Detection of Anxiety-Based Epileptic Seizures in EEG Signals Using Fuzzy Features and Parrot Optimization-Tuned LSTM
Abstract
:1. Introduction
- We augmented EEG signals to obtain anxiety-based ES signals and non-epileptic signals using EEG signals from the BONN dataset. High beta waves (typically 20–30 Hz) are closely associated with anxiety, panic attacks, and heightened stress responses. Increased beta activity is often observed in individuals with anxiety disorders, indicating a state of hyperarousal and agitation [39].
- We extracted features from data-augmented anxiety-based EEG signals using FCM and optimization algorithms, namely, (i) particle swarm optimization (PSO) and (ii) parrot optimization (PO), which were used to tune the hyperparameters of the LSTM layer and detect anxiety-related epileptic seizure signals.
- We classified EEG ES signals and non-ES signals derived from a random augmented EEG signal dataset.
2. Materials and Methods
2.1. Dataset
2.2. Pre-Processing
2.3. Data Augmentation of the BONN EEG Signals
Random Data Augmentation
3. Results
3.1. Feature Extraction via Fuzzy Classification
3.2. PSO-LSTM for Classification of EEG Signal
3.3. Hyperparameter Tuning Using PSO in LSTM
3.4. Proposed Methods
3.4.1. FCM-PS-LSTM
3.4.2. PS−LSTM
3.4.3. Parrot Optimization-LSTM
Mathematical Model of PO
3.5. Performance Metrics
4. Discussion
5. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Acknowledgments
Conflicts of Interest
References
- Ari, B.; Siddique, K.; Alcin, O.F.; Aslan, M.; Sengur, A.; Mehmood, R.M. Wavelet ELM-AE Based Data Augmentation and Deep Learning for Efficient Emotion Recognition Using EEG Recordings. IEEE Access 2022, 10, 72171–72181. [Google Scholar] [CrossRef]
- Rommel, C.; Paillard, J.; Moreau, T.; Gramfort, A. Data augmentation for learning predictive models on EEG: A systematic comparison. J. Neural Eng. 2022, 19, 066020. [Google Scholar] [CrossRef]
- Kalashami, M.P.; Pedram, M.M.; Sadr, H. EEG Feature Extraction and Data Augmentation in Emotion Recognition. Comput. Intell. Neurosci. 2022, 2022, 7028517. [Google Scholar] [CrossRef] [PubMed]
- Al-Shargie, F.; Kiguchi, M.; Badruddin, N.; Dass, S.C.; Hani, A.F.M.; Tang, T.B. Mental stress assessment using simultaneous measurement of EEG and fNIRS. Biomed. Opt. Express 2016, 7, 3882–3898. [Google Scholar] [CrossRef]
- Asif, A.; Majid, M.; Anwar, S.M. Human stress classification using EEG signals in response to music tracks. Comput. Biol. Med. 2019, 107, 182–196. [Google Scholar] [CrossRef] [PubMed]
- Chien, J.-H.; Colloca, L.; Korzeniewska, A.; Meeker, T.J.; Bienvenu, O.J.; Saffer, M.I.; Lenz, F.A. Behavioral, Physiological and EEG Activities Associated with Conditioned Fear as Sensors for Fear and Anxiety. Sensors 2020, 20, 6751. [Google Scholar] [CrossRef]
- Malviya, L.; Khandelwal, S.; Mal, S. Mental Stress Detection Using EEG Extracted Frequency Bands. Spinger Link Soft Comput. Theor. Appl. 2022, 425, 283–293. [Google Scholar] [CrossRef]
- Mishra, S.; Satapathy, S.K.; Mohanty, S.N.; Pattnaik, C.R. A DM-ELM based classifier for EEG brain signal classification for epileptic seizure detection. Commun. Integr. Biol. 2022, 16, 2153648. [Google Scholar] [CrossRef] [PubMed]
- Ibrahim, A.K.; Zhuang, H.; Tognoli, E.; Ali, A.M.; Erdol, N. Epileptic seizure prediction based on multiresolution convolutional neural networks. Front. Signal Process. 2023, 3, 1175305. [Google Scholar] [CrossRef]
