A robust approach toward recognizing valid arterial-blood-pressure pulses

Generative adversarial networks with fully connected layers to denoise PPG signals

Objective.The detection of arterial pulsating signals at the skin periphery with Photoplethysmography (PPG) are easily distorted by motion artifacts. This work explores the alternatives to the aid of PPG reconstruction with movement sensors (accelerometer and/or gyroscope) which to date have demonstrated the best pulsating signal reconstruction.Approach.A generative adversarial network with fully connected layers is proposed for the reconstruction of distorted PPG signals. Artificial corruption was performed to the clean selected signals from the BIDMC Heart Rate dataset, processed from the larger MIMIC II waveform database to create the training, validation and testing sets.Main results.The heart rate (HR) of this dataset was further extracted to evaluate the performance of the model obtaining a mean absolute error of 1.31 bpm comparing the HR of the target and reconstructed PPG signals with HR between 70 and 115 bpm.Significance.The model architecture is effective at reconstructing noisy PPG signals regardless the length and amplitude of the corruption introduced. The performance over a range of HR (70-115 bpm), indicates a promising approach for real-time PPG signal reconstruction without the aid of acceleration or angular velocity inputs.

Data analysis protocol for early autonomic dysfunction characterization after severe traumatic brain injury

Background

Traumatic brain injury (TBI) disrupts normal brain tissue and functions, leading to high mortality and disability. Severe TBI (sTBI) causes prolonged cognitive, functional, and multi-organ dysfunction. Dysfunction of the autonomic nervous system (ANS) after sTBI can induce abnormalities in multiple organ systems, contributing to cardiovascular dysregulation and increased mortality. Currently, detailed characterization of early autonomic dysfunction in the acute phase after sTBI is lacking. This study aims to use physiological waveform data collected from patients with sTBI to characterize early autonomic dysfunction and its association with clinical outcomes to prevent multi-organ dysfunction and improving patient outcomes.

Objective

This data analysis protocol describes our pre-planned protocol using cardiac waveforms to evaluate early autonomic dysfunction and to inform multi-dimensional characterization of the autonomic nervous system (ANS) after sTBI.

Methods

We will collect continuous cardiac waveform data from patients managed in an intensive care unit within a clinical trial. We will first assess the signal quality of the electrocardiogram (ECG) using a combination of the structural image similarity metric and signal quality index. Then, we will detect premature ventricular contractions (PVC) on good-quality ECG beats using a deep-learning model. For arterial blood pressure (ABP) data, we will employ a singular value decomposition (SVD)-based approach to assess the signal quality. Finally, we will compute multiple indices of ANS functions through heart rate turbulence (HRT) analysis, time/frequency-domain analysis of heart rate variability (HRV) and pulse rate variability, and quantification of baroreflex sensitivity (BRS) from high-quality continuous ECG and ABP signals. The early autonomic dysfunction will be characterized by comparing the values of calculated indices with their normal ranges.

Conclusion

This study will provide a detailed characterization of acute changes in ANS function after sTBI through quantified indices from cardiac waveform data, thereby enhancing our understanding of the development and course of eAD post-sTBI.

Generalizability of SuperAlarm via Cross-Institutional Performance Evaluation

Bedside patient monitors are ubiquitous tools in modern critical care units to provide timely patient status. However, current systems suffer from high volume of false alarms leading to alarm fatigue, one of top technical hazards in clinical settings. Many studies are racing to develop improved algorithms towards precision patient monitoring, while little has been done to investigate the aspect of algorithm generalizability across different health institutions. Our group has been developing an evolving framework termed SuperAlarm that extracts multivariate patterns in data streams (monitor alarms, electronic health records and physiologic waveforms) of modern health enterprise to predict patient deterioration and has demonstrated great potential in mitigating alarm fatigue. In this study, we further investigate the generalizability of SuperAlarm by designing a comprehensive approach to achieve performance comparison in predicting in-hospital code blue (CB) events across two health institutions. SuperAlarm model trained with alarm data in one institution is tested on both internal and external test sets. Results show comparable performance with sensitivity up to 80% within one-hour window of events and over 90% in reduction of false alarms in both institutions. Cross-institutional performance agreement can be further improved by predicting a more stringent CB subtype (cardiopulmonary arrest), with internal sensitivity lying within 95% confident interval of external one up to 8-hour before event onset. The cross-institutional performance comparison offers first-hand knowledge on both advantages and challenges in generalizing a prediction algorithm across different institutions, which hold key information to guide the design of model training and deployment strategy.

