Physiological Measurement

Heart rate variability (HRV) is a non-invasive biomarker of autonomic nervous system activity, commonly analyzed using a Poincare plot. This plot visualizes correlations between successive heartbeats (RRi vs. RRi+1) and quantifies autonomic regulation through SD1 and SD2 parameters. We introduce a second-order Poincare plot, a natural extension that clarifies serial dependencies by plotting successive differences in RR intervals ({Delta}RRi vs. {Delta}RRi+1). Applied to a PhysioNet dataset of 20 healthy individuals, this technique filtered out the slow HRV baseline of traditional elliptical plots to reveal distinct higher-order dynamics. These included ring-shaped structures indicating cardiorespiratory synchronization. A coupled-oscillator model, developed to simulate respiratory modulation, confirmed that these patterns are dictated by the respiratory frequency to heart rate ratio: slower breathing produces positive serial correlations in {Delta}RR, while faster breathing induces negative ones. By visualizing serial dependencies that conventional HRV metrics miss, the second-order Poincare plot extends the classical analysis framework. This tool provides a refined method for uncovering subtle dynamical features in HRV across diverse physiological and clinical states. HighlightsO_LISecond-order Poincare plots, plotting successive differences of RR intervals ({Delta}RRi vs. {Delta}RRi+1), extend traditional Poincare analysis to reveal rapid HRV dynamics. C_LIO_LIIn a dataset of 20 healthy individuals, second-order plots filtered out slow HRV components, highlighting respiratory modulation. C_LIO_LIRing-shaped patterns in some participants indicated strong cardiorespiratory coupling, while others showed positive or negative serial correlations linked to breathing rate. C_LIO_LIA coupled-oscillator model confirmed that the ratio of respiratory to heart rate frequency determines serial correlation patterns. C_LIO_LIThis method offers a novel tool for analyzing HRV dynamics, with potential applications in physiological and clinical research. C_LI

Pulse arrival time (PAT) is known to be correlated with blood pressure. Although PAT can be measured using electrocardiography (ECG), photoplethysmography (PPG), and other signals commonly available in clinical settings, recent literature has noted that devices recording these waveforms are often subject to many hardware-specific factors related to digital filtering, clock synchronization, temporal resolution, and latency. These factors can introduce relative timing errors between the ECG and PPG signals, resulting in a situation where traditional approaches for PAT measurement will not work as intended. In this work, we propose a methodology that accounts for these confounding factors and generates precise measurements of PAT using standard bedside monitoring equipment. This technique involves using heart rate variability to match heartbeats across waveforms and experimentally profiling the timing systems of bedside medical devices to correct various timing-related artifacts. To improve the precision of the resulting PAT measurements, we model temporal uncertainties stemming from the finite temporal resolution of the waveform samples. We apply this approach to a dataset with roughly 1.6 million hours of continuous ECG and PPG data from over 10,000 unique patients at a pediatric intensive care unit (ICU). After demonstrating that the observed timing artifacts are consistent across the entire dataset, we show that accounting for them results in more reasonable distributions of PAT measurements across age groups. It is our hope that this work will spur discussion around the standardization of PAT measurement using routinely collected signals in a clinical environment.

PurposeWhile bowel sound auscultation represents a key component of abdominal examination, its utility is limited because bowel sounds (BS) are intermittent, variable, and influenced by factors such as diet and digestive state. This renders it challenging to use them for a quantitative assessment of gastrointestinal health. MethodsBS signals were recorded from 84 subjects (39 patients and 45 healthy controls) using an acoustic SonicGuard sensor and categorized into four patterns. Metadata on physiological parameters were collected to examine their influence on BS characteristics and the differences between healthy and patient BS patterns. ResultsBowel sound patterns are significantly influenced by meal timing, caffeine consumption, and medication intake. Significant differences between healthy and patient groups were also observed in sound count, duration, energy, and waveform shape. These differences were mirrored in the performance of machine learning models finetuned for BS patterns classification, with performance depending on the group used for training and evaluation. ConclusionBS patterns present a promising quantitative indicators of gas-trointestinal health when analyzed alongside relevant physiological parameters.

