Differential diagnosis between dilated cardiomyopathy and ischemic cardiomyopathy based on variational mode decomposition and high order spectra analysis

The clinical manifestations of ischemic cardiomyopathy (ICM) bear resemblance to dilated cardiomyopathy (DCM). The definitive diagnosis of DCM necessitates the identification of invasive, costly, and contraindicated coronary angiography. Many diagnostic studies of cardiovascular disease have tried modal decomposition based on electrocardiogram (ECG) signals. However, these studies ignored the connection between modes and other fields, thus limiting the interpretability of modes to ECG signals and the classification performance of models. This study proposes a classification algorithm based on variational mode decomposition (VMD) and high order spectra, which decomposes the preprocessed ECG signal and extracts its first five modes obtained through VMD. After that, these modes are estimated for their corresponding bispectrums, and the feature vector is composed of fifteen features including bispectral, frequency, and nonlinear features based on this. Finally, a dataset containing 75 subjects (38 DCM, 37 ICM) is classified and compared using random forest (RF), decision tree, support vector machine, and K-nearest neighbor. The results show that, in comparison to previous approaches, the technique proposed provides a better categorization for DCM and ICM of ECG signals, which delivers 98.21% classification accuracy, 98.22% sensitivity, and 98.19% specificity. And mode 3 always has the best performance among single mode. The proposed computerized framework significantly improves automatic diagnostic performance, which can help relieve the working pressure on doctors, possible economic burden and health threaten.

Transformations between rotational and translational invariants formulated in reciprocal spaces

Correlation functions play an important role in the theoretical underpinnings of many disparate areas of the physical sciences: in particular, scattering theory. More recently, they have become useful in the classification of objects in areas such as computer vision and our area of cryoEM. Our primary classification scheme in the cryoEM image processing system, EMAN2, is now based on third order invariants formulated in Fourier space. This allows a factor of 8 speed up in the two classification procedures inherent in our software pipeline, because it allows for classification without the need for computationally costly alignment procedures. In this work, we address several formal and practical aspects of such multispectral invariants. We show that we can formulate such invariants in the representation in which the original signal is most compact. We explicitly construct transformations between invariants in different orientations for arbitrary order of correlation functions and dimension. We demonstrate that third order invariants distinguish 2D mirrored patterns (unlike the radial power spectrum), which is a fundamental aspects of its classification efficacy. We show the limitations of 3rd order invariants also, by giving an example of a wide family of patterns with identical (vanishing) set of 3rd order invariants. For sufficiently rich patterns, the third order invariants should distinguish typical images, textures and patterns.

Six artificial intelligence paradigms for tissue characterisation and classification of non-COVID-19 pneumonia against COVID-19 pneumonia in computed tomography lungs

Background

COVID-19 pandemic has currently no vaccines. Thus, the only feasible solution for prevention relies on the detection of COVID-19-positive cases through quick and accurate testing. Since artificial intelligence (AI) offers the powerful mechanism to automatically extract the tissue features and characterise the disease, we therefore hypothesise that AI-based strategies can provide quick detection and classification, especially for radiological computed tomography (CT) lung scans.

Methodology

Six models, two traditional machine learning (ML)-based (k-NN and RF), two transfer learning (TL)-based (VGG19 and InceptionV3), and the last two were our custom-designed deep learning (DL) models (CNN and iCNN), were developed for classification between COVID pneumonia (CoP) and non-COVID pneumonia (NCoP). K10 cross-validation (90% training: 10% testing) protocol on an Italian cohort of 100 CoP and 30 NCoP patients was used for performance evaluation and bispectrum analysis for CT lung characterisation.

Results

Using K10 protocol, our results showed the accuracy in the order of DL > TL > ML, ranging the six accuracies for k-NN, RF, VGG19, IV3, CNN, iCNN as 74.58 ± 2.44%, 96.84 ± 2.6, 94.84 ± 2.85%, 99.53 ± 0.75%, 99.53 ± 1.05%, and 99.69 ± 0.66%, respectively. The corresponding AUCs were 0.74, 0.94, 0.96, 0.99, 0.99, and 0.99 (p-values < 0.0001), respectively. Our Bispectrum-based characterisation system suggested CoP can be separated against NCoP using AI models. COVID risk severity stratification also showed a high correlation of 0.7270 (p < 0.0001) with clinical scores such as ground-glass opacities (GGO), further validating our AI models.

Conclusions

We prove our hypothesis by demonstrating that all the six AI models successfully classified CoP against NCoP due to the strong presence of contrasting features such as ground-glass opacities (GGO), consolidations, and pleural effusion in CoP patients. Further, our online system takes < 2 s for inference.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11548-021-02317-0.