A Supervised Approach to Robust Photoplethysmography Quality Assessment

Early detection of Atrial Fibrillation (AFib) is crucial to prevent stroke recurrence. New tools for monitoring cardiac rhythm are important for risk stratification and stroke prevention. As many of new approaches to long-term AFib detection are now based on photoplethysmogram (PPG) recordings from wearable devices, ensuring high PPG signal-to-noise ratios is a fundamental requirement for a robust detection of AFib episodes. Traditionally, signal quality assessment is often based on the evaluation of similarity between pulses to derive signal quality indices. There are limitations to using this approach for accurate assessment of PPG quality in the presence of arrhythmia, as in the case of AFib, mainly due to substantial changes in pulse morphology. In this paper, we first tested the performance of algorithms selected from a body of studies on PPG quality assessment using a dataset of PPG recordings from patients with AFib. We then propose machine learning approaches for PPG quality assessment in 30-s segments of PPG recording from 13 stroke patients admitted to the University of California San Francisco (UCSF) neuro intensive care unit and another dataset of 3764 patients from one of the five UCSF general intensive care units. We used data acquired from two systems, fingertip PPG (fPPG) from a bedside monitor system, and radial PPG (rPPG) measured using a wearable commercial wristband. We compared various supervised machine learning techniques including k-nearest neighbors, decisions trees, and a two-class support vector machine (SVM). SVM provided the best performance. fPPG signals were used to build the model and achieved 0.9477 accuracy when tested on the data from the fPPG exclusive to the test set, and 0.9589 accuracy when tested on the rPPG data.

A Simple Unsupervised, Real-time Clustering Method for Arterial Blood Pressure Signal Classification

Biomedical signal analysis often depends on methods to detect and distinguish abnormal or high noise/artifact signal from normal signal. A novel unsupervised clustering method suitable for resource constrained embedded computing contexts, classifies arterial blood pressure (ABP) beat cycles as normal or abnormal. A cycle detection algorithm delineates beat cycles, so that each cycle can be modeled by a continuous time Fourier series decomposition. The Fourier series parameters are a discrete vector representation for the cycle along with the cycle period. The sequence of cycle parameter vectors is a non-uniform discrete time representation for the ABP signal that provides feature input for a clustering algorithm. Clustering uses a weighted distance function of normalized cycle parameters to ignore cycle differences due to natural and expected physiological modulations, such as respiratory modulation, while accounting for differences due to other causes, such as patient movement artifact. Challenging cardiac surgery patient signal examples indicate effectiveness.

Deep learning approaches for plethysmography signal quality assessment in the presence of atrial fibrillation

OBJECTIVE

Photoplethysmography (PPG) monitoring has been implemented in many portable and wearable devices we use daily for health and fitness tracking. Its simplicity and cost-effectiveness has enabled a variety of biomedical applications, such as continuous long-term monitoring of heart arrhythmias, fitness, and sleep tracking, and hydration monitoring. One major issue that can hinder PPG-based applications is movement artifacts, which can lead to false interpretations. In many implementations, noisy PPG signals are discarded. Misinterpreted or discarded PPG signals pose a problem in applications where the goal is to increase the yield of detecting physiological events, such as in the case of paroxysmal atrial fibrillation (AF)-a common episodic heart arrhythmia and a leading risk factor for stroke. In this work, we compared a traditional machine learning and deep learning approaches for PPG quality assessment in the presence of AF, in order to find the most robust method for PPG quality assessment.

