Prediction of Epileptic Seizures: A Statistical Approach with DCT Compression

Electroencephalogram (EEG) signal compression is an essential process to speed-up the medical signal transmission with reduced storage requirements, costs, and required bandwidth. The main objective of the present work is to compress the EEG signals and study the effect of the compression process on the epileptic seizure prediction. A lossy EEG data compression scheme using Discrete Cosine Transform (DCT) is applied, followed by seizure prediction. The used dataset includes healthy, pre-ictal, and ictal signals with multiple channels. The EEG signals are segmented to segments of 10 sec length. Also, the probability density functions (PDFs) for seizure prediction are measured, including amplitude, derivative, local media, local variance, and local mean PDFs. During the testing phase, only the selected bins of PDFs are used in the prediction process to identify each signal segment as pre-ictal or normal. A method of equal benefit decision fusion is carried out in the final prediction stage leading to a single sequence of decisions representing the activities of all segments. Relative to a patient-specific estimation level, this series after being filtered with a moving average filter is compared.

I. Introduction

Different brain activities occur based on the individuals’ activities, such as being awake or in different stages of sleep, actions and eye movements [1], [2]. Abnormalities in the electroencephalography (EEG) signals, which measure the brain electrical activity, occur due to different reasons, including epilepsy or seizure disorders, bleeding or hemorrhage, sleep disorders, Encephalitis, tumors, migraines, or head injury [3]. According to the frequency bands of the brain waves, they can be categorized into Alpha waves at the frequencies of 8 to 13 Hz, Beta waves at the frequencies greater than 13 Hz, Theta waves at the frequencies of 4 to 7 Hz, Delta waves at the frequencies up to 4 Hz, and Gamma waves at the frequencies of 26 to 100 Hz [4]. Since the analysis of such signals takes a very long time and much resources, compression of EEG signals for further analysis has become a must. Typically, there are two main types of compression techniques, namely lossless compression techniques and lossy techniques that remove the irrelevant information. The minimum bandwidth for the EEG equipment is specified by the International Electro-technical Commission (IEC) standardization to be from 0.5 Hz- 50 Hz. For acquiring EEG signals, hundreds of channels are recorded, which requires high data rates for transmission. Compression has a vital role to reduce the power consumption, and achieve high data rates. Several transform methods can be employed to transform the signal into other domains for extracting dominant features [5]. Such transformations include the Discrete Wavelet Transform (DWT) and the Discrete Cosine Transform (DCT). These transformation techniques provide information about the signal that can be used by various encoding techniques for compression. The compression of EEG signals is performed by neglecting small and irrelevant coefficients, then quantization, and encoding techniques. Thus, the EEG compression system includes several steps, namely applying a transform technique, thresholding of the transformed coefficients, quantization, and finally encoding.