Parsimonious Waveform-derived Features consisting of Pulse Arrival Time and Heart Rate Variability Predicts the Onset of Septic Shock.

Sepsis is a major public health emergency and one of the leading causes of morbidity and mortality in critically ill patients. For each hour treatment is delayed, shock-related mortality increases, so early diagnosis and intervention is of utmost importance. However, earlier recognition of shock requires active monitoring, which may be delayed due to subclinical manifestations of the disease at the early phase of onset. Machine learning systems can increase timely detection of shock onset by exploiting complex interactions among continuous physiological waveforms. We use a dataset consisting of high-resolution physiological waveforms from intensive care unit (ICU) of a tertiary hospital system. We investigate the use of mean arterial blood pressure (MAP), pulse arrival time (PAT), heart rate variability (HRV), and heart rate (HR) for the early prediction of shock onset. Using only five minutes of the aforementioned vital signals from 239 ICU patients, our developed models can accurately predict septic shock onset 6 to 36 hours prior to clinical recognition with area under the receiver operating characteristic (AUROC) of 0.84 and 0.8 respectively. This work lays foundations for a robust, efficient, accurate and early prediction of septic shock onset which may help clinicians in their decision-making processes. This study introduces machine learning models that provide fast and accurate predictions of septic shock onset times up to 36 hours in advance. BP, PAT and HR dynamics can independently predict septic shock onset with a look-back period of only 5 mins.

A Comparison of Normalization Techniques for Individual Baseline-Free Estimation of Absolute Hypovolemic Status Using a Porcine Model.

Hypovolemic shock is one of the leading causes of death in the military. The current methods of assessing hypovolemia in field settings rely on a clinician assessment of vital signs, which is an unreliable assessment of hypovolemia severity. These methods often detect hypovolemia when interventional methods are ineffective. Therefore, there is a need to develop real-time sensing methods for the early detection of hypovolemia. Previously, our group developed a random-forest model that successfully estimated absolute blood-volume status (ABVS) from noninvasive wearable sensor data for a porcine model (n = 6). However, this model required normalizing ABVS data using individual baseline data, which may not be present in crisis situations where a wearable sensor might be placed on a patient by the attending clinician. We address this barrier by examining seven individual baseline-free normalization techniques. Using a feature-specific global mean from the ABVS and an external dataset for normalization demonstrated similar performance metrics compared to no normalization (normalization: R2 = 0.82 ± 0.025|0.80 ± 0.032, AUC = 0.86 ± 5.5 × 10-3|0.86 ± 0.013, RMSE = 28.30 ± 0.63%|27.68 ± 0.80%; no normalization: R2 = 0.81 ± 0.045, AUC = 0.86 ± 8.9 × 10-3, RMSE = 28.89 ± 0.84%). This demonstrates that normalization may not be required and develops a foundation for individual baseline-free ABVS prediction.

Real-Time Seismocardiogram Feature Extraction Using Adaptive Gaussian Mixture Models.

Wearable systems can provide accurate cardiovascular evaluations by estimating hemodynamic indices in real-time. Key hemodynamic parameters can be non-invasively estimated using the seismocardiogram (SCG), a cardiomechanical signal whose features link to cardiac events like aortic valve opening (AO) and closing (AC). However, tracking a single SCG feature is unreliable due to physiological changes, motion artifacts, and external vibrations. This work proposes an adaptable Gaussian Mixture Model (GMM) to track multiple AO/AC correlated features in quasi-real-time from the SCG. The GMM calculates the likelihood of an extremum being an AO/AC feature for each SCG beat. The Dijkstra algorithm selects heartbeat-related extrema, and a Kalman filter updates the GMM parameters while filtering features. Tracking accuracy is tested on a porcine hypovolemia dataset with varying noise levels. Blood volume loss estimation accuracy is also evaluated using the tracked features on a previously developed model. Experimental results show a 4.5 ms tracking latency and average root mean square errors (RMSE) of 1.47 ms for AO and 7.67 ms for AC at 10 dB noise, and 6.18 ms for AO and 15.3 ms for AC at -10 dB noise. When considering all AO/AC correlated features, the combined RMSE remains in similar ranges, specifically 2.70 ms for AO and 11.91 ms for AC at 10 dB noise, and 7.50 ms for AO and 16.35 ms for AC at -10 dB noise. The proposed algorithm offers low latency and RMSE for all tracked features, making it suitable for real-time processing. These systems enable accurate, timely extraction of hemodynamic indices for many cardiovascular monitoring applications, including trauma care in field settings.

