Automatic seizure detection is of great significance for epilepsy long-term monitoring, diagnosis, and rehabilitation, it the key to closed-loop brain stimulation. This paper presents a novel wavelet-based automatic method with high sensitivity. The proposed first conducts wavelet decomposition multi-channel intracranial EEG (iEEG) five scales, selects three frequency bands them subsequent processing. Effective features are extracted, such as relative energy, amplitude, coefficient variation...

Reliable prediction of forthcoming seizures will be a milestone in epilepsy research. A method capable timely predicting the occurrence could significantly improve quality life for patients and open new therapeutic approaches. Seizures are usually characterized by generalized spike wave discharges. With advent seizures, variation rate (SR) have different manifestations. In this study, seizure approach based on is proposed evaluated. Firstly, low-pass filter applied to remove high frequency...

Automatic seizure detection plays an important role in long-term epilepsy monitoring, and algorithms have been intensively investigated over the years. This paper proposes algorithm for using lacunarity Bayesian linear discriminant analysis (BLDA) intracranial EEG. Lacunarity is a measure of heterogeneity fractal. The proposed method first conducts wavelet decomposition on EEGs with five scales, selects coefficients at scale 3, 4, 5 subsequent processing. Effective features including...

Automatic seizure detection technology is of great significance for long-term electroencephalogram (EEG) monitoring epilepsy patients. The aim this work to develop a system with high accuracy. proposed was mainly based on multifractal analysis, which describes the local singular behavior fractal objects and characterizes structure using continuous spectrum. Compared computing single dimension, analysis can provide better description transient EEG time series during evolvement from interictal...

Automatic seizure detection technology can automatically mark the EEG by using epileptic algorithm, which is helpful to diagnosis and treatment of diseases. This paper presents an classification framework based on denoising sparse autoencoder. The autoencoder (DSAE) improved unsupervised deep neural network over autoencoder, learn closest representation data. sparsity constraint applied in hidden layer makes expression data as possible so obtain a more efficient signals. In addition,...

Automatic seizure detection is significant for the diagnosis of epilepsy and reducing massive workload reviewing continuous EEGs. In this work, a novel approach, combining Stockwell transform (S-transform) with deep Convolutional Neural Networks (CNN), proposed to detect onsets in long-term intracranial EEG recordings. Primarily, raw data filtered wavelet decomposition. Then, S-transform used obtain proper time-frequency representation each segment. After that, 15-layer CNN using dropout...

Automatic seizure detection plays a significant role in monitoring and diagnosis of epilepsy. This paper presents an efficient automatic method based on Stockwell transform (S-transform) bidirectional long short-term memory (BiLSTM) neural networks for intracranial EEG recordings. First, S-transform is applied to raw segments, the obtained matrix grouped into time-frequency blocks as inputs fed BiLSTM feature selecting classification. Afterwards, postprocessing adopted improve performance,...

Automatic seizure onset detection plays an important role in epilepsy diagnosis. In this paper, a novel method is proposed by combining empirical mode decomposition (EMD) of long-term scalp electroencephalogram (EEG) with common spatial pattern (CSP). First, wavelet transform (WT) and EMD are employed on EEG recordings respectively for filtering pre-processing time-frequency decomposition. Then CSP applied to reduce the dimension multi-channel representation, variance extracted as only...

Visual inspection of long-term electroencephalography (EEG) is a tedious task for physicians in neurology. Based on bidirectional gated recurrent unit (Bi-GRU) neural network, an automatic seizure detection method proposed this paper to facilitate the diagnosis and treatment epilepsy. Firstly, wavelet transforms are applied EEG recordings filtering pre-processing. Then relative energies signals several particular frequency bands calculated inputted into Bi-GRU network. Afterwards, outputs...

Epilepsy is a chronic neurological disorder characterized by sudden and apparently unpredictable seizures. A system capable of forecasting the occurrence seizures crucial could open new therapeutic possibilities for human health. This paper addresses an algorithm seizure prediction using novel feature - diffusion distance (DD) in intracranial Electroencephalograph (iEEG) recordings. Wavelet decomposition conducted on segmented electroencephalograph (EEG) epochs subband signals at scales 3, 4...

Imbalance data classification is a challenging task in automatic seizure detection from electroencephalogram (EEG) recordings when the durations of non-seizure periods are much longer than those activities. An imbalanced learning model proposed this paper to improve identification events long-term EEG signals. To better represent underlying microstructure distributions signals while preserving non-stationary nature, discrete wavelet transform (DWT) and uniform 1D-LBP feature extraction...

Automatic seizure detection is of great significance for epilepsy diagnosis and alleviating the massive burden caused by manual inspection long-term EEG. At present, most methods are highly patient-dependent have poor generalization performance. In this study, a novel patient-independent approach proposed to effectively detect onsets. First, multi-channel EEG recordings preprocessed wavelet decomposition. Then, Convolutional Neural Network (CNN) with proper depth works as an feature...

Electroencephalography (EEG) plays a crucial role in epilepsy analysis, and epileptic seizure prediction has significant value for clinical treatment of epilepsy. Currently, methods using Convolutional Neural Network (CNN) primarily focus on local features EEG, making it challenging to simultaneously capture the spatial temporal from multi-channel EEGs identify preictal state effectively. In order extract inherent relationships among while obtaining their correlations, this study proposed an...

In this study, we proposed and evaluated the use of Independent Component Analysis (ICA) combining EEG dipole model to automatically remove eye movement artifacts from without needing EOG as a reference. We separated data into independent components using ICA method, determined source localization these with single model. The signal was reconstructed by excluding those localized within preset EEGs 12 patients were analyzed. experimental results indicate that is very efficient at subtracting...

The automatic identification of epileptic EEG signals is significant in both relieving heavy workload visual inspection recordings and treatment epilepsy. This paper presents a novel method based on the theory sparse representation to identify EEGs. At first, raw epochs are preprocessed via Gaussian low pass filtering differential operation. Then, scheme classification (SRC), test sample sparsely represented training set by solving l 1 -minimization problem, residuals associated with ictal...

AbstractIn this work, we evaluated the differences between epileptic electroencephalogram (EEG) and interictal EEG by computing some non-linear features. Correlation dimension (CD) Hurst exponent (H) were calculated for 100 segments of EEG. A comparison was made in those parameters. Results show that mean values CD are 2·64 4·55 We also approximate entropy (ApEn) signals. The ApEn 0·90 generally lower than EEG, indicating less complexity signals during seizures. 0·19 0·29 exponents both...