A pinboard by
Natasa Reljin

postdoctoral fellow, University of Connecticut


Traumatic injury is the leading cause of death worldwide, with hemorrhage (blood loss) being accountable for about 40% of mortality. Therefore, accurate assessment of trauma patients for hemorrhage is very important in emergency and operative rooms, as well as on the battlefield. If not properly and promptly treated, massive bleeding, i.e., more than 30% of blood loss, may lead to hemorrhagic shock and death. The aim of my research is to automatically classify blood loss based on the photoplethysmographic signals collected with portable, noninvasive pulse oximeters placed on the three locations: finger, ear and forehead of a patient. From each signal four features (descriptors) were extracted at various time points in the recording. Then, a supervised machine learning algorithm is applied on the extracted features. Independently, every patient was assigned to one of two classes (Blood loss/No blood loss) by adjudication of all available clinical information compiled during the data collection by several physicians. The obtained results are encouraging, with high classification accuracy of distinguishing between blood loss and no blood loss.


Impact of lower body negative pressure induced hypovolemia on peripheral venous pressure waveform parameters in healthy volunteers.

Abstract: Lower body negative pressure (LBNP) creates a reversible hypovolemia by sequestrating blood volume in the lower extremities. This study sought to examine the impact of central hypovolemia on peripheral venous pressure (PVP) waveforms in spontaneously breathing subjects. With IRB approval, 11 healthy subjects underwent progressive LBNP (baseline, -30, -75, and -90 mmHg or until the subject became symptomatic). Each was monitored for heart rate (HR), finger arterial blood pressure (BP), a chest respiratory band and PVP waveforms which are generated from a transduced upper extremity intravenous site. The first subject was excluded from PVP analysis because of technical errors in collecting the venous pressure waveform. PVP waveforms were analyzed to determine venous pulse pressure, mean venous pressure, pulse width, maximum and minimum slope (time domain analysis) together with cardiac and respiratory modulations (frequency domain analysis). No changes of significance were found in the arterial BP values at -30 mmHg LBNP, while there were significant reductions in the PVP waveforms time domain parameters (except for 50% width of the respiration induced modulations) together with modulation of the PVP waveform at the cardiac frequency but not at the respiratory frequency. As the LBNP progressed, arterial systolic BP, mean BP and pulse pressure, PVP parameters and PVP cardiac modulation decreased significantly, while diastolic BP and HR increased significantly. Changes in hemodynamic and PVP waveform parameters reached a maximum during the symptomatic phase. During the recovery phase, there was a significant reduction in HR together with a significant increase in HR variability, mean PVP and PVP cardiac modulation. Thus, in response to mild hypovolemia induced by LBNP, changes in cardiac modulation and other PVP waveform parameters identified hypovolemia before detectable hemodynamic changes.

Pub.: 06 Jun '14, Pinned: 04 Jul '17

Time-varying methods for characterizing nonstationary dynamics of physiological systems.

Abstract: Accurate and early diagnosis of various diseases and pathological conditions require analysis techniques that can capture time-varying (TV) dynamics. In the pursuit of promising TV signal processing methods applicable to real-time clinical monitoring applications, nonstationary spectral techniques are of great significance.We present two potential practical applications of such techniques in quantifying TV physiological dynamics concealed in photoplethysmography (PPG) signals: early detection of blood-volume loss using a nonparametric approach known as variable frequency complex demodulation (VFCDM), and accurate detection of abrupt changes in respiratory rates using a parametric approach known as combined optimal parameter search and multiple mode particle filtering (COPS-MPF).The VFCDM technique has been tested using ear-PPG signals in two study models: mechanically ventilated patients undergoing surgery in operating room settings and spontaneously breathing conscious healthy subjects subjected to lower body negative pressure (LBNP) in laboratory settings. Extraction of respiratory rates has been tested using COPS-MPF technique in finger-PPG signals collected from healthy volunteers with abrupt changes in respiratory rate ranging from 0.1 to 0.4 Hz.VFCDM method showed promise to detect the blood loss noninvasively in mechanical ventilated patients well before blood losses become apparent to the physician. In spontaneously breathing subjects during LBNP experiments, the early detection and quantification of blood loss was possible at 40% of LBNP tolerance. COPS-MPF showed high accuracy in detecting the constant as well as sudden changes in respiratory rates as compared to other time-invariant methods.Integration of such robust algorithms into pulse oximeter device may have significant impact in real-time clinical monitoring and point-of-care healthcare settings.