Improving Respiratory Timing Estimation Using Quality Indexing and Electrocardiogram-Derived Respiration.

Numerous applications require accurate estimation of respiratory timings. Respiratory effort (RSP) measurement is a popular approach to accomplish this, especially when the tightness of the sensing belt around the chest can be ensured. In less controlled settings, however, belt looseness and artifacts from movement of the belt on the chest can corrupt the signal. This paper demonstrates that respiration quality indexing and outlier removal can help mitigate these issues, improving estimates of respiration rate (RR), inspiration time (Ti), and expiration time (Te)., In a sample of 15 healthy human participants undergoing a protocol of five controlled breathing exercises in four postures each, electrocardiogram (ECG) and RSP signals were collected. RSP signals were processed to extract breath-by-breath estimates of RR, Ti, and Te. These estimates were compared against ground truth spirometry-based estimates using Bland-Altman analysis. We find that incorporating quality indexing and outlier removal prior to feature extraction improves the 95% limits of agreement by 10-40%. We also find that by using ECG-derived respiration (EDR) during periods of RSP artifact, the data removal necessary for accurate respiratory timing estimation is significantly reduced ( for all postures). These findings encourage the use of quality assessment and EDR to enhance the robustness of RR, Ti, and Te estimation from RSP signals. Clinical Relevance- Detecting stimulus-induced or pathological changes in respiratory function can enhance our understanding and monitoring of respiratory health. Quality assessment and the use of EDR help accomplish this by enabling more accurate measurement of respiratory timings.

Classification of Blood Volume Decompensation State via Machine Learning Analysis of Multi-Modal Wearable-Compatible Physiological Signals.

This paper presents a novel computational algorithm to estimate blood volume decompensation state based on machine learning (ML) analysis of multi-modal wearable-compatible physiological signals. To the best of our knowledge, our algorithm may be the first of its kind which can not only discriminate normovolemia from hypovolemia but also classify hypovolemia into absolute hypovolemia and relative hypovolemia. We realized our blood volume classification algorithm by (i) extracting a multitude of features from multi-modal physiological signals including the electrocardiogram (ECG), the seismocardiogram (SCG), the ballistocardiogram (BCG), and the photoplethysmogram (PPG), (ii) constructing two ML classifiers using the features, one to classify normovolemia vs. hypovolemia and the other to classify hypovolemia into absolute hypovolemia and relative hypovolemia, and (iii) sequentially integrating the two to enable multi-class classification (normovolemia, absolute hypovolemia, and relative hypovolemia). We developed the blood volume decompensation state classification algorithm using the experimental data collected from six animals undergoing normovolemia, relative hypovolemia, and absolute hypovolemia challenges. Leave-one-subject-out analysis showed that our classification algorithm achieved an F1 score and accuracy of (i) 0.93 and 0.89 in classifying normovolemia vs. hypovolemia, (ii) 0.88 and 0.89 in classifying hypovolemia into absolute hypovolemia and relative hypovolemia, and (iii) 0.77 and 0.81 in classifying the overall blood volume decompensation state. The analysis of the features embedded in the ML classifiers indicated that many features are physiologically plausible, and that multi-modal SCG-BCG fusion may play an important role in achieving good blood volume classification efficacy. Our work may complement existing computational algorithms to estimate blood volume compensatory reserve as a potential decision-support tool to provide guidance on context-sensitive hypovolemia therapeutic strategy.

