Epilepsy seizure detection using complete ensemble empirical mode decomposition with adaptive noise

doi:10.1016/j.knosys.2019.105333

Knowledge-Based Systems

Volume 191, 5 March 2020, 105333

https://doi.org/10.1016/j.knosys.2019.105333 Get rights and content

Highlights

•
An automated epileptic seizure detection algorithm for EEG signals is proposed.
•
A novel signal processing technique, namely-CEEMDAN is employed.
•
We introduce adaptive boosting for computerized seizure diagnosis.
•
Efficacy of the method is confirmed by statistical and graphical analyses.
•
The performance of the proposed scheme, compared to the existing ones is promising.

Abstract

Background: Epileptic seizure detection is traditionally performed by visual observation of Electroencephalogram (EEG) signals. Owing to its onerous and time-consuming nature, seizure detection based on visual inspection hinders epilepsy diagnosis, monitoring, and large-scale data analysis in epilepsy research. So, there is a dire need of an automatic seizure detection scheme.

Method: An automated scheme for epileptic seizure identification is developed in this study. Here we utilize a signal processing technique, namely-complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) for epileptic seizure identification. First, we decompose segments of EEG signals into intrinsic mode functions by CEEMDAN. The mode functions are then modeled by normal inverse Gaussian (NIG) pdf parameters. In this work, NIG modeling is employed in conjunction with CEEMDAN for epileptic seizure detection for the first time. The efficacy of the NIG parameters in the CEEMDAN domain is demonstrated by intuitive, graphical, and statistical analyses. Adaptive Boosting, an eminent ensemble learning based classification model, is implemented to perform classification.

Results: Experimental outcomes suggest that the algorithmic performance of the proposed scheme is promising in all the cases of clinical significance. Comparative evaluation of algorithmic performance with the state-of-the-art schemes manifest that the seizure detection scheme proposed herein outperforms competing algorithms in terms of accuracy, sensitivity, specificity, and Cohen’s Kappa coefficient.

Conclusions: Upon its implementation in clinical practice, the proposed seizure detection scheme will eliminate the onus of medical professionals and expedite epilepsy research and diagnosis.

Introduction

Epilepsy is one of the most common neurological disorder that is characterized by recurrent occurrences of seizures, which are characterized by neuronal discharges with high and/or extended periods of strong spatial phase coherence. Epilepsy affects about 0.5–1.5% of the population of the world and deteriorates the quality of lives of the patients. In the United States alone, 150,000 new instances of epilepsy are being diagnosed every year [1]. The electrical activities in the neurons of the brain are represented by Electroencephalogram (EEG) signals which are obtained from electrodes placed on the scalp. EEG is an inexpensive modality and is commonly used to diagnose patients with neuro-pathologies. It is therefore widely used in epileptic seizure detection and epilepsy diagnosis [2]. Traditionally, epileptic seizure identification is performed from visual inspection of EEG signals by expert neurologists into various categories such as healthy, seizure (ictal), and non-seizure (inter-ictal).

Seizure detection based on visual observation has a surfeit of caveats associated with it. The feasibility of a device that can constantly and reliably monitor epilepsy patients at home, especially when the patients are asleep is reliant on a successful automated seizure detection algorithm. Moreover, seizure detection by visual scoring is onerous, time-consuming, reliant on expensive human resources, and subject to error due to fatigue [3]. Computer-assisted seizure detection algorithm can eliminate the onus of clinicians of visually analyzing a gargantuan bulk of data and expedite diagnosis. Besides, for the feasibility of a closed-loop and implantable neuro-stimulation devices for seizure suppression, developing an automated epileptic seizure detection scheme is a precondition [4], [5]. So it goes without saying that there is a dire need of an automated epileptic seizure detection algorithm.

Researchers have devised various algorithms for seizure recognition from EEG recordings. Prior works in the literature have employed empirical mode decomposition (EMD) [6], [7], [8], Hilbert marginal spectrum analysis [9], dual-tree complex wavelet transformation [10], multivariate autoregressive modeling [11], or extracted various features directly from the EEG signals [12], [13]. Even though EMD is data-adaptive, it suffers from mode mixing problem. Wavelet based seizure detection schemes, on the other hand, are reliant on the choice of a basis function. The best basis function varies across data-sets, making the detection scheme less data-driven.