Pub.: 28 Sep '10, Pinned: 04 Jul '17

A novel approach using time-frequency analysis of pulse-oximeter data to detect progressive hypovolemia in spontaneously breathing healthy subjects.

Abstract: Accurate and early detection of blood volume loss would greatly improve intraoperative and trauma care. This study has attempted to determine early diagnostic and quantitative markers for blood volume loss by analyzing photoplethysmogram (PPG) data from ear, finger and forehead sites with our high-resolution time-frequency spectral (TFS) technique in spontaneously breathing healthy subjects (n = 11) subjected to lower body negative pressure (LBNP). The instantaneous amplitude modulations present in heart rate (AM HR) and breathing rate (AMBR) band frequencies of PPG signals were calculated from the high-resolution TFS. Results suggested that the changes (P < 0.05) in AMBR and especially in AMHR values can be used to detect the blood volume loss at an early stage of 20% LBNP tolerance when compared to the baseline values. The mean percent decrease in AMHR values at 100% LBNP tolerance was 78.3%, 72.5%, and 33.9% for ear, finger, and forehead PPG signals, respectively. The mean percent increase in AMBR values at 100% LBNP tolerance was 99.4% and 19.6% for ear and finger sites, respectively; AMBR values were not attainable for forehead PPG signal. Even without baseline AMHR values, our results suggest that hypovolemia detection is possible with specificity and sensitivity greater than 90% for the ear and forehead locations when LBNP tolerance is 100%. Therefore, the TFS analysis of noninvasive PPG waveforms is promising for early diagnosis and quantification of hypovolemia at levels not identified by vital signs in spontaneously breathing subjects.

Pub.: 27 Apr '11, Pinned: 04 Jul '17

Using time-frequency analysis of the photoplethysmographic waveform to detect the withdrawal of 900 mL of blood.

Abstract: We designed this study to determine if 900 mL of blood withdrawal during spontaneous breathing in healthy volunteers could be detected by examining the time-varying spectral amplitude of the photoplethysmographic (PPG) waveform in the heart rate frequency band and/or in the breathing rate frequency band before significant changes occurred in heart rate or arterial blood pressure. We also identified the best PPG probe site for early detection of blood volume loss by testing ear, finger, and forehead sites.Eight subjects had 900 mL of blood withdrawn followed by reinfusion of 900 mL of blood. Physiological monitoring included PPG waveforms from ear, finger, and forehead probe sites, standard electrocardiogram, and standard blood pressure cuff measurements. The time-varying amplitude sequences in the heart rate frequency band and breathing rate frequency band present in the PPG waveform were extracted from high-resolution time-frequency spectra. These amplitudes were used as a parameter for blood loss detection.Heart rate and arterial blood pressure did not significantly change during the protocol. Using time-frequency analysis of the PPG waveform from ear, finger, and forehead probe sites, the amplitude signal extracted at the frequency corresponding to the heart rate significantly decreased when 900 mL of blood was withdrawn, relative to baseline (all P < 0.05); for the ear, the corresponding signal decreased when only 300 mL of blood was withdrawn. The mean percent decrease in the amplitude of the heart rate component at 900 mL blood loss relative to baseline was 45.2% (38.2%), 42.0% (29.2%), and 42.3% (30.5%) for ear, finger, and forehead probe sites, respectively, with the lower 95% confidence limit shown in parentheses. After 900 mL blood reinfusion, the amplitude signal at the heart rate frequency showed a recovery towards baseline. There was a clear separation of amplitude values at the heart rate frequency between baseline and 900 mL blood withdrawal. Specificity and sensitivity were both found to be 87.5% with 95% confidence intervals (47.4%, 99.7%) for ear PPG signals for a chosen threshold value that was optimized to separate the 2 clusters of amplitude values (baseline and blood loss) at the heart rate frequency. Meanwhile, no significant changes in the spectral amplitude in the frequency band corresponding to respiration were found.A time-frequency spectral method detected blood loss in spontaneously breathing subjects before the onset of significant changes in heart rate or blood pressure. Spectral amplitudes at the heart rate frequency band were found to significantly decrease during blood loss in spontaneously breathing subjects, whereas those at the breathing rate frequency band did not significantly change. This technique may serve as a valuable tool in intraoperative and trauma settings to detect and monitor hemorrhage.

Pub.: 01 May '12, Pinned: 04 Jul '17