A Novel Block Imaging Technique Using Nine Artificial Intelligence Models for COVID-19 Disease Classification, Characterization and Severity Measurement in Lung Computed Tomography Scans on an Italian Cohort

Computer Tomography (CT) is currently being adapted for visualization of COVID-19 lung damage. Manual classification and characterization of COVID-19 may be biased depending on the expert's opinion. Artificial Intelligence has recently penetrated COVID-19, especially deep learning paradigms. There are nine kinds of classification systems in this study, namely one deep learning-based CNN, five kinds of transfer learning (TL) systems namely VGG16, DenseNet121, DenseNet169, DenseNet201 and MobileNet, three kinds of machine-learning (ML) systems, namely artificial neural network (ANN), decision tree (DT), and random forest (RF) that have been designed for classification of COVID-19 segmented CT lung against Controls. Three kinds of characterization systems were developed namely (a) Block imaging for COVID-19 severity index (CSI); (b) Bispectrum analysis; and (c) Block Entropy. A cohort of Italian patients with 30 controls (990 slices) and 30 COVID-19 patients (705 slices) was used to test the performance of three types of classifiers. Using K10 protocol (90% training and 10% testing), the best accuracy and AUC was for DCNN and RF pairs were 99.41 ± 5.12%, 0.991 (p < 0.0001), and 99.41 ± 0.62%, 0.988 (p < 0.0001), respectively, followed by other ML and TL classifiers. We show that diagnostics odds ratio (DOR) was higher for DL compared to ML, and both, Bispecturm and Block Entropy shows higher values for COVID-19 patients. CSI shows an association with Ground Glass Opacities (0.9146, p < 0.0001). Our hypothesis holds true that deep learning shows superior performance compared to machine learning models. Block imaging is a powerful novel approach for pinpointing COVID-19 severity and is clinically validated.

Bispectral analysis of overnight airflow to improve the pediatric sleep apnea diagnosis

Pediatric Obstructive Sleep Apnea (OSA) is a respiratory disease whose diagnosis is performed through overnight polysomnography (PSG). Since it is a complex, time-consuming, expensive, and labor-intensive test, simpler alternatives are being intensively sought. In this study, bispectral analysis of overnight airflow (AF) signal is proposed as a potential approach to replace PSG when indicated. Thus, our objective was to characterize AF through bispectrum, and assess its performance to diagnose pediatric OSA. This characterization was conducted using 13 bispectral features from 946 AF signals. The oxygen desaturation index ≥3% (ODI3), a common clinical measure of OSA severity, was also obtained to evaluate its complementarity to the AF bispectral analysis. The fast correlation-based filter (FCBF) and a multi-layer perceptron (MLP) were used for subsequent automatic feature selection and pattern recognition stages. FCBF selected 3 bispectral features and ODI3, which were used to train a MLP model with ability to estimate apnea-hypopnea index (AHI). The model reached 82.16%, 82.49%, and 90.15% accuracies for the common AHI cut-offs 1, 5, and 10 events/h, respectively. The different bispectral approaches used to characterize AF in children provided complementary information. Accordingly, bispectral analysis showed that the occurrence of apneic events decreases the non-gaussianity and non-linear interaction of the AF harmonic components, as well as the regularity of the respiratory patterns. Moreover, the bispectral information from AF also showed complementarity with ODI3. Our findings suggest that AF bispectral analysis may serve as a useful tool to simplify the diagnosis of pediatric OSA, particularly for children with moderate-to-severe OSA.

Pathological discrimination of the phonocardiogram signal using the bispectral technique

Phonocardiography is a dynamic non-invasive and relatively low-cost technique used to monitor the state of the mechanical activity of the heart. The recordings generated by such a technique is called phonocardiogram (PCG) signals. When shown visually, PCG signals can provide more insights of heart sounds for medical doctors. Thus, several approaches have been proposed to analyse these sounds through PCG recordings. However, due to the complexity and the high nonlinear nature of these recordings, a computer-assisted technique based on higher-order statistics HOS is shown to be, among these techniques, an important tool in PCG signal processing. The third-order spectra technique is one of these techniques; known as bispectrum, it can provide significant information to support physicians with an accurate and objective interpretation of heart condition. This technique is implemented and discussed in this paper. The implemented technique is used for the analysis of heart severity on nine different PCG recordings. These are normal, innocent murmur, coarctation of the aorta, ejection click, atrial gallop, opening snap, aortic stenosis, drum rumble, and aortic regurgitation. A unique bispectrum representation is generated for each type of heart sounds signal. Then, based on the bispectrum analysis, fifteen higher-order spectra HOS features such as the bispectral amplitude, the entropies, the moments, and the weighted center are extracted from each PCG record. The obtained HOS-features showed a well-correlated evolution with the increasing importance of heart severity leading therefore to a high potential in discriminating pathological PCG signals. One should know that, generally, classification of pathological PCG signals refers to the distinction between the presence of a pathology from its absence (binary response) while the discrimination considered in this paper provides an analogue response (value) which can vary from one pathology to another in an increasing or decreasing way.