Background and ObjectiveHeart rate variability (HRV) is a promising clinical marker of health and disease. Although HRV methodology is relatively straightforward, accurate detection of R-peaks remains a significant methodological challenge; this is especially true for single-lead EKG signals, which are routinely collected alongside EEG monitoring and for which few software options exist. Most developed algorithms with favorable R-peak detection profiles require significant mathematical and computational proficiency for implementation, providing a significant barrier for clinical research. Our objective was to address these challenges by developing a simple, free, and open-source software package for HRV analysis of single-lead EKG signals. MethodsCardioPy was developed in python and optimized for short-term (5-minute) single-lead EKG recordings. CardioPys R-peak detection trades full automation and algorithmic complexity for an adaptive thresholding mechanism, manual artifact removal and parameter adjustment. Standard time and frequency domain analyses are included, such that CardioPy may be used as a stand-alone HRV analysis package. An example use-case of HRV across wakefulness and sleep is presented and results validated against the widely used Kubios HRV software. ResultsHRV analyses were conducted in 66 EKG segments collected from five healthy individuals. Parameter optimization was conducted or each segment, requiring ~1-3 minutes of manual inspection time. With optimization, CardioPys R-peak detection algorithm achieved a mean sensitivity of 100.0% (SD 0.05%) and positive predictive value of 99.8% (SD 0.20%). HRV results closely matched those produced by Kubios HRV, both by eye and by quantitative comparison; CardioPy power spectra explained an average of 99.7% (SD 0.50%) of the variance present in Kubios spectra. HRV analyses showed significant group differences between brain states; SDNN, low frequency power, and low frequency-to-high frequency ratio were reduced in slow wave sleep compared to wakefulness. ConclusionsCardioPy provides an accessible and transparent tool for HRV analyses. Manual parameter optimization and artifact removal allow granular control over data quality and a highly reproducible analytic pipeline, despite additional time requirements. Future versions are slated to include automatic parameter optimization and a graphical user interface, further reducing analysis time and improving accessibility.

The electrocardiogram (ECG) and impedance cardiography (ICG) are typically combined to estimate electromechanical features such as the pre-ejection period (PEP) and left ventricular ejection time (LVET); indicators of changes in the cardiac specific drive of the autonomic nervous system (ANS). Current methods of ICG are time intensive in subject preparation and the measurements are vulnerable to non-reproducible subject-specific electrode configuration. Furthermore, analysis of impedance waveforms can be time consuming and labeling of key time points can suffer from experimenter bias. Here we present a wearable heart monitor that includes ECG, but replaces the commonly used 8 ICG electrodes with a single accelerometer (ACC) placed at the suprasternal notch. The ACC indirectly measures movement of the arterial pulse wave as blood is ejected into the aorta and great vessels. The resulting ACC waveform is processed into two smooth and readily identified waves, corresponding to the timing of the opening and closing of the aortic valve. We tested the ACCs utility and reliability for tracking cardiac ANS tone by comparing PEP and LVET measurements obtained simultaneously with conventional ICG and the ACC. Participants were recorded in the sitting and supine position with ECG, ICG, and ACC. While seated, they engaged in a classic physical stress task known to modulate ANS activity. There were obvious and significant associations between ICG and ACC estimates of PEP and LVET derivatives with respect to time. These findings support ACC as a complementary method for tracking ANS that is robust, time efficient, and readily accessible to researchers.

BackgroundElectronic Health Records(EHR) are very crucial for Clinical Decision Support Systems and for proper care to be delivered to ICU heart failure patients, there is often missing data due to monitoring device errors thus the need for robust imputation methodologies. ObjectiveTo compare and evaluate three different methodologies for imputing missing data for heart failure patients from the MIMIC-III database: Denoising Autoencoder (DAE), Self-Attention Imputation for Time Series (SAITS), and Multiple Imputation by Chained Equations (MICE) with LightGBM. MethodsAnalysis of 14,090 ICU admissions for patients with heart failure was performed using data from the MIMIC-III database. Features were selected based off of clinical relevance, and 19 clinical features were selected through a combination of Random Forest analysis, correlation analysis, and Mutual Information. The introduction of artificial missing values of 20%, 30%, and 50% was applied to the data set, and then 3 imputation methodologies were evaluated with the DAE, SAITS, and MICE+LightGBM. The performance of each imputation methodology was evaluated using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Normalized Root Mean Square Error (NRMSE). ResultsBoth DAE and SAITS had superior performance on the imputation of missing values across all percentages of missing values. At 20% missingness, DAE had mean MAE = 0.004967, RMSE = 0.005217, and NRMSE = 3.260893 while SAITS had mean MAE = 0.005461, RMSE = 0.005797, and NRMSE = 3.244695; thus MICE+LightGBM resulted in a higher number of errors. At 50% missingness, the SAITS methodology demonstrated the best performance followed by DAE and MICE+LightGBM methods demonstrated decreased performance. The deep learning methodologies maintained a consistent level of accuracy between the clinical variables measured. ConclusionsOur analysis indicates that deep learning-based imputation methodologies significantly outperform traditional methodologies for imputing missing values in ICU heart failure data thus supporting the implementation of these methodologies into Clinical Decision Support Systems for heart failure patients.