APPROACH

The training data set was composed of 78 278 30 s long PPG recordings from 3764 patients using bedside patient monitors. Two different representations of PPG signals were employed-a time-series based (1D) one and an image-based (2D) one. Trained models were tested on an independent set of 2683 30 s PPG signals from 13 stroke patients.

MAIN RESULTS

ResNet18 showed a higher performance (0.985 accuracy, 0.979 specificity, and 0.988 sensitivity) than SVM and other deep learning approaches. 2D-based models were generally more accurate than 1D-based models.

SIGNIFICANCE

2D representation of PPG signal enhances the accuracy of PPG signal quality assessment.

CHEMOMETRIC ANALYSIS OF DIFFUSE REFLECTANCE SPECTRAL DATA USING SINGULAR VALUE DECOMPOSITION FOR BLOOD GLUCOSE DETECTION

Diabetes mellitus is a metabolic disorder that affects the production or usage of insulin by the body. Diabetes prevails in the body as a long-term condition which causes several other disorders if left unnoticed. Proper control of Diabetes needs continuous monitoring. The current measurement technique is invasive in nature and requires the withdrawal of blood from the body. Periodic quantification of blood glucose leads to pain and discomfort for the subject. This paper presents a non-invasive glucose measuring system using near-infrared diffuse reflectance spectroscopy (DRS). This work attempts to determine the blood glucose value from the diffuse reflected spectra in the NIR region. The study is executed with the spectral signatures of 33 diabetic subjects collected non-invasively using diffuse reflectance spectrometer from a diabetic centre. Blood glucose level of the same subjects are also recorded using the clinical method. The spectral information is subjected to standard normal variate (SNV) preprocessing method to remove baseline drift and then dimension reduction using singular value decomposition (SVD) is applied to the preprocessed data. The extracted singular values when compared with the clinically measured blood glucose is found to have a proportional relationship. The proposed study using singular value decomposition paves us the way for estimating the blood glucose value non-invasively with the obtained set of clinical blood glucose and the corresponding singular value table as a standard reference set.

Errors, Omissions, and Outliers in Hourly Vital Signs Measurements in Intensive Care

OBJECTIVE

To empirically examine the prevalence of errors, omissions, and outliers in hourly vital signs recorded in the ICU.

DESIGN

Retrospective analysis of vital signs measurements from a large-scale clinical data warehouse (Multiparameter Intelligent Monitoring in Intensive Care III).

SETTING

Data were collected from the medical, surgical, cardiac, and cardiac surgery ICUs of a tertiary medical center in the United States.

PATIENTS

We analyzed data from approximately 48,000 ICU stays including approximately 28 million vital signs measurements.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

We used the vital sign day as our unit of measurement, defined as all the recordings from a single patient for a specific vital sign over a single 24-hour period. Approximately 30-40% of vital sign days included at least one gap of greater than 70 minutes between measurements. Between 3% and 10% of blood pressure measurements included logical inconsistencies. With the exception of pulse oximetry vital sign days, the readings in most vital sign days were normally distributed. We found that 15-38% of vital sign days contained at least one statistical outlier, of which 6-19% occurred simultaneously with outliers in other vital signs.

CONCLUSIONS

We found a significant number of missing, erroneous, and outlying vital signs measurements in a large ICU database. Our results provide empirical evidence of the nonrepresentativeness of hourly vital signs. Additional studies should focus on determining optimal sampling frequencies for recording vital signs in the ICU.

Use of the Kalman Filter for Aortic Pressure Waveform Noise Reduction

Clinical applications that require extraction and interpretation of physiological signals or waveforms are susceptible to corruption by noise or artifacts. Real-time hemodynamic monitoring systems are important for clinicians to assess the hemodynamic stability of surgical or intensive care patients by interpreting hemodynamic parameters generated by an analysis of aortic blood pressure (ABP) waveform measurements. Since hemodynamic parameter estimation algorithms often detect events and features from measured ABP waveforms to generate hemodynamic parameters, noise and artifacts integrated into ABP waveforms can severely distort the interpretation of hemodynamic parameters by hemodynamic algorithms. In this article, we propose the use of the Kalman filter and the 4-element Windkessel model with static parameters, arterial compliance C, peripheral resistance R, aortic impedance r, and the inertia of blood L, to represent aortic circulation for generating accurate estimations of ABP waveforms through noise and artifact reduction. Results show the Kalman filter could very effectively eliminate noise and generate a good estimation from the noisy ABP waveform based on the past state history. The power spectrum of the measured ABP waveform and the synthesized ABP waveform shows two similar harmonic frequencies.

