Sequential Probability Assignment for Outlier Detection in Heartbeat Timings.

Wearable physiological monitors have become increasingly popular, often worn during people’s daily life, collecting data 24 hours a day, 7 days a week. In the last decade, these devices have attracted the attention of the scientific community as they allow us to automatically extract information about user physiology (e.g., heart rate, sleep quality and physical activity) enabling inference on their health. However, the biggest issue about the data recorded by wearable devices is the missing values due to motion and mechanical artifacts induced by external stimuli during data acquisition. This missing data could negatively affect the assessment of heart rate (HR) response and estimation of heart rate variability (HRV), that could in turn provide misleading insights concerning the health status of the individual. In this study, we focus on healthy subjects with normal heart activity and investigate the effects of missing variation of the timing between beats (RR-intervals) caused by motion artifacts on HRV features estimation by randomly introducing missing values within a five min time windows of RR-intervals obtained from the nsr2db PhysioNet dataset by using Gilbert burst method. We then evaluate several strategies for estimating HRV in the presence of missing values by interpolating periods of missing values, covering the range of techniques often deployed in the literature, via linear, quadratic, cubic, and cubic spline functions. We thereby compare the HRV features obtained by handling missing data in RR-interval time series against HRV features obtained from the same data without missing values. Finally, we assess the difference between the use of interpolation methods on time (i.e., the timestamp when the heartbeats happen) and on duration (i.e., the duration of the heartbeats), in order to identify the best methodology to handle the missing RR-intervals. The main novel finding of this study is that the interpolation of missing data on time produces more reliable HRV estimations when compared to interpolation on duration. Hence, we can conclude that interpolation on duration modifies the power spectrum of the RR signal, negatively affecting the estimation of the HRV features as the amount of missing values increases. We can conclude that interpolation in time is the optimal method among those considered for handling data with large amounts of missing values, such as data from wearable sensors.

Heart rate is a vital sign, whereas heart rate variability is an important quantitative measure of cardiovascular regulation by the autonomic nervous system. Although the design of algorithms to compute heart rate and assess heart rate variability is an active area of research, none of the approaches considers the natural point-process structure of human heartbeats, and none gives instantaneous estimates of heart rate variability. We model the stochastic structure of heartbeat intervals as a history-dependent inverse Gaussian process and derive from it an explicit probability density that gives new definitions of heart rate and heart rate variability: instantaneous R-R interval and heart rate standard deviations. We estimate the time-varying parameters of the inverse Gaussian model by local maximum likelihood and assess model goodness-of-fit by Kolmogorov-Smirnov tests based on the time-rescaling theorem. We illustrate our new definitions in an analysis of human heartbeat intervals from 10 healthy subjects undergoing a tilt-table experiment. Although several studies have identified deterministic, nonlinear dynamical features in human heartbeat intervals, our analysis shows that a highly accurate description of these series at rest and in extreme physiological conditions may be given by an elementary, physiologically based, stochastic model.

Non-contact heart rate and heart rate variability measurements have applications in healthcare and affective computing. Recently, a system utilizing a five-band camera (RGBCO: red, green, blue, cyan, orange) was proposed, and shown to improve both remote measurement of heart rate and heart rate variability over an RGB camera. In this paper, we propose an improved method for video-based estimation of heart rate variability. We introduce three advancements over previous work utilizing five-band cameras: (1) an adaptive non-rectangular region of interest identified using automatically detected facial feature points, (2) improved peak detection within the blood volume pulse (BVP) signal, and (3) improved HRV calculation using the Welch periodogram. We apply our method to a test dataset of subjects at rest and under cognitive stress and show qualitative improvements in the stability of HRV spectrogram estimation. Although we evaluate our method using a five-band camera, the method could be applied to video recorded with any camera.

To evaluate the effects of record length selection on the accuracy of spectral estimates of heart rate variability (HRV), a simulation study was carried out using a set of 58 signals obtained by autoregressive (AR) fitting a representative sample of real HRV signals. Four record lengths of 180, 300, 420, and 540 s were considered. Spectral estimation was performed by both the Blackman-Tukey (B-T) and AR methods. Accuracy was assessed for: 1) point spectral estimates, by computing the normalized averaged bias (NAB) and variance (NAV); and 2) the most commonly used spectral parameters [total power (TP) and the powers in the bands: very low frequency (VLF) (0 divided by 0.04 Hz), low frequency (LF) (0.04 divided by 0.15 Hz), and high frequency (HF) (0.15 divided by 0.45 Hz)], by computing the normalized bias (NB) and variance (NV). The results are: whatever the record length considered, the 90th percentiles (90P) of the NAB were < 10%, whereas those of the NB were < 9% for TP, LF, and HF powers, and < 14% for the VLF power, in both methods. The NAV was proportional to the reciprocal of record length, showing high 90P values for the shortest record length (26.4% for B-T and 44.2% for AR). The NV showed the same trend but 90P values were much lower (< 8% for TP, LF, and HF powers and < 19% for VLF power, in both methods). In the final part of the paper a procedure for the computation of approximate upper bounds of the relative absolute error of spectral measures at each record length, based on the knowledge of the NB and NV, is presented.

