Segmentation of EEG during sleep using time-varying autoregressive modeling

Time-varying AR modeling is applied to sleep EEG signal, in order to perform parameter estimation and detect changes in the signal characteristics (segmentation). Several types of basis functions have been analyzed to determine how closely they can approximate parameter changes characteristics of the EEG signal. The TV-AR model was applied to a large number of simulated signal segments, in order to examine the behaviour of the estimation under various conditions such as variations in the EEG parameters and in the location of segment boundaries, and different orders of the basis functions. The set of functions that is the basis for the Discrete Cosine Transform (DCT), and the Walsh functions were found to be the most efficient in the estimation of the model parameters. A segmentation algorithm based on an “Identification function” calculated from the estimated model parameters is suggested.     

Robust algorithm for arrhythmia classification in ECG using extreme learning machine

Background  

Recently, extensive studies have been carried out on arrhythmia classification algorithms using artificial intelligence pattern recognition methods such as neural network. To improve practicality, many studies have focused on learning speed and the accuracy of neural networks. However, algorithms based on neural networks still have some problems concerning practical application, such as slow learning speeds and unstable performance caused by local minima.     

Cardiac disease modeling using induced pluripotent stem cell-derived human cardiomyocytes

Causative mutations and variants associated with cardiac diseases have been found in genes encoding cardiac ion channels, accessory proteins, cytoskeletal components, junctional proteins, and signaling molecules. In most cases the functional evaluation of the genetic alterationhas been carried out by expressing the mutated proteins in in-vitro heterologous systems. While these studies have provided a wealth of functional details that have greatly enhanced the understanding of the pathological mechanisms, it has always been clear that heterologous expression of the mutant protein bears the intrinsic limitation of the lack of a proper intracellular environment and the lack of pathological remodeling. The results obtained from the application of the next generation sequencing technique to patients suffering from cardiac diseases have identified several loci, mostly in non-coding DNA regions, which still await functional analysis. The isolation and culture of human embryonic stem cells has initially provided a constant source of cells from which cardiomyocytes(CMs) can be obtained by differentiation. Furthermore, the possibility to reprogram cellular fate to a pluripotent state, has opened this process to the study of genetic diseases. Thus induced pluripotent stem cells(i PSCs) represent a completely new cellular model that overcomes the limitations of heterologous studies. Importantly, due to the possibility to keep spontaneously beating CMs in culture for several months, during which they show a certain degree of maturation/aging, this approach will also provide a system in which to address the effect of long-term expression of the mutated proteins or any other DNA mutation, in terms of electrophysiological remodeling. Moreover, since i PSC preserve the entire patients’ genetic context, the system will help the physicians in identifying the most appropriate pharmacological intervention to correct the functional alteration. This article summarizes the current knowledge of cardiac genetic diseases modelled with i PSC.     

Local fractal dimension based ECG arrhythmia classification

We propose a local fractal dimension based nearest neighbor classifier for ECG based classification of arrhythmia. Local fractal dimension (LFD) at each sample point of the ECG waveform is taken as the feature. A nearest neighbor algorithm in the feature space is used to find the class of the test ECG beat. The nearest neighbor is found based on the RR-interval-information-biased Euclidean distance, proposed in the current work. Based on the two algorithms used for estimating the LFD, two classification algorithms are validated in the current work, viz. variance based fractal dimension estimation based nearest neighbor classifier and power spectral density based fractal dimension estimation based nearest neighbor classifier. Their performances are evaluated based on various figures of merit. MIT-BIH (Massachusetts Institute of Technology - Boston’s Beth Israel Hospital) Arrhythmia dataset has been used to validate the algorithms. Along with showing good performance against all the figures of merit, the proposed algorithms also proved to be patient independent in the sense that the performance is good even when the test ECG signal is from a patient whose ECG is not present in the training ECG dataset.     

