ABSTRACT
Electroencephalograms or EEGs are recordings that represent the dynamics of the electrical activity in the brain over time. These types of signals are frequently used by neuroscientists to study the brain behavior in clinical and nonclinical settings. Typically, EEG activity is broken up into four frequency bands, referred to as the delta (0–4 Hz), theta (4–8 Hz), alpha (8–13 Hz), and beta (above 13 Hz) bands. Different kinds of activity/behavior may induce brain waves in one or more of these frequency bands. For instance, the normal resting EEG usually consists of brain activity in the alpha and beta bands (Dyro, 1989). In the context of characterizing cognitive fatigue, previous EEG studies have suggested that the fatigue is associated with an increase in the theta band power in mid-frontal locations, accompanied by an increase in the alpha band power in parietal locations (Trejo et al., 2006). Therefore, developing and implementing models and methodology that allow us to infer the latent quasiperiodic structure underlying EEG signals over time are the key to understanding brain activity.