The analysis of electrophysiological signals from the human body has become increasingly crucial, especially given the widespread adoption of wearable technologies and the growing trend of remote and online monitoring. In situations where demographic patient data is unavailable, the evaluation of such information from electrophysiological signals becomes imperative for making well-informed diagnostic and therapeutic decisions, particularly in ambulatory and urgent cases. This study underscores the significance of this necessity by utilizing intracardiac electrograms to predict patient weight. Intracardiac electrograms were recorded from 44 patients (14 female, with an average age of 59.2{+/-}11.5 years) using a 64-pole basket catheter over a duration of 60 seconds. A dataset comprising 2,816 unipolar electrogram signal segments, each lasting 4 seconds, was utilized. Weight, considered as a continuous variable, underwent discretization into k bins with uniformly distributed widths, where various values of k were experimented with. As the value of k increases, class imbalance also increases. The state-of-the-art time series classification algorithm, Minirocket, was employed alongside the popular machine learning algorithm eXtreme Gradient Boosting (XGBoost). Minirocket consistently demonstrates superior performance compared to XGBoost across all class number scenarios and across all evaluation metrics, such as accuracy, F1 score, and Area Under the Curve (AUC) values, achieving scores of approximately 0.96. Conversely, XGBoost shows signs of overfitting, particularly noticeable in scenarios with higher class imbalance. Tuning probability thresholds for classes could potentially mitigate this issue. Additionally, XGBoosts performance improves with reduced bin numbers, emphasizing the importance of balanced classes. This study provides novel insights into the predictive capabilities of these algorithms and their implications for personalized medicine and remote health monitoring.

BackgroundDespite continuous progress in medical treatment, heart failure (HF) is the leading cause of hospitalizations with a high all-cause mortality in patients. Patients with a left ventricular ejection fraction (LVEF) below 50% are characterized by the highest risk of cardiovascular complications. The objective of this study was to examine how LVEF below 50% and aging impact cardiovascular physiology. MethodsSixteen males with physician diagnosed coronary artery disease and LVEF = 42 {+/-} 6% (age 62 {+/-} 6 years, BMI 29.1 {+/-} 3.8kg/m2) and 10 healthy controls (9 male and 1 female, age 28.5 {+/-} 9.1 years, BMI = 24.1 {+/-} 1.2kg/m2) were recruited in our study. Finger photoplethysmography for blood pressure (BP) and electrocardiogram (ECG) were recorded while participants rested in a supine position. Wavelet transformations were used to analyze the amplitudes, phase coherence and phase difference of BP and ECG. The frequency intervals were separated as follows: I (0.6-2Hz), II (0.145- 0.6Hz), III (0.052-0.145Hz), and IV (0.021-0.052Hz). ResultsHF patients showed a decrease (p<0.05) in BP wavelet amplitude intervals III and IV in comparison to controls, and interval I for ECG. A decrease in phase coherence (p<0.01) at interval I is also found in HF patients compared to controls. ConclusionsA decrease in smooth muscle cell activity and smooth muscle autonomic innervation (intervals III and IV) contributions to BP, along with a decrease in cardiac activity as shown by the wavelet amplitude in ECG, suggests altered BP and ECG function in aging HF patients. Furthermore, a decrease in the cardiac interval represents an impairment in the BP and ECG relationship in HF patients. The wavelet transform has the potential to expand our understanding of LVEF and improve diagnostic procedures and patient prognosis.

