University Of Bonn Research Articles

Fruitfly optimizer with deep neural network for the detection of brain tumours using EEG signals

Presently, the brain signals are recorded by the various techniques like MRI, FMRI, CT SCAN and EEG. Here we are using EEG technique to record the brain signals in this study. As it is a common technique used extensively for the detection of brain tumours. The motive behind this study is to examine the variations of the brain tumours. Here two different database of the EEG recordings are considered; Bern Barcelona (BB), and Bonn University (BU). The Butterworth filter is applied to take out the artifacts from the EEG signals. The decomposition technique is used to decompose the denoised EEG signals. This technique of decomposition helps to determine the necessary information, for the complete modeling of the multifarious time series data. This Fruitfly optimizer is used to extract the features from the main data signals. Finally, the data is classified into two different environments that is one being seizure and the other one is healthy. From the experimental simulation, proposed model has better accuracy in comparison, with the previous models. The significant improvement is observed in terms of accuracy is 5. 49%, sensitivity is 5%, specificity is 6% and positive predictive value is about 4 %.

Зв"язки КДУ ім.Т.Г. Шевченка з освітніми й науковими установами країн Заходу (1944 - 1975 рр.)

The article reveals the international cooperation of Taras Shevchenko Kyiv State University with scientific and educational institutions of Western countries in 1944–1975. It was noted that at the end of the Second World War (1944–1945), Kyiv University couldn’t establish ties with educational and scientific institutions of Western countries due to the reconstruction of the city and the university itself. During the period of post-war reconstruction (1946–1950), the Soviet-Western confrontation was added to the mentioned problems, which then turned into the Cold War. However, the liberal social and political changes in the USSR associated with de-Stalinization (1953–1956) and the Khrushchev «Thaw» (1956–1964) had a positive impact on the international activities of the Soviet higher school and KSU named T. G. Shevchenko. It is indicated that since the mid-1950s, delegations and individual scientists from France, Austria, Belgium, and Sweden began to visit Kyiv University. Since the second half of the 1950s, teachers and scientists from Finland and Great Britain, as well as Communist Party leaders, and representatives of student and trade union organizations from Western countries visited Kyiv University to give lectures and deliver scientific reports. However, in 1959–1960, plans for the teaching work of Taras Shevchenko Kyiv State University ​in the Great Britain universities remained unrealized. Nevertheless, since then, teachers and scientists of Kyiv University have actively participated in international scientific events held in Western countries (Madrid, Paris, London, Vienna, and Stockholm). Some teachers completed internships at universities in Italy, France, and Great Britain. Students also did internships in these countries. Mostly, these were senior-year students of the Faculty of Philology who were studying foreign languages. It is noted that the scientific works and teachers of Taras Shevchenko Kyiv State University were published abroad. Among them were Professor Mytropolskyi Yu. (in Great Britain and Sweden), Professor Vsekhsvyatskyi S. (in Great Britain and Belgium), Professor Bileckyi A. (in Greece), Professor Marynych O. (in Great Britain and Sweden) works. Scientists of Kyiv University worked with colleagues from universities and scientific institutions of the West on common scientific themes. The international book exchange of Kyiv State University, as of July 1, 1965, was held with such universities as the Taylor Institute at the University of Oxford, the University of Oslo, the Mathematical Institute at the University of Bonn, Liège (Belgium), Besanson and Cannes (France) universities, and also by the academies of sciences of Denmark and Ireland. The emergence of an international détente in the relations between the West and the USSR at the end of the 1960s had a positive effect on the ties of Taras Shevchenko Kyiv State University with the countries of the West. The number of their youth at Kyiv University continued to grow. Thus, if in 1969 one representative of a Belgian and a Frenchman studied at the university, then as of January 1, 1975, 60 students from the «capitalist countries» studied at the university. In turn, the cooperation of Taras Shevchenko Kyiv State University with educational and scientific institutions of Western countries in 1975–1991 remains understudied. However, this is the subject of the next scientific research.