Application of higher order statistics for atrial arrhythmia classification

Atrial fibrillation (AF) and atrial flutter (AFL) are the two common atrial arrhythmia encountered in the clinical practice. In order to diagnose these abnormalities the electrocardiogram (ECG) is widely used. The conventional linear time and frequency domain methods cannot decipher the hidden complexity present in these signals. The ECG is inherently a non-linear, non-stationary and non-Gaussian signal. The non-linear models can provide improved results and capture minute variations present in the time series. Higher order spectra (HOS) is a non-linear dynamical method which is highly rugged to noise. In the present study, the performances of two methods are compared: (i) 3rd order HOS cumulants and (ii) HOS bispectrum. The 3rd order cumulant and bispectrum coefficients are subjected to dimensionality reduction using independent component analysis (ICA) and classified using classification and regression tree (CART), random forest (RF), artificial neural network (ANN) and k-nearest neighbor (KNN) classifiers to select the best classifier. The ICA components of cumulant coefficients have provided the average accuracy, sensitivity, specificity and positive predictive value of 99.50%, 100%, 99.22% and 99.72% respectively using KNN classifier. Similarly, the ICA components of HOS bispectrum coefficients have yielded the average accuracy, sensitivity, specificity and PPV of 97.65%, 98.16%, 98.75% and 99.53% respectively using KNN. So, the ICA performed on the 3rd order HOS cumulants coupled with KNN classifier performed better than the HOS bispectrum method. The proposed methodology is robust and can be used in mass screening of cardiac patients.     

Cardiac arrhythmia and thyroid dysfunction: A novel genetic link

Inherited Long QT Syndrome (LQTS), a cardiac arrhythmia that predisposes to the often lethal ventricular fibrillation, is commonly linked to mutations in KCNQ1. The KCNQ1 voltage-gated K⁺ channel α subunit passes ventricular myocyte K⁺ current that helps bring a timely end to each heart-beat. KCNQ1, like many K⁺ channel α subunits, is regulated by KCNE β subunits, inherited mutations in which also associate with LQTS. KCNQ1 and KCNE mutations are also associated with atrial fibrillation. It has long been known that thyroid status strongly influences cardiac function, and that thyroid dysfunction causes abnormal cardiac structure and rhythm. We recently discovered that KCNQ1 and KCNE2 form a thyroid-stimulating hormone-stimulated K⁺ channel in the thyroid that is required for normal thyroid hormone biosynthesis. Here, we review this novel genetic link between cardiac and thyroid physiology and pathology, and its potential influence upon future therapeutic strategies in cardiac and thyroid disease.     

Cardiac arrhythmia mechanisms in rats with heart failure induced by pulmonary hypertension

Pulmonary hypertension provokes right heart failure and arrhythmias. Better understanding of the mechanisms underlying these arrhythmias is needed to facilitate new therapeutic approaches for the hypertensive, failing right ventricle (RV). The aim of our study was to identify the mechanisms generating arrhythmias in a model of RV failure induced by pulmonary hypertension. Rats were injected with monocrotaline to induce either RV hypertrophy or failure or with saline (control). ECGs were measured in conscious, unrestrained animals by telemetry. In isolated hearts, electrical activity was measured by optical mapping and myofiber orientation by diffusion tensor-MRI. Sarcoplasmic reticular Ca(2+) handling was studied in single myocytes. Compared with control animals, the T-wave of the ECG was prolonged and in three of seven heart failure animals, prominent T-wave alternans occurred. Discordant action potential (AP) alternans occurred in isolated failing hearts and Ca(2+) transient alternans in failing myocytes. In failing hearts, AP duration and dispersion were increased; conduction velocity and AP restitution were steeper. The latter was intrinsic to failing single myocytes. Failing hearts had greater fiber angle disarray; this correlated with AP duration. Failing myocytes had reduced sarco(endo)plasmic reticular Ca(2+)-ATPase activity, increased sarcoplasmic reticular Ca(2+)-release fraction, and increased Ca(2+) spark leak. In hypertrophied hearts and myocytes, dysfunctional adaptation had begun, but alternans did not develop. We conclude that increased electrical and structural heterogeneity and dysfunctional sarcoplasmic reticular Ca(2+) handling increased the probability of alternans, a proarrhythmic predictor of sudden cardiac death. These mechanisms are potential therapeutic targets for the correction of arrhythmias in hypertensive, failing RVs.     