Wearable Sensors and Machine Learning for Hypovolemia Problems in Occupational, Military and Sports Medicine: Physiological Basis, Hardware and Algorithms.

Hypovolemia is a physiological state of reduced blood volume that can exist as either (1) absolute hypovolemia because of a lower circulating blood (plasma) volume for a given vascular space (dehydration, hemorrhage) or (2) relative hypovolemia resulting from an expanded vascular space (vasodilation) for a given circulating blood volume (e.g., heat stress, hypoxia, sepsis). This paper examines the physiology of hypovolemia and its association with health and performance problems common to occupational, military and sports medicine. We discuss the maturation of individual-specific compensatory reserve or decompensation measures for future wearable sensor systems to effectively manage these hypovolemia problems. The paper then presents areas of future work to allow such technologies to translate from lab settings to use as decision aids for managing hypovolemia. We envision a future that incorporates elements of the compensatory reserve measure with advances in sensing technology and multiple modalities of cardiovascular sensing, additional contextual measures, and advanced noise reduction algorithms into a fully wearable system, creating a robust and physiologically sound approach to manage physical work, fatigue, safety and health issues associated with hypovolemia for workers, warfighters and athletes in austere conditions.

Estimation of Tidal Volume Using Load Cells on a Hospital Bed.

Although respiratory failure is one of the primary causes of admission to intensive care, the importance placed on measurement of respiratory parameters is commonly overshadowed compared to cardiac parameters. With the increased demand for unobtrusive yet quantifiable respiratory monitoring, many technologies have been proposed recently. However, there are challenges to be addressed for such technologies to enable widespread use. In this work, we explore the feasibility of using load cell sensors embedded on a hospital bed for monitoring respiratory rate (RR) and tidal volume (TV). We propose a globalized machine learning (ML)-based algorithm for estimating TV without the requirement of subject-specific calibration or training. In a study of 15 healthy subjects performing respiratory tasks in four different postures, the outputs from four load cell channels and the reference spirometer were recorded simultaneously. A signal processing pipeline was implemented to extract features that capture respiratory movement and the respiratory effects on the cardiac (i.e., ballistocardiogram, BCG) signals. The proposed RR estimation algorithm achieved a root mean square error (RMSE) of 0.6 breaths per minute (brpm) against the ground truth RR from the spirometer. The TV estimation results demonstrated that combining all three axes of the low-frequency force signals and the BCG heartbeat features best quantifies the respiratory effects of TV. The model resulted in a correlation and RMSE between the estimated and true TV values of 0.85 and 0.23 L, respectively, in the posture independent model without electrocardiogram (ECG) signals. This study suggests that load cell sensors already existing in certain hospital beds can be used for convenient and continuous respiratory monitoring in general care settings.

Reducing the Impact of External Vibrations on Fiducial Point Detection in Seismocardiogram Signals.

OBJECTIVE: Wearable systems that enable continuous non-invasive monitoring of hemodynamic parameters can aid in cardiac health evaluation in non-hospital settings. The seismocardiogram (SCG) is a non-invasively acquired cardiovascular biosignal for which timings of fiducial points, like aortic valve opening (AO) and aortic valve closing (AC), can enable estimation of key hemodynamic parameters. However, SCG is susceptible to motion artifacts, making accurate estimation of these points difficult when corrupted by high-g or in-band vibration artifacts. In this paper, a novel denoising pipeline is proposed that removes vehicle-vibration artifacts from corrupted SCG beats for accurate fiducial point detection. METHODS: The noisy SCG signal is decomposed with ensemble empirical mode decomposition (EEMD). Corrupted segments of the decomposed signal are then identified and removed using the quasi-periodicity of the SCG. Signal quality assessment of the reconstructed SCG beats then removes unreliable beats before feature extraction. The overall approach is validated on simulated vehicle-corrupted SCG generated by adding real subway collected vibration signals onto clean SCG. RESULTS: SNR increased by 8.1dB in the AO complex and 11.5dB in the AC complex of the SCG signal. Hemodynamic timing estimation errors reduced by 16.5% for pre-ejection period (PEP), 67.2% for left ventricular ejection time (LVET), and 57.7% for PEP/LVET-a feature previously determined in prior work to be of great importance for assessing blood volume status during hemorrhage. CONCLUSION: These findings suggest that usable SCG signals can be recovered from vehicle-corrupted SCG signals using the presented denoising framework, allowing for accurate hemodynamic timing estimation. SIGNIFICANCE: Reliable hemodynamic estimates from vehicle-corrupted SCG signals will enable the adoption of the SCG in outside-of-hospital settings.

