Nonlinear measures of respiration: Respiratory irregularity and increased chaos of respiration in patients with panic disorder

Background. Respiratory irregularity has been previously reported in patients with panic disorder using time domain measures. However, the respiratory signal is not entirely linear and a few previous studies used approximate entropy (APEN), a measure of regularity of time series. We have been studying APEN and other nonlinear measures including a measure of chaos, the largest Lyapunov exponent (LLE) of heart rate time series, in some detail. In this study, we used these measures of respiration to compare normal controls (n = 18) and patients with panic disorder (n = 22) in addition to the traditional time domain measures of respiratory rate and tidal volume. Methods: Respiratory signal was obtained by the Respitrace system using a thoracic and an abdominal belt, which was digitized at 500 Hz. Later, the time series were constructed at 4 Hz, as the highest frequency in this signal is limited to 0.5 Hz. We used 256 s of data (1,024 points) during supine and standing postures under normal breathing and controlled breathing at 12 breaths/min. Results: APEN was significantly higher in patients in standing posture during normal as well as controlled breathing (p = 0.002 and 0.02, respectively). LLE was also significantly higher in standing posture during normal breathing (p = 0.009). Similarly, the time domain measures of standard deviations and the coefficient of variation (COV) of tidal volume (TV) were significantly higher in the patient group (p = 0.02 and 0.004, respectively). The frequency of sighs was also higher in the patient group in standing posture (p = 0.02). In standing posture, LLE (p < 0.05) as well as APEN (p < 0.01) contributed significantly toward the separation of the two groups over and beyond the linear measure, i.e. the COV of TV. Conclusion: These findings support the previously described respiratory irregularity in patients with panic disorder and also illustrate the utility of nonlinear measures such as APEN and LLE as additional measures toward a better understanding of the abnormalities of respiratory physiology in similar patient populations as the correlation between LLE, APEN and some of the time domain measures only explained up to 50-60% of the variation. Copyright (C) 2002 S. Karger AG, Basel.

Effect of nortriptyline and paroxetine on measures of chaos of heart rate time series in patients with panic disorder

Tricyclic antidepressants have notable cardiac side effects, and this issue has become important due to the recent reports of increased cardiovascular mortality in patients with depression and anxiety. Several previous studies indicate that serotonin reuptake inhibitors (SRIs) do not appear to have such adverse effects. Apart from the effects of these drugs on routine 12-lead ECG, the effects on beat-to-beat heart rate (HR) and QT interval time series provide more information on the side effects related to cardiac autonomic function. In this study, we evaluated the effects of two antidepressants, nortriptyline (n = 13), a tricyclic, and paroxetine (n = 16), an SRI inhibitor, on HR variability in patients with panic disorder, using a measure of chaos, the largest Lyapunov exponent (LLE) using pre- and posttreatment HR time series. Our results show that nortriptyline is associated with a decrease in LLE of high frequency (HF: 0.15-0.5 Hz) filtered series, which is most likely due to its anticholinergic effect, while paroxetine had no such effect. Paroxetine significantly decreased sympathovagal ratios as measured by a decrease in LLE of LF/HF. These results suggest that paroxetine appears to be safer in regards to cardiovascular effects compared to nortriptyline in this group of patients. (C) 2003 Elsevier Inc. All rights reserved.

Nonlinear measures of QT interval series: novel indices of cardiac repolarization lability: MEDqthr and LLEqthr

In this study, we investigated nonlinear measures of chaos of QT interval time series in 28 normal control subjects, 36 patients with panic disorder and 18 patients with major depression in supine and standing postures. We obtained the minimum embedding dimension (MED) and the largest Lyapunov exponent (LLE) of instantaneous heart rate (HR) and QT interval series. MED quantifies the system's complexity and LLE predictability. There was a significantly lower MED and a significantly increased LLE of QT interval time series in patients. Most importantly, nonlinear indices of QT/HR time series, MEDqthr (MED of QT/HR) and LLEqthr (LLE of QT/HR), were highly significantly different between controls and both patient groups in either posture. Results remained the same even after adjusting for age. The increased LLE of QT interval time, series in patients with anxiety and depression is in line with our previous findings of higher QTvi (QT variability index, a log ratio of QT variability corrected for mean QT squared divided by heart rate variability corrected for mean heart rate squared) in these patients, using linear techniques. Increased LLEqthr (LLE of QT/HR) may be a more sensitive tool to study cardiac repolarization and a valuable addition to the time domain measures such as QTvi. This is especially important in light of the finding that LLEqthr correlated poorly and nonsignificantly with QTvi. These findings suggest an increase in relative cardiac sympathetic activity and a decrease in certain aspects of cardiac vagal function in patients with anxiety as well as depression. The lack of correlation between QTvi and LLEqthr suggests that this nonlinear index is a valuable addition to the linear measures. These findings may also help to explain the higher incidence of cardiovascular mortality in patients with anxiety and depressive disorders. (C) 2002 Elsevier Science Ireland Ltd. All rights reserved.