ABSTRACTIt is argued that: if linear trends are present in heart rate (HR) data, due to sinus arrhythmic variation and the gradual nature of the HR response to an external treatment, the time‐series statistic successive difference mean square (SDMS) will be a logically more appropriate estimate of HR variability than is variance. A fifty‐subject study assessed both the empirical import of this argument as well as the question of the appropriate scale of measurement, i.e., either beats per minute (BPM) or interbeat intervals (IBI). Separate analyses to determine changes in HR variability from baseline to treatment conditions were carried out using each combination of statistic and scale to define HR variability. These analyses showed that HR variability either did not change or was significantly decreased by the treatment depending on the combination of statistic and scale used to define HR variability. This differential result, depending on the type of analysis, was due to the fact that nonrandom linear trends in the HR data spuriously inflated variance relative to SDMS and affected the BPM scale significantly more than the IBI scale. The major result from this study is that the combination of SDMS as the variability statistic, and IBI as the scale of measurement for HR, is both the best and most appropriate technique for assessing changes in HR variability. In addition, the IBI scale should be considered when analyzing mean HR scores if differential linear trends exist between conditions being compared, as in this case, since the IBI scale should increase statistical precision for detecting differences between mean HR scores during those conditions. Finally, when the more appropriate SDMS statistic and IBI scale define HR variability, this study supports the claims of previous investigators who have reported that increases in cognitive functioning led to decreases in HR variability.

Heart Rate Variability (HRV) evaluates the autonomic nervous system regulation and can be used as a monitoring tool in conditions such as cardiovascular diseases, neuropathies and sleep staging. It can be extracted from the electrocardiogram (ECG) and the photoplethysmogram (PPG) signals. Typically, the HRV is obtained from the ECG processing. Being the PPG sensor widely used in clinical setups for physiological parameters monitoring such as blood oxygenation and ventilatory rate, the question arises regarding the PPG adequacy for HRV extraction. There is not a consensus regarding the PPG being able to replace the ECG in the HRV estimation. This work aims to be a contribution to this research area by comparing the HRV estimation obtained from simultaneously acquired ECG and PPG signals from forty subjects. A peak detection method is herein introduced based on the Hilbert transform: Hilbert Double Envelope Method (HDEM). Two other peak detector methods were also evaluated: Pan-Tompkins and Wavelet-based. HRV parameters for time, frequency and the non-linear domain were calculated for each algorithm and the Pearson correlation, T-test and RMSE were evaluated. The HDEM algorithm showed the best overall results with a sensitivity of 99.07% and 99.45% for the ECG and the PPG signals, respectively. For this algorithm, a high correlation and no significant differences were found between HRV features and the gold standard, for the ECG and PPG signals. The results show that the PPG is a suitable alternative to the ECG for HRV feature extraction.

Patients with transthyretin amyloidosis (ATTR) polyneuropathy, a hereditary fatal disease, often report defects in both thermal perception and autonomic nervous system function as their first clinical symptoms. While elevated thermal perception thresholds (TPT) for cold and warmth only recently have been shown as an early marker of small nerve fiber dysfunction in these patients, heart rate variability (HRV) has frequently been used to quantify autonomic neuropathy. The main purpose with this report was to elucidate a possible relationship between estimates of HRV and TPT in a selected group of early and late-onset Swedish Val30Met ATTR patients. The results show significantly more pronounced elevation of TPT in early compared to late-onset patients. Significant correlations between HRV and TPT were found among late-onset cases, indicating a possible relationship between loss of thin nerve fibers in somatic and autonomic nerves, while generally no such relationships were found among early-onset cases. This observation emphasizes the importance of testing both HRV and TPT to ensure optimal early detection of neuropathic changes in an as wide as possible range of small nerve fibers in suspected ATTR patients. This is of particular importance as the phenotype of the ATTR disease varies between groups with different age of onset.