Guo et al. [12] employed line length feature in conjunction with artificial neural network for epileptic seizure identification. Song et al. [13] implemented optimized sample entropy and extreme learning machine classifier for automated seizure detection. Bajaj et al. [7] computed modulation bandwidth features in the empirical mode decomposition (EMD) domain and used LS-SVM to perform classification. Alam et al. [8] used EMD and artificial neural network to identify epileptic seizure. Both the aforementioned methods suffer from mode-mixing problem. Peker et al. [10] utilized dual-tree complex wavelet transformation and complex-valued neural network to classify epileptic seizure. Wang et al. [11] implemented multivariate autoregressive model, partial directed coherence and svm classifier for automatic seizure detection. Samiee et al. [14] put forward a rational discrete short-time Fourier transform and statistical feature based feature extraction scheme for epileptic classification. Riaz et al. [6] utilized temporal and spectral statistics in the EMD domain and SVM to devise an automated seizure detection scheme. Fu et al. [9] implemented Hilbert marginal spectrum analysis and SVM for seizure detection from EEG data.

A schematic outline of the proposed framework is given in Fig. 1. The computerized epileptic seizure detection algorithm proposed in this work implements complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) for feature generation. This is the first implementation of CEEMDAN for epileptic seizure detection to the best of the authors’ knowledge. Ar first, using CEEMDAN, EEG epochs or signal segments are decomposed into intrinsic mode functions. Epileptic seizures are associated with aberrant synchronization of neural rhythms. Since CEEMDAN is ideal to analyze nonlinear and nonstationary signals, we hypothesize that we will observe statistical differences in the distributions of CEEMDAN modes between healthy and epileptic classes. The estimated NIG parameters are then computed from each of the mode functions as we can see from Fig. 1. These parameters are used to construct the train and the test matrices, which are eventually fed into the classifier. The classifier used in the proposed scheme is Adaptive Boosting (AdaBoost). Nevertheless, the performance of NIG parameter based features in the CEEMDAN domain is analyzed for various classification models. The efficacy of our feature extraction scheme is validated by intuitive, graphical, and statistical analysis. 10-fold cross-validation is employed both for making optimal choices of AdaBoost parameters and classification model’s prediction capability assessment. The outcomes of our experiments reveal that the performance of the proposed seizure detection scheme is better or comparable than that of the state-of-the-art works in the literature.

The rest of the paper is organized as follows. Section 2 gives a description of our experimental recordings, presents the proposed seizure detection framework, and demonstrates the effectiveness of NIG parameter based features in the CEEMDAN domain. We then present our experimental results and explicate their significance in Section 3. Finally, Section 4 discusses some of the possible future extensions of this work and brings the article to conclusion.

Section snippets

Experimental data

To conduct our experiments, we utilize university of Bonn’s widely used and publicly available epilepsy data-set in this work [15]. The data-set contains 500 single-channel EEG signal segments. Each of the segment is 23.6-s in duration. The recordings have been grouped into five sets, namely A, B, C, D, and E, each of which contains 100 EEG epochs. Both sets A and B contain surface EEG recordings that have been collected from five healthy subjects in relaxed and awake state. International

Experimental results and discussions

To evaluate the proposed seizure detection scheme, we conducted various experiments. Here we describe our experiments, present the outcomes, and explicate their significance. The performance metrics used in this work are accuracy, sensitivity, specificity, and Cohen’s Kappa co-efficient. Cohen’s kappa measures inter-annotator agreement for categorical items statistically. It has been considered as a more robust measurement of performance than simple percent agreement calculation owing to the

Conclusions and future directions

In this study, the problem of automatic seizure detection from single-channel EEG signals has been tackled by performing NIG modeling of CEEMDAN mode functions. AdaBoost is implemented to perform classification. The efficacy of NIG parameters as features has been demonstrated. The performance of NIG parameters in the CEEMDAN domain has been inspected for various common classification models. Performance comparison with competing methods in the literature suggests that our method gives

CRediT authorship contribution statement

Ahnaf Rashik Hassan: Conceptualization, Methodology, Software, Writing - original draft. Abdulhamit Subasi: Supervision, Writing - review & editing. Yanchun Zhang: Supervision, Writing - review & editing.

