Is entheses ultrasound reliable? A reading Latin American exercise.

The objective of this study is to evaluate inter-reader entheses ultrasound (US) reliability and the influence of the type of image or degree of sonographer experience on US reliability in patients with spondyloarthritis (SpA). Eighteen Latin American ultrasonographers with different experience took part in an US reading exercise evaluating 60 entheseal images (50 % static images and 50 % videos) from healthy controls and SpA patients. The following sonographic lesions were assessed: structure, thickness, bone proliferation/tendon calcification, erosions, bursitis, and Doppler signal. Another group of three experts with significant experience in entheses US read all images too. Inter-reader reliability among participants and experts was calculated by the Cohen's kappa coefficient. Thresholds for kappa values were <0.2 poor, 0.21-0.4 fair, 0.41-0.6 moderate, 0.61-0.8 good, and 0.81-1 excellent. Furthermore, the results for the expert group were stratified based on the type of image. Kappa correlation coefficients among participants, showed variability depending on the type of lesion, being fair for structure and thickness, moderate for calcifications, erosions, and bursitis, and excellent for Doppler signal. Inter-reader reliability among experts was higher, being moderate for structure and thickness, good for calcifications and bursitis, and excellent for erosions and Doppler. Inter-reader reliability for assessing calcification and structure using static images was significantly higher than for videos. Overall inter-reader reliability for assessing entheses by US in SpA is moderate to excellent for most of the lesions. However, special training seems fundamental to achieve better inter-reader reliability. Moreover, the type of image influenced these results, where evaluation of entheses by videos was more difficult than by static images.

Effects of stray capacitance to ground in three electrode monopolar needle bioimpedance measurements.

Positive phase angle is documented and analyzed in a three electrode monopolar needle measurement. Inductance equivalent behavior of the stray capacitance to ground is described as error source in a non-inductive sample measurement.

On time interval measurements using BCG.

Time intervals measured between the electrocardiogram (ECG), the photoplethysmogram (PPG) or the impedance plethysmogram (IPG), have long been used to noninvasively assess cardiovascular function. Recently, the ballistocardiogram (BCG) has been proposed as an alternative physiological signal to be used in time interval measurements for the same purpose. In this work, we study the behavior of the RJ interval, defined as the time between the R wave of the electrocardiogram (ECG) and the J wave of the BCG, under fast pressure changes induced by paced respiration and tracked by a beat-to-beat blood pressure (SBP and DBP) waveform monitor. The aim of this work is to gain a deeper understanding of these newly proposed time intervals and to further assess their usefulness to determine cardiovascular performance.

Pulse arrival time estimation from the impedance plethysmogram obtained with a handheld device.

This paper describes a novel method to estimate pulse arrival time (PAT) from the electrocardiogram (ECG) and the impedance plethysmogram (TPG) obtained by using a compact and easy-to-use handheld device with only four electrodes. A proof-of-concept has been carried out where PAT values obtained with the proposed device have been compared to PAT values measured between the ECG and the photoplethysmogram (PPG) during three experiments of paced respiration to induce controlled PAT changes. The results show that both methods yield equivalent PAT values in within ± 7 ms (95 % confidence interval), which is less than typical deviations reported for common PAT measurements.

Can driven-right-leg circuits increase interference in ECG amplifiers?

The Driven-Right-Leg (DRL) circuit has been used for about 50 years to reduce interference due to common-mode voltage in biopotential amplifiers in scenarios that range from fixed equipment supplied from power lines to battery-supplied ambulatory monitors, and for systems that use gelled, dry, textile, and capacitive electrodes. However, power-line interference models predict that for isolation amplifiers, currently mandated by safety standards, power-line interference can often couple mostly in differential mode rather than in common mode. In this work we analyze the effect of the DRL circuit in different ECG leads to elucidate its actual effect on power-line interference reduction. It turns out that that the DRL circuit, which effectively reduces common-mode interference, affects differential-mode interference in an unpredictable way and can increase interference.

Heart rate detection from an electronic weighing scale.

