Clustering Continuous Wavelet Transform Characteristics of Heart Rate Variability through Unsupervised Learning.

The analysis and interpretation of physiological signals acquired non-invasively are increasingly important in Smart Health, precision medicine, and medical research. However, this analysis is hampered due to the length, complexity, and inter-subject variation of these signals, and, consequently, dimensionality reduction and clustering offer substantial benefits. Machine learning, used widely in biomedicine, is increasingly being applied to physiological time series. Among the applications of unsupervised learning, clustering is one of the most important. In this paper, an unsupervised autoen-coder architecture, deep convolutional embedded clustering, is presented as a data-driven approach to study time-frequency characteristics of heart rate variability records. An autoen-coder network is trained on continuous wavelet transforms of heart rate variability signals calculated from publicly-available annotated ECG records with a wide variety of conditions. The latent variables learned by the clustering autoencoder are low-dimensional representations of wavelet transform characteristics that can be visualized and further analyzed. The results indicate that the learned clusters correspond to beat morphologies in the electrocardiogram in many cases, but also that the reduced dimensions of the time-frequency features can potentially provide additional insights into cardiac activity and the autonomic nervous system.

Quantitative feature analysis of continuous analytic wavelet transforms of electrocardiography and electromyography.

Theoretical and practical advances in time-frequency analysis, in general, and the continuous wavelet transform (CWT), in particular, have increased over the last two decades. Although the Morlet wavelet has been the default choice for wavelet analysis, a new family of analytic wavelets, known as generalized Morse wavelets, which subsume several other analytic wavelet families, have been increasingly employed due to their time and frequency localization benefits and their utility in isolating and extracting quantifiable features in the time-frequency domain. The current paper describes two practical applications of analysing the features obtained from the generalized Morse CWT: (i) electromyography, for isolating important features in muscle bursts during skating, and (ii) electrocardiography, for assessing heart rate variability, which is represented as the ridge of the main transform frequency band. These features are subsequently quantified to facilitate exploration of the underlying physiological processes from which the signals were generated.This article is part of the theme issue 'Redundancy rules: the continuous wavelet transform comes of age'.

Analysis of free moment and center of pressure frequency components during quiet standing using magnitude squared coherence.

To date, no postural studies have investigated the specific relationship between linear (anteroposterior (AP) and mediolateral (ML)) postural sway and the free moment (FM) over the range of biomechanically important frequencies. The goal of the current paper is to study the relationship between FM and the AP/ML movements during quiet standing with respect to individual frequencies. Mean squared coherence, which measures the degree of the relationship between two signals as a function of frequency, is employed to address this question. The results showed that, in two conditions (eyes opened and eyes closed), at very low frequencies (<0.5Hz), AP and FM were strongly correlated (>0.8) while there was a weak correlation between ML and FM (∼0.2). The situation reversed from (0.5 to 1.5Hz), with AP/FM correlation decreasing, and ML/FM correlation peaking slightly below 1.0Hz. Both conditions were only weakly correlated beyond 1.5Hz. It is suggested that these observations arise from differences in ankle activation between the left and right sides, whereas at higher frequencies, high coherence between ML and FM is a hip control strategy.

From Regional to National Clouds: TV Coverage in the Czech Republic.

Media, and particularly TV media, have a great impact on the general public. In recent years, spatial patterns of information and the relevance of intangible geographies have become increasingly important. Gatekeeping plays a critical role in the selection of information that is transformed into media. Therefore, gatekeeping, through national media, also co-forms the generation of mental maps. In this paper, correspondence analysis (a statistical method) combined with cloud lines (a new visual analytics technique) is used to analyze how individual major regional events in one of the post-communist countries, the Czech Republic, penetrate into the media on a national scale. Although national news should minimize distortions about regions, this assumption has not been verified by our research. Impressions presented by the media of selected regions that were markedly influenced by one or several events in those regions demonstrate that gatekeepers, especially news reporters, functioned as a filter by selecting only a few specific, and in many cases, unusual events for dissemination.

Assessing heart rate variability through wavelet-based statistical measures.

Because of its utility in the investigation and diagnosis of clinical abnormalities, heart rate variability (HRV) has been quantified with both time and frequency analysis tools. Recently, time-frequency methods, especially wavelet transforms, have been applied to HRV. In the current study, a complementary computational approach is proposed wherein continuous wavelet transforms are applied directly to ECG signals to quantify time-varying frequency changes in the lower bands. Such variations are compared for resting and lower body negative pressure (LBNP) conditions using statistical and information-theoretic measures, and compared with standard HRV metrics. The latter confirm the expected lower variability in the LBNP condition due to sympathetic nerve activity (e.g. RMSSD: p=0.023; SDSD: p=0.023; LF/HF: p=0.018). Conversely, using the standard Morlet wavelet and a new transform based on windowed complex sinusoids, wavelet analysis of the ECG within the observed range of heart rate (0.5-1.25Hz) exhibits significantly higher variability, as measured by frequency band roughness (Morlet CWT: p=0.041), entropy (Morlet CWT: p=0.001), and approximate entropy (Morlet CWT: p=0.004). Consequently, this paper proposes that, when used with well-established HRV approaches, time-frequency analysis of ECG can provide additional insights into the complex phenomenon of heart rate variability.

