Search

Diabetes mellitus (DM) or diabetes is an incurable, chronic, and genetic link health problem that occurs due to the higher glucose level in the blood. Continuous monitoring and automatic screening of diabetic patients will significantly improve the quality of the medical management system. DM induces cardiovascular autonomic neuropathy, which alters the morphology of electrocardiogram (ECG) signals. Hence, in this experiment, a new single-lead ECG signal database of 86 subjects (35 diabetic and 51 normal) is recorded. For the automatic screening of DM, an intrinsic time-scale decomposition (ITD) and machine learning-based framework is developed. In the first stage, denoised recorded signals are segmented into fragments of 5-s and decomposed into rotational components using the ITD algorithm. In the second stage, four features, namely, Hjorth complexity, Shannon entropy, log energy entropy, and log energy, are extracted from the ITD components. In the third stage, the Kruskal–Wallis (K–W) test is applied to select the most distinguishable features and fed to a decision tree classifier (DTC) with three different kernel functions for the automatic detection of diabetic patients. The fine tree (FT) kernel function provides the highest classification accuracy (ACC) of 86.9%. The proposed framework is developed using a 10-fold validation strategy. The developed framework is patient-centered suitable for screening in resource-limited environments and ready to be tested with more databases. [ABSTRACT FROM AUTHOR]

Objective: Bundle Branch Block (BBB) is the most serious asymptomatic, cardiovascular complication caused due to obstruction or delay across the pathways of electrical signals. Vectorcardiography (VCG) is the most useful method for measuring heart activities, containing information about the different physiological states of the heart. It is a powerful tool for detecting blockage inside the lower chamber of the heart. Clinically, posterior sensing electrodes are used to record VCG signals. Manual inspection of BBB-induced VCG changes is a time-consuming, labor-intensive, and error-prone procedure. This is the first work that integrates a deep learning technique with S-T spectrograms of smoothed VCG signals for the accurate diagnosis of BBB. A novel filter-fusion technique has also been proposed, to remove the various artifacts from the signals, which considerably increased the BBB diagnosis performance. Methods: VCG signals are divided into fragments of 4- ${s}$. A newly introduced filter-fusion technique is utilized to remove baseline wander and smooth the signal. Stock-well transform (S-T) is applied to extract spectrograms from de-noised smoothed VCG fragments. Obtained S-T spectrograms are fed to pre-trained Alex-Net, Squeeze-Net, and a newly developed less complex deep learning model with data augmentation technique utilizing a 10-fold cross-validation scheme. Results: Proposed improved deep learning model (IDLM) yielded the highest detection accuracy of 98.80%, true positive rate of 98.74%, and true negative rate of 98.80%. Conclusion: The findings show that developed DLM is computationally fast and efficient than benchmark models. The proposed IDLM is available for testing with other datasets. [ABSTRACT FROM AUTHOR]

Parkinson’s disease (PD) is a neurodegenerative ailment which causes changes in the neuronal, behavioral, and physiological structures. During the early stages of PD, these changes are very subtle and hence accurate diagnosis is challenging. Pathological and neurological experts assess the PD patients by examining their drawing, writing, walking, tremor, facial expressions, and speech. The manual analysis performed by specialists is time-consuming and prone to errors. Electroencephalogram (EEG) signals represent changes in the brain activities, but it is difficult to manually analyze these signals due to their non-linear, non-stationary, and complex nature. Traditional machine learning methods require several manual steps, such as decomposition, extraction of features, and classification. To overcome these limitations, automated PD detection using smoothed pseudo-Wigner Ville distribution (SPWVD) coupled with convolutional neural networks (CNN) called Parkinson’s disease CNN (PDCNNet) is proposed. First the EEG signals are subjected to SPWVD to obtain time-frequency representation (TFR). Then these two-dimensional plots are fed to an CNN. The proposed model is developed using two public databases. We have obtained an accuracy of 100% and 99.97% for dataset 1 and 2, respectively in detecting PD automatically using our proposed PDCNNet model. Our developed prototype has outperformed all existing state-of-the-art techniques and ready to be validated with more diverse datasets. [ABSTRACT FROM AUTHOR]