This paper describes a method for estimation of heart rate (HR) and heart rate variability (HRV) with accelerometers and gyroscopes. We denote this joint seismocardiography (SCG) and gyrocardiography (GCG) approach as SCG/GCG. In principle, SCG which is a well known method measures the linear mechanical movements of the heart and GCG is a new technique which measures angular motion due to the chest micro-vibrations caused by myocardial rotation. As electrocardiography (ECG), they can also be performed in non-invasive manner using a device in contact to subjects skin, for example. Our method extracts HRV parameters based on single-axis and multi-axes autocorrelation analysis (1-AC and 6-AC) of all simultaneously captured SCG/GCG axes. The results of each axes are combined to maintain reliable HR- and HRV. We validate our results with a comparison study between simultaneous ECG and SCG/GCG recordings using a study group of 29 healthy male volunteers. The study provides a promising approach for HRV estimation with modern wearable devices.

With the rapid development of science and technology, the living habits of people have also changed from those in the past; the diet, living environment, various life pressures, etc., all overwhelm the body and mind, meaning that, nowadays, more people are suffering from mental illness and cardiovascular disease year on year. Therefore, a non-contact measurement of heart rate and heart rate variability (HRV) is proposed to assist physicians in diagnosing symptoms related to mental illness and cardiovascular disease. In this paper, continuous images are obtained by general network cameras with non-contact, facial feature points and regions of interest (ROI) employed to track faces and regional images; HRV parameters were analyzed with the green wavelength of RGB color space. The artifact signal is eliminated by a hybrid algorithm of independent component analysis (ICA) and particle swarming optimization (PSO). Finally, the values of heart rate and HRV are obtained with signal processes of using band-pass filter, fast Fourier transform (FFT), and power spectrum analysis in the time and frequency domains, respectively. The non-contact measurement performance of the proposed method can effectively not only avoid infection doubts and obtain heart rate and HRV quickly, but also provide better physiological parameters, root mean square error (RMSE), and mean absolute percentage error (MAPE), than those of recent published papers.

We propose a method for septic shock prediction by real time monitoring of heart rate variability (HRV) for patients in intensive care unit (ICU). An objective criterion that is related to septic occurrence and can be measured automatically in real time is required to detect septic occurrence. Although HRV is a candidate of such criteria, an automatic estimation of HRV has been difficult because there are many arrhythmias in ICU patients. In this study, we employed a stochastic model on R-R interval series, namely heartbeat interval series, in order to detect abnormal heartbeats in normal R-R interval series automatically. Abnormal heartbeats and very low frequency components are eliminated from R-R series by using the stochastic model. Applying the elimination process before HRV estimation, we identified a distinctive V-shaped temporal pattern in HRV series before septic shock was occurred. The V-shaped pattern is applicable as a signal of septic occurrence. We finally propose a real-time monitoring for sepsis occurrence by using the V-shaped pattern in HRV.KeywordsHeart rate variabilityArrhythmiaTrendSeptic shockMonitoring

Independent of other established risk factors, depressed heart rate variability (HRV) has been shown to be a powerful predictor of cardiac events after MI. Unfortunately, the need of 24-hour ECG recording and subsequent laborious editing of Holter data limits the clinical use of long-term HRV. In order to perform post-MI risk stratification more efficiently, we evaluated the value of short-term HRV estimates for preselection of patients who might benefit from long-term HRV assessment. Two measures were assessed from 24-hour ambulatory ECGs recorded in 729 survivors of acute MI prior to hospital discharge. In addition to a complete 24-hour HRV index, a standard deviation of normal-to-normal RR intervals (SDNN) was obtained from the first stationary and ectopic free 5-minute segment of the Holter recording. Predictive power (relation between positive predictive accuracy and sensitivity) of a complete 24-hour HRV index in identifying patients who suffered from cardiac mortality or arrhythmic events during a 2-year follow-up was compared to the predictive power of assessing the 24-hour HRV index limited to 50%, 40%, or 20% of patients with the lowest values of 5-minute SDNN. The HRV index was significantly lower in patients who died (19 +/- 11 units) or had an arrhythmic event (AE) (18 +/- 11 units) compared to those who survived without an event (28 +/- 10 resp. 27 +/- 11 units; P < 0.001). Similarly, 5-minute SDNN was significantly lower in patients who died (25 +/- 12 ms) or suffered an AE (26 +/- 13 ms) compared to survivors (40 +/- 19 ms resp. 39 +/- 19 ms; P < 0.001). When limited to patients with depressed 5-minute SDNN, assessment of the HRV index performed better than 5-minute SDNN alone in positive prediction of cardiac events. Preselected assessment of the lowest HRV index in 50% to 20% of the total population yielded a 2-year cardiac event prediction rate as high as analysis of the HRV index in all patients. Long-term HRV assessment for risk stratification after MI in patients preselected by depressed short-term SDNN is safe and efficient, and allows a practical identification of patients with the highest likelihood of cardiac events during long-term follow-up.