BackgroundHeart rate variability (HRV) is a common measure of autonomic and cardiovascular system function assessed via electrocardiography (ECG). Consumer wearables, commonly employed in epidemiological research, use photoplethysmography (PPG) to report HRV metrics (PRV), although these may not be equivalent. One potential cause of dissociation between HRV and PRV is the variability in pulse transit time (PTT). This study sought to determine if PPG-derived HRV (i.e., PRV) is equivalent to ECG-derived HRV and ascertain if PRV measurement error is sufficient for a biomarker separate from HRV. MethodsThe ECG data from 1,084 subjects were obtained from the PhysioNet Autonomic Aging dataset, and individual PTT variances for both the wrist (n=42) and finger (n=49) were derived from Mol et al. A Bayesian simulation was constructed whereby the individual arrival times of the PPG wave were calculated by placing a Gaussian prior on the individual QRS-wave timings of each ECG series. The standard deviation of the prior corresponds to the PTT variances. This was simulated 10,000 times for each PTT variance. The root mean square of successive differences (RMSSD) and standard deviation of N-N intervals (SDNN) were calculated for both HRV and PRV. The Region of Practical Equivalence bounds (ROPE) were set a priori at {+/-}0.2% of true HRV. The Highest Density Interval (HDI) width, encompassing 95% of the posterior distribution, was calculated for each PTT variance. ResultsThe lowest PTT variance (2.0 SD) corresponded to 88.4% within ROPE for SDNN and 21.4% for RMSSD. As the SD of PTT increases, the equivalence of PRV and HRV decreases for both SDNN and RMSSD. Thus, between PRV and HRV, RMSSD is nearly never equivalent and SDNN is only somewhat equivalent under very strict circumstances. The HDI interval width increases with increasing PTT variance, with the HDI width increasing at a higher rate for RMSSD than SDNN. ConclusionsFor individuals with greater PTT variability, PRV is not a surrogate for HRV. When considering PRV as a unique biometric measure, our findings reveal that SDNN has more favorable measurement properties than RMSSD, though both exhibit a non-uniform measurement error.

BackgroundNon-code dose boluses of epinephrine are utilized in critically ill pediatric patients during periods of hemodynamic deterioration, often with the hopes of preventing a cardiac arrest. Data regarding the physiologic effects of these administrations are limited. The primary aim of this study was to use high fidelity physiologic data to characterize the effects of intravenous non-code dose bolus epinephrine. MethodsPediatric patients in the cardiac intensive care unit who received non-code dose bolus epinephrine were identified. Those who received fluid boluses or chest compressions within 2 minutes of bolus epinephrine were excluded. ARIMAX analyses were conducted to characterize the time-dependent changes in hemodynamic indices. Cluster analyses were then conducted to determine patterns in hemodynamic changes associated with bolus epinephrine. ResultsA total of 71 non-code dose bolus epinephrine administrations were included in the final analyses. Heart rate, blood pressure, and renal near infrared spectroscopy all demonstrated statistically significant changes after bolus epinephrine administration. Peak change in each was 40%, 52%, and 9%, respectively, with peaks occurring between 60-seconds and 120-seconds after administration. Three response-based clusters were identified. ConclusionNon-code dose bolus epinephrine is associated with a significant increase in heart rate, blood pressure, and systemic oxygen delivery. Cluster analysis using the peak change identified distinct clinical clusters.

Cardiovascular Diseases (CVDs) are the leading cause of mortality worldwide, necessitating early and accurate diagnosis to prevent severe outcomes such as Heart Failure (HF). Despite the widespread use of Electrocardiogram (ECG) for cardiac monitoring, traditional methods often miss subtle preclinical changes. In this paper, we present an automated digital biomarker discovery pipeline that leverages explainable artificial intelligence (XAI) to enhance the interpretability and clinical applicability of ECG-based biomarkers for CVDs. Using an inter-pretable feature extractor combined with unsupervised clustering and Particle Swarm Optimisation (PSO), our method identifies both known and novel ECG features associated with high CVD risk. These include established markers like RR Interval Sample Entropy and the discovery of novel biomarkers such as T-Wave Multiscale Entropy, which we found to be significantly associated with CVD risk. Our pipeline enhances early detection by bridging Artificial Intelligence (AI) methods with clinical relevance, providing interpretable insights that align with physiological principles. This transparency promotes clinician trust and supports the integration of AI into routine medical practice. Our results demonstrate that this approach can significantly improve the prediction and understanding of heart diseases, thus offering a powerful tool for reducing the global burden of CVDs.

