Search

Aging is accompanied by changes in the cardiovascular physiology that promote the development of age-related diseases. This paper presents a modern approach to quantify the physiological effects of age on the cardiovascular system by applying modern machine learning techniques to several indicators of autonomic cardiovascular function. In 885 healthy subjects, 33 different indices were calculated on resting state electrocardiogram and continuous blood pressure recordings. Based on those parameters, five different approaches were applied in order to reconstruct the calendar age of healthy subjects, i.e., linear regression (LR), neural network (NN), Gaussian process regression (GPR), support vector regression (SVR), and relevance vector regression (RVR). Hyper parameters of machine learning methods were optimized via grid search. After 20 repetitions of a five-fold cross-validation, the mean absolute error (MAE) was computed between the calendar and estimated age to assess the accuracy of each method. The results show that the lowest error for age estimation was achieved using SVR with a MAE of 5.49 years. GPR performed comparably well to SVR with a MAE of 5.55 years, while NN led to a MAE of 5.72 years. RVR and LR revealed MAE of more than six years (6.21 and 6.34 years). The error of age estimation could be further reduced by applying outlier correction to the input data leading to a minimum MAE of 4.53 years with GPR. In conclusion, our results suggest that machine learning can be used to quantify the effects of healthy aging on cardio-vascular function.

The ability to predict when a person is about to fall asleep is an important challenge in recent biomedical research and has various possible applications. Sleepiness and fatigue are known to increase pupillary fluctuations and the occurrence of eye blinks. In this study, we evaluated the use of the pupil diameter to forecast sleep episodes of short duration (>1s). We conducted multi-channel physiological and pupillometric recordings (diameter, gaze position) in 91 healthy volunteers at rest in supine position. Although they were instructed to keep their eyes open, short sleep episodes were detected in 20 participants (16 males, age: 26.2±5.6 years), 53 events in total. Before each sleep event, pupil size was extracted in a window of 30s (without additional sleep event). Mean pupil diameter and its standard deviation, Shannon entropy and wavelet entropy in the first half (15s) were compared to the second half of the window (15s). Linear and nonlinear measures demonstrated an elevation of pupil size variability before sleep onset. Most obviously, WE and SD increased significantly from 0.054±0.056 and 0.38±0.16 mm to 0.113±0.103 (T(102)=2.44, p

Autonomic control of the heart rate was demonstrated to be complex and nonlinear. Respiratory sinus arrhythima plays a crucial role in heart rate vagal modulation. Here we present an approach of assessing nonlinear causal relationships in bivariate time series, called mutual compression entropy (MCE). We applied MCE to cardiorespiratory data of 29 patients with acute schizophrenia and 29 matched controls. The method is based on data compression and estimates to which extend a (target) time series can be compressed regarding another time series (driver). In schizophrenia an elevated sympathetic and reduced parasympathetic heart rate modulation was found. The nonlinear influence of respiration on heart rate variability was demonstrated by a highly significant reduction of MCE (0.816 vs. 0.808, p¡0.01). In healthy subjects MCE was mainly related to sympathovagal balance. We conclude, that this index has the potential to uncover physiological information beyond linear measures.

Evaluating individual aerobic exercise capacity is fundamental in sports and exercise medicine but associated with organizational and instrumental effort. Here, we extract an index related to common performance markers, the aerobic and anaerobic thresholds enabling the estimation of exercise capacity from a conventional sports watch supporting beatwise heart rate tracking. Therefore, cardiac vagal threshold (CVT) was determined in 19 male subjects performing an incremental maximum exercise test. CVT varied around the anaerobic threshold AnT with mean deviation of 7.9 ± 17.7 W. A high correspondence of the two thresholds was indicated by Bland-Altman plots with limits of agreement −27.5 W and 43.4 W. Additionally, CVT was strongly correlated AnT (rp = 0.86, p < 0.001) and reproduced this marker well (rc = 0.81). We conclude, that cardiac vagal threshold derived from compression entropy time course can be useful to assess physical fitness in an uncomplicated way.

Monitoring of cardiac and respiratory activity, is essential in several clinical interventions like bicycle ergometries. The respiration signal can be derived from the ECG if it is not recorded itself (ECG derived respiration, EDR). In this study, we tried to reconstruct breathing rates (BR) from stress test high resolution 12-channel-ECGs in nine healthy subjects using higher order central moments. A mean absolute error per subjects of 2.9/min and relatively high correlation (rp = 0.85) and concordance coefficient (rc = 0.79) indicated a quite accurate reproduction of respiratory activity. The analysis of the different test stages revealed an increase of BR errors while subjects were effortful cycling compared to rest. During incremental cycling exercise test the mean absolute error per subjects was 3.4/min. Compared to the results reported in other studies at rest in supine position, this seems adequately accurate. In conclusion, our results indicate that EDR using higher order central moments is suited for monitoring BR during physical activity.

The arterial blood pressure (ABP) holds a lot of information about the cardiovascular system. To analyse the ABP signal, at first the pulse wave has to be detected correctly. Due to influences, e.g. of noise, an accurate beat detection is a challenging task. In this work, a novel real-time ABP detector based on higher-order statistics is presented. The method uses the 4th order cumulants for ABP-peak detection. To evaluate this detector, the Fantasia database freely available at physionet was used. In this database, 142,936 systolic peaks were available to check the efficiency of the beat detector. A sensitivity of 99.91% and a positive predictive value of 99.50% were achieved with the novel method.

For a variety of clinical applications like magnetic resonance imaging (MRI) the monitoring of vital signs is a common standard in clinical daily routine. Besides the electrocardiogram (ECG), the respiratory activity is an important vital parameter and might reveal pathological changes. Thoracic movement and the resulting impedance change between ECG electrodes enable the estimation of the respiratory signal from the ECG. This ECG-derived respiration (EDR) can be used to calculate the breathing rate without the need for additional devices or monitoring modules. In this paper a new method is presented to estimate the respiratory signal from a single-lead ECG. The 4th order central moments was used to estimate the EDR signal exploiting the change of the R-wave slopes induced by respiration. This method was compared with two approaches by analyzing the Fantasia database from www.physionet.org. Furthermore, the ECG signals of 24 healthy subjects placed in an 3 T MR-scanner were acquired.