Feature extraction from EEG spectrograms for epileptic seizure detection

Seizure detection using nonlinear measures over EEG frequency bands and deep learning classifiers

Epilepsy is a brain disorder that causes patients to suffer from convulsions, which affects their behavior and way of life. Epilepsy can be detected with electroencephalograms (EEGs), which record brain neural activity. Traditional approaches for detecting epileptic seizures from an EEG signal are time-consuming and annoying. To supersede these traditional methods, a myriad of automated seizure detection frameworks based on machine learning techniques have recently been developed. Feature extraction and classification are the two essential phases for seizure detection. The classifier assigns the proper class label after feature extraction lowers the input pattern space while maintaining useful features. This paper proposes a new feature extraction method based on calculating nonlinear features from the most relevant EEG frequency bands. The EEG signal is first decomposed into smaller time segments from which a vector of nonlinear features is computed and supplied to machine learning (ML) and deep learning (DL) classifiers. Experiments on the Bonn dataset reveals an accuracy of 99.7% reached in classifying normal and ictal EEG signals; and an accuracy of 98.8% in the discrimination of ictal and interictal EEG signals. Furthermore, a performance of 100% is achieved on the Hauz Khas dataset. The classification results of the proposed approach were compared to those from published state of the art techniques. Our results are equivalent to or better than some recent studies appeared in the literature.

Classifying marine mammals signal using cubic splines interpolation combining with triple loss variational auto-encoder

In practical applications of passive sonar principles for extracting characteristic frequencies of acoustic signals, scientists typically employ traditional time-frequency domain transformation methods such as Mel-frequency, Short time Fourier transform (STFT), and Wavelet transform (WT). However, these solutions still face limitations in resolution and information loss when transforming data collected over extended periods. In this paper, we present a study using a two-stage approach that combines pre-processing by Cubic-splines interpolation (CSI) with a probability distribution in the hidden space with Siamese triple loss network model for classifying marine mammal (MM) communication signals. The Cubic-splines interpolation technique is tested with the STFT transformation to generate STFT-CSI spectrograms, which enforce stronger relationships between characteristic frequencies, enhancing the connectivity of spectrograms and highlighting frequency-based features. Additionally, stacking spectrograms generated by three consecutive methods, Mel, STFT-CSI, and Wavelet, into a feature spectrogram optimizes the advantages of each method across different frequency bands, resulting in a more effective classification process. The proposed solution using an Siamese Neural Network-Variational Auto Encoder (SNN-VAE) model also overcomes the drawbacks of the Auto-Encoder (AE) structure, including loss of discontinuity and loss of completeness during decoding. The classification accuracy of marine mammal signals using the SNN-VAE model increases by 11% and 20% compared to using the AE model (2013), and by 6% compared to using the Resnet model (2022) on the same actual dataset NOAA from the National Oceanic and Atmospheric Administration - United State of America.

Nonlinear and chaos features over EMD/VMD decomposition methods for ictal EEG signals detection

The detection and identification of epileptic seizures attracted considerable relevance for the neurophysiologists. In order to accomplish the detection of epileptic seizures or equivalently ictal EEG states, this paper proposes the use of nonlinear and chaos features not computed over the raw EEG signals as it was commonly experienced, but instead over intrinsic mode functions (IMFs) extracted subsequently to the application of newly time-frequency signal decomposition methods on the basis of empirical mode decomposition (EMD) and variational mode decomposition (VMD) methods. The first step within the proposed methodology is to excerpt the various components of the IMFs by EMD and VMD decomposition methods on time EEG segments. The Hjorth parameters, the Hurst exponent, the Recurrence Quantification Analysis (RQA), the detrended fluctuation analysis (DFA), the Largest Lyapunov Exponent (LLE), The Higuchi and Katz fractal dimensions (HFD and KFD), seven nonlinear and chaos features computed over the IMFs were investigated and their classification performances evaluated using the k-nearest neighbor (KNN) and the multilayer perceptron neural network (MLPNN) classifiers. Furthermore, the combination of the best nonlinear features has also been examined in terms of sensitivity, specificity and overall classification accuracy. The publicly available Bonn EEG dataset has been has been employed to validate the efficiency of the proposed method for detecting ictal EEG signals from normal or interictal EEG segments. Among the several experiments involved in the current study, the ultimate results establish that the overall classification accuracy can achieve 100%, 99.45%, 99.8%, 99.8%, 98.6% and 99.1% for six different epileptic seizure detection case problems studied, confirming the ability of the proposed methodology in helping the clinic practitioners in the epilepsy detection care units to classify seizure events with a great confidence.