- Liu, Z.; Zhu, B.; Hu, M.; Deng, Z.; Zhang, J. Revised Tunable Q-Factor Wavelet Transform for EEG-Based Epileptic Seizure Detection. IEEE Trans. Neural Syst. Rehabil. Eng. 2023, 31, 1707–1720. [Google Scholar] [CrossRef]
- Jing, J.; Pang, X.; Pan, Z.; Fan, F.; Meng, Z. Classification and identification of epileptic EEG signals based on signal enhancement. Biomed. Signal Process. Control 2021, 71, 103248. [Google Scholar] [CrossRef]
- Chen, W.; Wang, Y.; Ren, Y.; Jiang, H.; Du, G.; Zhang, J.; Li, J. An automated detection of epileptic seizures EEG using CNN classifier based on feature fusion with high accuracy. BMC Med. Informatics Decis. Mak. 2023, 23, 96. [Google Scholar] [CrossRef] [PubMed]
- Khayretdinova, M.; Shovkun, A.; Degtyarev, V.; Kiryasov, A.; Pshonkovskaya, P.; Zakharov, I. Predicting age from resting-state scalp EEG signals with deep convolutional neural networks on TD-brain dataset. Front. Aging Neurosci. 2022, 14, 1019869. [Google Scholar] [CrossRef]
- Zhang, P.; Zhang, X.; Liu, A. Effects of Data Augmentation with the BNNSMOTE Algorithm in Seizure Detection Using 1D-MobileNet. J. Health Eng. 2022, 2022, 4114178. [Google Scholar] [CrossRef]
- Ling, H.; Luyuan, Y.; Xinxin, L.; Bingliang, D. Staging study of single-channel sleep EEG signals based on data augmentation. Front. Public Health 2022, 10, 1038742. [Google Scholar] [CrossRef]
- Martins, F.M.; Suárez, V.M.G.; Flecha, J.R.V.; López, B.G. Data Augmentation Effects on Highly Imbalanced EEG Datasets for Automatic Detection of Photoparoxysmal Responses. Sensors 2023, 23, 2312. [Google Scholar] [CrossRef]
- Lee, S.-H. Classification of epileptic seizure using feature selection based on fuzzy membership from EEG signal. Technol. Health Care 2021, 29, 519–529. [Google Scholar] [CrossRef] [PubMed]
- Shoeibi, A.; Ghassemi, N.; Khodatars, M.; Moridian, P.; Alizadehsani, R.; Zare, A.; Khosravi, A.; Subasi, A.; Acharya, U.R.; Gorriz, J.M. Detection of epileptic seizures on EEG signals using ANFIS classifier, autoencoders and fuzzy entropies. Biomed. Signal Process. Control 2021, 73, 103417. [Google Scholar] [CrossRef]
- Shirsagar, K.; Akojwar, S. Optimization of BPNN parameters using PSO for EEG signals. Adv. Intell. Syst. Res. 2017, 137, 385–394. [Google Scholar] [CrossRef]
- Sun, Q.; Liu, Y.; Li, S.; Wang, C. Automatic Epileptic Seizure Detection Using PSO-Based Feature Selection and Multilevel Spectral Analysis for EEG Signals. J. Sens. 2022, 2022, 6585800. [Google Scholar] [CrossRef]
- George, S.T.; Subathra, M.; Sairamya, N.; Susmitha, L.; Premkumar, M.J. Classification of epileptic EEG signals using PSO based artificial neural network and tunable-Q wavelet transform. Biocybern. Biomed. Eng. 2020, 40, 709–728. [Google Scholar] [CrossRef]
- Kapoor, B.; Nagpal, B.; Jain, P.K.; Abraham, A.; Gabralla, L.A. Epileptic Seizure Prediction Based on Hybrid Seek Optimization Tuned Ensemble Classifier Using EEG Signals. Sensors 2022, 23, 423. [Google Scholar] [CrossRef]
- Gini, A.P.; Queen, D.F. An Improved Optimization Algorithm for Epileptic Seizure Detection in EEG Signals Using Random Forest Classifier. Webology 2021, 18, 327–340. [Google Scholar] [CrossRef]
- Mohapatra, S.K.; Patnaik, S. ESA-ASO: An enhanced search ability based atom search optimization algorithm for epileptic seizure detection. Meas. Sens. 2022, 24, 100519. [Google Scholar] [CrossRef]
- Liu, C.; Chen, W.; Zhang, T. Wavelet-Hilbert transform based bidirectional least squares grey transform and modified binary grey wolf optimization for the identification of epileptic EEGs. Biocybern. Biomed. Eng. 2023, 43, 442–462. [Google Scholar] [CrossRef]