Unifying the Estimation of Blood Volume Decompensation Status in a Porcine Model of Relative and Absolute Hypovolemia Via Wearable Sensing.

Hypovolemia remains the leading cause of preventable death in trauma cases. Recent research has demonstrated that using noninvasive continuous waveforms rather than traditional vital signs improves accuracy in early detection of hypovolemia to assist in triage and resuscitation. This work evaluates random forest models trained on different subsets of data from a pig model (n = 6) of absolute (bleeding) and relative (nitroglycerin-induced vasodilation) progressive hypovolemia (to 20% decrease in mean arterial pressure) and resuscitation. Features for the models were derived from a multi-modal set of wearable sensors, comprised of the electrocardiogram (ECG), seismocardiogram (SCG) and reflective photoplethysmogram (RPPG) and were normalized to each subject.s baseline. The median RMSE between predicted and actual percent progression towards cardiovascular decompensation for the best model was 30.5% during the relative period, 16.8% during absolute and 22.1% during resuscitation. The least squares best fit line over the mean aggregated predictions had a slope of 0.65 and intercept of 12.3, with an R2 value of 0.93. When transitioned to a binary classification problem to identify decompensation, this model achieved an AUROC of 0.80. This study: a) developed a global model incorporating ECG, SCG and RPPG features for estimating individual-specific decompensation from progressive relative and absolute hypovolemia and resuscitation; b) demonstrated SCG as the most important modality to predict decompensation; c) demonstrated efficacy of random forest models trained on different data subsets; and d) demonstrated adding training data from two discrete forms of hypovolemia increases prediction accuracy for the other form of hypovolemia and resuscitation.

Accurate Ballistocardiogram Based Heart Rate Estimation Using an Array of Load Cells in a Hospital Bed.

The ballistocardiogram (BCG), a cardiac vibration signal, has been widely investigated for continuous monitoring of heart rate (HR). Among BCG sensing modalities, a hospital bed with multi-channel load-cells could provide robust HR estimation in hospital setups. In this work, we present a novel array processing technique to improve the existing HR estimation algorithm by optimizing the fusion of information from multiple channels. The array processing includes a Gaussian curve to weight the joint probability according to the reference value obtained from the previous inter-beat-interval (IBI) estimations. Additionally, the probability density functions were selected and combined according to their reliability measured by q-values. We demonstrate that this array processing significantly reduces the HR estimation error compared to state-of-the-art multi-channel heartbeat detection algorithms in the existing literature. In the best case, the average mean absolute error (MAE) of 1.76 bpm in the supine position was achieved compared to 2.68 bpm and 1.91 bpm for two state-of-the-art methods from the existing literature. Moreover, the lowest error was found in the supine posture (1.76 bpm) and the highest in the lateral posture (3.03 bpm), thus elucidating the postural effects on HR estimation. The IBI estimation capability was also evaluated, with a MAE of 16.66 ms and confidence interval (95%) of 38.98 ms. The results demonstrate that improved HR estimation can be obtained for a bed-based BCG system with the multi-channel data acquisition and processing approach described in this work.

Harnessing the Manifold Structure of Cardiomechanical Signals for Physiological Monitoring During Hemorrhage.