Enhanced Gradient Boosting Machines Fusion based on the Pattern of Majority Voting for Automatic Epilepsy Detection

Automatic detection of epilepsy based on EEG signals is one of the interesting fields to be developed in medicine to provide an alternative method for detecting epilepsy. High accuracy values are very important for accurate diagnosis in detecting epilepsy and avoid errors in diagnosing patients. Therefore, this study proposes the Enhanced Gradient Boosting Machines Fusion (Enhanced GBM Fusion) for automatically detecting epilepsy based on electroencephalographic (EEG) signals. Enhanced part of GBM Fusion is the pattern of majority voting evaluation based on the fusion of five-class and two-class GBM, called Enhanced GBM Fusion. The raw signal is extracted using Discrete Fourier Transform (DFT) and Discrete Wavelet Transform (DWT), then feature is selected by using Genetic Algorithm (GA) before classification. This proposed method was evaluated using five classes (normal in open eyes, normal in close eyes, interictal with hippocampal, interictal, and ictal) from the University of Bonn. The experimental results show that the proposed Enhanced GBM Fusion can increase the accuracy of GBM Fusion of 99.8% to classify five classes of epilepsy based on EEG signal. However, the performance of Enhanced GBM Fusion cannot be generalized to other datasets.

A combination of statistical parameters for epileptic seizure detection and classification using VMD and NLTWSVM

The epileptic seizure detection and classification is of great significance for clinical diagnosis and treatment. To realize the detection and classification of epileptic seizure, this paper proposes a method based on the combination of signal decomposition and statistical methods. First, the algorithm of variational mode decomposition (VMD) is applied to extract the components of intrinsic mode functions (IMFs) by decomposing the EEG signals. Then the statistical method is utilized to calculate the eight features of maximum, minimum, average, variance, skewness, kurtosis, coefficient of variation and volatility index for each extracted IMF component. Finally, the best combinations of extracted features are fed into the non-linear twin support vector machine (NLTWSVM) to classify the epileptic signals. The EEG database from University of Bonn is used to confirm the effectiveness of the proposed method for epileptic seizure detection. The final experimental results demonstrate that the classification accuracy can reach 98.86%, 98.37%, 99.02%, 99.41% and 99.57% for the database of C-E, D-E, CD-E, ABCD-E and AB-CD-E, respectively. The TUSZ corpus in the TUH EEG corpus is also used to classify epileptic seizure types using the method in this article. The result is expressed by the confusion matrix and the weighted F1 score is 0.923, which shows this method has potential to help experienced neurophysiologists classify epileptic seizure types in the clinic.

Epileptic Seizure Detection Using Brain-Rhythmic Recurrence Biomarkers and ONASNet-Based Transfer Learning.

The electroencephalogram (EEG) tool has great potential for real-time monitoring of abnormal brain activities, such as preictal and ictal seizures. Developing an EEG-based detection system for patients with epilepsy is vital for clinical management and targeted therapy. This paper proposes a single-channel seizure detection system using brain-rhythmic recurrence biomarkers (BRRM) and an optimized model (ONASNet). BRRM is a direct mapping of the recurrence morphology of brain rhythms in phase space; it reflects the nonlinear dynamics of original EEG signals. The architecture of ONASNet is determined through a modified neural network searching strategy. Then, we exploited transfer learning to apply ONASNet to our EEG data. The combination of BRRM and ONASNet leverages the multiple channels of a neural network to extract features from different brain rhythms simultaneously. We evaluated the efficiency of BRRM-ONASNet on the real EEG recordings derived from Bonn University. In the experiments, different transfer-learning models (TLMs) are respectively constructed using ONASNet and seven well-known neural network structures (VGG16/VGG19/ResNet50/InceptionV3/DenseNet121/Xception/NASNet). Moreover, we compared those TLMs by model size, computing complexity, learning capability, and prediction latency. ONASNet outperforms other structures by strong learning capability, high stability, small model size, short latency, and less requirement of computing resources. Comparing BRRM-ONASNet with other existing methods, our work performs better than others with 100% accuracy under the identical dataset and same detection task. Contributions: The proposed method in this study, analyzing nonlinear features from phase-space representations using a deep neural network, provides new insights for EEG decoding. The successful application of this method in epileptic-seizure detection contributes to computationally medical assistance for epilepsy.

