Related Experiment Videos
Real-time ocular artifact suppression using recurrent neural network for electro-encephalogram based brain-computer
1Department of Biomedical Engineering, Faculty of Electrical Engineering, Iran University of Science and Technology, Narmak, Tehran-16844, Iran. erfanian@iust.ac.ir
Medical & Biological Engineering & Computing
|May 4, 2005
Summary
This study introduces an artificial neural network-based adaptive noise canceller (ANC) to remove electro-oculogram (EOG) interference from electro-encephalogram (EEG) signals, significantly improving signal quality.
Area of Science:
- Biomedical Engineering
- Signal Processing
- Artificial Intelligence
Background:
- Electro-encephalogram (EEG) signals are often contaminated by electro-oculogram (EOG) artifacts, primarily from eye movements and blinks.
- Conventional adaptive noise canceller (ANC) filters rely on linear models, which are suboptimal for predicting complex biomedical signal interferences.
- Accurate artifact removal is crucial for reliable EEG analysis in clinical and research settings.
Purpose of the Study:
- To develop and evaluate an artificial neural network (ANN)-based ANC filter for real-time removal of EOG interference from EEG signals.
- To compare the performance of an ANN-based ANC with conventional linear ANC methods for ocular artifact suppression.
- To demonstrate the effectiveness of the proposed method for both simulated and real-time EEG data.
Main Methods:
- A recurrent neural network (RNN) was employed to model the interference signals, capturing non-linear dynamics.
- Reference signals were generated using electrodes placed on the forehead and temple, capturing eye artifact data.
- The reference signal underwent low-pass filtering (moving averaged filter) before being applied to the ANC, and Matlab Simulink was used for implementation.
Main Results:
- The ANN-based ANC achieved a significant average improvement in signal-to-noise ratio (SNR) of up to 27 dB.
- Simulation results validated the technique's effectiveness across various SNRs of the primary EEG signal.
- Real-data experiments confirmed the successful removal of ocular artifacts, preserving the integrity of EEG recordings.
Conclusions:
- The proposed recurrent neural network-based adaptive noise canceller effectively removes ocular artifacts from EEG signals.
- This method offers a significant improvement in SNR compared to conventional techniques.
- The developed scheme is suitable for real-time and short-time electro-encephalogram recordings, enhancing diagnostic and research capabilities.