Schizophrenia (SZ) is a psychiatric disorder characterized by cognitive dysfunctions, hallucinations, and delusions, which may lead to lifetime disability. Detection and diagnosis of SZ by visual inspection is subjective, difficult, and time-consuming. Electroencephalogram (EEG) signals are widely used to detect the SZ as they reflect the conditions of the brain. Conventional machine learning methods involve many lengthy manual steps, such as decomposition, feature extraction, feature selection, and classification. In this article, automated identification of SZ is proposed using a combination of time–frequency analysis and convolutional neural network (CNN) to overcome the limitations of feature extraction-based methods. Three press button tasks are analyzed to segregate normal subjects and SZ patients. The EEG signals are subjected to continuous wavelet transform, short-time Fourier transform, and smoothed pseudo-Wigner–Ville distribution (SPWVD) techniques to obtain scalogram, spectrogram, and SPWVD-based time–frequency representation (TFR) plots, respectively. These 2-D plots are fed to pretrained AlexNet, VGG16, ResNet50, and CNN. We have obtained an accuracy of 93.36% using the SPWVD-based TFR and CNN model. In comparison to the benchmark AlexNet, ResNet50, and VGG16 networks, the developed CNN model with four convolutional layers not only requires fewer learnable parameters but also is computationally efficient and fast. This clearly indicates that our proposed method combining the SPWVD-CNN model has performed better than the state-of-the-art transfer learning techniques. Our developed model is ready to be tested with more EEG data and can aid psychiatrists in their diagnosis. [ABSTRACT FROM AUTHOR]

Background: Drivers drowsiness is one of the prime reasons for road accidents. Electroencephalogram (EEG) signals provide crucial information regarding drowsy state due to neurological changes in the brain. But the complex nature of EEG signals makes it difficult to study these changes. A detailed analysis of the EEG signal can be done if it is decomposed into multi-modes. Method: In this paper, adaptive variational mode decomposition (AVMD) is used for accurate analysis and synthesis of EEG signals. The number of modes (${J}$) and quadratic penalty factor ($\alpha $) is selected adaptively to find out representative information from EEG signals. Selection of ${J}$ and $\alpha $ is done by minimizing the reconstruction error using the Jaya optimization algorithm. Features are extracted from the adaptively decomposed modes. Entropy-based features selected by statistical analysis are classified with different classification algorithms. Eight performance parameters are evaluated to test the system’s effectiveness. Results: The reconstruction error of $4.035\times 10^{-09}$ and $1.564\times 10^{-09}$ for the alert and drowsy state shows that the proposed method gives a better synthesis of signals. An accuracy, sensitivity, specificity, F-1 score, Kappa, false-positive rate, error, and precision of 97.19%, 97.01%, 97.46%, 0.976, 94.23%, 2.54%, 2.81%, and 98.18% shows that the proposed method provides representative modes for analysis. Conclusion: The comparison shows that AVMD is superior over conventional and existing methods by about 7% and 1%, respectively. The solution provided in this paper takes a step ahead for efficient synthesis and analysis of EEG signals to detect the drowsy state of drivers. [ABSTRACT FROM AUTHOR]

Emotion is a neuronic transient that drives a person to a certain action. Emotion recognition from electroencephalogram (EEG) signals plays a vital role in the development of a brain–computer interface (BCI). Extracting the important information from raw EEG signals is difficult due to its nonstationary nature. Fixing a factual predefined basis function for efficient decomposition using a tunable $Q$ wavelet transform is an arduous task. In this article, an adaptive tunable $Q$ wavelet transform is proposed for the automatic selection of tuning parameters. Optimum tuning parameters are obtained using gray wolf optimization (GWO). Tuning parameters obtained by GWO are used to decompose the EEG signals into subbands (SBs). The set of time-domain features elicited from the SBs are used as an input to multiclass least-squares support vector machine. Classification accuracy of four basic emotions, namely, happy, fear, sad, and relax, is tested and compared with existing methods. An accuracy of 95.70% is achieved with a radial basis function kernel that is about 5% more than the existing methods using the same data set. This article proposes the development of a nonparameterized decomposition method for efficient decomposition of EEG signals. This method can be used with machine learning algorithms to take a step forward in the development of BCI systems. [ABSTRACT FROM AUTHOR]

Cognitive workload assessment plays an important role to examine the mental status of brain for emergency situations faced by air traffic controllers, military personals and many more. It is also significant for brain disorder diagnosis and psychological health monitoring. Mental workload induced by various factors needs to be critically examined for better decisions. This paper presents an automated model for cognitive workload assessment to provide accurate categorization on intensity of workload induced by multitasking situations. Statistical wavelets corresponding to brain frequencies act as features for categorization task. The brain analysis needs technical expertise to identify the relevant information but, due to lack of sufficient number of experts, feature selection model helps to identify the relevant features responsible to categorize cognitive workload. Teaching Learning based Optimization (TLBO) method has been modified using ‘Logical Operators’ for binary feature selection problem in the current study and evaluated based on different parameters. Results on Logical TLBO reached upto 100% for binary and multiclass classification of cognitive workload. The results also show the proposed method is converging in minimum iteration for multiclass cognitive workload. PROMETHEE based multi-criteria decision-making approach has adopted for ranking of algorithm to select the best algorithm when multiple alternatives are available. Based on study of current literature, this approach appears to be first time implemented to provide ranking of optimization algorithms. The PROMETHEE method also suggests that the Logical TLBO is best among all the other compared approaches. [ABSTRACT FROM AUTHOR]