Siamese Network‐Based Feature Transformation for Improved Automated Epileptic Seizure Detection

Epilepsy is a common electrophysiological disorder of the brain, detected mainly by electroencephalogram (EEG) signals. Since correctly diagnosing epilepsy seizures by monitoring the EEG signal is very tedious and time‐consuming for a neurologist, a growing number of studies have been conducted on developing automated epileptic seizure detection (AESD). In this work, a novel supervised model is proposed for AESD. Initially, the EEG signals are collected from Bonn University EEG (BU‐EEG) database. Then, empirical mode decomposition and feature extraction (combination of entropy, frequency, and statistical features) are applied to extract the features. Furthermore, Siamese network is utilized to lessen the number of extracted features and obtain the most discriminative features. Then, these features are exploited to classify seizure and non‐seizure EEG signals by using a support vector machine classifier. This paper examines the Siamese network’s contribution as a learning‐based feature transformation in improving seizure detection performance. The numerical results confirm that the proposed framework can achieve a perfect classification performance (100% accuracy). This approach can constructively help doctors to detect epileptic seizure activity and reduce their workload.

An automated approach for electroencephalography-based seizure detection using machine learning algorithms

Epilepsy is a chronic neurological disease that causes recurrent life-threatening seizures due to irregular brain activity. The purpose of seizure detection algorithms is to detect seizures from electroencephalography (EEG) recordings accurately. The main goal of our work is to facilitate an automated seizure detection system using machine learning algorithms. We proposed a model that evaluates eight machine learning algorithms on the Bonn University dataset. Two classifiers, Random Forest and Gaussian Naive Bayes, achieve the highest accuracy of 100% with 100% sensitivity, 100% specificity, 0.01 FPR, and 0.99 AUC with feature extraction. These two algorithms also work better without using feature extraction. This performance is superior to existing seizure detection approaches and comparable to deep learning approaches. Our work comprehensively compares the traditional machine learning algorithms and reinforces the effectiveness of feature extraction. Our work contributes to aiding neurologists in making faster and more precise decisions for epilepsy treatment.

Epilepsy detection from electroencephalogram signal using singular value decomposition and extreme learning machine classifier

Automatic detection of seizure plays an important role for both long-term monitoring and diagnosis of epilepsy. In this work, the proposed singular value decomposition-extreme learning machine (SVD-ELM) classifier technique provide good generalised performance with a remarkable fast learning speed in comparison to existing conventional techniques. Here, both feature extraction and classification of EEG signal has been done for detection of epileptic seizure of human brain, taking Bonn University dataset. Proposed method is based upon the multi-scale eigenspace analysis of the matrices generated using discrete wavelet transform (DWT) of EEG signal by SVD at substantial scale and are classified using extracted singular value features and extreme learning machine (ELM) with dissimilar activation functions. The proposed SVD-ELM technique has been applied for the first time on EEG signal for epilepsy detection using five class classification which produces overall accuracy of 95% (p < 0.001) with sine and radbas activation function.

Automatic seizure detection with different time delays using SDFT and time-domain feature extraction.

Automatic seizure detection is important for fast detection of the seizure because the way that the expert denotes and searches for seizure in the long signal takes time. The most common way to detect seizures automatically is to use an electroencephalogram (EEG). Many studies have used feature extraction that needs time for calculation. In this study, sliding discrete Fourier transform (SDFT) was applied for conversion to a frequency domain without using a window, which was compared with using window for feature selection. SDFT was calculated for each time series sample directly without any delay by using a simple infinite impulse response (IIR) structure. The EEG database of Bonn University was used to test the proposed method, and two cases were defined to examine a two-classifier feedforward neural network and an adaptive network-based fuzzy inference system. Results revealed that the maximum accuracies were 93% without delay and 99.8% with a one-second delay. This delay accrued because the average was taken for the results with a one-second window.