We propose a novel technique for beat-to-beat heart rate detection based on the ballistocardiographic (BCG) force signal from a subject standing on a common electronic weighing scale. The detection relies on sensing force variations related to the blood acceleration in the aorta, works even if wearing footwear and does not require any sensors attached to the body because it uses the load cells in the scale. We have devised an approach to estimate the sensitivity and frequency response of three commercial weighing scales to assess their capability to detect the BCG force signal. Static sensitivities ranged from 490 nV V(-1) N(-1) to 1670 nV V(-1) N(-1). The frequency response depended on the subject's mass but it was broad enough for heart rate estimation. We have designed an electronic pulse detection system based on off-the-shelf integrated circuits to sense heart-beat-related force variations of about 0.24 N. The signal-to-noise ratio of the main peaks of the force signal detected was higher than 30 dB. A Bland-Altman plot was used to compare the RR time intervals estimated from the ECG and BCG force signals for 17 volunteers. The error was +/-21 ms, which makes the proposed technique suitable for short-term monitoring of the heart rate.

Heart rate detection from an electronic weighing scale.

We propose a novel technique for heart rate detection on a subject that stands on a common electronic weighing scale. The detection relies on sensing force variations related to the blood acceleration in the aorta, works even if wearing footwear, and does not require any sensors attached to the body. We have applied our method to three different weighing scales, and estimated whether their sensitivity and frequency response suited heart rate detection. Scale sensitivities were from 490 nV/V/N to 1670 nV/V/N, all had an underdamped transient response and their dynamic gain error was below 19% at 10 Hz, which are acceptable values for heart rate estimation. We also designed a pulse detection system based on off-the-shelf integrated circuits, whose gain was about 70x10(3) and able to sense force variations about 240 mN. The signal-to-noise ratio (SNR) of the main peaks of the pulse signal detected was higher than 48 dB, which is large enough to estimate the heart rate by simple signal processing methods. To validate the method, the ECG and the force signal were simultaneously recorded on 12 volunteers. The maximal error obtained from heart rates determined from these two signals was +/-0.6 beats/minute.

Heart rate detection from plantar bioimpedance measurements.

The heart rate is a basic health indicator, useful in both clinical measurements and home health care. Current home care systems often require the attachment of electrodes or other sensors to the body, which can be cumbersome to the patient. Moreover, some measurements are sensitive to movement artifacts, are not user-friendly and require a specialized supervision. In this paper, a novel technique for heart rate measurement for a standing subject is proposed, which is based on plantar bioimpedance measurements, such as those performed by some bathroom weighting scales for body composition analysis. Because of the low level of heart-related impedance variations, the measurement system has a gain of 1400. We have implemented a fully differential AC amplifier with a common-mode rejection ratio (CMRR) of 105 dB at 10 kHz. Coherent demodulation based on synchronous sampling yields a signal-to-noise ratio (SNR) of 55 dB. The system has a sensitivity of 1.9 V/Omega. The technique has been demonstrated on 18 volunteers, whose bioimpedance signal and ECG were simultaneously measured to validate the results. The average cross-correlation coefficient between the heart rates determined from these two signals was 0.998 (std. dev. 0.001).

A novel robust index to assess beat-to-beat variability in heart rate time-series analysis.

A new index is proposed to estimate the variance of the differentiated heart rate (RR) time series from its truncated histogram. The index is more robust to artifacts than the standard deviation of the differentiated RR time series (rMSDD) and, unlike the pNN50, does not saturate for very high or very low heart rate variability.

Variations in breathing patterns increase low frequency contents in HRV spectra.

This paper shows that variations in breathing patterns broaden heart rate variability (HRV) spectral bands and increase the power amplitude of low-frequency bands. Because of these influences, spectral markers for HRV signals, such as the quotient between spectral power at different frequency bands, should be compared only under controlled breathing conditions or after considering the effect of variations in breathing patterns.

Ag-AgCl electrode noise in high-resolution ECG measurements.

The authors measured the noise and impedance from face-to-face Ag-AgCl electrode pairs, as well as the noise from Ag-AgCl electrodes placed on the human body surface, in the frequency band from 0.5 Hz to 500 Hz, which corresponds to high-resolution ECG measurements. Electrode noise and electrode impedance were measured simultaneously to compare electrode noise with the thermal noise from the real part of electrode impedance. The results show that electrode noise depends on electrode area, electrolytic gel, the patient, and the placement site. In the frequency band from 0.5 Hz to 500 Hz, root-mean-square electrode noise is typically less than 1 microV for electrodes placed face-to-face and ranges from 1 microV to 15 microV for electrodes on the body surface. The noise spectral density increases at low frequencies as 1/fa and it is always higher than the thermal noise from the real part of the electrode impedance. There is a high correlation between electrode dc offset voltage and electrode noise. Thus, offset voltage measurements allow identification of noise from low-noise electrodes.

