Classification of Epileptic EEG Signals using Time-Delay Neural Networks and Probabilistic Neural Networks

Abstract

The aim of this paper is to investigate the performance of time delay neural networks (TDNNs) and the probabilistic neural networks (PNNs) trained with nonlinear features (Lyapunov exponents and Entropy) on electroencephalogram signals (EEG) in a specific pathological state. For this purpose, two types of EEG signals (normal and partial epilepsy) are analyzed. To evaluate the performance of the classifiers, mean square error (MSE) and elapsed time of each classifier are examined. The results show that TDNN with 12 neurons in hidden layer result in a lower MSE with the training time of about 19.69 second. According to the results, when the sigma values are lower than 0.56, the best performance in the proposed probabilistic neural network structure is achieved. The results of present study show that applying the nonlinear features to train these networks can serve as useful tool in classifying of the EEG signals

Computational Intelligence in Magnetic Resonance Imaging of the Human Brain: The Classic-Curvature and the Intensity-Curvature Functional in a Tumor Case Study

“…Our research intersects with biomedical informatics [13] as well as with biomedical image processing [14], and biomedical feature extraction from the pathological human brain [15].…”

Calculation of the Classic-Curvature and the Intensity-Curvature Term Before Interpolation of a Bivariate Polynomial

“…However, each case was given different values of the ‗a' and ‗b' constants (thus ten different numbers have been used for the ‗a' and ‗b' constants: ten for ‗a' and ten for ‗b'). The ‗a' and ‗b' constants appear in the bivariate polynomial model function in (1), and, consistently with the mathematical procedure, they also appear in the relevant formulae of the theory and methods section of the paper. Five ‗a' constants and five ‗b' constants are negative and the remaining five ‗a' constants and five ‗b' constants are positive.…”

“…For instance, resting-state electroencephalogram signal is very useful for the neurologists, the clinicians and the physicians in order to determine the baseline of the human brain EEG. Time delay neural networks (TDNNs) and probabilistic neural networks (PNNs) trained with nonlinear features (Lyapumov exponents and Entropy) have been demonstrated useful in order to classify electroencephalogram signals (EEG) of the human brain and so to discriminate between normal controls and subjects affected by partial epilepsy [1].…”

Automated Pre-Seizure Detection for Epileptic Patients Using Machine Learning Methods

“…Image, Graphics and Signal Processing, 2017, 7, 1-9 kNN and linear classifiers [5]. In [8], EEG signals of two normal and partial epilepsy subjects are analyzed with time delay neural networks (TDNNs) and the probabilistic neural networks (PNNs). Then their performances are compared.…”

“…Feature selection with regression tree, classification with Radial Basis Function (RBF) and SVM is performed. In [4], 8 features determined in time domain for EEG data are selected by Principal Component Analysis (PCA) and classified using kNN and k-Means algorithms. Classification was made to separate the pre-seizure period from the seizure.…”