- Wang, B.; Yang, X.; Li, S.; Wang, W.; Ouyang, Y.; Zhou, J.; Wang, C. Automatic epileptic seizure detection based on EEG using a moth-flame optimization of one-dimensional convolutional neural networks. Front. Neurosci. 2023, 17, 1291608. [Google Scholar] [CrossRef] [PubMed]
- Cherukuvada, S.; Kayalvizhi, R. Modified Gorilla Troops Optimization with Deep Learning Based Epileptic Seizure Prediction Model on EEG Signals. Trait. Signal 2023, 40, 589–599. [Google Scholar] [CrossRef]
- Xu, G.; Ren, T.; Chen, Y.; Che, W. A One-Dimensional CNN-LSTM Model for Epileptic Seizure Recognition Using EEG Signal Analysis. Front. Neurosci. 2020, 14, 578126. [Google Scholar] [CrossRef] [PubMed]
- Hu, X.; Yuan, S.; Xu, F.; Leng, Y.; Yuan, K.; Yuan, Q. Scalp EEG classification using deep Bi-LSTM network for seizure detection. Comput. Biol. Med. 2020, 124, 103919. [Google Scholar] [CrossRef]
- Khan, P.; Khan, Y.; Kumar, S.; Khan, M.S.; Gandomi, A.H. HVD-LSTM based recognition of epileptic seizures and normal human activity. Comput. Biol. Med. 2021, 136, 104684. [Google Scholar] [CrossRef]
- Singh, K.; Malhotra, J. Two-layer LSTM network-based prediction of epileptic seizures using EEG spectral features. Complex. Intell. Syst. 2022, 8, 2405–2418. [Google Scholar] [CrossRef]
- Tuncer, E.; Bolat, E.D. Channel based epilepsy seizure type detection from electroencephalography (EEG) signals with machine learning techniques. Biocybern. Biomed. Eng. 2022, 42, 575–595. [Google Scholar] [CrossRef]
- Palanichamy, I.; Ahamed, F.B.B. Prediction of Seizure in the EEG Signal with Time Aware Recurrent Neural Network. Rev. D’Intell. Artif. 2022, 36, 717–724. [Google Scholar] [CrossRef]
- Al-Jumaili, S.; Duru, A.D.; Ibrahim, A.A.; Uçan, O.N. Investigation of Epileptic Seizure Signatures Classification in EEG using Supervised Machine Learning Algorithms. Trait. Signal 2023, 40, 43–54. [Google Scholar] [CrossRef]
- Shah, S.Y.; Larijani, H.; Gibson, R.M.; Liarokapis, D. Epileptic Seizure Classification Based on Random Neural Networks Using Discrete Wavelet Transform for Electroencephalogram Signal Decomposition. Appl. Sci. 2024, 14, 599. [Google Scholar] [CrossRef]
- Attar, E.T. Review of electroencephalography signals approaches for mental stress assessment. Neurosciences 2022, 27, 209–215. [Google Scholar] [CrossRef]
- Byeon, J.; Moon, J.Y.; Je, S.R.; Park, S.H.; Kim, J.W. Quantitative electroencephalographic biomarker of pharmacological treatment response in patients with anxiety disorder: A retrospective study. Sci. Rep. 2023, 13, 3802. [Google Scholar] [CrossRef] [PubMed]
- Newson, J.J.; Thiagarajan, T.C. EEG Frequency Bands in Psychiatric Disorders: A Review of Resting State Studies. Front. Hum. Neurosci. 2019, 12, 521. [Google Scholar] [CrossRef]
- Leaf, C.; Turner, R.; Wasserman, C.; Paulson, R.; Kopooshian, N.; Lynch, G.; Leaf, A. Psycho-neuro-biological Correlates of Beta Activity. NeuroRegulation 2023, 10, 11–20. [Google Scholar] [CrossRef]
- George, O.; Smith, R.; Madiraju, P.; Yahyasoltani, N.; Ahamed, S.I. Data augmentation strategies for EEG-based motor imagery decoding. Heliyon Sci. Direct 2022, 8, e10240. [Google Scholar] [CrossRef]
- Satapathy, S.K.; Dehuri, S.; Jagadev, A.K. EEG signal classification using PSO trained RBF neural network for epilepsy identification. Inform. Med. Unlocked 2017, 6, 1–11. [Google Scholar] [CrossRef]
