An EEG Based Computational Model for Seizure Detetcion
- 1. An EEG-Based Computational Model for Epileptic Seizure Detection and Preictal Time Prediction Presented by D.SHYAM SUNDAR (3122213002098)
- 2. Problem Statement Develop an accurate and timely computational model using EEG data to classify the kinds of seizures and predict epileptic seizures in advance, improving intervention and optimizing epilepsy management.
- 3. Outline • Objective • Motivation • Literature Survey • Introduction • Dataset Description • Work • Results • Conclusion • References
- 4. Objective • Accurate Seizure Prediction: Develop a model to predict epileptic seizures with high accuracy using EEG data. • Timely Detection: Identify the preictal phase to enable early intervention and improve patient safety. • Feature Extraction: Utilize advanced techniques for extracting relevant features from EEG signals to enhance prediction reliability. • Clinical Integration: Create a model that can be easily integrated into clinical practice for effective epilepsy management.
- 5. Motivation • Improve Safety: Early seizure prediction reduces the risk of injury for epilepsy patients. • Better Clinical Outcomes: Enhance treatment and management through reliable seizure predictions. • Advance Understanding: Gain insights into preictal brain activity patterns and seizure mechanisms. • Overcome Limitations: Address the accuracy and invasiveness issues of current epilepsy management technologies.
- 6. Literature Survey Inference Authors Title of the Paper S.No Feature Extraction from EEG signals in time domain and frequency domain K. Kannadasan B. Shameedha Begum V. Sridevi An EEG-Based Computational Model for Decoding Emotional Intelligence, Personality, and Emotions 1 Early detection of epileptic seizures by detecting the spikes in the EEG waveform Khakon Das Partha Pratim Roy Atri Chatterjee Shankar Prasad Sha Seizure prediction by the detection of EEG waveform from the pre- ictal phase of EEG Signals 2 Seizure detection based on the frequency bands in the brain A.M. Chan F.T. Sun E.H. Boto Automated seizure onset detection for accurate onset time determination in intracranial EEG 3 Seizure detection in the preictal phase S.S.P. Kumar L. Ajitha Early detection of Epilepsy using EEG Signals 4
- 7. Introduction • Epilepsy is a neurological disorder with recurrent seizures that significantly affect patients' quality of life. • Existing seizure prediction methods lack accuracy and timely detection, limiting effective intervention. • This project focuses on accurate preictal time prediction, providing early warnings for better seizure management. • The project aims to develop a computational model using EEG data to improve seizure prediction, enhance patient safety, and advance clinical practices.
- 9. Dataset Description • The Bonn EEG Dataset is a collection of EEG (electroencephalogram) recordings used for neuroscience and machine learning research. • This dataset contains 5 signal files each with 4096 samples with classes like normal, partial seizure, general seizure, focal seizure, and myoclonic seizure. • The sampling rate of the data was 173.61Hz for better performance. In EEG recordings, sampling is necessary because it allows for the digital processing and analysis of the signals. • The application of a low-pass filter, with fc = 40 Hz is the first step involved in processing to remove the noise and unwanted components.
- 10. Work - Methodology Global Architecture of the Proposed Model
- 11. Work – Data Visualisation • Time Domain Visualisation: Continuous-time Signal Plot of the EEG Waveform from each class of the dataset • Frequency Domain Visualisation: Usage of Continuous Wavelet Transform (CWT) for scalogram and Short Time Fourier Transform (STFT) for spectrogram to represent the frequency content of the signal Time Domain – Signal Plot Frequency Domain – Scalogram
- 12. Data Preprocessing • Up-sampling: The data is interpolated (up-sampled) to 200 Hz to increase the temporal resolution of the EEG signals. • Down-sampling: Followed by interpolation, the data is decimated (down- sampled) to 100 Hz to reduce the computational load and retain essential information • Filtering: A low-pass filter with a cut-off frequency of 40Hz is applied to eliminate high frequencies • Noise Removal: Usage of Haar wavelet for artifact removal techniques. Filtered Signal
- 13. Data Analysis • Descriptive Analysis: The mean, standard deviation, median, and percentiles (25th and 75th) are calculated to understand the central tendency and variability of EEG signals. • EEG Analysis: The Welch method is used to analyze frequency components, identifying dominant frequencies and power distribution across different bands. • Spectral Analysis: Generate brain topographic maps to visualize power distribution across brain regions, focusing on Delta, Theta, Alpha, and Beta bands for seizure localization.