Gradient boosting machines fusion for automatic epilepsy detection from EEG signals based on wavelet features

Automatic epilepsy detection from electroencephalogram (EEG) signals is an alternative to manual detection performed by a human expert. High classification performance is needed in automatic epilepsy detection from EEG signals to avoid miss detection. This study aims to propose a classification method for automatic epilepsy detection from EEG signals. The original EEG signals were processed using discrete Fourier transform (DFT) and discrete wavelet transform (DWT) prior to feature extraction. A fusion of 2-class and 3 class gradient boosting machines (GBM), called GBMs fusion, was used to classify EEG signals based on some statistical features and crossing frequency features. In addition, a genetic algorithm was used to select the prominent features before classification. The proposed method has been evaluated using three classes EEG signals (normal-interictal-ictal) included in EEG dataset from University of Bonn. The experimental result shows that the proposed GBMs fusion can improve the performance of a single GBM in classifying EEG signals. Furthermore, the proposed GBMs fusion can perfectly detect epilepsy from EEG signals with an accuracy of 100%. However, the performance of GBMs fusion may not be generalized to the other EEG dataset.

Detection of epileptic seizures on EEG signals using ANFIS classifier, autoencoders and fuzzy entropies

Epileptic seizures are one of the most crucial neurological disorders, and their early diagnosis will help the clinicians to provide accurate treatment for the patients. The electroencephalogram (EEG) signals are widely used for epileptic seizures detection, which provides specialists with substantial information about the functioning of the brain. In this paper, a novel diagnostic procedure using fuzzy theory and deep learning techniques is introduced. The proposed method is evaluated on the Bonn University dataset with six classification combinations and also on the Freiburg dataset. The tunable-Q wavelet transform (TQWT) is employed to decompose the EEG signals into different sub-bands. In the feature extraction step, 13 different fuzzy entropies are calculated from different sub-bands of TQWT, and their computational complexities are calculated to help researchers choose the best set for various tasks. In the following, an autoencoder (AE) with six layers is employed for dimensionality reduction. Finally, the standard adaptive neuro-fuzzy inference system (ANFIS), and also its variants with grasshopper optimization algorithm (ANFIS-GOA), particle swarm optimization (ANFIS-PSO), and breeding swarm optimization (ANFIS-BS) methods are used for classification. Using our proposed method, ANFIS-BS method has obtained an accuracy of 99.74% in classifying into two classes and an accuracy of 99.46% in ternary classification on the Bonn dataset and 99.28% on the Freiburg dataset, reaching state-of-the-art performances on both of them.

Epileptic seizure classification using level-crossing EEG sampling and ensemble of sub-problems classifier

Epilepsy is a disorder of the brain characterized by seizures and requires constant monitoring particularly in serious patients. Electroencephalogram (EEG) signals are frequently used in epilepsy diagnosis and monitoring. A new paradigm of battery packed wearable gadgets has recently gained popularity which constantly monitor a patient’s signals. These gadgets acquire the data and transmit it to the cloud for further processing. Power consumption due to data transmission is a major issue in these devices. Moreover, in a constant monitoring environment, the number of classes to be identified are usually higher and overlapping. Existing techniques either require the entire data to be transmitted, such as in deep learning, or suffer from reduced accuracy. In this context, we propose a new framework for EEG based epilepsy detection which requires a low data transmission while maintaining high accuracy for multiclass classification. At the device-end, we use a preprocessing mechanism that uses adaptive rate sampling, modified activity selection, filtering, and wavelet decomposition to extract only a handful of highly discriminatory features to be transmitted instead of the entire EEG waveform. For multiclass classification, we propose a novel ensemble of sub-problems-based classification paradigm to achieve high accuracy using the reduced data. Our proposed solution shows many-fold increase in computational gains and an accuracy of 100% and 99.38% on the 2-class problem when tested on the popular University of Bonn and CHB-MIT datasets, respectively. An accuracy of 99.6% on 3-class, 96% on 4-class, and 92% on 5-class problems is obtained for the University of Bonn dataset.

Application of computer-generated images to train pattern recognition used in semiquantitative immunohistochemistry scoring.