EEG based dynamic RDS recognition with frequency domain selection and bispectrum feature optimization

BACKGROUND

Stereopsis plays a vital role in many aspects of human daily life. Random-dot stereogram (RDS) is often used to detect stereoacuity and perform research on visual cognition. Electroencephalogram (EEG) is one of the commonly adopted visual cognition techniques due to its noninvasive collection.

NEW METHOD

In this study, a methodology named WPT-BED based on wavelet packet transform (WPT) and bispectral eigenvalues of differential signals (BED) is proposed, which can classify the three-pattern EEG signals evoked by dynamic RDS (DRDS). Specifically, the signals are decomposed into different frequency bands by WPT. The appropriate sub-bands are selected for reconstruction. Finally, the optimized bispectrum features are extracted for classification to achieve higher accuracy.

RESULTS

The classification performance of the proposed method in different periods of signal processing are investigated. The method WPT-BED has the highest classification accuracy 84.38%, and the average classification accuracy is 73.98%. The active channels with higher accuracy are focused on the visual pathway in the human cerebral cortex.

COMPARISON WITH EXISTING METHODS

Comparison with other methods for EEG signals classification is performed to identify the effectiveness of the proposed methodology.

CONCLUSIONS

The proposed methodology can effectively distinguish the EEG signals evoked by DRDS. It demonstrates the feasibility of DRDS recognition based on EEG.

Utility of bispectrum in the screening of pediatric sleep apnea-hypopnea syndrome using oximetry recordings

BACKGROUND AND OBJECTIVE

The aim of this study was to assess the utility of bispectrum-based oximetry approaches as a complementary tool to traditional techniques in the screening of pediatric sleep apnea-hypopnea syndrome (SAHS).

METHODS

298 blood oxygen saturation (SpO₂) signals from children ranging 0-13 years of age were recorded during overnight polysomnography (PSG). These recordings were divided into three severity groups according to the PSG-derived apnea hypopnea index (AHI): AHI < 5 events per hour (e/h), 5 ≤ AHI < 10 e/h, AHI ≥ 10 e/h. For each pediatric subject, anthropometric variables, 3% oxygen desaturation index (ODI3) and spectral features from power spectral density (PSD) and bispectrum were obtained. Then, the fast correlation-based filter (FCBF) was applied to select a subset of relevant features that may be complementary, excluding those that are redundant. The selected features fed a multiclass multi-layer perceptron (MLP) neural network to build a model to estimate the SAHS severity degrees.

RESULTS

An optimum subset with features from all the proposed methodological approaches was obtained: variables from bispectrum, as well as PSD, ODI3, Age, and Sex. In the 3-class classification task, the MLP model trained with these features achieved an accuracy of 76.0% and a Cohen's kappa of 0.56 in an independent test set. Additionally, high accuracies were reached using the AHI cutoffs for diagnosis of moderate (AHI = 5 e/h) and severe (AHI = 10 e/h) SAHS: 81.3% and 85.3%, respectively. These results outperformed the diagnostic ability of a MLP model built without using bispectral features.

CONCLUSIONS

Our results suggest that bispectrum provides additional information to anthropometric variables, ODI3 and PSD regarding characterization of changes in the SpO₂ signal caused by respiratory events. Thus, oximetry bispectrum can be a useful tool to provide complementary information for screening of moderate-to-severe pediatric SAHS.

Paroxysmal atrial fibrillation prediction method with shorter HRV sequences

This paper proposes a method that predicts the onset of paroxysmal atrial fibrillation (PAF), using heart rate variability (HRV) segments that are shorter than those applied in existing methods, while maintaining good prediction accuracy. PAF is a common cardiac arrhythmia that increases the health risk of a patient, and the development of an accurate predictor of the onset of PAF is clinical important because it increases the possibility to stabilize (electrically) and prevent the onset of atrial arrhythmias with different pacing techniques. We investigate the effect of HRV features extracted from different lengths of HRV segments prior to PAF onset with the proposed PAF prediction method. The pre-processing stage of the predictor includes QRS detection, HRV quantification and ectopic beat correction. Time-domain, frequency-domain, non-linear and bispectrum features are then extracted from the quantified HRV. In the feature selection, the HRV feature set and classifier parameters are optimized simultaneously using an optimization procedure based on genetic algorithm (GA). Both full feature set and statistically significant feature subset are optimized by GA respectively. For the statistically significant feature subset, Mann-Whitney U test is used to filter non-statistical significance features that cannot pass the statistical test at 20% significant level. The final stage of our predictor is the classifier that is based on support vector machine (SVM). A 10-fold cross-validation is applied in performance evaluation, and the proposed method achieves 79.3% prediction accuracy using 15-minutes HRV segment. This accuracy is comparable to that achieved by existing methods that use 30-minutes HRV segments, most of which achieves accuracy of around 80%. More importantly, our method significantly outperforms those that applied segments shorter than 30 minutes.