Diagnosis of schizophrenia (SZ) is traditionally performed through patient’s interviews by a skilled psychiatrist. This process is time-consuming, burdensome, subject to error and bias. Hence the aim of this study is to develop an automatic SZ identification scheme using electroencephalogram (EEG) signals that can eradicate the aforementioned problems and support clinicians and researchers. This study introduces a methodology design involving empirical mode decomposition (EMD) technique for diagnosis of SZ from EEG signals to perfectly handle the behavior of non-stationary and nonlinear EEG signals. In this study, each EEG signal is decomposed into intrinsic mode functions (IMFs) by the EMD algorithm and then twenty-two statistical characteristics/features are calculated from these IMFs. Among them, five features are selected as significant feature applying Kruskal Wallis test. The performance of the obtained feature set is tested through several renowned classifierson a SZ EEG dataset. Among the considered classifiers, theensemble bagged tree performed as the best classifier producing 93.21% correct classification rate for SZ, with an overall accuracy of 89.59% for IMF 2. These results indicate that EEG signals discriminate SZ patients from healthy control (HC) subjects efficiently and have the potential to become a tool for the psychiatrist to support the positive diagnosis of SZ. [ABSTRACT FROM AUTHOR]

Hyperbox classifier is an efficient tool for modern pattern classification problems due to its transparency and rigorous use of Euclidian geometry. Fuzzy min–max (FMM) network efficiently implements the hyperbox classifier, and has been modified several times to yield better classification accuracy. However, the obtained accuracy is not up to the mark. Therefore, in this paper, a new improved FMM (IFMM) network is proposed to increase the accuracy rate. In the proposed IFMM network, a modified constraint is employed to check the expandability of a hyperbox. It also uses semiperimeter of the hyperbox along with k-nearest mechanism to select the expandable hyperbox. In the proposed IFMM, the contraction rules of conventional FMM and enhanced FMM (EFMM) are also modified using semiperimeter of a hyperbox in order to balance the size of both overlapped hyperboxes. Experimental results show that the proposed IFMM network outperforms the FMM, k-nearest FMM, and EFMM by yielding more accuracy rate with less number of hyperboxes. The proposed methods are also applied to histopathological images to know the best magnification factor for classification. [ABSTRACT FROM AUTHOR]

Sleep apnea is a sleeping disorder, which adversely affects the health of humans. The diagnosis of sleep apnea is possible by the detection of apnea events using electroencephalogram (EEG) recordings. This paper introduces an adaptive decomposition for the detection of apnea events using EEG signals. In introduced decomposition, the evolutionary techniques (ETs) optimized Hermite functions (HFs) represent the EEG signals. In tested ETs, the artificial bee colony (ABC) algorithm provides the least reconstruction error for the representation of EEG signals. The ABC is considered for Hermite coefficients-based feature extraction. From the extracted features, a highly discriminative feature set is obtained using the Fisher-score ranking test. The apnea detection performance of ranking-based selected features is evaluated using the extreme learning machine and least-squares support vector machine classifiers. The proposed method obtained performance measures, sensitivity 99.47%, specificity 99.58%, and accuracy 99.53%, are better as compared to the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Early detection of driver drowsiness and the development of a functioning driver alertness system may support the prevention of numerous vehicular accidents worldwide. Wearable sensors and camera-based systems are generally employed in the driver drowsiness detection. Electroencephalogram (or EEG) is considered another effective option for the driver drowsiness detection. Various EEG-based drowsiness detection systems have been proposed to date. In this paper, EEG signals are also used for the detection of drowsiness, with the proposed method being composed of three main building blocks. Both raw EEG signals and their corresponding spectrograms are used in the proposed building blocks. In the first building block, while energy distribution and zero-crossing distribution features are calculated from the raw EEG signals, spectral entropy and instantaneous frequency features are extracted from the EEG spectrogram images. In the second building block, deep feature extraction is employed directly on the EEG spectrogram images using pre-trained AlexNet and VGGNet. In the third building block, the tunable Q-factor wavelet transform (TQWT) is used to decompose the EEG signals into related sub-bands. The spectrogram images of the obtained sub-bands and statistical features, such as mean and standard deviation of the sub-bands’ instantaneous frequencies, are then calculated. Each feature group from each building block is fed to a long-short term memory (LSTM) network for the purposes of classification. The obtained results from the LSTM networks are then fused with a majority voting layer. The MIT-BIH Polysomnographic database was used in the experimental works. The evaluation of the proposed method was carried out with ten-fold cross validation test and the average accuracy represented accordingly. The obtained average accuracy score was 94.31%. The obtained result was also compared with other results to be found in the literature. The comparison shows that the proposed method’s achievement was found to be better than the compared results. [ABSTRACT FROM AUTHOR]