This study aimed to clarify whether the pattern recognition involved in scoring proliferation fractions can be trained by abstract computerized images of virtual tissues. Twenty computer-generated images with randomly distributed blue or red dots were scored by 12 probands (all co-workers or collaborators of the Institute of Pathology, University of Bonn). Afterward, the probands underwent a training phase during which they received an immediate feedback on the actual rate of positivity after each image. Finally, the initial testing series was rescored. In a second round with 15 different probands, 20 Ki-67 immunohistochemistry images of tonsil tissue were scored, followed by the same training phase with computer-generated images, before the immunohistochemistry slides were scored again. Paired t-tests were used to compare the differences in mean rates pre- and post-training. Concerning computerized images, untrained probands scored the percentages of positive dots with a mean deviation from the true rates of 8.2%. Following training, the same testing series was scored significantly better with a mean deviation of 4.9% (mean improvement 3.3%, p < 0.001). Scoring real immunohistochemistry slides, the training with computerized images also improved correct estimations, albeit to a lesser degree (mean improvement 1%, p = 0.03). Abstract computerized images of virtual tissues may be a useful tool to train and improve the accuracy of pattern recognition involved in semiquantitative scoring of immunohistochemistry slides. As a side results, this study highlights the value of computer-generated images to verify the performance of image-analysis software.

BIOM-24. PROTEIN SURFACE SIGNATURE ON SERUM EXTRACELLULAR VESICLES FOR NON-INVASIVE DETECTION OF TUMOR PROGRESSION IN GLIOBLASTOMA PATIENTS

Abstract INTRODUCTION Detection of tumor progression in glioblastoma patients remains a major challenge for clinicians due to equivocal MRI results. Extracellular vesicles (EVs) are potential biomarkers and can be detected in the blood of tumor patients. In this study, we evaluated the potential of serum-derived EVs from glioblastoma patients to serve as a marker for tumor progression in adjunction with MRI assessment. METHODS Glioblastoma patients from two independent cohorts, one from the multicenter Phase III CeTeG/NOA-09 trial (n=36) and the other from patients treated at the University of Bonn (n=31), were included in this study. EVs from serum of glioblastoma patients and healthy volunteers were separated by size exclusion chromatography and ultracentrifugation. EV markers were defined by using a proximity-extension assay and bead-based flow cytometry. Tumor progression was defined according to modified RANO criteria. RESULTS EVs from the serum of glioblastoma patients (n=67) showed an upregulation of CD29 (p=0.08), CD44 (p&amp;lt; 0.0001), CD81 (p&amp;lt; 0.0001), CD146 (p&amp;lt; 0.0001), C1QA (p=0.003), and histone H3 (p&amp;lt; 0.0001) as compared to serum EVs from healthy volunteers. For both independent cohorts of glioblastoma patients, we noted upregulation of C1QA, CD44, and histone H3 upon tumor progression, but not in patients with stable disease. Notably, six patients with worse survival compared to the median survival of the cohort did not fulfill RANO criteria at the time of suspected progression, yet showed an elevation of at least one out of these three markers. In a multivariable logistic regression analysis, a combination of CD29, CD44, CD81, C1QA, and histone H3 correlated with RANO-defined tumor progression with an AUC of 0.76. CONCLUSION Measurement of CD29, CD44, CD81, C1QA, and histone H3 in serum-derived EVs of glioblastoma patients, along with standard MRI assessment, could improve detection of true tumor progression and thus be a useful tool for clinical decision making.

Classification of Epileptic Seizure From EEG Signal Based on Hilbert Vibration Decomposition and Deep Learning.

A convolution neural network (CNN) architecture has been designed to classify epileptic seizures based on two-dimensional (2D) images constructed from decomposed mono-components of electroencephalogram (EEG) signals. For the decomposition of EEG, Hilbert vibration decomposition (HVD) has been employed. In this work, four brain rhythms - delta, theta, alpha, and beta have been utilized to obtain the mono-components. Certainly, the data-driven CNN model is most efficient for 2D image processing and recognition. Therefore, 2D images have been generated from one-dimensional (1D) decomposed mono-components by employing continuous wavelet transform (CWT). Next, simultaneous multiple input images in parallel have been directly fed into the CNN pipeline for feature extraction and classification. For evaluation, the EEG dataset provided by the Bonn University has been taken into consideration. Further, a 5-fold cross-validation technique has been applied to obtain generalized and robust classification performance. The average classification accuracy, sensitivity, and specificity reached up to 98.6%, 97.2%, and 100% respectively. The results show that the proposed idea is very much efficient in seizure classification. The proposed idea resourcefully combines the advantages of HVD and CNN to classify epileptic seizures from EEG signal.