References (70)

PachoriR.B. et al.
Analysis of normal and epileptic seizure eeg signals using empirical mode decomposition
Comput. Methods Programs Biomed.
(2011)
FuK. et al.
Hilbert marginal spectrum analysis for automatic seizure detection in eeg signals
Biomed. Signal Process. Control
(2015)
WangG. et al.
Epileptic seizure detection based on partial directed coherence analysis
IEEE J. Biomed. Health Inform.
(2015)
GuoL. et al.
Automatic epileptic seizure detection in eegs based on line length feature and artificial neural networks
J. Neurosci. Methods
(2010)
SongY. et al.
Automatic epileptic seizure detection in eegs based on optimized sample entropy and extreme learning machine
J. Neurosci. Methods
(2012)
AndrzejakR. et al.
The epileptic process as nonlinear deterministic dynamics in a stochastic environment: an evaluation on mesial temporal lobe epilepsy
Epilepsy Res.
(2001)
HassanA.R. et al.
A decision support system for automatic sleep staging from eeg signals using tunable q-factor wavelet transform and spectral features
J. Neurosci. Methods
(2016)
HassanA.R. et al.
Computer-aided gastrointestinal hemorrhage detection in wireless capsule endoscopy videos
Comput. Methods Programs Biomed.
(2015)
HassanA.R. et al.
Computer-aided sleep staging using complete ensemble empirical mode decomposition with adaptive noise and bootstrap aggregating
Biomed. Signal Process. Control
(2016)
HassanA.R.
Computer-aided obstructive sleep apnea detection using normal inverse gaussian parameters and adaptive boosting
Biomed. Signal Process. Control
(2016)

HassanA.R. et al.

Computer-aided obstructive sleep apnea screening from single-lead electrocardiogram using statistical and spectral features and bootstrap aggregating

Biocybernet. Biomed. Eng.

(2016)

HassanA.R. et al.

Automatic sleep scoring using statistical features in the emd domain and ensemble methods

Biocybernet. Biomed. Eng.

(2016)

ZhangX. et al.

Image denoising in contourlet domain based on a normal inverse gaussian prior

Digit. Signal Process.

(2010)

FreundY. et al.

A decision-theoretic generalization of on-line learning and an application to boosting

J. Comput. System Sci.

(1997)

OrhanU. et al.

Eeg signals classification using the k-means clustering and a multilayer perceptron neural network model

Expert Syst. Appl.

(2011)

KannathalN. et al.

Characterization of eeg—a comparative study

Comput. Methods Programs Biomed.

(2005)

SubasiA. et al.

Classification of eeg signals using neural network and logistic regression

Comput. Methods Programs Biomed.

(2005)

SubasiA. et al.

Eeg signal classification using wavelet feature extraction and a mixture of expert model

Expert Syst. Appl.

(2007)

PolatK. et al.

Classification of epileptiform eeg using a hybrid system based on decision tree classifier and fast fourier transform

Appl. Math. Comput.

(2007)

KayaY. et al.

1d-local binary pattern based feature extraction for classification of epileptic eeg signals

Appl. Math. Comput.

(2014)

UbeyliE.D.

Least square support vector machine employing model-based methods coefficients for analysis of eeg signals

Expert Syst. Appl.

(2010)

ChandakaS. et al.

Cross-correlation aided support vector machine classifier for classification of eeg signals

Expert Syst. Appl.

(2009)

NicolaouN. et al.

Detection of epileptic electroencephalogram based on permutation entropy and support vector machines

Expert Syst. Appl.

(2012)

SubasiA. et al.

Effect of photic stimulation for migraine detection using random forest and discrete wavelet transform

Biomed. Signal Process. Control

(2019)

RahmanM.M. et al.

Sleep stage classification using single-channel eog

Comput. Biol. Med.

(2018)

HassanA.R. et al.

Automated identification of sleep states from eeg signals by means of ensemble empirical mode decomposition and random under sampling boosting

Comput. Methods Programs Biomed.

(2017)

HassanA.R. et al.

A decision support system for automated identification of sleep stages from single-channel eeg signals

Knowl.-Based Syst.

(2017)

HassanA.R. et al.

An expert system for automated identification of obstructive sleep apnea from single-lead ecg using random under sampling boosting

Neurocomputing

(2017)

HassanA.R. et al.

An automated method for sleep staging from eeg signals using normal inverse gaussian parameters and adaptive boosting

Neurocomputing

(2017)

HassanA.R.

