Automated Detection of Aberrant Episodes in Epileptic Conditions: Leveraging EEG and Machine Learning Algorithms

Bioengineering, Journal Year: 2025, Volume and Issue: 12(4), P. 355 - 355

Published: March 29, 2025

Epilepsy is a neurologic condition characterized by recurring seizures resulting from aberrant brain activity. It crucial to promptly and precisely detect epileptic ensure efficient treatment. The gold standard electroencephalography (EEG) accurately records the brain’s electrical activity in real time. intent of this study episodes leveraging machine learning deep algorithms on EEG inputs. proposed approach aims evaluate feasibility developing novel technique that utilizes Hurst exponent identify signal properties could be for classification. idea posits prolonged duration patients those who are not experiencing can differentiate between two groups. To achieve this, we analyzed long-term memory characteristics employing time-dependent analysis. Together, Daubechies 4 discrete wavelet transformation constitute basis unique feature extraction. We utilize ANOVA test random forest regression as selection techniques. Our creates evaluates support vector machine, classifier, long short-term network models classify using highlight our research it examines efficacy aforementioned classifying utilizing single-channel with minimally handcrafted features. classifier outperforms other options, an accuracy 97% sensitivity 97.20%. Additionally, model’s capacity generalize unobserved data evaluated CHB-MIT scalp database, showing remarkable outcomes. Since framework computationally efficient, implemented edge hardware. This strategy redefine epilepsy diagnoses hence provide individualized regimens improve patient

Language: Английский

---

Advancing task recognition towards artificial limbs control with ReliefF-based deep neural network extreme learning

Computers in Biology and Medicine, Journal Year: 2023, Volume and Issue: 169, P. 107894 - 107894

Published: Dec. 22, 2023

Language: Английский

---

Early Detection of Stress and Anxiety Based Seizures in Position Data Augmented EEG Signal Using Hybrid Deep Learning Algorithms

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 35351 - 35365

Published: Jan. 1, 2024

Epilepsy is a neurological problem due to aberrant brain activity. diagnose through Electroencephalography (EEG) signal. Human interpretation and analysis of EEG signal for earlier detection epilepsy subjected error. Detection Epileptic seizures stress anxiety the major problem. seizure size, shape changes from person based on their level. Stress epileptic signals vary in amplitude, width, combination width amplitude. In this paper, Seizures different size are synthesized using data augmentation Different such as (i) position (PDA) (ii) random (RDA) applied BONN dataset synthetizations signals. Augment analyzed proposed methods i) FCM-PSO-LSTM ii) PSO-LSTM anxiety-based seizures. The algorithms perform better predicted accuracy about i)98.5% 97%, PDA RDA 98% 98.5%, respectively.

Language: Английский

---

Residual and bidirectional LSTM for epileptic seizure detection

Frontiers in Computational Neuroscience, Journal Year: 2024, Volume and Issue: 18

Published: June 17, 2024

Electroencephalogram (EEG) plays a pivotal role in the detection and analysis of epileptic seizures, which affects over 70 million people world. Nonetheless, visual interpretation EEG signals for epilepsy is laborious time-consuming. To tackle this open challenge, we introduce straightforward yet efficient hybrid deep learning approach, named ResBiLSTM, detecting seizures using signals. Firstly, one-dimensional residual neural network (ResNet) tailored to adeptly extract local spatial features Subsequently, acquired are input into bidirectional long short-term memory (BiLSTM) layer model temporal dependencies. These output further processed through two fully connected layers achieve final seizure detection. The performance ResBiLSTM assessed on datasets provided by University Bonn Temple Hospital (TUH). achieves accuracy rates 98.88–100% binary ternary classifications dataset. Experimental outcomes recognition across seven types TUH corpus (TUSZ) dataset indicate that attains classification 95.03% weighted F1 score with 10-fold cross-validation. findings illustrate outperforms several recent state-of-the-art approaches.

Language: Английский

---

A New Approach to Automatic Epilepsy Detection from EEG Signals Using Archimedean Spiral and Swin Transformer

Research Square (Research Square), Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 20, 2025

Epilepsy, a neurological disorder marked by recurrent and unpredictable seizures due to abnormal brain electrical activity, is studied using electroencephalography (EEG) which measures activity. The EEG signals are commonly employed diagnose monitor conditions like epilepsy, sleep disorders, injuries. This research work introduces an effective hybrid approach based on Archimedean spiral coding (ASC) swin transformer-based convolutional neural network (CNN) techniques detect epilepsy automatically signals. proposed ASC method transforms into visually informative 3D matrix employs CNN architecture for classification. It yields accuracy of 97.98% 88.22% sample- subject-based ten-fold cross-validation, respectively the public database 121 populations. developed system ready be tested with more patients from different races validate performance. produced better score as compared existing results.

Language: Английский

---

Unveiling encephalopathy signatures: A deep learning approach with locality-preserving features and hybrid neural network for EEG analysis

Neuroscience Letters, Journal Year: 2025, Volume and Issue: 849, P. 138146 - 138146

Published: Jan. 31, 2025

Language: Английский

---

Eeg Driven Seizure Classification Framework Leveraging Variational Mode Decomposition Technique and Entropy Features Based Bayesian Optimized SVM

Published: Jan. 1, 2025

Language: Английский

---

Optimized seizure detection leveraging band-specific insights from limited EEG channels

Health Information Science and Systems, Journal Year: 2025, Volume and Issue: 13(1)

Published: March 19, 2025

Language: Английский

---

PI Net: An End-to-End Semantic Decoding Model for EEG Signals in Perception and Imagination Tasks

Digital Signal Processing, Journal Year: 2025, Volume and Issue: unknown, P. 105250 - 105250

Published: April 1, 2025

Language: Английский

---

Wavelet-Hilbert transform based bidirectional least squares grey transform and modified binary grey wolf optimization for the identification of epileptic EEGs

Journal of Applied Biomedicine, Journal Year: 2023, Volume and Issue: 43(2), P. 442 - 462

Published: April 1, 2023

Language: Английский

---