Depressed heart rate variability (HRV) after a myocardial infarction is associated with increased mortality. This is thought to be due to reduced parasympathetic activity and heightened sympathetic activity. Aging is associated with depressed HRV, but little is known of the affect of aging on parasympathetic activity. This study examined 56 healthy subjects (age range 40-102 years; 39 women). None had a history of heart disease or were on medication that would affect cardiac function. All had normal resting ECGs, normal heart size on chest X ray, and normal electrolytes. In all subjects, 24-hour Holter recordings were performed and used to measure HRV. In particular, the study examined the affect of age on HRV triangular index, which gives an estimate of overall HRV, and on RMSSD (square root of the mean squared differences of successive normal-to-normal RR intervals), which gives an estimate of short-term components of HRV and is thought to reflect the overall extent of vagal modulations of heart rates. Both these parameters were compared in patients younger and older than 70 years. Each recording lasted at least 17 hours; the majority of recordings were longer than 20 hours. There was a significant decrease in HRV triangular index with age (r = -0.4, P < 0.05) and no significant change in RMSSD with age(r = -0.08, P = NS). There was a significant difference in HRV index in those > 70 years compared with those < 70 years (38.0 +/- 9.3 vs 31.0 +/- 11, respectively, P < 0.02). There was no significant difference in RMSSD between the two age groups (26.7 +/- 8.2 ms vs 28.4 +/- 11.3 ms, respectively, P = NS). Thus, the study concludes that aging reduces the global measure of HRV and may reflect reduced responsiveness of autonomic activity to external environmental stimuli with age. However, the time-domain short-term components of HRV are not affected by age and, therefore, the fast and presumably vagal modulations of heart rate appear to be maintained.

The performance of a heart rate (HR) and heart rate variability (HRV) monitor based on photoplethysmography (PPG) can be significantly degraded by the variations in contacting force applied on the sensor. This paper presents a wavelet approach to evaluate and reduce the influence of the force variations on the estimation of HR and HRV. The experiment was conducted on eighteen healthy subjects under three different force conditions - low, medium and high, respectively. The preliminary results indicate that reasonably accurate parameter estimation can be achieved from considerably distorted PPG signals even at a single scale of continuous wavelet transform (CWT).

Photoplethysmograph (PPG) is low cost, portable and simple optical technique used to monitor cardiovascular related parameters such as heart rate, blood pressure, cardiac output, oxygen saturation of blood, respiratory rate and many such parameters. In recent times, PPG has gained popularity due to its ease of measuring from the user's fingertips and its wearable nature. The measurements are taken preferably on the fingertips of the skin surface which has maximum nerve endings. In this paper, the PPG signal is obtained from the human subjects with the data acquisition system using Arduino board and Maxim Integrated sensor. Then signal processing is performed for removing baseline wander, motion artifacts and filtering of the signal for peak detection using the algorithm developed in MATLAB. The heart rate is calculated for 5 subjects using the maxim sensor kept on the finger tip. The capnobase database of various subjects of different ages is obtained for estimation of Heart Rate Variability and Signal to Noise Ratio. The peak detection algorithm shows maximum accuracy which is validated with capnobase labels.

The Kolmogorov-Smirnov (KS) test is an empirical distribution function (EDF) based goodness-of-fit test that requires the underlying hypothesized density to be continuous and completely specified. When the parameters are unknown and must be estimated from the data, standard tables of the KS test statistic are not valid. Approximate upper tail percentage points of the KS statistic for the inverse Gaussian (IG) distribution with unknown parameters are tabled in this paper. A study of the power of the KS test for the IG distribution indicates that the test is able todiscriminate between the IG distribution and distributions such as the uniform and exponentialdistributions that are very different in shape, but is relatively unable to discriminate between the IG distribution and distributions that are similar in shape such as the lognormal and Weibull distributions. In modeling settings the former distinction is typically more important to make than the latter distinction.