Computer-aided obstructive sleep apnea detection using normal inverse gaussian parameters and adaptive boosting

Biomed. Signal Process. Control

(2016)

SubasiA. et al.

Effect of photic stimulation for migraine detection using random forest and discrete wavelet transform

Biomed. Signal Process. Control

(2019)

Krook-MagnusonE. et al.

Beyond the hammer and the scalpel: selective circuit control for the epilepsies

Nature Neurosci.

(2015)

PekerM.

A new approach for automatic sleep scoring: combining taguchi based complex-valued neural network and complex wavelet transform

Comput. Methods Programs Biomed.

(2016)

SalamM.T. et al.

A novel low-power-implantable epileptic seizure-onset detector

IEEE Trans. Biomed. Circuits Syst.

(2011)

SalamM.T. et al.

An implantable closedloop asynchronous drug delivery system for the treatment of refractory epilepsy

IEEE Trans. Neural Syst. Rehabil. Eng.

(2012)

Cited by (118)

Classification of epileptic EEG signals with the utilization of Bonferroni mean based fuzzy pattern tree
2024, Expert Systems with Applications
The Electroencephalogram (EEG) is an essential tool used to detect and investigate multiple neurological disorders within the human brain. However, examination and visualization of any such abnormalities in the brain through inspecting these EEG signals is a time-consuming and repetitive task for a neurologist. Hence, there is a need to develop a model that can automatically classify brain disorders based on the EEG signals. In this study, a novel fuzzy classification model is proposed for classifying the EEG signals. First, the discrete wavelet transform (DWT) technique is employed to decompose the signal into time–frequency sub-bands. Then, by utilizing these sub-bands, seven meaningful statistical features are extracted effectively for additional investigation purposes. Then, the retrieved features are provided to the fuzzy pattern tree classifier to obtain meaningful inferences from the data. Here, a distinctive structure of a fuzzy pattern tree is proposed by utilizing the interrelationship handling aggregation operators. More precisely, we utilize the Bonferroni mean (BM) aggregation operator for the enlargement or expansion of the tree while considering the concatenation between the parent and the slave tree. To the best of our knowledge, this is the first work that presents such an interrelationship handling fuzzy logic-based classifier model for EEG signal classification. The proposed algorithm is evaluated using two publicly available datasets, namely the Bonn University (Bonn) epileptic datasets and the Temple University Hospital (TUH) EEG subcorpus of abnormal (TUAB) and epilepsy (TUEP) datasets. This study considers different frames of the binary problem (i.e., healthy vs. seizure detection) and multi-class problem (i.e., healthy vs. seizure free vs. seizure) over the Bonn dataset and achieves more than 99% accuracy over the binary model by testing over all possible combinations of the available experimental dataset and 97.8% accuracy in multi-class model. In the case of TUAB and TUEP categories, the model achieves an accuracy of 88% in predicting normal to abnormal activities and 86% in detecting epilepsy and non-epilepsy within the signals. The outcomes derived from the proposed model outperform some existing studies available in the literature for EEG signal classification. Also, the model is efficient as it considers the uncertainty and association among the features while framing the classifier model.
Shorter latency of real-time epileptic seizure detection via probabilistic prediction
2024, Expert Systems with Applications
Although recent studies have proposed seizure detection algorithms with good sensitivity performance, there is a remained challenge that they were hard to achieve significantly short detection latency in real-time scenarios. In this manuscript, we propose a novel deep learning framework intended for shortening epileptic seizure detection latency via probabilistic prediction. We are the first to convert the seizure detection task from traditional binary classification to probabilistic prediction by introducing a crossing period from seizure-oriented EEG recording and proposing a labeling rule using soft-label for crossing period samples. And, a novel multiscale STFT-based feature extraction method combined with 3D-CNN architecture is proposed to accurately capture predictive probabilities of samples. Furthermore, we also propose rectified weighting strategy to enhance predictive probabilities, and accumulative decision-making rule to achieve significantly shorter detection latency. We implement the proposed framework on two prevalent datasets — CHB-MIT scalp EEG dataset and SWEC-ETHZ intracranial EEG dataset in patient-specific leave-one-seizure-out cross-validation scheme. Eventually, the proposed algorithm successfully detected 94 out of 99 seizures during crossing period and 100% seizures detected after EEG onset, averaged 14.84% rectified predictive ictal probability (RPIP) errors of crossing samples, 2.3 s detection latency, 0.08/h false detection rate (FDR) on CHB-MIT dataset. Meanwhile, 84 out of 89 detected seizures during crossing period, 100% detected seizures after EEG onset, 16.17% RPIP errors, 4.7 s detection latency, and 0.08/h FDR are achieved on SWEC-ETHZ dataset. The obtained detection latencies are at least 50% shorter than state-of-the-art results reported in previous studies.
MRP-Net: Seizure detection method based on modified recurrence plot and additive attention convolution neural network
2023, Biomedical Signal Processing and Control
Electroencephalographic (EEG) signals play an important role in the detection of seizures in epilepsy, and accurate detection of seizures can buy patients valuable treatment time. However, most seizure detection methods ignore the nonlinear, implicit characteristics of EEG, which has some impact on the accuracy of detection. Therefore, in this paper, we propose an EEG epilepsy detection network (MRP-Net) based on modified recurrence plot (MRP) and additive attention convolution neural network. This network can fully take into account the nonlinear, occult characteristics of EEG which can be mapped to the two-dimensional plane and served as the input of additive attention convolution neural network to automatically learn, analyze, and extract the EEG characteristics of seizures. The performance of the proposed method was evaluated on the Bonn University dataset and the SWEC-ETHZ short-term dataset. Sensitivity, specificity and accuracy were 100% in multiple detection tasks in the University of Bonn single-channel EEG dataset. The sensitivity, specificity and accuracy of SWEC’s short-term, multi-channel EEG dataset were 99.77%, 99.57% and 99.69%, respectively, higher than the latest methods (3.76%, 4.73% and 4.27%). The results of the experiments show that the network in this paper is superior and universal in epilepsy detection.