Using shallow neural networks with functional connectivity from EEG signals for early diagnosis of Alzheimer's and frontotemporal dementia

Introduction

Dementia is a neurological disorder associated with aging that can cause a loss of cognitive functions, impacting daily life. Alzheimer's disease (AD) is the most common cause of dementia, accounting for 50-70% of cases, while frontotemporal dementia (FTD) affects social skills and personality. Electroencephalography (EEG) provides an effective tool to study the effects of AD on the brain.

Methods

In this study, we propose to use shallow neural networks applied to two sets of features: spectral-temporal and functional connectivity using four methods. We compare three supervised machine learning techniques to the CNN models to classify EEG signals of AD / FTD and control cases. We also evaluate different measures of functional connectivity from common EEG frequency bands considering multiple thresholds.

Results and discussion

Results showed that the shallow CNN-based models achieved the highest accuracy of 94.54% with AEC in test dataset when considering all connections, outperforming conventional methods and providing potentially an additional early dementia diagnosis tool.

Epileptic Seizure Detection Using Machine Learning: Taxonomy, Opportunities, and Challenges

Epilepsy is a life-threatening neurological brain disorder that gives rise to recurrent unprovoked seizures. It occurs due to abnormal chemical changes in our brains. For many years, studies have been conducted to support the automatic diagnosis of epileptic seizures for clinicians’ ease. For that, several studies entail machine learning methods for early predicting epileptic seizures. Mainly, feature extraction methods have been used to extract the right features from the EEG data generated by the EEG machine. Then various machine learning classifiers are used for the classification process. This study provides a systematic literature review of the feature selection process and classification performance. This review was limited to finding the most used feature extraction methods and the classifiers used for accurate classification of normal to epileptic seizures. The existing literature was examined from well-known repositories such as MDPI, IEEE Xplore, Wiley, Elsevier, ACM, Springer link, and others. Furthermore, a taxonomy was created that recapitulates the state-of-the-art used solutions for this problem. We also studied the nature of different benchmark and unbiased datasets and gave a rigorous analysis of the working of classifiers. Finally, we concluded the research by presenting the gaps, challenges, and opportunities that can further help researchers predict epileptic seizures.

Fault Detection and Classification in Transmission Lines Connected to Inverter-Based Generators Using Machine Learning