A new method for automated blood pressure measurement.

We propose a new technique for automated indirect blood pressure measurement based on the auscultatory method. Systolic, diastolic and blood mean pressure are identified by looking at trend changes in the spectral energy dispersion of Korotkoff sounds. The detection is solely based on patient measurements, not on population studies. By comparing the automatic detection with common auditory detection, in 286 measurements taken in 15 subjects there was agreement (+/- 1 sound--400 Pa error) in 278 cases for the systole and 276 cases for the diastole.

Multi-frequency static imaging in electrical impedance tomography: Part 1. Instrumentation requirements.

Static images of the human body using electrical impedance tomography techniques can be obtained by measuring at two or more different frequencies. The frequencies used depend on the application, and their selection depends on the frequency behaviour of the impedance for the target tissue. An analysis using available data and theoretical models for tissue impedance yields the expected impedance and boundary voltage changes, therefore setting the measurement instrument specifications. The instrument errors produced by different sources are analysed, and, from this analysis it is possible to determine the feasibility of building the instrument, the limit values for some parameters (or components) and indications on the most suitable design of critical parts. This analysis also shows what kinds of error can be expected in the reconstructed images. It is concluded that it is possible to build an instrument with limited errors, allowing static images to be obtained. An instrument has been built that meets some of the design requirements and fails in others because of technological problems. In vivo images obtained with this instrument will be presented in Part 2 of this work.

A multifrequency multichannel electrical impedance data acquisition system for body fluid shift monitoring.

This paper discusses some important issues for the design of electrical impedance measurement systems intended for body fluid shift monitoring, in particular during dialysis treatments. We have studied two common signal generation systems: digital synthesis and carrier recovery. We have found that in prolonged measurement applications, digital synthesis yields the best performance. On the demodulation side, we balance the demodulator errors between the real and imaginary parts by rotating the demodulation axes. We use segmental multifrequency impedance measurements to estimate the values of intracellular and extracellular impedance by adjusting the parameters of a Cole-Cole model for each segment measured. We stress the need to perform segmental measurements in order to accurately measure the segments of interest, in particular the trunk during dialysis treatments. Our results show that there is a sharp disequilibrium between the intracellular and extracellular compartments in the very first dialysis period. This fact generates the need to continuously measure segmental impedance instead of comparing initial and final values.

Errors in prolonged electrical impedance measurements due to electrode repositioning and postural changes.

Long-term electrical impedance measurements are affected by specific errors. Electrode failure, changes in its impedance due to aging, and postural changes are among the most important. We analyse errors due to electrode replacement and body postural changes. Electrode replacement errors can cause impedance changes up to 5% of basal value. This is one of the most important factors in data reproducibility. Body postural changes also contribute to impedance variations. We have proposed the use of a reference position to carry out impedance measurements as the one that shows the smallest impedance sensitivity to postural changes. In general, we observed that this is achieved with arms and legs slightly separated from the body. We propose the use of a ratio of impedance at two different frequencies to discern the origin of impedance changes, whether from physiological phenomena or postural errors.

Bioelectric impedance measurements using synchronous sampling.

We have applied synchronous sampling to the demodulation of bioelectric impedance signals. This overcomes the need for analog demodulators in bioimpedance measurements. The sampling rate is determined by signal bandwidth, rather than by the highest frequency component before demodulation.

AC instrumentation amplifier for bioimpedance measurements.

We analyze the input impedance and CMRR requirements for an amplifier for bioimpedance measurements when considering the capacitive components of the electrode-skin contact impedance. We describe an ac-coupled instrumentation amplifier (IA) that, in addition to fulfilling those requirements, both provides interference and noise reduction, and yields a zero phase shift over a wide frequency band without using broadband op amps.

Walsh function current patterns and data synthesis for electrical impedance tomography.

A data collection method which uses Walsh functions as injection current patterns is presented. This method can satisfy two conditions: the optimality of current patterns in every iteration and the single-time data measurement condition. The use of Walsh functions simplifies the design of current sources since only two levels of current (+1 and -1) are required, whereas sinusoidal injection requires a digital-to-analog converter to produce many different values of currents. Compared to diagonal or neighboring type of pulses as injection current patterns, Walsh injection current patterns provide more information about the interior of the subject since Walsh function simulate low and high spatial frequency patterns. Therefore, Walsh function injection uses the simplicity of pulse type injection and yields the better distinguishability or SNR of sinusoidal injection.