Modelling the EEG based event-related brainwaves using statistical times series

Abstract

Objective: The aim of this study was to test whether the alpha oscillation in the auditory event-related activity of the brain can be explained using an Autoregressive Integrated Moving Average (ARIMA) model. Methods: The sample consisted of 17 (11 males, 6 females), right-handed, healthy, volunteer participants. Event-related activity was recorded under the omitted stimulus paradigm. Alpha response was obtained through digital filtering. The alpha response of each participant was separately analyzed. Results: The ARIMA (0,2,1) model was found to be applicable to alpha response that was evoked by stimulus omission. ARIMA (0,2,1) shows that this activity is short-lasting and it fluctuates stochastically. Conclusion: The present study showed that the alpha responses of single participants obtained under the omitted stimulus paradigm were best modeled by the ARIMA (0, 2, 1) type of time series model.