Integrating inverter-based generators in power systems introduces several challenges to conventional protection relays. The fault characteristics of these generators depend on the inverters’ control strategy, which matters in the detection and classification of the fault. This paper presents a comprehensive machine-learning-based approach for detecting and classifying faults in transmission lines connected to inverter-based generators. A two-layer classification approach was considered: fault detection and fault type classification. The faults were comprised of different types at several line locations and variable fault impedance. The features from instantaneous three-phase current and voltages and calculated swing-center voltage (SCV) were extracted in time, frequency, and time–frequency domains. A photovoltaic (PV) and a Doubly-Fed Induction Generator (DFIG) wind farm plant were the considered renewable resources. The unbalanced data problem was investigated and mitigated using the synthetic minority class oversampling technique (SMOTE). The hyperparameters of the evaluated classifiers, namely decision trees (DT), Support Vector Machines (SVM), k-Nearest Neighbors (k-NN), and Ensemble trees, were optimized using the Bayesian optimization algorithm. The extracted features were reduced using several methods. The classification performance was evaluated in terms of the accuracy, specificity, sensitivity, and precision metrics. The results show that the data balancing improved the specificity of DT, SVM, and k-NN classifiers (DT: from 99.86% for unbalanced data to 100% for balanced data; SVM: from 99.28% for unbalanced data to 99.93% for balanced data; k-NN: from 99.64% for unbalanced data to 99.74% for balanced data). The forward feature selection combined with the Bag ensemble classifier achieved 100% accuracy, sensitivity, specificity, and precision for fault detection (binary classification), while the Adaboost ensemble classifier had the highest accuracy (99.4%), compared to the other classifiers when using the complete set of features. The classification models with the highest performance were further tested using a new dataset test case. They showed high detection and classification capabilities. The proposed approach was compared with the previous methodologies from the literature.

Unsupervised seizure identification on EEG

BACKGROUND AND OBJECTIVE

Epilepsy is one of the most common neurological disorders, whose development is typically detected via early seizures. Electroencephalogram (EEG) is prevalently employed for seizure identification due to its routine and low expense collection. The stochastic nature of EEG makes manual seizure inspections laborsome, motivating automated seizure identification. The relevant literature focuses mostly on supervised machine learning. Despite their success, supervised methods require expert labels indicating seizure segments, which are difficult to obtain on clinically-acquired EEG. Thus, we aim to devise an unsupervised method for seizure identification on EEG.

METHODS

We propose the first fully-unsupervised deep learning method for seizure identification on raw EEG, using a variational autoencoder (VAE). In doing so, we train the VAE on recordings without seizures. As training captures non-seizure activity, we identify seizures with respect to the reconstruction errors at inference time. Moreover, we extend the traditional VAE training loss to suppress EEG artifacts. Our method does not require ground-truth expert labels indicating seizure segments or manual feature extraction.

RESULTS

We implement our method using the PyTorch library and execute experiments on an NVIDIA V100 GPU. We evaluate our method on three benchmark EEG datasets: (i) intracranial recordings from the University of Pennsylvania and the Mayo Clinic, (ii) scalp recordings from the Temple University Hospital of Philadelphia, and (iii) scalp recordings from the Massachusetts Institute of Technology and the Boston Children's Hospital. To assess performance, we report accuracy, precision, recall, Area under the Receiver Operating Characteristics Curve (AUC), and p-value under the Welch t-test for distinguishing seizure vs. non-seizure EEG windows. Our approach can successfully distinguish seizures from non-seizure activity, with up to 0.83 AUC on intracranial recordings. Moreover, our algorithm has the potential for real-time inference, by processing at least 10 s of EEG in a second.

CONCLUSION

We take the first successful steps in deep learning-based unsupervised seizure identification on raw EEG. Our approach has the potential of alleviating the burden on clinical experts regarding laborsome EEG inspections for seizures. Furthermore, aiding the identification of early seizures via our method could facilitate successful detection of epilepsy development and initiate antiepileptogenic therapies.

EEGFuseNet: Hybrid Unsupervised Deep Feature Characterization and Fusion for High-Dimensional EEG With an Application to Emotion Recognition

