[Paper Review] A hybrid automated detection of epileptic seizures in EEG based on wavelet and machine learning techniques

Asmaa Hamad, Aboul Ella Hassanien+2 more|arXiv (Cornell University)|Jul 13, 2018

EEG and Brain-Computer Interfaces4 citations

TL;DR

This study proposes a hybrid EEG-based epileptic seizure detection system using Discrete Wavelet Transform (DWT) for feature extraction and three machine learning classifiers—SVMRBF, KNN, and Naive Bayes—for classification. The method achieves 100% accuracy in distinguishing normal from epileptic EEG signals using SVMRBF, outperforming prior methods and demonstrating high reliability across multiple test cases.

ABSTRACT

Epilepsy is a neurological condition such that it affects the brain and the nervous system. It is characterized by recurrent seizures, which are physical reactions to sudden, usually brief, excessive electrical discharges in a group of brain cells. Hence, seizure identification has great importance in clinical therapy of epileptic patients. Electroencephalogram (EEG) is one of the main biomarker that can measure voltage fluctuations of the brain and EEG data analysis helps to investigate the patient with epilepsy syndrome as epilepsy leaves their signature in EEG signals. In this paper, the Discrete Wavelet Transform (DWT) is applied to EEG signals to pre-processing, decompose it till the 4th level of decomposition tree.Various features like Entropy, Min, Max, Mean, Median, Standard deviation, Variance, Skewness, Energy and Relative Wave Energy (RWE) were computed in terms of detailed coefficients and the approximation coefficients of the last decomposition level.Then, the extracted features are evaluated by three modern machine-learning classifiers such as Radial Basis Function based Support Vector Machine (SVMRBF), k-Nearest Neighbor (KNN) and Naive Bayes (NB). The experimental results demonstrate that the highest classification accuracy (100\\%) for normal subject data versus epileptic data is obtained by SVMRBF. the corresponding accuracy between normal subject data and epileptic data using KNN and NB is obtained as 99.50\\% and 99\\% for the eyes open and eyes closed conditions, respectively. The similar accuracies, while comparing the interictal and ictal data, are obtained as 99\\%, 97.50\\% and 98.50\\% using the SVMRBF, KNN and NB classifiers, respectively. These accuracies are quite higher than earlier results published.

Motivation & Objective

• To develop an automated, accurate, and efficient method for detecting epileptic seizures in EEG signals to reduce reliance on time-consuming visual interpretation.
• To improve classification performance by combining wavelet-based feature extraction with advanced machine learning models.
• To evaluate and compare the effectiveness of SVMRBF, KNN, and Naive Bayes classifiers in distinguishing between normal, interictal, and ictal EEG states.
• To establish a robust, reproducible framework for EEG-based seizure detection that can support clinical decision-making.

Proposed method

• Discrete Wavelet Transform (DWT) is applied to EEG signals up to the 4th decomposition level to extract detailed and approximation coefficients.
• Ten features—Entropy, Min, Max, Mean, Median, Standard Deviation, Variance, Skewness, Energy, and Relative Wave Energy (RWE)—are computed from the DWT coefficients.
• The extracted features are used as input to three machine learning classifiers: Radial Basis Function-based Support Vector Machine (SVMRBF), k-Nearest Neighbor (KNN), and Naive Bayes (NB).
• 10-fold cross-validation is employed to train and test the classifiers, with 90% of data used for training and 10% for testing, repeated 10 times to ensure statistical reliability.
• The classification performance is evaluated using metrics including accuracy, sensitivity, specificity, precision, and F-measure.
• The method is validated across four test cases comparing normal (eyes open/closed), interictal, and ictal EEG data.

Experimental results

Research questions

• RQ1 Can DWT-based feature extraction effectively capture discriminative patterns in EEG signals for epileptic seizure detection?
• RQ2 How do SVMRBF, KNN, and Naive Bayes classifiers compare in classifying normal versus epileptic EEG states?
• RQ3 What is the classification accuracy of the proposed hybrid system across different EEG signal types (e.g., eyes open, eyes closed, interictal, ictal)?
• RQ4 Does the integration of wavelet features with machine learning models significantly improve detection accuracy compared to existing methods?

Key findings

• The SVMRBF classifier achieved 100% accuracy in distinguishing normal EEG (eyes open) from epileptic ictal activity, with 100% sensitivity and specificity.
• For the eyes closed condition, SVMRBF maintained 100% accuracy, while KNN and Naive Bayes achieved 99.5% and 99% accuracy, respectively.
• In classifying interictal versus ictal EEG signals, SVMRBF achieved 97% accuracy, KNN 96.5%, and Naive Bayes 96.5%, demonstrating strong performance across all cases.
• The proposed method outperformed existing approaches, with reported accuracies exceeding earlier studies such as 85.16% (SVM with PE), 95.2% (wavelet + ANN), and 99.2% (GP + KNN).
• Relative Wave Energy (RWE) and other statistical features extracted from DWT coefficients proved highly effective in capturing seizure-related patterns in EEG signals.
• The 10-fold cross-validation results show consistent and high performance across all test cases, confirming the robustness and generalizability of the proposed framework.