- Lian, J.; Hui, G.; Ma, L.; Zhu, T.; Wu, X.; Heidari, A.A.; Chen, Y.; Chen, H. Parrot optimizer: Algorithm and applications to medical problems. Comput. Biol. Med. 2024, 172, 108064. [Google Scholar] [CrossRef] [PubMed]
- Riccio, C.; Martone, A.; Zazzaro, G.; Pavone, L. Training Datasets for Epilepsy Analysis: Preprocessing and Feature Extraction from Electroencephalography Time Series. Data 2024, 9, 61. [Google Scholar] [CrossRef]
- Wang, B.; Xu, Y.; Peng, S.; Wang, H.; Li, F. Detection Method of Epileptic Seizures Using a Neural Network Model Based on Multimodal Dual-Stream Networks. Sensors 2024, 24, 3360. [Google Scholar] [CrossRef] [PubMed]
Sets | Subjects | ||||
---|---|---|---|---|---|
Patient Phase | Electrode Kind/Location | No. of Study | Number of Data | Length of Sections | |
Set-A | Eye open | Surface (subject skin) | 5 | 100 | 4097 |
Set-B | Eye close | Surface (subject skin) | 5 | 100 | 4097 |
Set-C | Seizure free (Non-Epileptic) | Intracranial (skull) | 5 | 100 | 4097 |
Set-D | Seizure Free | Intracranial (skull) | 5 | 100 | 4097 |
Set-E | Seizure Activity (Epileptic) | Intracranial (skull) | 5 | 100 | 4097 |
Description | Values Used | Reference | |
---|---|---|---|
Particle Swarm Optimization (PSO) | |||
Size of swarm | 9 | - | - |
Maximum value of repetitions | 100 | - | - |
C1 (Cognition Coefficient), C2 (Social Coefficient) | C1 = C2 = 2 | 1.4962 [21] | 0.9 [41] |
LSTM | |||
Gradient Threshold | 0.01 | - | - |
Learning rate | 0.0001 | 0.001 [29] | 0.005 [30] |
No. of hidden units | 100 | 64 [29] | - |
Input layer | Sequence layer | - | - |
Activation function | tanh(state), Sigmoid(gate) | SoftMax [29] | - |
Output layer | Regression layer | - | - |
Drop out | 0.5 | - | - |
Parrot Optimization | |||
Maximum Iteration | 1000 | [42] | |
Lower Bound (lb) | −100 | ||
Upper Bound (ub) | 100 | ||
Number of search agents | 30 | ||
Dimension of search space | 30 |
Methods | FCM-PS-LSTM | PS-LSTM | PO-LSTM |
---|---|---|---|
Accuracy (%) | |||
Before Data Augmentation (BDA) | 94.45 | 95 | 94 |
After Random Data Augmentation (ARDA) | 98 | 98.5 | 96 |
Methods | BDA | ARDA | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Classification Effectiveness | LR | GNB | MLR | FCM-PS-LSTM | PS-LSTM | PO-LSTM | LR | GNB | MLR | FCM-PS-LSTM | PS-LSTM | PO-LSTM |
Accuracy | 93.33 | 68.3 | 91.66 | 94.45 | 95 | 94 | 96 | 85.5 | 97.5 | 98 | 98.5 | 96 |
Sensitivity | 93.7 | 82.35 | 90.9 | 94.94 | 94.11 | 95 | 97 | 84.84 | 97 | 97 | 98 | 97 |
Specificity | 92.8 | 62.79 | 92.59 | 94 | 95.91 | 93 | 95 | 86.13 | 97.97 | 98.9 | 98 | 95 |
Precision | 93.75 | 46.66 | 93.75 | 94 | 96 | 93.13 | 95.09 | 85.7 | 98 | 98.9 | 98 | 95.09 |
F1-Score | 93.75 | 59.57 | 92.3 | 94.47 | 95.04 | 94.05 | 96.03 | 85.27 | 97.51 | 98 | 98.4 | 96.03 |
MCC | 86.6 | 40.6 | 83.2 | 89.0 | 90.0 | 88.01 | 66.4 | 54.0 | 68.3 | 68.9 | 69.1 | 92.01 |
Kappa | 86.6 | 36.6 | 83.2 | 89 | 90 | 88 | 92.1 | 72 | 95.03 | 96 | 97 | 92 |
CSI | 88.2 | 42.4 | 85.7 | 89.5 | 90.5 | 88.7 | 92.3 | 74.3 | 95.1 | 96.1 | 97 | 92.3 |
FM Index | 93.72 | 61.98 | 92.3 | 94.46 | 95.04 | 94.06 | 96.04 | 85.26 | 97.49 | 98.44 | 98 | 96.04 |
Title | Purpose | Database | Strategy | Evaluation Metrics | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
PP | Features | Classifier | ||||||||||||
EEG Feature Extraction and Data Augmentation in Emotion Recognition [3] | Detection of Arousal and valence | DEAP dataset (Emotional) | CWGAN for data augmentation | Average PSD, Zero Crossing rate, Mean, variance for traits | SVM, DNN | Accuracy 71.9% | ||||||||
