or separately to NREM and REM sleep periods.
Takeaway
This study created a tool that helps find problems in brain wave recordings during sleep, making it easier to understand sleep patterns.
Methodology
The algorithm preprocesses EEG signals, applies multitaper spectral analysis, computes correlations between channels, and identifies outlier segments iteratively.
Potential Biases
Potential bias if the algorithm's performance is evaluated on a limited number of expert-annotated examples.
Limitations
The method may perform poorly if all channels experience artifacts simultaneously due to a common disturbance.
Participant Demographics
The study included data from 9641 in-laboratory PSG studies conducted between January 2019 and March 2023.
Statistical Information
P-Value
0.001
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website