How to effectively and efficiently extract valid and reliable features from high-dimensional electroencephalography (EEG), particularly how to fuse the spatial and temporal dynamic brain information into a better feature representation, is a critical issue in brain data analysis. Most current EEG studies work in a task driven manner and explore the valid EEG features with a supervised model, which would be limited by the given labels to a great extent. In this paper, we propose a practical hybrid unsupervised deep convolutional recurrent generative adversarial network based EEG feature characterization and fusion model, which is termed as EEGFuseNet. EEGFuseNet is trained in an unsupervised manner, and deep EEG features covering both spatial and temporal dynamics are automatically characterized. Comparing to the existing features, the characterized deep EEG features could be considered to be more generic and independent of any specific EEG task. The performance of the extracted deep and low-dimensional features by EEGFuseNet is carefully evaluated in an unsupervised emotion recognition application based on three public emotion databases. The results demonstrate the proposed EEGFuseNet is a robust and reliable model, which is easy to train and performs efficiently in the representation and fusion of dynamic EEG features. In particular, EEGFuseNet is established as an optimal unsupervised fusion model with promising cross-subject emotion recognition performance. It proves EEGFuseNet is capable of characterizing and fusing deep features that imply comparative cortical dynamic significance corresponding to the changing of different emotion states, and also demonstrates the possibility of realizing EEG based cross-subject emotion recognition in a pure unsupervised manner.

A novel multi-class imbalanced EEG signals classification based on the adaptive synthetic sampling (ADASYN) approach

BACKGROUND

Brain signals (EEG-Electroencephalography) are a gold standard frequently used in epilepsy prediction. It is crucial to predict epilepsy, which is common in the community. Early diagnosis is essential to reduce the treatment process of the disease and to keep the process healthier.

METHODS

In this study, a five-classes dataset was used: EEG signals from different individuals, healthy EEG signals from tumor document, EEG signal with epilepsy, EEG signal with eyes closed, and EEG signal with eyes open. Four different methods have been proposed to classify five classes of EEG signals. In the first approach, the EEG signal was first divided into four different bands (beta, alpha, theta, and delta), and then 25 time-domain features were extracted from each band, and the main EEG signal and these extracted features were combined to obtain 125-time domain features (feature extraction). Using the Random Forests classifier, EEG activities were classified into five classes. In the second approach, each One-Against-One (OVO) approach with 125 attributes was split into ten parts, pairwise, and then each piece was classified with the Random Forests classifier. The majority voting scheme was used to combine decisions from the ten classifiers. In the third proposed method, each One-Against-All (OVA) approach with 125 attributes was divided into five parts, and then each piece was classified with the Random Forests classifier. The majority voting scheme was used to combine decisions from the five classifiers. In the fourth proposed approach, each One-Against-All (OVA) approach with 125 attributes was divided into five parts. Since each piece obtained had an imbalanced data distribution, an adaptive synthetic (ADASYN) sampling approach was used to stabilize each piece. Then, each balanced piece was classified with the Random Forests classifier. To combine the decisions obtanied from each classifier, the majority voting scheme has been used.

RESULTS

The first approach achieved 71.90% classification success in classifying five-class EEG signals. The second approach achieved a classification success of 91.08% in classifying five-class EEG signals. The third method achieved 89% success, while the fourth proposed approach achieved 91.72% success. The results obtained show that the proposed fourth approach (the combination of the ADASYN sampling approach and Random Forest Classifier) achieved the best success in classifying five class EEG signals. This proposed method could be used in the detection of epilepsy events in the EEG signals.

Feature extraction of EEG signals based on functional data analysis and its application to recognition of driver fatigue state

OBJECTIVE

Our objective is to study how to obtain features which can reflect the continuity and internal dynamic changes of electroencephalography (EEG) signals and study an effective method for fatigued driving state recognition based on the obtained features.

APPROACH

A method of EEG signalfeature extraction based on functional data analysis is proposed. Combined with kernel principal component analysis method, the obtained features are applied to the recognition of driver fatigue state, and a corresponding recognition model of fatigued driving state is constructed.

MAIN RESULTS

The recognition model is tested on the real collected driver fatigue EEG signals by selecting a suitable classifier. The test results show that the proposed driver fatigue state recognition method has good recognition effect, especially on the classifier based on decision tree, with an average accuracy of 99.50%.

SIGNIFICANCE

The extracted features well reflect the continuityand internal dynamic changes of the EEG signals, and it is of great significance and application value to study an effective method of fatigued driver state recognition based on the features.