Unleashing JupyterHub: Exploiting Resources Without Inbound Network Connectivity Using HTCondor

In recent years Jupyter notebooks have conquered class rooms and some scientists also enjoy their convenience to quickly evaluate ideas and check whether a more detailed study is justified. To lower the threshold for getting started with Jupyter notebooks and to ease sharing and collaborative use, offering a JupyterHub service is tempting. However, offering such a service for a larger science class also requires a compute backend with sufficient resources such that hundreds of notebooks can be run simultaneously. Since resource usage for teaching activities typically fluctuates significantly over the year, dedicated compute resources seem inefficient. In this paper we present an alternative by exploiting an existing high throughput computing cluster (BAF2) at the University of Bonn, which comes with the additional advantage that scientific users may use the very same software and data environment they also select for their batch jobs. To implement this, we used a novel approach which allowed us to integrate BAF2 execute nodes although they do not have inbound network connectivity. Therefore, it does not touch the security concept of the cluster. The very same technique can be used to integrate any compute resources without inbound network connectivity and thus allows one to overcome usual firewall restrictions. This design also simplifies exploiting remote resources e.g. offered by resource federations or cloud providers.

The impact of COVID-19-related distress on levels of depression, anxiety and quality of life in psychogeriatric patients.

Within the elderly population, psychogeriatric patients may be particularly susceptible to negative mental health effects of the coronavirus crisis. Detailed information about the psychosocial well-being of psychogeriatric patients during the pandemic is still sparse. Here we examined which aspects of subjective experience of the COVID-19 pandemic especially affect levels of depression, anxiety and quality of life in psychogeriatric patients with and without cognitive impairment. A cross-sectional paper survey was conducted during the first German lockdown among patients with a diagnosed psychiatric disorder (≥ 60 years) or a diagnosed neurodegenerative disease (regardless of their age) from the department for neurodegenerative diseases and geriatric psychiatry at the University of Bonn. The WHO-5-, GAD-7- and WHOQOL-old score were used to determine levels of depression, anxiety and quality of life. The second part obtained information about the subjective experience of the COVID-19 pandemic. Statistical analysis included among others principal component analysis and multiple linear regression analysis. COVID-19-related, immediate distress was a strong predictor of elevated symptoms of depression, anxiety and a reduced quality of life. COVID-19-related concerns regarding health and financial security, however, were not significantly associated with negative mental health outcomes. The overall prevalence of symptoms of depression (50.8% [95% CI 43.8–57.6%]) and anxiety (32.7% [95% CI 26.4–39.2%]) among psychogeriatric patients was high. Our findings indicate that psychogeriatric patients are not significantly affected by COVID-19-related concerns but are primarily suffering from emotional consequences resulting from changed living conditions due to the pandemic.Supplementary InformationThe online version contains supplementary material available at 10.1007/s00406-021-01340-1.

George A. Mashour, MD, PhD, Recipient of the 2021 Excellence in Research Award

Article| October 2021 George A. Mashour, MD, PhD, Recipient of the 2021 Excellence in Research Award Kevin K. Tremper, PhD, MD Kevin K. Tremper, PhD, MD Search for other works by this author on: This Site PubMed Google Scholar ASA Monitor October 2021, Vol. 85, 22. https://doi.org/10.1097/01.ASM.0000795084.56772.28 Views Icon Views Article contents Figures & tables Video Audio Supplementary Data Peer Review Share Icon Share Twitter LinkedIn Cite Icon Cite Get Permissions Search Site Citation Kevin K. Tremper; George A. Mashour, MD, PhD, Recipient of the 2021 Excellence in Research Award. ASA Monitor 2021; 85:22 doi: https://doi.org/10.1097/01.ASM.0000795084.56772.28 Download citation file: Ris (Zotero) Reference Manager EasyBib Bookends Mendeley Papers EndNote RefWorks BibTex toolbar search Search Dropdown Menu nav search search input Search input auto suggest search filter All ContentAll PublicationsASA Monitor Search Advanced Search Topics: awards and prizes It is my great pleasure to congratulate Dr. George A. Mashour on receiving the 2021 Excellence in Research Award from the American Society of Anesthesiologists (Schaumburg, Illinois). Dr. Mashour is internationally renowned in the fields of anesthesiology, neuroscience, and translational science, and is highly deserving of this recognition. After his undergraduate study of philosophy and classics, Dr. Mashour received his MD and PhD in neuroscience from Georgetown University (Washington, D.C.) in 2001 and was a two-time Fulbright Scholar in the neurosciences (Max Delbruck Center for Molecular Medicine, Berlin, Germany, 1998-1999; University of Bonn, Bonn, Germany, 2002-2003). He completed his residency and chief residency in anesthesiology at the Massachusetts General Hospital/Harvard Medical School (Boston, Massachusetts) in 2006 and a fellowship in neurosurgical anesthesiology at the University of Michigan (Ann Arbor, Michigan) in 2007. He is currently the Robert B. Sweet Professor and Chair of the Department of Anesthesiology at the University... You do not currently have access to this content.