Epileptic seizure detection on a compressed EEG signal using energy measurement
2023, Biomedical Signal Processing and Control
Epilepsy is a common neurological disorder affecting both children and adults. It can trigger seizures without any stimuli. An accurate diagnosis of epilepsy is essential to the treatment and management of the patient’s condition. Electroencephalography EEG is a non-invasive technique for the diagnosis of epilepsy. Data analysis and processing can be challenging due to the high dimensionality of EEG signals. In this study, the dimensionality reduction technique using compressive sensing is applied to EEG signals to optimize the process and improve the accuracy of the classification in the epileptic EEG signals. The CS method is applied to minimize processed data by reducing EEG signal dimensions, and energy is calculated on the compressed signal as a feature. The performance of the proposed method is evaluated using SVM with Linear, Quadratic, Cubic, and Radian Basis Function kernels through 5- and 10-fold cross-validation. The results showed 100% accuracy in the classification of ictal and inter-ictal EEG signals with compression ratios of 1/2, 1/4, and 1/8. The results demonstrate that CS is effective in analyzing EEG and diagnosing epilepsy. In addition, this study has significant implications for further development of dimensionality reduction techniques in various EEG-based applications, such as epilepsy classification, sleep stage analysis, and brain–computer interface.
Pre-processed and combined EEG data for epileptic seizure classification using deep learning
2023, Biomedical Signal Processing and Control
Epilepsy is a neurological disease that affects nearly 60 million people around the world. Electroencephalography (EEG) and Machine Learning (ML) can be used to detect epileptic seizures, where ML algorithms depend on hand crafted features. Recently, deep learning (DL) methods became popular due to automatic feature learning. However, current DL based approaches mainly focuses on different DL architecture design, especially deeper DL networks. Only few works experiment modified EEG data as an input to DL models. In this work, we propose modified, pre-processed and combined EEG signals as an input to different DL models for epilepsy seizure classification rather than using raw EEG signals. A variety of pre-processed and combined EEG signals have been evaluated with three different DL architectures: A simple Deep Neural Network (DNN) model and a moderately complex 1D-Convolutional Neural Network (CNN) model and a complex 1D-CNN model. We perform several experiments for two-class and three-class epileptic seizure classification on UCI-Bonn dataset using the proposed EEG signals: (1)Original EEG signal of UCI-Bonn dataset, (2)standardized EEG signal, (3)original EEG signal combined with squared EEG signal, and then standardized, (4)original EEG signal combined with differentiated EEG signal, and then standardized, and (5)original EEG signal combined with magnitude of Fast Fourier Transform (FFT) of EEG signal, and then standardized (original + FFT). Extensive evaluations and comparisons show that the best results are achieved with original + FFT combination on all DL architectures. Using original + FFT EEG signal, competitive results can be achieved with simple and moderately complex DL models, by utilizing less computational resources.
Phase space reconstruction, geometric filtering based Fisher discriminant analysis and minimum distance to the Riemannian means algorithm for epileptic seizure classification
2023, Expert Systems with Applications
Most clinicians observe the changes of the electroencephalograms (EEGs) via the visual inspection of the recorded signals of the patients. However, the visual inspection is time consuming. Also, it is subjective and empirical. Therefore, designing an accurate and an automatic epilepsy classification system has a profound significance for reducing the workload of the medical officers.
First, this paper applies the phase space reconstruction (PSR) method for constructing the covariance matrix for characterizing the chaotic states of the EEGs. The covariance matrices are symmetric positive definite (SPD) matrices used as the feature descriptors of the EEGs. This set of the covariance matrices is in fact a kind of the Riemannian manifold. Denote this Riemannian manifold as $M$ . Secondly, the tangent space at the Riemannian geometric mean is computed via the logarithmic mapping operator. Denote this tangent space as $T_{M}$ . This tangent space $T_{M}$ is a kind of the Euclidean space containing a point in the Riemannian manifold $M$ . That is, the SPD matrices in the Riemannian manifold $M$ are mapped to the matrices in the tangent space $T_{M}$ at the Riemannian geometric mean via the logarithmic mapping operator. Then, these matrices in the tangent space $T_{M}$ are further mapped to the tangent feature vectors. The set of the tangent feature vectors is a kind of the vector space. Denote this vector space as $V_{T}$ . Thirdly, the transformation matrix based on the Fisher discrimination analysis (FDA) is used to map the high dimensional vectors in the vector space $V_{T}$ to the low dimensional vectors in another vector space. Then, the discriminative vectors are projected back to the Riemannian manifold $M$ via the exponential mapping operator. Finally, the minimum distance to the Riemannian means (MDRM) algorithm is applied for performing the epileptic seizure classification.
The computer numerical simulations are conducted based on three well known EEGs epileptic seizure datasets. Overall, our proposed method achieves the best performances for three datasets compared to the state of the art methods. Also, the computational complexity of our proposed method is low.
Our proposed method is effective and robust for performing the seizure classification. Also, our proposed method can be implemented in the real time.

