Mireya Fernández-Chimeno

Heart rate variability analysis using a seismocardiogram signal

Seismocardiography is a simple and non invasive method of recording cardiac activity from the movements of the body caused by heart pumping. In this preliminary study we use a smartphone to record this acceleration and estimate the heart rate. We compare the heart rate variability parameters from the seismocardiogram and ECG reference signal. The results show a great similarity and are strongly influenced by the instability in the sampling frequency of the device. The differences between RR series are lower than 10 ms.

Errors in the Estimation of Approximate Entropy and Other Recurrence-Plot-Derived Indices Due to the Finite Resolution of RR Time Series

An analysis of the errors due to the finite resolution of RR time series in the estimation of the approximate entropy (ApEn) is described. The quantification errors in the discrete RR time series produce considerable errors in the ApEn estimation (bias and variance) when the signal variability or the sampling frequency is low. Similar errors can be found in indices related to the quantification of recurrence plots. An easy way to calculate a figure of merit [the signal to resolution of the neighborhood ratio (SRN)] is proposed in order to predict when the bias in the indices could be high. When SRN is close to an integer value n, the bias is higher than when near n-1/2 or n+1/2. Moreover, if SRN is close to an integer value, the lower this value, the greater the bias is.

Bias and uncertainty in heart rate variability spectral indices due to the finite ECG sampling frequency

Spectral analysis of the heart rate variability is becoming a usual tool as a marker of the autonomic nervous system. The final output of the spectral analysis is a set of indices that are always estimators due to technical limitations. In this work, the bias and the uncertainty in the VLF, LF, HF and LF/HF indices due to the finite sampling frequency of the ECG are analysed. The results indicate that for low sampling frequency (125 Hz), the bias and uncertainty in the HF and LF/HF indices can blur the results of the analysis, especially if the RR time series has low variability. The HF index is overestimated and, accordingly, the LF/HF index is underestimated. Then, results from RR time series with low sampling frequency must be used with care. The uncertainty of the spectral indices is proportional to the inverse of the sampling frequency and the bias is proportional to the inverse of the square sampling frequency.

A new index for the analysis of heart rate variability dynamics: characterization and application

A new index for the analysis of the heart rate variability (HRV) dynamics is presented. The proposed index (acceleration change index (ACI)) characterizes the sign of the differences of a time series. A theoretical study shows an expression that relates ACI and the autocorrelation function of the time series. This formula is tested and validated with different simulated time series (uncorrelated noise, sinusoidal and fractional Brownian motion). Next, ACI is applied to RR time series from healthy subjects showing that ACI decreases with periodic controlled breathing, increases during exercise, and it has a lower value at night than during the day. In a preliminary study, ACI has been shown to be lower in healthy subjects than in patients who had suffered a myocardial infarction one month previously. ACI can be employed as a fast and robust new marker of the HRV dynamics.

Differences in QRS Locations due to ECG Lead: Relationship with Breathing

Results of heart rate variability analysis depend on the accuracy of the RR time series that is measured only in one lead of the ECG. RR time series can subtly change from lead to lead. We have obtained the ECG of 21 healthy subjects in a quiet measurement, sampled at 5 kHz and from the I and II standard leads. For each subject a total of 60 minutes was measured. The QRS complexes in both leads have been detected using a conventional QRS detector plus a further refinement using a matching template. Differences between the locations of the QRS complexes have been quantified and compared with the breathing signal of each subject as well as the derived RR time series. The typical uncertainty in the fiducial points and RR time series is usually below 1 ms and the errors are modulated by breathing.

Drowsiness detection by thoracic effort signal analysis in real driving environments

Detection of drowsiness while driving is a leading objective in advanced driver assistance systems. This work presents a new index to assess the alertness state of drivers based on the respiratory dynamics derived from an inductive band. More than 100 hours of driving in real environments from 13 healthy subjects were analyzed. The proposed method has a sensitivity of 93.7% and specificity of 86.3% in detecting full awake drivers while it has a sensitivity of 83.1% and specificity of 95.3% in detecting drowsy drivers. The results show that the proposed index may be promising to assess the alertness state of real drivers.

An application of fractional differintegration to heart rate variability time series

Fractional differintegration is used as a new tool to characterize heart rate variability time series. This paper proposes and focuses in two indexes (αc and fnQ) derived from the fractional differintegration operator. Both indexes are applied to fractional Gaussian noise (fGn) and actual RR time series in order to test their behavior. In the analysis of monofractal time series, αc is linearly related with the Hurst exponent and the estimation of the exponent by the proposed index has lower variance than by using Detrended Fluctuation Analysis (DFA) or the periodogram. The other index fnQ quantifies how the time series adjust to a monofractal time series. Age, postural changes and paced breathing cause significant changes on fnQ while αc only shows significant changes due to posture. In the analyzed actual HRV time series, αc shows good correlation with the short term scaling exponent obtained by DFA, LF/HF and RMSSD while no correlations have been found for fnQ.

Understanding electrosurgical unit perturbations in order to address hospital operating room electromagnetic compatibility

We analyze the radiated emissions from a low power electrosurgical unit(ESU). Measurements at a 3-meter distance are performed in order to find a suitable measurement setup intended to reproduce realistically the real working ESU behavior in a test lab. The broad-band and narrow-band characteristics of the perturbation are studied in order to address the other equipment immunity.

Mobile phones electromagnetic interference in medical environments: A review

This paper presents a review of the effect of electromagnetic interferences of mobile phones in medical environment. The review is focussed in two areas, EMI in implantable devices like pacemakers or ICD, and EMI in medical equipment in hospital areas with life supporting equipment like intensive care units (ICU) or operating rooms (OR). The applied methodology for interference testing is different for all the analyzed studies, and the results range from 100% of interfered devices to no effect. The majority of authors recommend maintaining a safety distance both between cellular phones and medical devices. Also, it is necessary to establish clear policies for the use of mobile phones in critical areas in hospitals. Finally, it is necessary to develop a family of EMC standards for testing in-situ the effects of non-medical devices inside medical facilities. This will permit medical staff to take profit from new technologies in their daily work in a safe way.

The effect of electrocardiographic lead choice on RR time series

Results of heart rate variability analysis depend on the quality of the initial RR time series that is measured only in one lead of the ECG. This work shows that RR time series can subtly change from lead to lead so the choice of the analyzed lead is another source of uncertainty. The standard deviation of the differences of two RR time series obtained from different leads can change from 0.5 ms to more than 20 ms depending on the amount of noise, the morphological changes of the QRS complexes, the strategies of fiducial point determination and the measured subject. This source of uncertainty is in healthy subjects greater than that associated to the sampling frequency of the ECG for sampling frequencies greater than 400 Hz.