Nonlinear neural dynamics

Abstract

Brain activities measured by EEG exhibit complex behavior with nonlinear dynamic properties. Therefore, methods derived from nonlinear theory could contribute to the understanding of the EEG dynamics and the underlying brain processes. Until now, a number of nonlinear dynamic methods have been proposed. These methods reveal various nonlinear properties of the EEG signals. Among them, “complexity” and “entropy” are the widely used concept in the EEG analysis. Moreover, entropy-based measures have been applied into clinical practice of monitoring depth of anesthesia. This study selects three classical types of nonlinear dynamic measures (total 12), introducing their basic theory and giving examples of applying them in real EEG analysis. All the MATLAB codes can be downloaded from https://pan.baidu.com/s/1ro2VEJWdsb5X6dCtueGUOA with password: hf4x. Although previous studies compared the performance of various nonlinear dynamic measures at different situations, it is still improper to determine which measure is the best one. The measure selection in your study should take a number of factors into account, such as the parameter selection, robust to artifacts, compute consumption, the correlation of the nonlinear characteristics with the underlying neural process, and so on.