This paper introduces a methodology based on deep convolutional neural networks (DCNN) for motor imagery (MI) tasks recognition in the brain-computer interface (BCI) system. More specifically, the DCNN is used for classification of the right hand and right foot MI-tasks based electroencephalogram (EEG) signals. The proposed method first transforms the input EEG signals into images by applying the time-frequency (T-F) approaches. The used T-F approaches are short-time-Fourier-transform (STFT) and continuous-wavelet-transform (CWT). After T-F transformation the images of MI-tasks EEG signals are applied to the DCNN stage. The pre-trained DCNN model, AlexNet is explored for classification. The efficiency of the proposed method is evaluated on IVa dataset of BCI competition-III. The evaluation metrics such as accuracy, sensitivity, specificity, F1-score, and kappa value are used for measuring the proposed method results quantitatively. The obtained results show that the CWT approach yields better results than the STFT approach. In addition, the proposed method obtained 99.35% accuracy score is the best one among the existing methods accuracy scores. [ABSTRACT FROM AUTHOR]

In this paper, a new design method for digital bandpass and bandstop infinite impulse response filters with nearly linear-phase response is proposed. In this method, the phase response of an all-pass filter is optimized in the frequency domain to yield less passband error ${\text{Er}}_{p}$ and stopband error ${\text{Er}}_{s}$ with optimal stopband attenuation $A_{s}$. To achieve high accuracy in passband and stopband regions, fractional derivative (FD) constraints are evaluated in the respective regions, and the filter coefficients are computed using the Lagrange multiplier method. The behavior of fidelity parameters measured in terms of ${\text{Er}}_{p}{,\text{Er}}_{s}$ , and phase error ${\text{Er}}_{{\text{ph}}}$ is multimodel w.r.t. FD values. Therefore, modern heuristic technique, known as cuckoo search optimization is used for determining the optimal value of FDs and reference frequency simultaneously to minimize the fitness function, which is constructed as a sum of the squared error in passband and stopband. The designed filter yields up to 60% reduction in ${\text{Er}}_{p}$ and ${\text{Er}}_{{\text{ph}}}$ in the case of bandpass filter. Meanwhile, the response of filter is not degraded due to the finite word length effect. [ABSTRACT FROM AUTHOR]

Automatic drowsiness detection system plays a vital role to prevent the road accidents caused by drowsiness. In this regard, the electroencephalogram (EEG) signal provides valuable information of brain physiology for detection of drowsiness. EEG signals exhibit non-stationary nature which is tough to explore by prior defined and fixed number of basis functions. In this paper, an adaptive Hermite decomposition is proposed for the detection of drowsiness EEG signals. In the proposed decomposition, Hermite functions are used as basis functions, which are adaptively selected for each EEG signal by evolutionary optimization algorithms (EOAs). The mean square error of decomposition is proposed as an objective function to EOAs. The minimum decomposition error provided artificial bee colony EOA is considered to Hermite coefficients-based feature extraction for EEG signals. The extracted features are tested with the extreme learning machine (ELM), decision tree, $k$ -nearest neighbor, least squares support vector machine, artificial neural network, and naive Bayes for detection of alertness and drowsiness EEG signals. The proposed method with ELM classifier obtained 95.45% and 87.92% detection rates for alertness and drowsiness states, respectively, which are better as compared to other existing methods. [ABSTRACT FROM AUTHOR]

Electrocardiograph (ECG) denoising is the most important step in diagnosis of heart-related diseases, as the diagnosis gets influenced with noises. In this paper, a new method for ECG denoising is proposed, which incorporates empirical mode decomposition algorithm with Riegmann Liouvelle (RL) fractional integral filtering and Savitzky–Golay (SG) filtering. In the proposed method, noisy ECG signal is decomposed into its intrinsic mode functions (IMFs), from which noisy IMFs, corrupted with high-frequency (HF) and low-frequency (LF) noises, are identified by proposed noisy-IMFs identification methodologies. To denoise the signal, RL fractional integral filtering and SG filtering are applied on noisy IMFs corrupted with HF and LF noises, respectively; ECG signal is reconstructed with denoised IMFs and remaining signal dominant IMFs to obtain noise-free ECG signal. Proposed methodology is tested with MIT-BIH arrhythmia database. Its performance, in terms of signal-to-noise ratio and mean square error, is compared with other related ECG denoising methods based on fractional integral, empirical mode decomposition, and ensemble empirical mode decomposition. The obtained results by proposed method prove that the proposed method gives efficient noise removal performance. [ABSTRACT FROM AUTHOR]