Variability analysis of visual evoked potentials in humans by pattern recognition in phase domain

Area of Science:

• Neuroscience
• Signal Processing
• Biomedical Engineering

Background:

• Single trial analysis of visually evoked potentials (VEP) is crucial for understanding brain responses.
• Variability in VEPs presents challenges for accurate analysis and interpretation.
• Existing models may not fully capture the dynamic nature of post-stimulus brain electrical activity.

Purpose of the Study:

• To present a novel approach for single trial VEP variability analysis.
• To introduce and experimentally verify a new convolution model for post-stimulus brain electrical activity.
• To propose and demonstrate a pattern recognition method in the signal phase domain for detecting time-locked transient signals.

Main Methods:

• Development of a new convolution model for post-stimulus brain electrical activity.
• Experimental verification using flash stimulus effects on electroencephalogram (EEG) amplitude and phase spectra.
• Application of a clustering algorithm in the two-dimensional unwrapped phase of EEG Fourier transform space for pattern recognition.

Main Results:

• The proposed convolution model was experimentally verified.
• The signal phase domain proved effective for detecting time-locked transient signals.
• Clustering in the EEG phase space successfully identified occipitally recorded VEPs in human subjects.

Conclusions:

• The novel approach offers a robust method for single trial VEP variability analysis.
• Pattern recognition in the signal phase domain is a promising technique for transient signal detection in EEG.
• The findings advance the understanding and analysis of brain electrical activity in response to visual stimuli.