Physics of Biology


Identification of Low Dimensional Chaos in Epileptogenic Seizure and Non-Seizure EEG using Solitary Wavelet Analysis

Authors: Sai Venkatesh Balasubramanian

The present work purports to the spectral and nonlinear analysis of intracranially recorded Electroencephalogram (EEG) time series data resulting in the identification and characterization of a low-dimensional chaos therein. Specifically, three categories of EEG time series data are considered: those recorded in non-epileptogenic regions during normal activity (‘healthy’ EEG), those recorded in epileptogenic regions during seizure activity (‘Seizure’ EEG), and those recorded in epileptogenic regions during normal non-seizure activity. In order to perform the spectral analysis efficiently, a new kind of wavelet, the solitary wavelet is proposed based on the hyperbolic secant function, where it is seen that this wavelet possesses vanishing higher moments with a negative logarithmic slope, thus translating to efficient detection of burst type signals without multiple levels of decomposition and reconstruction. The spectral analyses of the three categories of EEG data performed using Fourier and Wavelet analyses reveal that seizure EEG spectra, to a greater extent, and non-seizure epileptogenic EEG spectra, to a lesser extent, display prominent high frequency peaks, suggestive of resonant behavior. Nonlinear analysis using phase portraits too confirm the resonance oriented periodic orbits seen in Seizure EEG. Finally, quantitative analysis performed using Largest Lyapunov Exponents (LLE) and Fractal Dimension (D) reveal that both D and LLE values for seizure EEG are much lower than the healthy counterparts, with non-seizure epileptogenic EEG D and LLE values seen somewhere in between. The essence of these results is the observation of a distinct, uniquely identifiable low-dimensional chaotic behavior in EEG taken from epileptogenic region during non-seizure activities. This information can be used both as a preventive epilepsy diagnostic technique, as well as a post-surgical recovery assessment tool, and this forms the novelty of the present work.

Comments: 16 Pages.

Download: PDF

Submission history

[v1] 2015-10-25 08:06:53

Unique-IP document downloads: 42 times

Add your own feedback and questions here:
You are equally welcome to be positive or negative about any paper but please be polite. If you are being critical you must mention at least one specific error, otherwise your comment will be deleted as unhelpful.

comments powered by Disqus