OBJECTIVE: Local oscillation of the chest wall in response to events during the cardiac cycle may be captured using a sensing modality called seismocardiography (SCG), which is commonly used to infer cardiac time intervals (CTIs) such as the pre-ejection period (PEP). An important factor impeding the ubiquitous application of SCG for cardiac monitoring is that morphological variability of the signals makes consistent inference of CTIs a difficult task in the time-domain. The goal of this work is therefore to enable SCG-based physiological monitoring during trauma-induced hemorrhage using signal dynamics rather than morphological features. METHODS: We introduce and explore the observation that SCG signals follow a consistent low-dimensional manifold structure during periods of changing PEP induced in a porcine model of trauma injury. Furthermore, we show that the distance traveled along this manifold correlates strongly to changes in PEP ( ∆PEP). RESULTS: ∆PEP estimation during hemorrhage was achieved with a median R2 of 92.5% using a rapid manifold approximation method, comparable to an ISOMAP reference standard, which achieved an R2 of 95.3%. CONCLUSION: Rapidly approximating the manifold structure of SCG signals allows for physiological inference abstracted from the time-domain, laying the groundwork for robust, morphology-independent processing methods. SIGNIFICANCE: Ultimately, this work represents an important advancement in SCG processing, enabling future clinical tools for trauma injury management.

Localizing Placement of Cardiomechanical Sensors during Dynamic Periods via Template Matching.

Captured with a chest-mounted sensor, the seismo- cardiogram (SCG) is a useful signal for assessing cardiomechanical function. However, the reliability of information obtained from this signal often depends upon sensor location. This has important practical implications, as consistent placement is not guaranteed in at-home and other uncontrolled settings. Building on prior research that localized SCG sensor placement when the patient was at rest - which may not be the case in practical settings - this work presents a more robust method which is able to localize sensor placement during dynamic periods, specifically exercise recovery. This was accomplished via a template-based signal quality index (SQI), which was used to infer sensor location using a variety of classifiers. While prior work generated synthetic templates for this task using an averaging method, it is shown that selecting representative templates from the training set instead enables, for the first time, SCG sensor localization during dynamic periods without patient-specific calibration. With this method, a peak accuracy of 83.32% was achieved for correctly classifying sensor position among five tested positions, with avenues for improvement of these results also presented.

Mitigating Hypovolemia-Induced Miscalibration of Photoplethysmogram-Derived Blood Pressure.

Pulse transit time (PTT) is a hemodynamic indicator that may be obtained non-invasively using photoplethysmogram (PPG) signals for continuous blood pressure (BP) monitoring. Among the most promising applications of this technology are military and civilian trauma cases, where reduced blood volume due to hemorrhage, or absolute hypovolemia, is the leading preventable cause of death. However, the drawback of this method is that it requires calibration for each patient; additionally, changes in physiological state may affect PTT calibration. In this work, a porcine model (n = 6) was used to demonstrate that changes in blood volume lead to miscalibration of PTT for BP estimation. To mitigate hypovolemia-induced miscalibration, this work first defines a template-based signal quality index (SQI) for characterizing the morphology of PPG signals; it is then shown that the subject-specific calibration of SQI to BP is more robust to changes in blood volume than PTT. Though changes in PPG signal quality are not necessarily specific to changes in BP, these results suggest that PPG-based monitoring systems may benefit from incorporating morphological information for cuffless BP estimation in trauma settings.

Enabling the assessment of trauma-induced hemorrhage via smart wearable systems.

As the leading cause of trauma-related mortality, blood loss due to hemorrhage is notoriously difficult to triage and manage. To enable timely and appropriate care for patients with trauma, this work elucidates the externally measurable physiological features of exsanguination, which were used to develop a globalized model for assessing blood volume status (BVS) or the relative severity of blood loss. These features were captured via both a multimodal wearable system and a catheter-based reference and used to accurately infer BVS in a porcine model of hemorrhage (n = 6). Ultimately, high-level features of cardiomechanical function were shown to strongly predict progression toward cardiovascular collapse and used to estimate BVS with a median error of 15.17 and 18.17% for the catheter-based and wearable systems, respectively. Exploring the nexus of biomedical theory and practice, these findings lay the groundwork for digital biomarkers of hemorrhage severity and warrant further study in human subjects.