Beat detection is a key step in the analysis of photo-plethysmogram (PPG) signals. The MSPTD algorithm was recently identified as one of the most accurate beat detection algorithms, but its current open-source implementation is substantially more computationally expensive than other leading algorithms such as qppgfast. The aim of this work was to develop a more efficient, open-source implementation of the MSPTD algorithm. Five potential improvements were identified to increase efficiency. Each potential improvement was evaluated in turn, and an optimal algorithm configuration named MSPTDfast was developed which incorporated all of the improvements found to reduce algorithm execution time whilst not substantially reducing the accuracy of beat detection. Performance was assessed using data collected from young adults during a lunchbreak in the PPG-DaLiA dataset. The data consisted of wrist PPG signals acquired using an Empatica E4 device, alongside simultaneous ECG signals from which reference heartbeat timings were obtained. MSPTDfast was found to be substantially more efficient than MSPTD (a reduction in execution time of 72.3%), with minimal difference in beat detection accuracy (F1-score 87.8% vs. 87.7%). In addition, the performance of MSPTDfast was much closer to that of the state-of-the-art qppgfast algorithm than the MSPTD algorithm, with a comparable F1-score (87.4% vs. 87.7%), and an execution time which was only 19.2% longer than that of qppgfast (vs. 330.8% longer for MSPTD). In conclusion, MSPTD-fast is an efficient and accurate open-source PPG beat detection algorithm with a substantially faster execution time than MSPTD. It is available under the permissive MIT licence.

We examine the long-range temporal structure of forecasts produced by Time-Series Foundation Models (TSFMs) on heartbeat dynamics using the MIT-BIH Normal Sinus Rhythm Database (NSRDB). Our findings indicate that these models do not adequately capture long-range dependencies, as reflected in growing errors in RR-interval predictions over longer forecast horizons. Code is available at https://github.com/SubramanianLab/ecg-tsfm-benchmark.

Purposeto describe HRV metrics in Covid-19 patients at admission in the ICU and its relationship with mortality and invasive mechanical ventilation (IMV). Heart rate variability (HRV) in sitting position in critically ill patients has not been explored. Methodswe conducted a prospective single-centre observational study. Adult patients admitted in the ICU with respiratory failure due to RT-PCR-confirmed SARS-CoV-2 but not under IMV were included. Electrocardiogram was recorded at least for 15 minutes at 500 Hz during a stable sitting condition. Power spectrum was obtained using wavelets. Very low frequency (VLF), low frequency (LF) and high frequency (HF) powerbands were calculated and then normalized to total power (VLFn%, LFn% and HFn%). We also analyzed non-linear HRV dynamics. Results27 patients were included. LFn% was lower in non-survivors (4.5 vs 8 %, p=0.015) and related to 28-day mortality (OR 0.61; 95% CI: 0.32 to 0.9, P=0.05). In a robust generalized Gamma linear model, we found that detrended fluctuation analysis alpha 2 (DFA2) (RR 0.092; bootstrapped 95% CI: 0.031-0.361, P=0.003), admission APACHE II score (RR 1.084; bootstrapped 95% CI: 1.046 to 1.123, P=0.002) and SAF index (RR 0.985; bootstrapped 95% CI: 0.982 - 0.990, P<0.001) were associated with longer ICU LOS. Conclusionsdiminished normalized LF power was associated with 28-day mortality in univariate analysis among critically ill COVID-19 patients on spontaneous ventilation, potentially reflecting an altered autonomic response in early severe COVID-19. Normalized and absolute VLF power should be considered when analyzing HRV in ICU patients. DFA2 was the HRV variable with the strongest association with ICU LOS. These exploratory results could be helpful to design newer tools for early prognostication in COVID-19 patients.

