What are Proper Statistics for Steady-State Visual Evoked Potentials

Steady-state visual evoked potential (SSVEP) provides a low cost approach to learn about rapid neural activities in our brain. Furthermore, SSVEP provides a high signal-to-noise ratio and enables monitoring the neural responses to multiple simultaneously-visible stimuli. There are SSVEP experiments at which the SSVEP signal fluctuates independently of noise. The signal modulation will correlate to a brain process like the signal in experiments on attention which fluctuate probably as a function of the attention level. The response amplitude and phase of such experiments might have significantly different variances, and they might correlate with each other. Analyzing such SSVEP data using current statistics like phase coherence , Circ Toolbox , and T_circ^2 statistics might result in losing some information provided by amplitude and/or phase. We will show how to model the variation of Fourier components for further statistical analyses particularly when the signal fluctuates. We apply our approach to experiments on visual attention. We conducted an experiment with two visual stimuli flickering at 10 and 12 Hz frequencies, and the subject was asked to only attend to one stimulus in each trial. Our data show that there are occasions when the response amplitude changes in multiple electrodes probably as a function of the level of attention whereas the phase remains relatively unchanged. Any amplitude or phase fluctuation should not be entirely considered as a real noise. We show how to test T_circ^2 statistic, and how to model SSVEP signal with an ellipse using a three-parameter fit: the amplitude and the phase of the mean signal, and the angle that represents the correlation between amplitude and phase. The resultant fit is useful to represent the signal modulation corresponding to neural phenomena in one EEG channel and/or across multiple channels. It may provide insights into the neural mechanism(s) involved in the neural phenomena.

Keywords:

Visual evoked potentials , Fourier analysis , Phase , Attention