LZW Based Distance Measures for Spoken Language Identification

We present a new approach to spoken language modeling for language identification (LID) using the Lempel-Ziv-Welch (LZW) algorithm. The LZW technique is applicable to any kind of tokenization of the speech signal. Because of the efficiency of LZW algorithm to obtain variable length symbol strings in the training data, the LZW codebook captures the essentials of a language effectively. We develop two new deterministic measures for LID based on the LZW algorithm namely: (i) Compression ratio score (LZW-CR) and (ii) weighted discriminant score (LZW-WDS). To assess these measures, we consider error-free tokenization of speech as well as artificially induced noise in the tokenization. It is shown that for a 6 language LID task of OGI-TS database with clean tokenization, the new model (LZW-WDS) performs slightly better than the conventional bigram model. For noisy tokenization, which is the more realistic case, LZW-WDS significantly outperforms the bigram technique

Combination of similarity measures for time series classification using genetic algorithms

Time series classification deals with the problem of classification of data that is multivariate in nature. This means that one or more of the attributes is in the form of a sequence. The notion of similarity or distance, used in time series data, is significant and affects the accuracy, time, and space complexity of the classification algorithm. There exist numerous similarity measures for time series data, but each of them has its own disadvantages. Instead of relying upon a single similarity measure, our aim is to find the near optimal solution to the classification problem by combining different similarity measures. In this work, we use genetic algorithms to combine the similarity measures so as to get the best performance. The weightage given to different similarity measures evolves over a number of generations so as to get the best combination. We test our approach on a number of benchmark time series datasets and present promising results.

Estimation of voice-onset time in continuous speech using temporal measures

This paper proposes an automatic acoustic-phonetic method for estimating voice-onset time of stops. This method requires neither transcription of the utterance nor training of a classifier. It makes use of the plosion index for the automatic detection of burst onsets of stops. Having detected the burst onset, the onset of the voicing following the burst is detected using the epochal information and a temporal measure named the maximum weighted inner product. For validation, several experiments are carried out on the entire TIMIT database and two of the CMU Arctic corpora. The performance of the proposed method compares well with three state-of-the-art techniques. (C) 2014 Acoustical Society of America

Holistic Measures for Evaluating Prediction Models in Smart Grids

The performance of prediction models is often based on ``abstract metrics'' that estimate the model's ability to limit residual errors between the observed and predicted values. However, meaningful evaluation and selection of prediction models for end-user domains requires holistic and application-sensitive performance measures. Inspired by energy consumption prediction models used in the emerging ``big data'' domain of Smart Power Grids, we propose a suite of performance measures to rationally compare models along the dimensions of scale independence, reliability, volatility and cost. We include both application independent and dependent measures, the latter parameterized to allow customization by domain experts to fit their scenario. While our measures are generalizable to other domains, we offer an empirical analysis using real energy use data for three Smart Grid applications: planning, customer education and demand response, which are relevant for energy sustainability. Our results underscore the value of the proposed measures to offer a deeper insight into models' behavior and their impact on real applications, which benefit both data mining researchers and practitioners.

Global response sensitivity analysis using probability distance measures and generalization of Sobol's analysis

The study introduces two new alternatives for global response sensitivity analysis based on the application of the L-2-norm and Hellinger's metric for measuring distance between two probabilistic models. Both the procedures are shown to be capable of treating dependent non-Gaussian random variable models for the input variables. The sensitivity indices obtained based on the L2-norm involve second order moments of the response, and, when applied for the case of independent and identically distributed sequence of input random variables, it is shown to be related to the classical Sobol's response sensitivity indices. The analysis based on Hellinger's metric addresses variability across entire range or segments of the response probability density function. The measure is shown to be conceptually a more satisfying alternative to the Kullback-Leibler divergence based analysis which has been reported in the existing literature. Other issues addressed in the study cover Monte Carlo simulation based methods for computing the sensitivity indices and sensitivity analysis with respect to grouped variables. Illustrative examples consist of studies on global sensitivity analysis of natural frequencies of a random multi-degree of freedom system, response of a nonlinear frame, and safety margin associated with a nonlinear performance function. (C) 2015 Elsevier Ltd. All rights reserved.