Related Experiment Video
Updated: May 23, 2025
10:35
Best Current Practice for Obtaining High Quality EEG Data During Simultaneous fMRI
Published on: June 3, 2013
32.3K
How much data is enough? Optimization of data collection for artifact detection in EEG recordings
Lu Wang-Nöth1,2, Philipp Heiler1, Hai Huang2
1brainboost GmbH, Augsburgerstraße 4, 80337 Munich, Germany.
Journal of Neural Engineering
|March 10, 2025
Summary
This study optimizes data collection for electroencephalography (EEG) artifact removal using deep learning. We reduced necessary electromyography (EMG) artifact tasks, improving data efficiency for cleaner EEG signals.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Signal Processing
Background:
- Electroencephalography (EEG) is cost-effective but suffers from artifacts, reducing signal quality.
- Electromyography (EMG) artifacts are challenging biological artifacts impacting EEG analysis.
- Current data collection for artifact cleaning lacks systematic justification and efficient utilization.
Purpose of the Study:
- To propose an optimized data collection design for EEG artifact cleaning using deep learning.
- To minimize data collection efforts while maintaining artifact cleaning efficiency.
- To provide a systematic, quantitative approach for selecting artifact types and quantities.
Main Methods:
- Applied binary classification to differentiate artifact and non-artifact epochs using three neural architectures.
- Developed a deep learning-based artifact detection system.
- Focused on optimizing electromyography (EMG) artifact data collection.
Main Results:
- Reduced the number of required EMG artifact tasks from twelve to three.
- Decreased repetitions of isometric contraction tasks from ten to as few as one.
- Demonstrated efficient data collection for improved artifact cleaning.
Conclusions:
- The proposed method offers a systematic and dynamic quantitative approach to biological data collection.
- Provides clear justifications for artifact selection and quantity, guiding future research.
- Aims to enhance the effectiveness and economy of EEG and EMG research data collection.