Electrocardiogram (ECG) provides a non-invasive method for identifying cardiac issues, particularly arrhythmias or irregular heartbeats. In recent years, the fields of artificial intelligence and machine learning have made significant inroads into various healthcare applications, including the development of arrhythmia classifiers using deep learning techniques. However, a persistent challenge in this domain is the limited availability of large, well-annotated ECG datasets, which are crucial for building and evaluating robust machine learning models. To address this limitation, we propose a novel deep transfer learning framework designed to perform effectively on small training datasets. Our approach involves fine-tuning ResNet-18, a general-purpose image classifier, using the MIT-BIH arrhythmia dataset. This method aims to leverage the power of transfer learning to overcome the constraints of limited data availability. Furthermore, this paper conducts a critical examination of existing deep learning models in the field of ECG analysis. Our investigation reveals that many of these models suffer from methodological flaws, particularly in terms of data leakage. This issue potentially leads to overly optimistic performance estimates and raises concerns about the reliability and generalizability of these models in real-world clinical applications. By addressing these challenges, our work contributes to the advancement of more robust and reliable ECG analysis techniques, potentially improving the accuracy and applicability of automated arrhythmia detection in clinical settings.

BackgroundBurnout is a significant concern among healthcare professionals, particularly in high-stress environments such as intensive care units (ICUs). While prior research has linked burnout to self-reported stress and psychological distress, objective physiological markers like heart rate variability (HRV) may offer a more reliable assessment of occupational stress and burnout risk. Our previous pilot study suggested an association between HRV and stress; however, it did not incorporate standardized burnout assessments. This study aims to bridge that gap by examining the relationship between HRV, self-reported stress, and validated burnout scales. Additionally, it seeks to identify key predictors of burnout and develop a predictive model for early risk detection. MethodsThis cross-sectional observational study included 57 nurses and nursing assistants working in ICUs and general hospital wards. Participants completed validated burnout assessments, including the Cuestionario para la Evaluacion del Sindrome de Quemarse por el Trabajo (CESQT; Spanish Burnout Inventory), the Maslach Burnout Inventory (MBI), the Professional Quality of Life Scale (ProQOL), and the State-Trait Anxiety Inventory (STAI). HRV parameters were recorded using a Biosignals Plux system for 10 minutes at rest before the start of the work shift and analyzed with the OpenSignals software. Extracted HRV metrics included the root mean square of successive differences (rMSSD), low-frequency to high-frequency ratio (LF/HF), Standard Deviation 1 and 2 Ratio (SD1/SD2 ratio), and Poincare area. Statistical analyses involved descriptive statistics, correlation analysis, and group comparisons to examine differences in burnout across workplace conditions, shift types, and shift durations. A logistic regression model with 10-fold cross-validation was developed to predict burnout risk, integrating HRV parameters, psychological distress, and occupational factors. ResultsHRV parameters were significantly associated with self-reported stress and burnout indicators, reinforcing their potential role as objective biomarkers of occupational stress. Night shift workers and those with extended work hours exhibited higher burnout levels and greater autonomic dysregulation. The predictive model demonstrated strong accuracy in identifying individuals at risk of burnout. The model integrating HRV parameters, psychological distress, and occupational factors (Model 2) achieved an AUC-ROC of 0.832 (95% CI: 0.735- 0.929) and an accuracy of 79.1%, outperforming the model based solely on demographic and psychometric data (Model 1, AUC-ROC = 0.791, 95% CI: 0.685-0.897, accuracy = 76.3%). HRV and psychological stress emerged as key contributing factors. ConclusionThese findings highlight HRV as a promising tool for the objective assessment of burnout risk in healthcare professionals. The predictive model developed provides a framework for early identification of high-risk individuals, enabling targeted interventions to improve well-being and staff retention in healthcare settings. Future research should validate these findings in larger cohorts and assess the long-term applicability of HRV-based monitoring systems in occupational health programs.