Staging Study of Single-Channel Sleep EEG signals Based on Data Augmentation [15] | Detection of sleep period(wake,N1,N2,N3,REM) | SC subset of Sleep-EDF Database | RDB-DCGAN data augmentation model. Wavelet time frequency transform | CNN | Accuracy 76% | |||||||||
Classification of Epileptic EEG Signals Using PSO-Based Artificial Neural Network and Tunable-Q Wavelet Transform [21] | Categorization of Epileptic EEG signals(Normal/Focal/Generalized) | KIT TUH | TQWT | Non-linear attributes such as log energy entropy, Shannon entropy and Stein’s unbiased risk estimate entropy. PSO | ANN | Accuracy: (i) normal–focal (95.1%), (ii) normal–generalised (97.4%), (iii) normal–focal + generalized (96.2%), and (iv) normal–focal generalized (88.8%) for TUH | ||||||||
Epileptic Seizure Prediction Based on Hybrid Seek Optimization Tuned Ensemble Classifier Using EEG signals [22] | Prediction of Epileptic seizure | Siena database CHB-MIT | BPF | Statistical, Wavelet and Entropy-based attributes | DT, RF & AdaBoost classifier | Siena | CHB-MIT | |||||||
Accuracy 95.3% Sensitivity 93.17% Specificity90.06% | Accuracy 96.6% Sensitiivty 94.67% Specificity 91.36% | |||||||||||||
Prediction of Seizure in the EEG Signal with Time Aware Recurrent Neural Network [33] | Prediction of EEG seizure (Inter-ictal) | CHB-MIT BONN VIRGO EEG | Time Aware CNN and Recurrent Neural Network (TA-CNN-RNN) Model | LSTM | CHB-MIT | BONN | VIRGO EEG | |||||||
Accuracy 89% Precision 88.3% Recall 91.3% F-measure 89.8% | Accuracy 88.6% Precision 87.7% Recall 90.9% F-measure 89.2% | Accuracy 88.7% Precision89.4% Recall 92.4% F-measure 90.7% | ||||||||||||
Training Datasets for Epilepsy Analysis: Preprocessing and Feature Extraction from Electroencephalography Time Series [43] | Seizure Prediction | Freiburg Databas. | Sliding window Techniques and multiple features are extracted TrBtool used | NA | NA | |||||||||
Detection Method of Epileptic Seizures Using a Neural Network Model Based on Multimodal Dual-Stream Networks [44] | Seizure detection | Bonn EEG dataset New Delhi Dataset | Short time Fourier transform | signal differential attributes, frequency domain amplitude spectrum and phase spectrum methods | Multimodal dual stream networks | 99.69%Accuracy, 99.44%Precision, 1%Recall, 99.72%F1-score for Bonn EEG dataset | ||||||||
Proposed Methods | EEG Epileptic seizure (anxiety based) | BONN EEG dataset | BPF, Median and RADWT | Statistical | LSTM (i) PFCM-PS-LSTM (ii) PS-LSTM (iii) PO-LSTM | ARDA(%) | ||||||||
Accuracy | Sensitivity | Speficity | Precision | F1-score | MCC | Kappa | CSI | FM Index | ||||||
98 | 97 | 98.9 | 98.9 | 98 | 68.9 | 96 | 96.1 | 98.44 | ||||||
98.5 | 98 | 98 | 98 | 98.4 | 69.1 | 97 | 97 | 98 | ||||||
96 | 97 | 95 | 95.09 | 96.03 | 92.01 | 92 | 92.3 | 96.04 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Palanisamy, K.K.; Rengaraj, A. Detection of Anxiety-Based Epileptic Seizures in EEG Signals Using Fuzzy Features and Parrot Optimization-Tuned LSTM. Brain Sci. 2024, 14, 848. https://doi.org/10.3390/brainsci14080848
Palanisamy KK, Rengaraj A. Detection of Anxiety-Based Epileptic Seizures in EEG Signals Using Fuzzy Features and Parrot Optimization-Tuned LSTM. Brain Sciences. 2024; 14(8):848. https://doi.org/10.3390/brainsci14080848
Chicago/Turabian StylePalanisamy, Kamini Kamakshi, and Arthi Rengaraj. 2024. "Detection of Anxiety-Based Epileptic Seizures in EEG Signals Using Fuzzy Features and Parrot Optimization-Tuned LSTM" Brain Sciences 14, no. 8: 848. https://doi.org/10.3390/brainsci14080848