- 14. Cont.. Range of Operation Frequency band 0.5 - 4 Hz Delta Band 4 - 8 Hz Theta Band 8 -12 Hz Alpha Band 13 - 30 Hz Beta Band Table 1 Frequency bands operating ranges
- 15. Feature Extraction • Time Domain: Features like Normalized First Difference (NFF), Hjorth Mobility (HM), Hjorth Complexity (HC), and Higher Order Crossings (Hoc) provide information on signal dynamics and patterns. • Frequency Domain: Derived from Power Spectral Density (PSD) of theta, delta, alpha, and beta bands using the Welch method to capture energy within specific frequency ranges. • Time-Frequency Domain: Combines time and frequency aspects with features like Root Mean Square (RMS) and Recursive Energy Efficiency (REE) to quantify energy distribution using Continuous Wavelet Transform (CWT) for detailed signal representation. • Feature Compilation: Extract features, organize them into a structured dataset, save them in CSV format, and prepare for analysis and machine learning model training.
- 16. Cont… Table 2 Time Domain Features Expression Parameter Maximum amplitude / Mean amplitude Normalized First difference (NFF) (var(y’(t))/var(y(t))0.5 Hjorth Mobility (HM) mobility(y’(t))/mobility(y(t)) Hjorth Complexity (HC) Number of times a signal crosses its mean value Higher-order Crossings (HoC) After the extraction of all the required features in the Time domain, the Frequency domain, and the Time-Frequency domain, the extracted features are compiled into a feature set and converted to a Comma Separated Values file (.csv file) for the classification and preictal time prediction.
- 17. Seizure Classification • To classify the different kinds of seizures, machine learning classification models like K-Nearest Neighbours(KNN), Support Vector Machine(SVM), Logistic Regression(LR), Linear Discriminant Analysis(LDA), Decision Tree(DT) and Naive Bayes(NB) are used. • The seizures are classified as Healthy (Normal – No seizure), Focal seizures, General seizures, Partial Seizures, and Myoclonic Seizures. • Each kind of seizure is characterized by its unique features, EEG patterns, location of origination in the brain, different kinds of symptoms and neuronal activity in the brain.
- 18. Symptoms EEG Pattern Origin Type of Seizure motor, sensory, autonomic, or psychic symptoms. Focal epileptiform discharges, localized abnormalities. Specific area of the brain Partial Seizures tonic-clonic movements, brief lapses in awareness, sudden muscle tone loss Generalized spike- and-wave, polyspike-and-wave discharges Both hemispheres from onset General Seizures Similar to partial seizures. Focal spikes, sharp waves, localized slowing. Specific area (similar to Partial Seizures) Focal Seizures Sudden , brief , involuntary muscle jerks Generalized, bilateral, synchronous, polyspike – and – wave discharges Diffuse, generalized (specific syndromes) Myoclonic Seizures Cont…
- 19. Cont…
- 20. Cont…
- 21. Preictal time prediction • Feature Extraction: The scalogram features using CWT and band power features from PSD in specific frequency bands (delta, theta, alpha, beta) are extracted from the EEG files. • Threshold Definition: Specific thresholds for each frequency band are established to differentiate preictal periods from interictal and ictal states, identifying significant changes indicative of an impending seizure. • RNN Model: The extracted features are fed into a Recurrent Neural Network (RNN) model to capture temporal dependencies in EEG signals and predict preictal time by learning patterns associated with the preictal phase. • SMS Alert: An SMS alert is generated and sent to the patient’s mobile, providing an advanced warning of an upcoming seizure for timely intervention and management.