A review of methods for the signal quality assessment to improve reliability of heart rate and blood pressures derived parameters

The assessment of signal quality has been a research topic since the late 1970s, as it is mainly related to the problem of false alarms in bedside monitors in the intensive care unit (ICU), the incidence of which can be as high as 90 %, leading to alarm fatigue and a drop in the overall level of nurses and clinicians attention. The development of efficient algorithms for the quality control of long diagnostic electrocardiographic (ECG) recordings, both single- and multi-lead, and of the arterial blood pressure (ABP) signal is therefore essential for the enhancement of care quality. The ECG signal is often corrupted by noise, which can be within the frequency band of interest and can manifest similar morphologies as the ECG itself. Similarly to ECG, also the ABP signal is often corrupted by non-Gaussian, nonlinear and non-stationary noise and artifacts, especially in ICU recordings. Moreover, the reliability of several important parameters derived from ABP such as systolic blood pressure or pulse pressure is strongly affected by the quality of the ABP waveform. In this work, several up-to-date algorithms for the quality scoring of a single- or multi-lead ECG recording, based on time-domain approaches, frequency-domain approaches or a combination of the two will be reviewed, as well as methods for the quality assessment of ABP. Additionally, algorithms exploiting the relationship between ECG and pulsatile signals, such as ABP and photoplethysmographic recordings, for the reduction in the false alarm rate will be presented. Finally, some considerations will be drawn taking into account the large heterogeneity of clinical settings, applications and goals that the reviewed algorithms have to deal with.

Toward Ubiquitous Blood Pressure Monitoring via Pulse Transit Time: Theory and Practice

Ubiquitous blood pressure (BP) monitoring is needed to improve hypertension detection and control and is becoming feasible due to recent technological advances such as in wearable sensing. Pulse transit time (PTT) represents a well-known potential approach for ubiquitous BP monitoring. The goal of this review is to facilitate the achievement of reliable ubiquitous BP monitoring via PTT. We explain the conventional BP measurement methods and their limitations; present models to summarize the theory of the PTT-BP relationship; outline the approach while pinpointing the key challenges; overview the previous work toward putting the theory to practice; make suggestions for best practice and future research; and discuss realistic expectations for the approach.

Automated detection of perturbed cardiac physiology during oral food allergen challenge in children

This paper investigates the fully automated computer-based detection of allergic reaction in oral food challenges using pediatric ECG signals. Nonallergic background is modeled using a mixture of Gaussians during oral food challenges, and the model likelihoods are used to determine whether a subject is allergic to a food type. The system performance is assessed on the dataset of 24 children (15 allergic and 9 nonallergic) totaling 34 h of data. The proposed detector correctly classified all nonallergic subjects (100% specificity) and 12 allergic subjects (80% sensitivity) and is capable of detecting allergy on average 17 min earlier than trained clinicians during oral food challenges, the gold standard of allergy diagnosis. Inclusion of the developed allergy classification platform during oral food challenges recorded would result in a 30% reduction of doses administered to allergic subjects. The results of study introduce the possibility to halt challenges earlier which can safely advance the state of clinical art of allergy diagnosis by reducing the overall exposure to the allergens.