Study of cardiac arrhythmia using zero-crossing analysis

It has been known for some time that the variability of the R-R intervals in the electrocardiogram yields valuable information concerning the types of arrhythmia which might be present. In this paper, an investigation is made into the application of zero-crossing analysis to the study of such variability. The number of times that the R-R interval crosses its mean value over a specified interval of time is counted, and may be associated with a particular characteristic frequency, related to the dominant frequency components of the power spectrum of R-R intervals. Higher order crossing counts may be computed by taking combinations of sum and difference operations on the original time series. The advantage of using zero-crossing analysis over spectral analysis is the computational simplicity of the former. It is demonstrated, by analysing data taken from the MIT-BIH Arrhythmia database, that zero crossing analysis can sometimes be used to distinguish between different arrhythmias, but forethought concerning the number of sum and difference operations to be taken on the original data set is required when computing the higher order crossing counts.     

Reliability analysis of microarray data using fuzzy c-means and normal mixture modeling based classification methods

MOTIVATION: A serious limitation in microarray analysis is the unreliability of the data generated from low signal intensities. Such data may produce erroneous gene expression ratios and cause unnecessary validation or post-analysis follow-up tasks. Therefore, the elimination of unreliable signal intensities will enhance reproducibility and reliability of gene expression ratios produced from microarray data. In this study, we applied fuzzy c-means (FCM) and normal mixture modeling (NMM) based classification methods to separate microarray data into reliable and unreliable signal intensity populations. RESULTS: We compared the results of FCM classification with those of classification based on NMM. Both approaches were validated against reference sets of biological data consisting of only true positives and true negatives. We observed that both methods performed equally well in terms of sensitivity and specificity. Although a comparison of the computation times indicated that the fuzzy approach is computationally more efficient, other considerations support the use of NMM for the reliability analysis of microarray data. AVAILABILITY: The classification approaches described in this paper and sample microarray data are available as Matlab( TM ) (The MathWorks Inc., Natick, MA) programs (mfiles) and text files, respectively, at http://rc.kfshrc.edu.sa/bssc/staff/MusaAsyali/Downloads.asp. The programs can be run/tested on many different computer platforms where Matlab is available. CONTACT: asyali@kfshrc.edu.sa.     

Preprocessing, classification modeling and feature selection using flow injection electrospray mass spectrometry metabolite fingerprint data

Metabolome analysis by flow injection electrospray mass spectrometry (FIE-MS) fingerprinting generates measurements relating to large numbers of m/z signals. Such data sets often exhibit high variance with a paucity of replicates, thus providing a challenge for data mining. We describe data preprocessing and modeling methods that have proved reliable in projects involving samples from a range of organisms. The protocols interact with software resources specifically for metabolomics provided in a Web-accessible data analysis package FIEmspro (http://users.aber.ac.uk/jhd) written in the R environment and requiring a moderate knowledge of R command-line usage. Specific emphasis is placed on describing the outcome of modeling experiments using FIE-MS data that require further preprocessing to improve quality. The salient features of both poor and robust (i.e., highly generalizable) multivariate models are outlined together with advice on validating classifiers and avoiding false discovery when seeking explanatory variables.     

Class discovery and classification of tumor samples using mixture modeling of gene expression data--a unified approach

MOTIVATION: The DNA microarray technology has been increasingly used in cancer research. In the literature, discovery of putative classes and classification to known classes based on gene expression data have been largely treated as separate problems. This paper offers a unified approach to class discovery and classification, which we believe is more appropriate, and has greater applicability, in practical situations. RESULTS: We model the gene expression profile of a tumor sample as from a finite mixture distribution, with each component characterizing the gene expression levels in a class. The proposed method was applied to a leukemia dataset, and good results are obtained. With appropriate choices of genes and preprocessing method, the number of leukemia types and subtypes is correctly inferred, and all the tumor samples are correctly classified into their respective type/subtype. Further evaluation of the method was carried out on other variants of the leukemia data and a colon dataset.     

Use of predictive modeling for Propionibacterium strain classification

Computed data analysis of biochemical or molecular profiles is currently used in studies of microbial taxonomy, epidemiology, and microbial diversity. We assessed the use of Partial Least Squares (PLS) regression for multivariate data analysis in bacteriology. We identified clear relationships between RAPD profiles of propionibacteria strains and their species classification, autolytic capacities, and their origins. The PLS regression also predicted species identity of some strains with RAPD profiles partially related to those of reference strains. The PLS analysis also allowed us to identify key characteristics to use to classify strains. PLS regression is particularly well adapted to i) describing a collection of bacterial isolates, ii) justifying bacterial groupings using several sets of data, and iii) predicting phenotypic characters of strains that have been classified by routine typing methods.