- 22. Results – ML Models The performance metrics of the ML classification models used for the classification of different kinds of seizures are shown in the below table F1 score Recall Precision Accuracy Model 0.87 0.96 0.80 0.81 KNN 0.85 0.88 0.81 0.80 LDA 0.82 0.80 0.84 0.80 LR 0.74 1.00 0.68 0.74 NB 0.84 0.92 0.77 0.78 SVM 0.90 0.92 0.88 0.82 DT
- 23. Results – RNN Model • Extracted scalogram and power band features are fed into the RNN model. • Adam optimizer is employed for efficient training over 50 epochs. • The dataset is split into 70% training, 15% validation, and 15% testing for robust evaluation. • -Model achieves 93% training accuracy and 89% validation accuracy, indicating effective learning and generalization. • Accuracy and loss function plots visualize training dynamics, confirming the model's capability in predicting preictal time.
- 24. Results - Training and Testing Plot of Accuracy Curve Plot of Loss function Curve
- 25. Results – Alert Generation
- 26. Results – Alert Generation
- 27. Results – Clinical Validation • The predicted preictal time results were validated by a consultant neurosurgeon, on a scale of 0-5 where 0 represents “Poor prediction” and 5 for “Prefect prediction”. • The average of the results was taken and statistical analysis of the results is performed.
- 28. Novelty of the work Novelty Methods used Title of the Paper Focus on the classification of seizures and partial preictal prediction Machine Learning classification models Early detection of Epilepsy using EEG signals by Selvin Pradeep Kumar et. al Classification of preictal and interictal stages of epilepsy DCNN and Bidirectional RNN Deep Learning-based Reliable Early Epileptic Seizure Predictor by Hisham Daoud et.al Classification of normal and epileptic EEG in the preictal stage Feature extraction and Pattern matching Epileptic seizure prediction by the detection of seizure waveform from the pre-ictal phase of EEG signal Multidomain feature extraction, classification of seizure, and preictal time prediction and alert generation ML classifiers and RNN with LSTM This work
- 29. Conclusion • This study presents a comprehensive model for epilepsy detection and preictal time prediction using machine learning, leveraging features from multiple domains and an RNN for precise intervention timing. • The future vision is to develop a wearable device, like a smartwatch, to predict seizures 20 minutes in advance, enhancing epilepsy management and patient safety. Wristband wearable for epilepsy alert generation
- 30. SPECIAL THANKS Dr . Nikunj Arunkumar Bhagat Assistant Professor, Department of Electrical Engineering, Department of Biological Sciences and Biosciences, Indian Institute of Technology, Kanpur Dr. Vighneshwar Ravishankar Consultant Neurosurgeon, Institute of Neuroscience, Apollo Hospitals
- 31. References • K. Kannadasan, J. Shukla, S. Veerasingam, B. S. Begum, and N. Ramasubramanian, "An EEG-Based Computational Model for Decoding Emotional Intelligence, Personality, and Emotions," IEEE Transactions on Instrumentation and Measurement, vol. 73, Art. no. 2505413, 2024. • A. M. Chan, F. T. Sun, E. H. Boto, and B. M. Wingeier, "Automated seizure onset detection for accurate onset time determination in intracranial EEG," Clin. Neurophysiol., vol. 119, no. 11, pp. 2572-2579, Nov. 2008. doi: 10.1016/j.clinph.2008.08.025. • E. Chesktor, K. Das, D. Daschakladar, P. P. Roy, A. Chatterjee, and S. P. Saha, "Epileptic seizure prediction by the detection of seizure waveform from the pre-ictal phase of EEG signal," Biomedical Signal Processing and Control, vol. 57, p. 101720, 2020. • T. Jhansi Rani and D. Kavitha, "A study on EEG signals for epileptic seizure detection using machine learning classifiers," in Proceedings of the 6th International Conference on Communication and Electronics Systems (ICCES-2021), IEEE Xplore Part Number: CFP21AWO-ART, ISBN: 978-0-7381-1405-7.
- 32. Thank You!