Evaluating the Relationship Between Muscle Activation and Spine Kinematics Through Wavelet Coherence.

Advances in time-frequency analysis can provide new insights into the important, yet complex relationship between muscle activation (ie, electromyography [EMG]) and motion during dynamic tasks. We use wavelet coherence to compare a fundamental cyclical movement (lumbar spine flexion and extension) to the surface EMG linear envelope of 2 trunk muscles (lumbar erector spinae and internal oblique). Both muscles cohere to the spine kinematics at the main cyclic frequency, but lumbar erector spinae exhibits significantly greater coherence than internal oblique to kinematics at 0.25, 0.5, and 1.0 Hz. Coherence phase plots of the 2 muscles exhibit different characteristics. The lumbar erector spinae precedes trunk extension at 0.25 Hz, whereas internal oblique is in phase with spine kinematics. These differences may be due to their proposed contrasting functions as a primary spine mover (lumbar erector spinae) versus a spine stabilizer (internal oblique). We believe that this method will be useful in evaluating how a variety of factors (eg, pain, dysfunction, pathology, fatigue) affect the relationship between muscles' motor inputs (ie, activation measured using EMG) and outputs (ie, the resulting joint motion patterns).

Why Do Some First Nations Communities Have Safe Water and Others Not? Socioeconomic Determinants of Drinking Water Risk.

Securing safe and adequate drinking water is an ongoing issue for many Canadian First Nations communities despite nearly 15 years of reports, studies, policy changes, financial commitments, and regulations. The federal drinking water evaluation scheme is narrowly scoped, ignoring community level social factors, which may play a role in access to safe water in First Nations. This research used the 2006 Aboriginal Affairs and Northern Development Canada First Nations Drinking Water System Risk Survey data and the Community Well-Being Index, including labour force, education, housing, and income, from the 2006 Census. Bivariate analysis was conducted using the Spearman's correlation, Kendall's tau correlation, and Pearson's correlation. Multivariable analysis was conducted using an ordinal (proportional or cumulative odds) regression model. Results showed that the regression model was significant. Community socioeconomic indicators had no relationship with drinking water risk characterization in both the bivariate and multivariable models, with the sole exception of labour force, which had a significantly positive effect on drinking water risk rankings. Socioeconomic factors were not important in explaining access to safe drinking water in First Nations communities. Improvements in the quality of safe water data as well as an examination of other community processes are required to address this pressing policy issue.

The Assessment of Muscular Effort, Fatigue, and Physiological Adaptation Using EMG and Wavelet Analysis.

Peroxisome proliferator-activated receptor gamma coactivator 1 alpha (PGC-1α) is a transcription factor co-activator that helps coordinate mitochondrial biogenesis within skeletal muscle following exercise. While evidence gleaned from submaximal exercise suggests that intracellular pathways associated with the activation of PGC-1α, as well as the expression of PGC-1α itself are activated to a greater extent following higher intensities of exercise, we have recently shown that this effect does not extend to supramaximal exercise, despite corresponding increases in muscle activation amplitude measured with electromyography (EMG). Spectral analyses of EMG data may provide a more in-depth assessment of changes in muscle electrophysiology occurring across different exercise intensities, and therefore the goal of the present study was to apply continuous wavelet transforms (CWTs) to our previous data to comprehensively evaluate: 1) differences in muscle electrophysiological properties at different exercise intensities (i.e. 73%, 100%, and 133% of peak aerobic power), and 2) muscular effort and fatigue across a single interval of exercise at each intensity, in an attempt to shed mechanistic insight into our previous observations that the increase in PGC-1α is dissociated from exercise intensity following supramaximal exercise. In general, the CWTs revealed that localized muscle fatigue was only greater than the 73% condition in the 133% exercise intensity condition, which directly matched the work rate results. Specifically, there were greater drop-offs in frequency, larger changes in burst power, as well as greater changes in burst area under this intensity, which were already observable during the first interval. As a whole, the results from the present study suggest that supramaximal exercise causes extreme localized muscular fatigue, and it is possible that the blunted PGC-1α effects observed in our previous study are the result of fatigue-associated increases in muscle acidosis. This should be explored in future research using further combinations of EMG and muscle biochemistry and histology.

Emotion-specific modulation of early visual perception.