An Improved Optimization Algorithm for Epileptic Seizure Detection in EEG Signals Using Random Forest Classifier

Epilepsy is a psychiatric condition that has serious consequences for the human brain. The Electroencephalogram (EEG) may reveal a pattern that tells physicians whether an epileptic seizure is likely to occur again. EEG testing may also help the physician exclude other conditions that mimic epilepsy as a reason for the seizure. Now-a-days the researchers are showing much interest in these seizure detection because of its significance in epileptic detection. This paper is addressing an efficient soft computing framework for seizure detection from the EEG signal. The proposed pipeline of work is having the state-of-art as the possibility of achieving the maximum accuracy. The spectral features extracted from the Intrinsic mode functions (IMF) of EEG samples and it is directing the proposed flow towards the efficient detection of seizure and also the random forest algorithm based a convulsion classification is reliable for because of its learning behavior from the huge number of known dataset. The feature selection algorithm in this proposed work is stimulating the overall work towards the maximum true positive rate. This work is implemented on MATLAB platform and dataset were downloaded from the universal database such as Bonn university database. The results obtained from the proposed approach is showing the truthfulness of the approach introduced here.

Epileptic-seizure classification using phase-space representation of FBSE-EWT based EEG sub-band signals and ensemble learners

Electroencephalogram (EEG) signals are non-linear and non-stationary in nature. The phase-space representation (PSR) method is useful for analysing the non-linear characteristics of EEG signals. Hence non-linear features based on a phase-space representation of EEG signals are effective in epileptic-seizure classification. In the past, various machine learning methods are used for classifying seizure EEG signals. However, they might fail to classify accurately due to complex data, so with effective non-linear features, ensemble learning classifiers can be investigated to improve the accuracy of an automated epileptic-seizure detection system. In this paper, the EEG signals are first decomposed in sub-bands using empirical wavelet transform (EWT) based on the Fourier Bessel series expansion (FBSE) which is termed as FBSE-EWT. Then, these sub-bands are reconstructed into three-dimensional (3D) PSR. Next, entropy-based features like line-length (LL), log-energy-entropy (LEEnt), and norm-entropy (NEnt) are computed from Euclidean distances of 3D PSR of sub-band signals. The extracted features are ranked based on p-values obtained from the Kruskal-Wallis statistical test for reducing the feature space. Experiments have been conducted using obtained ranked features on different ensemble learning classifiers, and the five best-performing classifiers are reported here, which are random forest (RF), extra tree (ET), extremegradientboosting tree (xgBT), bagged-SVM (B-SVM), and bagged-k-nearest neighbours (B-k-NN), for classifying epileptic-seizure EEG signals. The performance of the proposed framework is evaluated using a publically accessible Bonn university EEG database for classifying epileptic-seizure EEG signals on well-known classification problems such as C1 (seizure, normal), C2, and C3 (seizure, normal, and seizure-free). This dataset consists recording of 100 single-EEG signals from five seizure, seizure-free, and normal (healthy) subjects each. Model is trained and tested using 10-fold cross-validation to stave off from overfitting. The performance is also compared with other state-of-art methods. Obtained results confirm the superior performance of the proposed framework by achieving maximum classification accuracy of 100 %, 98.3 %, 97.8 % with ET, B-SVM, and ET classifiers from each studied classification problem, respectively. Hence, the proposed framework can assist medical professionals in analysing epileptic-seizure EEG signals more accurately.