An Automatic Prediction of Epileptic Seizures Using Cloud Computing and Wireless Sensor Networks

Epilepsy is one of the most common neurological disorders which is characterized by the spontaneous and unforeseeable occurrence of seizures. An automatic prediction of seizure can protect the patients from accidents and save their life. In this article, we proposed a mobile-based framework that automatically predict seizures using the information contained in electroencephalography (EEG) signals. The wireless sensor technology is used to capture the EEG signals of patients. The cloud-based services are used to collect and analyze the EEG data from the patient's mobile phone. The features from the EEG signal are extracted using the fast Walsh-Hadamard transform (FWHT). The Higher Order Spectral Analysis (HOSA) is applied to FWHT coefficients in order to select the features set relevant to normal, preictal and ictal states of seizure. We subsequently exploit the selected features as input to a k-means classifier to detect epileptic seizure states in a reasonable time. The performance of the proposed model is tested on Amazon EC2 cloud and compared in terms of execution time and accuracy. The findings show that with selected HOS based features, we were able to achieve a classification accuracy of 94.6 %.

Third order spectral analysis robust to mixing artifacts for mapping cross-frequency interactions in EEG/MEG

We present a novel approach to the third order spectral analysis, commonly called bispectral analysis, of electroencephalographic (EEG) and magnetoencephalographic (MEG) data for studying cross-frequency functional brain connectivity. The main obstacle in estimating functional connectivity from EEG and MEG measurements lies in the signals being a largely unknown mixture of the activities of the underlying brain sources. This often constitutes a severe confounder and heavily affects the detection of brain source interactions. To overcome this problem, we previously developed metrics based on the properties of the imaginary part of coherency. Here, we generalize these properties from the linear to the nonlinear case. Specifically, we propose a metric based on an antisymmetric combination of cross-bispectra, which we demonstrate to be robust to mixing artifacts. Moreover, our metric provides complex-valued quantities that give the opportunity to study phase relationships between brain sources. The effectiveness of the method is first demonstrated on simulated EEG data. The proposed approach shows a reduced sensitivity to mixing artifacts when compared with a traditional bispectral metric. It also exhibits a better performance in extracting phase relationships between sources than the imaginary part of the cross-spectrum for delayed interactions. The method is then applied to real EEG data recorded during resting state. A cross-frequency interaction is observed between brain sources at 10Hz and 20Hz, i.e., for alpha and beta rhythms. This interaction is then projected from signal to source level by using a fit-based procedure. This approach highlights a 10-20Hz dominant interaction localized in an occipito-parieto-central network.

Entropy analysis of muscular near-infrared spectroscopy (NIRS) signals during exercise programme of type 2 diabetic patients: quantitative assessment of muscle metabolic pattern

Diabetes mellitus (DM) is a metabolic disorder that is widely rampant throughout the world population these days. The uncontrolled DM may lead to complications of eye, heart, kidney and nerves. The most common type of diabetes is the type 2 diabetes or insulin-resistant DM. Near-infrared spectroscopy (NIRS) technology is widely used in non-invasive monitoring of physiological signals. Three types of NIRS signals are used in this work: (i) variation in the oxygenated haemoglobin (O2Hb) concentration, (ii) deoxygenated haemoglobin (HHb), and (iii) ratio of oxygenated over the sum of the oxygenated and deoxygenated haemoglobin which is defined as: tissue oxygenation index (TOI) to analyze the effect of exercise on diabetes subjects. The NIRS signal has the characteristics of non-linearity and non-stationarity. Hence, the very small changes in this time series can be efficiently extracted using higher order statistics (HOS) method. Hence, in this work, we have used sample and HOS entropies to analyze these NIRS signals. These computer aided techniques will assist the clinicians to diagnose and monitor the health accurately and easily without any inter or intra observer variability. Results showed that after a one-year of physical exercise programme, all diabetic subjects increased the sample entropy of the NIRS signals, thus revealing a better muscle performance and an improved recruitment by the central nervous system. Moreover, after one year of physical therapy, diabetic subjects showed a NIRS muscular metabolic pattern that was not distinguished from that of controls. We believe that sample and bispectral entropy analysis is need when the aim is to compare the inner structure of the NIRS signals during muscle contraction, particularly when dealing with neuromuscular impairments.