Exposure to fearful facial expressions enhances vision at low spatial-frequencies and impairs vision at high spatial-frequencies. This perceptual trade-off is thought to be a consequence of a fear-related activation of the magnocellular visual pathway to the amygdala. In this study we examined the generality of the effect of emotion on low-level visual perception by assessing participants' orientation sensitivity to low and high spatial-frequency targets following exposure to disgust, fear, and neutral facial expressions. The results revealed that exposure to fear and disgust expressions have opposing effects on early vision: fearful expressions enhanced low spatial-frequency vision and impaired high spatial-frequency vision, while disgust expressions, like neutral expressions, impaired low spatial-frequency vision and enhanced high spatial-frequency vision. Thus we show the effect of exposure to fear on visual perception is not a general emotional effect, but rather one that may that depend on amygdala activation, or one that may be specific to fear.

Microarray image compression using a variation of singular value decomposition.

Microarray images are becoming increasingly important in bioinformatics, proteomics, and in the development of patient-specific therapies. The compression, processing, and analysis of these images are relatively new topics of research. In this paper, we focus on microarray image compression using singular value decomposition (SVD), a well known information compaction method. Although the SVD algorithm produces significant compression results, modifications may lead to further improvements. In an attempt to increase the compression ratio while maintaining a high peak signal-to-noise ratio, we adopt a subdivision scheme wherein the modified SVD is applied on each subimage. Experimental results indicate that SVD approaches are promising in compression, and may also lead to improved post-processing operations and analysis techniques.

High-performance medical image registration using new optimization techniques.

Optimization of a similarity metric is an essential component in intensity-based medical image registration. The increasing availability of parallel computers makes parallelizing some registration tasks an attractive option to increase speed. In this paper, two new deterministic, derivative-free, and intrinsically parallel optimization methods are adapted for image registration. DIviding RECTangles (DIRECT) is a global technique for linearly bounded problems, and multidirectional search (MDS) is a recent local method. The performance of DIRECT, MDS, and hybrid methods using a parallel implementation of Powell's method for local refinement, are compared. Experimental results demonstrate that DIRECT and MDS are robust, accurate, and substantially reduce computation time in parallel implementations.

A real time finite element based tissue simulation method incorporating nonlinear elastic behavior.

This paper presents a new finite element simulation approach for surgical simulators. Based on the solution of the algebraic equations derived from a nonlinear elastic model, we propose a real time simulation rule based on the implicit relation between the displacements of contacted and free nodes. This rule is an analytic expression in the linear case, and an approximation of the implicit relation in the non-linear case. We also remove some of the restrictions on flexibility exhibited by previous linear and nonlinear approaches. In the linear case, real time reconfiguration of the contacted nodes and the boundary constraints is realized using the simulation rule, while in the nonlinear case, a similar result is obtained by employing affine mapping. These methods allow nonlinear material properties to be applied to real time tissue simulation, with an efficiency comparable to that of the tensor matrix method for linear elastic models.

Registration of two-dimensional cardiac images to preprocedural three-dimensional images for interventional applications.

PURPOSE: To evaluate the accuracy and efficiency of rigid-body registration of two-dimensional fast cine and real-time cardiac images to high-resolution and SNR three-dimensional preprocedural reference volumes for application during MRI-guided interventional procedures. MATERIALS AND METHODS: Mutual information (MI) and correlation ratio (CR) similarity measures were evaluated. The dependence of registration accuracy and efficiency on different resolution and SNR parameters, and also on cardiac-phase differences was evaluated in a porcine model. Two-dimensional images were initially misoriented at distances (d) of 2-10 mm, and rotations of +/-5 degrees about all axes. Registration error and computation time were evaluated, and performance was also assessed visually. RESULTS: The maximum registration error using MI (<2.7 mm and <3.6 degrees ) occurred for d = 10 mm, misrotation of +/-5 degrees , and relative SNR = 1. The computation time was 15 seconds for MI and 10 seconds for CR. CONCLUSION: Registration accuracy was not highly dependent on the relative timing, within the cycle, between the two-dimensional and three-dimensional images. Registration using CR was faster than that using MI, although accuracy was marginally higher with MI. J.

Enhanced pre-computed finite element models for surgical simulation.

Soft tissue modeling is an important component in effective surgical simulation systems. A pre-computed finite element method based on elastic models is well suited to modeling soft tissue deformation. This paper addresses two principal issues: the flexibility of the pre-computed FE method and the approximation approach to non-linear elastic models. We describe a dynamic mechanism of the reconfiguration of the contacted nodes and the fixed boundary, without re-computing the inverse of the global stiffness matrix. The flexibility of the pre-computed models is described for both linear and non-linear elastic models.

Estimation of K distribution parameters using neural networks.

The K distribution is an accurate model for ultrasonic backscatter. A neural approach is developed to estimate K distribution parameters. Accuracy and consistency of the estimates from simulated K and envelope data compare favorably with other techniques. Neural networks can potentially be used as a complementary technique for tissue characterization.