Empirical mode decomposition improves detection of SSVEP

Area of Science:

• Neuroscience
• Signal Processing
• Biomedical Engineering

Background:

• Steady State Visual Evoked Potentials (SSVEPs) are crucial for quantifying attention-related neural activity to visual stimuli.
• Accurate detection of SSVEPs is essential for brain-computer interfaces and cognitive state monitoring.
• Traditional methods may face limitations in SSVEP detection accuracy and efficiency.

Purpose of the Study:

• To investigate the efficacy of Empirical Mode Decomposition (EMD) in improving SSVEP detection accuracy and rate.
• To compare the performance of EMD-assisted SSVEP detection using Gabor transform and Canonical Correlation Analysis (CCA).
• To explore the utility of harmonic responses in enhancing SSVEP detection.

Main Methods:

• Electroencephalogram (EEG) signals were decomposed into Intrinsic Mode Functions (IMFs) using EMD.
• IMF components relevant to SSVEPs were selected for target frequency detection.
• Target frequencies were identified using Gabor transform and CCA, with and without EMD pre-processing.

Main Results:

• EMD significantly improved SSVEP recognition accuracy when combined with the Gabor transform, even with reduced Gaussian window lengths.
• EMD showed minimal impact on the detection performance of Canonical Correlation Analysis (CCA).
• Incorporating harmonic responses of the target frequency enhanced SSVEP detection for both Gabor transform and CCA methods.

Conclusions:

• EMD is a valuable pre-processing technique for enhancing SSVEP detection accuracy, particularly when using the Gabor transform.
• Harmonic frequency analysis offers a robust strategy to boost SSVEP detection performance across different analytical methods.
• The findings suggest potential improvements for attention quantification and brain-computer interface applications utilizing SSVEPs.