Independent component analysis-based method for electroencephalogram analysis during photic stimulation

Area of Science:

• Neuroscience
• Signal Processing

Background:

• Intermittent photic stimulation (IPS) induces phase- and time-locked electroencephalogram (EEG) activities.
• Averaging techniques capture evoked responses, but analyzing time-locked changes in ongoing EEG for event-related synchronization/desynchronization (ERS/ERD) remains challenging.
• Existing methods provide only spectral estimates of time-locked activities.

Purpose of the Study:

• To introduce a new independent component analysis (ICA)-based method for analyzing EEG during IPS.
• To assess the capability of the proposed method in estimating ongoing EEG activity in both time and frequency domains.
• To enable precise separation of evoked responses from time-locked changes in ongoing EEG for ERS/ERD studies.

Main Methods:

• An ICA-based approach was developed for EEG analysis during IPS.
• Artificially generated sinusoidal functions at stimulation frequency and harmonics were combined with acquired EEG signals.
• The combined signal vector was processed using the ICA algorithm.

Main Results:

• The proposed ICA method was validated using simulated data.
• Application to occipital EEG signals demonstrated accurate estimation of ongoing EEG activity.
• The method effectively provided estimations in both time and frequency domains.

Conclusions:

• The ICA method precisely estimates the power spectrum of ongoing EEG activity, comparable to coherence-based methods.
• A key advantage is its ability to also obtain the time-domain ongoing EEG activity.
• This method facilitates ERS/ERD studies by distinguishing phase-locked evoked responses from time-locked ongoing EEG changes.