Nonlinear EEG dynamics during imagined self-paced movements

The majority of studies devoted to reveal electrophysiological correlates of words and sentences comprehension, imageability and remembering are based on the event-related potentials and frequency synchronization in different narrow frequency bands. These linear methods reveal some patterns of EEG activity in time and frequency domain. Having in mind that the activation of many cortical structures is a result of mass of nonlinearly interconnected neurons, the linear methods seem to be insufficient to discover the complexity of the information transfer. We revealed recently nonlinear dynamic transients in EEG, long before real performance of goal-directed voluntary movements with different temporal and spatial distributions over frontal, sensorimotor and parietal cortical areas (Popivanov and Dushanova, 1999). The aim of this study was to establish whether similar behavior of the nonlinear characteristics exists when the subject imagines movements of a given type. The Kolmogorov entropy computed over time after the sentence end proved to be an useful characteristic that complement the linear methods.