View all citing articles on Scopus

^☆: No author associated with this paper has disclosed any potential or pertinent conflicts which may be perceived to have impending conflict with this work. For full disclosure statements refer to https://doi.org/10.1016/j.knosys.2019.105333.

View full text

Epilepsy seizure detection using complete ensemble empirical mode decomposition with adaptive noise☆

Highlights

Abstract

Introduction

Section snippets

Experimental data

Experimental results and discussions

Conclusions and future directions

CRediT authorship contribution statement

Comput. Methods Programs Biomed.

Biomed. Signal Process. Control

IEEE J. Biomed. Health Inform.

J. Neurosci. Methods

J. Neurosci. Methods

Epilepsy Res.

J. Neurosci. Methods

Comput. Methods Programs Biomed.

Biomed. Signal Process. Control

Biomed. Signal Process. Control

Biocybernet. Biomed. Eng.

Biocybernet. Biomed. Eng.

Digit. Signal Process.

J. Comput. System Sci.

Expert Syst. Appl.

Comput. Methods Programs Biomed.

Comput. Methods Programs Biomed.

Expert Syst. Appl.

Appl. Math. Comput.

Appl. Math. Comput.

Expert Syst. Appl.

Expert Syst. Appl.

Expert Syst. Appl.

Biomed. Signal Process. Control

Comput. Biol. Med.

Comput. Methods Programs Biomed.

Knowl.-Based Syst.

Neurocomputing

Neurocomputing

Biomed. Signal Process. Control

Biomed. Signal Process. Control

Beyond the hammer and the scalpel: selective circuit control for the epilepsies

Nature Neurosci.

A new approach for automatic sleep scoring: combining taguchi based complex-valued neural network and complex wavelet transform

Comput. Methods Programs Biomed.

A novel low-power-implantable epileptic seizure-onset detector

IEEE Trans. Biomed. Circuits Syst.

An implantable closedloop asynchronous drug delivery system for the treatment of refractory epilepsy

IEEE Trans. Neural Syst. Rehabil. Eng.