Continuous detection of cerebral vasodilatation and vasoconstriction using intracranial pulse morphological template matching

Although accurate and continuous assessment of cerebral vasculature status is highly desirable for managing cerebral vascular diseases, no such method exists for current clinical practice. The present work introduces a novel method for real-time detection of cerebral vasodilatation and vasoconstriction using pulse morphological template matching. Templates consisting of morphological metrics of cerebral blood flow velocity (CBFV) pulse, measured at middle cerebral artery using Transcranial Doppler, are obtained by applying a morphological clustering and analysis of intracranial pulse algorithm to the data collected during induced vasodilatation and vasoconstriction in a controlled setting. These templates were then employed to define a vasodilatation index (VDI) and a vasoconstriction index (VCI) for any inquiry data segment as the percentage of the metrics demonstrating a trend consistent with those obtained from the training dataset. The validation of the proposed method on a dataset of CBFV signals of 27 healthy subjects, collected with a similar protocol as that of training dataset, during hyperventilation (and CO₂ rebreathing tests) shows a sensitivity of 92% (and 82%) for detection of vasodilatation (and vasoconstriction) and the specificity of 90% (and 92%), respectively. Moreover, the proposed method of detection of vasodilatation (vasoconstriction) is capable of rejecting all the cases associated with vasoconstriction (vasodilatation) and outperforms other two conventional techniques by at least 7% for vasodilatation and 19% for vasoconstriction.

Service oriented architecture to support real-time implementation of artifact detection in critical care monitoring

The quality of automated real-time critical care monitoring is impacted by the degree of signal artifact present in clinical data. This is further complicated when different clinical rules applied for disease detection require source data at different frequencies and different signal quality. This paper proposes a novel multidimensional framework based on service oriented architecture to support real-time implementation of clinical artifact detection in critical care settings. The framework is instantiated through a Neonatal Intensive Care case study which assesses signal quality of physiological data streams prior to detection of late-onset neonatal sepsis. In this case study requirements and provisions of artifact and clinical event detection are determined for real-time clinical implementation, which forms the second important contribution of this paper.

Lack of consistent intracranial pressure pulse morphological changes during episodes of microdialysis lactate/pyruvate ratio increase

Lactate/pyruvate ratio (LPR) from microdialysis is a well-established marker of cerebral metabolic crisis. For brain injury patients, abnormally high LPR could indicate cerebral ischemia or failure of O(2) uptake. However, there is a debate on the primary factor responsible for LPR increase. Exploiting the potential of using the morphology of a high temporal resolution signal such as intracranial pulse (ICP) to characterize cerebrovascular changes, a data analysis experiment is taken to test whether consistent changes in ICP pulse morphological metrics accompany the LPR increase. We studied 3517 h of LPR and continuous ICP data from 19 severe traumatic brain injury patients. Our morphological clustering and analysis of intracranial pressure (MOCAIP) algorithm was applied to ICP pulses, which were matched in time to the LPR measurements, and 128 pulse morphological metrics were extracted. We automatically identified the episodes of LPR increases using a moving time window of 10-20 h. We then studied the trending patterns of each of the 128 ICP MOCAIP metrics within these identified periods and determined them to be one of the following three types: increasing, decreasing or no trend. A binomial test was employed to investigate whether any MOCAIP metrics show a consistent trend among all episodes of LPR increase per patient. Regardless of the selected values for different parameters of the proposed method, for the majority of the subjects in the study (78%), none of the ICP metrics show any consistent trend during the episodes of LPR increase. Even for the few subjects who have at least one ICP metric with a consistent trend during the LPR increase episodes, the number of such metrics is small and varies from subject to subject. Given the fact that ICP pulse morphology is influenced by the cerebral vasculature, our results suggest that a dominant cerebral vascular cause may be behind the changes in LPR when LPR trends correlate with ICP pulse morphological changes. However, the incidence of such correlation seems to be low.

Consistent changes in intracranial pressure waveform morphology induced by acute hypercapnic cerebral vasodilatation

BACKGROUND

Intracranial pressure (ICP) remains a pivotal physiological signal for managing brain injury and subarachnoid hemorrhage (SAH) patients in neurocritical care units. Given the vascular origin of the ICP, changes in ICP waveform morphology could be used to infer cerebrovascular changes. Clinical validation of this association in the setting of brain trauma, and SAH is challenging due to the multi-factorial influences on, and uncertainty of, the state of the cerebral vasculature.