Modeling Consistent Dynamics of Cardiogenic Vibrations in Low-Dimensional Subspace.

The seismocardiogram (SCG) measures the movement of the chest wall in response to underlying cardiovascular events. Though this signal contains clinically-relevant information, its morphology is both patient-specific and highly transient. In light of recent work suggesting the existence of population-level patterns in SCG signals, the objective of this study is to develop a method which harnesses these patterns to enable robust signal processing despite morphological variability. Specifically, we introduce seismocardiogram generative factor encoding (SGFE), which models the SCG waveform as a stochastic sample from a low-dimensional subspace defined by a unified set of generative factors. We then demonstrate that during dynamic processes such as exercise-recovery, learned factors correlate strongly with known generative factors including aortic opening (AO) and closing (AC), following consistent trajectories in subspace despite morphological differences. Furthermore, we found that changes in sensor location affect the perceived underlying dynamic process in predictable ways, thereby enabling algorithmic compensation for sensor misplacement during generative factor inference. Mapping these trajectories to AO and AC yielded R2 values from 0.81-0.90 for AO and 0.72-0.83 for AC respectively across five sensor positions. Identification of consistent behavior of SCG signals in low dimensions corroborates the existence of population-level patterns in these signals; SGFE may also serve as a harbinger for processing methods that are abstracted from the time domain, which may ultimately improve the feasibility of SCG utilization in ambulatory and outpatient settings.

A Unified Framework for Quality Indexing and Classification of Seismocardiogram Signals.

The seismocardiogram (SCG) is a noninvasively-obtained cardiovascular bio-signal that has gained traction in recent years, however is limited by its susceptibility to noise and motion artifacts. Because of this, signal quality must be assured before data are used to inform clinical care. Common methods of signal quality assurance include signal classification or assignment of a numerical quality index. Such tasks are difficult with SCG because there is no accepted standard for signal morphology. In this paper, we propose a unified method of quality indexing and classification that uses multi-subject-based methods to overcome this challenge. Dynamic-time feature matching is introduced as a novel method of obtaining the distance between a signal and reference template, with this metric, the signal quality index (SQI) is defined as a function of the inverse distance between the SCG and a large set of template signals. We demonstrate that this method is able to stratify SCG signals on held-out subjects based on their level of motion-artifact corruption. This method is extended, using the SQI as a feature for classification by ensembled quadratic discriminant analysis. Classification is validated by demonstrating, for the first time, both detection and localization of SCG sensor misplacement, achieving an F1 score of 0.83 on held-out subjects. This paper may provide a necessary step toward automating the analysis of SCG signals, addressing many of the key limitations and concerns precluding the method from being widely used in clinical and physiological sensing applications.

Characterization of Anisotropic Leaky Mode Modulators for Holovideo.

Holovideo displays are based on light-bending spatial light modulators. One such spatial light modulator is the anisotropic leaky mode modulator. This modulator is particularly well suited for holographic video experimentation as it is relatively simple and inexpensive to fabricate. Some additional advantages of leaky mode devices include: large aggregate bandwidth, polarization separation of signal light from noise, large angular deflection and frequency control of color. In order to realize these advantages, it is necessary to be able to adequately characterize these devices as their operation is strongly dependent on waveguide and transducer parameters. To characterize the modulators, the authors use a commercial prism coupler as well as a custom characterization apparatus to identify guided modes, calculate waveguide thickness and finally to map the device's frequency input and angular output of leaky mode modulators. This work gives a detailed description of the measurement and characterization of leaky mode modulators suitable for full-color holographic video.