ObjectiveSkin sympathetic nerve activity (SKNA) has emerged as a promising non-invasive surrogate measure of sympathetic drive, but its relevant physiological characteristics remain ill-defined. This observational study aims to investigate its regulatory patterns during rest and Valsalva maneuver (VM) in healthy participants. MethodUsing a two-layer strategy integrating signal analysis and physiological modelling, we analyzed data recorded from 41 subjects performing repeated VMs. The observational layer includes time-domain feature comparisons using linear mixed-effect models, and time-varying spectral coherence analysis. The mechanistic layer proposes a mathematical model to investigate whether baroreflex and respiratory modulation are sufficient to reproduce the observed HR and average SKNA (aSKNA) dynamics. Main ResultsMean integrated SKNA (iSKNA) showed more significant change than HRV for VM induced effects. We also found mean iSKNA increase during VM varies with BMI and sex. The coherence analysis indicated that iSKNA strongly synchronized with EDR under resting conditions. The proposed model successfully reproduced main characteristics of aSKNA dynamics, yielding a high median Pearson correlation coefficient of 0.80 ([Q1, Q3] = [0.60, 0.91]). In contrast, HR dynamics were only partially captured, with a median PCC of 0.37 ([Q1, Q3] = [0.16, 0.55]). These results likely suggest SKNA provides a more direct representation of sympathetic burst dynamics during VM in healthy subjects. SignificanceThis study provides convergent evidence that SKNA reflects known autonomic regulatory influences in healthy subjects. These findings strengthen the physiological interpretability of SKNA while clarifying its appropriate use as a practical biomarker of sympathetic function.

Volitional motor activity is associated with a feedforward cardiorespiratory response to actual or impending movements. We have previously shown in the CRC study that the expectation of physical exercise causes a decrease in cardiorespiratory coherence that scales with the anticipated load. The present work uses a modeling approach to investigate the mechanisms that can cause a fall in cardiorespiratory coherence (CRC). We devised a Hodgkin-Huxley model of a cardiac pacemaker cell using the NEURON module. We simulated the effect of autonomic tone, sympathetic & respiratory-vagal modulation, and respiratory irregularity on pacemaker cell output by injecting efflux/influx current to model the parasympathetic/sympathetic effects, respectively. The vago-sympathetic tone was modeled by altering the direct current bias of the injected current and the respiratory-vagal effect by the periodic modulation of the injected current at a frequency of 0.2 Hz, corresponding to a respiratory rate of 12 breaths/min. Sympathetic modulation was simulated by injecting a low-frequency current close to Mayer wave frequency (0.08 Hz). We computed the coherence between the instantaneous pacemaker rate and respiratory-vagal modulation current as a model analog to experimental CRC. We found that sympathetic modulation, low vagal tone/high sympathetic tone, and respiratory irregularity can cause a decrease in CRC. We corroborated the model results with the actual data from the CRC study. In conclusion, we employ a novel approach combining insights from the experimental study and a physiologically plausible modeling framework to understand the mechanisms underlying the fall of cardiorespiratory coherence induced by the expectation of exercise. NEW & NOTEWORTHY Cardiorespiratory coherence is diminished in response to respiratory irregularity, low vagal/high sympathetic tone, and prominent low-frequency sympathetic modulation. Expectation of physical activity induces respiratory irregularity and increased sigh frequency and that contributes to diminished cardiorespiratory coherence in expectation of exercise. There is a greater fall of coherence with the non-linear (logistic) transformation of injected current, indicating the non-linear nature of cardiorespiratory interactions preceding the onset of exercise.