METHODS

To gain a more controlled setting, in this articel, we study ICP signals recorded in four uninjured patients undergoing a CO2 inhalation challenge in which hypercapnia induced acute cerebral vasodilatation. We apply our morphological clustering and analysis of intracranial pressure (MOCAIP) algorithm to identify six landmarks on individual ICP pulses (based on the three established ICP sub-peaks; P1, P2, and P3) and extract 128 ICP morphological metrics. Then by comparing baseline, test, and post-test data, we assess the consistency and rate of change for each individual metric.

RESULTS

Acute vasodilatation causes consistent changes in a total of 72 ICP pulse morphological metrics and the P2 sub-region responds to cerebral vascular changes in the most consistent way with the greatest change as compared to P1 and P3 sub-regions.

CONCLUSIONS

Since the dilation/constriction of the cerebral vasculature resulted in detectable consistent changes in ICP MOCIAP metrics, by an extended monitoring practice of ICP that includes characterizing ICP pulse morphology, one can potentially detect cerebrovascular changes, continuously, for patients under neurocritical care.

A subspace decomposition approach toward recognizing valid pulsatile signals

Following recent studies, the automatic analysis of intracranial pressure (ICP) pulses appears to be a promising tool for the prediction of critical intracranial and cerebrovascular pathophysiological variations during the management of many neurological disorders. A pulse analysis framework has been recently developed to automatically extract morphological features of ICP pulses. The algorithm is capable of enhancing the quality of ICP signals, recognizing valid (not contaminated with noise or artifacts) ICP pulses and designating the locations of the three ICP sub-peaks in a pulse. This paper extends the algorithm by proposing a singular value decomposition (SVD) technique to replace the correlation-based approach originally utilized in recognizing valid ICP pulses. The validation of the proposed method is conducted on a large database of ICP signals built from 700 h of recordings from 67 neurosurgical patients. A comparative analysis of the valid ICP recognition using the proposed SVD technique and the correlation-based method demonstrates a significant improvement in terms of (1) accuracy (61.96% reduction in the false positive rate while keeping the true positive rate as high as 99.08%) and (2) computational time (91.14% less time consumption), all in favor of the proposed method. Finally, this SVD-based valid pulse recognition can be potentially applied to process pulsatile signals other than ICP because no proprietary ICP features are incorporated in the algorithm.

A comparison of vital signs charted by nurses with automated acquired values using waveform quality indices

OBJECTIVE

(1) To investigate if there exist any discrepancies between the values of vital signs charted by nurses and those recorded by bedside monitors for a group of patients admitted for neurocritical care. (2) To investigate possible interpretations of discrepancies by exploring information in the alarm messages and the raw waveform data from monitors.

METHODS

Each charted vital sign value was paired with a corresponding value from data collected by an archival program of bedside monitors such that the automatically archived data preceded the charted data and had minimal time lag to the charted value. Next, the absolute differences between the paired values were taken as the discrepancy between charted and automatically-collected data. Archived alarm messages were searched for technical alarms of sensor/lead failure types. Additionally, 7-min waveform data around the place of large discrepancy were analyzed using signal abnormality indices (SAI) for quantifying the quality of recorded signals.

RESULTS

About 31,145 pairs of systolic blood pressure (BP-S) and 67,097 pairs of SpO(2) were investigated. Seven and a half percent of systolic blood pressure pairs had a discrepancy greater than 20 mmHg and less than one percent of the SpO2 pairs had a discrepancy greater than 10. We could not find any technical alarms from the monitors that could explain the large difference. However, SAI calculated for the waveforms associated with this group of cases was significantly larger than the SAI values calculated for the control waveform data of the same patients with small discrepancies.

CONCLUSION

Charted vital signs reflect in large the raw data as reported by bedside monitors. Poor signal quality could partially explain the existence of cases of large discrepancies.