IntroductionIntracranial pressure (ICP) monitoring is commonly used in neuro-intensive care, but its utility may be limited by a suboptimal use. The brain pressure-volume relationship, a potential predictor of neurological health, is now approached using time-domain methods, which can be challenging to implement. Frequency-domain methods may offer an alternative, but their relationship with time-domain metrics remains unclear. This study compares time- and frequency-domain methods for assessing brain compliance and evaluates their real-time usability. MethodologyA monocentric, prospective observational study was conducted in the neurological ICU of the Hospices Civils de Lyon, France, to evaluate markers of brain compliance. Adult patients with brain lesions requiring multimodal monitoring were included. Continuous high-density physiological data, including ICP, arterial blood pressure (ABP), end-tidal CO2, and electrocardiogram (ECG), were collected for analysis. Some spontaneous ICP rise events were automatically detected based on heuristic criteria and used as brain compliance challenges to compare the co-evolution of metrics across multiple time windows. Time-domain (pulse shape-related metrics) and frequency-domain analyses (examining heart and respiratory components in ICP) were computed to assess the intracranial pressure-volume state. Statistical analyses were performed using linear mixed-effects modeling, adjusting for vasoactive and sedative medications and Spearman correlations. ResultsThe study included 66 patients, with a mean age of 49.16 [38.38, 57.58] years. A total of 518 spontaneous ICP rise events were detected in 56 patients. Our findings revealed that: 1) frequency-domain metrics strongly correlated with time-domain metrics during these challenges (r > 0.8, p < 0.001), 2) frequency-domain metrics were significantly drastically less computationally demanding, and 3) the impact of the heart on ICP showed a significant correlation with the P2/P1 ratio (r = 0.391, p < 0.001) and other potential markers of brain compliance. In contrast, the impact of respiration on ICP was only marginally correlated with these markers. ConclusionsFrequency-domain analyses exploring the impact of cardiac activity on ICP map provide similarly informative value to more complex machine learning-based tools, but with the advantage of being much less computationally demanding. This makes this approach particularly suitable and intuitive for real-time clinical monitoring in hospital settings, where computational resources are often limited. Authors statementInitials: VG, NE, GP, LB, FG, RC, CB, FC, AM, AR, RM, FD, TR, SG, BB O_LIConception and Design: VG, NE, BB C_LIO_LIData Collection : VG, NE, LB, GP, TR, FG, RC, CB, FC, FD, BB C_LIO_LIData Analysis and Interpretation: VG, BB C_LIO_LIMethodology Development: VG, BB, SG C_LIO_LIManuscript Drafting and Writing: VG, BB C_LIO_LISupervision and Project Oversight: BB C_LIO_LIFunding Acquisition: BB, VG C_LIO_LIVisualization: VG, BB C_LIO_LICritical Review and Editing: NE, GP, LB, FG, RC, CB, FC, AM, AR, RM, FD, TR, SG, BB C_LIO_LIApproval of the Final Version: All C_LIO_LISoftware Development: VG, SG C_LIO_LIEthical Approval and Regulatory Compliance: BB C_LI

RationaleAutonomic dysfunction is a hallmark of sepsis pathophysiology, yet its quantification remains challenging. Multiscale entropy (MSE) derived from heart rate variability (HRV) offers a dynamic measure of physiological complexity and may serve as a biomarker of early deterioration associated with subsequent organ failure, vasopressor escalation, or mortality. ObjectiveTo determine whether MSE computed across multiple temporal scales during the first 24 hours of Intensive Care Unit (ICU) admission is associated with short-term mortality and longer-term organ dysfunction in patients with sepsis, and whether these relationships vary across vasopressor exposure. Unlike prior studies that focused on short-term HRV metrics, we applied MSE across multiple temporal scales and incorporated these features into machine learning models to evaluate their prognostic utility in septic shock. MethodsThis retrospective cohort study included adult ICU sepsis patients at Emory University Hospital from January 2016 to December 2019. Of 2,076 eligible patients, 958 were propensity matched into two cohorts: fluids-only and fluids-plus-vasopressor, with norepinephrine as the primary vasopressor. High-resolution electrocardiogram (ECG) waveforms were analyzed to compute MSE across 20 temporal scales. Machine learning models using (1) MSE features alone and (2) MSE combined with demographic and vital sign data (MSE-DV) were compared against traditional HRV measures based model and severity of illness scores for predicting outcomes. Model performance was assessed using the area under the receiver operating characteristic curve (AUROC), with a primary outcome of mortality at day 7 and secondary outcome of persistent organ dysfunction at day 28. ResultsIn the fluids-plus-vasopressor cohort, MSE-based models demonstrated superior predictive performance for 7-day mortality (AUROC 0.84) compared to severity of illness scores (AUROC 0.64). MSE-DV models also predicted organ dysfunction including 28-day renal (AUROC 0.75), neurological (AUROC 0.79), and respiratory (AUROC 0.71) dysfunction. Patients receiving second-line and third-line vasopressors and corticosteroids exhibited progressively lower MSE values, particularly at mid-range and long-range scales. ConclusionMSE features in the first 24 hours of ICU stay predict mortality and organ dysfunction with higher discrimination than traditional severity of illness scores. Future work should validate these findings, assess longitudinal MSE trends, and race-specific autonomic patterns to refine predictive models.