Physiological measurement最新文献

Physics-informed neural networks (PINNs) represent a transformative approach to data models by incorporating known physical laws into neural network training, thereby improving model generalizability, reduce data dependency, and enhance interpretability. Like many other fields in engineering and science, the analysis of physiological signals has been influenced by PINNs in recent years. This manuscript provides a comprehensive overview of PINNs from various perspectives in the physiological signal analysis domain. After exploring the literature and screening the search results, more than 40 key studies in the related domain are selected and categorized based on both practically and theoretically significant perspectives, including input data types, applications, physics-informed models, and neural network architectures. While the advantages of PINNs in tackling forward and inverse problems in physiological signal contexts are highlighted, challenges such as noisy inputs, computational complexity, loss function types, and overall model configuration are discussed, providing insights into future research directions and improvements. This work can serve as a guiding resource for researchers exploring PINNs in biomedical and physiological signal processing, paving the way for more precise, data-efficient, and clinically relevant solutions.

Objective.The existing literature lacks a comprehensive analysis of the clinical evolution of septic patients, which is highly heterogeneous and patient-dependent. The aim of this study is to develop machine learning models capable of predicting the clinical evolution of septic patients and to evaluate the predictive ability of features.Approach. Data from intensive care unit septic patients were extracted from the freely available HiRID database and a comprehensive pipeline for time series analysis of critical care data was developed. Predictive models of cardiovascular deterioration (based on mean pressure and lactate values) and global organ dysfunction (based on SOFA score) were developed, and the addition of variability, such as entropies, cross-entropies and cross-correlation of heart rate and blood pressure (BP), was tested against the use of standard metrics alone.Main results.The best model achieved an area under the ROC curve value of 0.9671, with SOFA score values and trends being the most important features in the model, followed by features related to lactate, fluid balance, therapy and entropy values of BP.Significance.The results show that the dynamics of vital signs and their cross-coupling, as captured by the proposed variability indices, can provide additional insights into the physiological responses to the therapy administered.

Objective.This study evaluates the potential of photoplethysmography imaging (PPGI) with automated facial tracking for detecting hemodynamic and autonomic changes induced by lower-body negative pressure (LBNP). The goal is to assess whether PPGI-derived facial perfusion variations are related with stroke volume (SV), systemic vascular resistance (SVR), heart rate variability (HRV), and autonomic responses to progressive hypovolemia.Approach.Twenty-four healthy adults (8 females, 16 males; aged 28.7 ± 3.5 years) underwent a seven-stage LBNP protocol (-15 to -60 mmHg, recovery). Facial perfusion was recorded using cross-polarized PPGI, along with SV, SVR, HR, and mean arterial pressure. Facial landmark tracking (MediaPipe) was used to extract region-specific PPGI signals. Wavelet synchrosqueezing transform enabled spectral analysis, and HRV was assessed with NeuroKit2.Main Results.At -60 mmHg, the LBNP-intolerant group showed a 25.2% decrease in SV (p< 0.0001) and a 19% increase in SVR (p= 0.041). At -30 mmHg recovery, SV remained reduced by 21% (p< 0.001), with SVR elevated by 30.1% (p= 0.002). In contrast, the tolerant group exhibited SV increases of 12% and 18% at these stages (bothp< 0.0001), and a HR reduction of up to 5% (p< 0.05), with a decreasing SVR trend. HRV analysis indicated greater sympathetic activation in the intolerant group, with reduced HF power (p= 0.037) and increased LF/HF ratio (3.5 at -60 mmHg,p= 0.020). First harmonic PPGI amplitudes significantly declined in the intolerant group, most notably in the cheeks (-44.2%,p= 0.005).Significance.These findings suggests that PPGI, combined with AI-based face tracking and wavelet analysis, enables non-invasive, spatially resolved monitoring of vascular and autonomic responses. PPGI differentiates tolerant and intolerant groups, supporting its potential for real-time cardiovascular assessment in critical care and emergency settings.

Electrical impedance tomography (EIT) is an emerging imaging technology that has garnered increasing attention in recent years, particularly in the medical field and the diagnosis and treatment of respiratory diseases. Fascinating developments were achieved after the previous review focusing on clinical applications in Chinese hospitals. Over hundred publications in SCI journals related to thoracic EIT clinical research and daily applications have been recorded in the past five years. As EIT devices become more accessible and portable, clinical application scenarios include not only ICU, but also chronic disease management, and health screening. We were excited to welcome more than 10 local companies manufacturing their own EIT devices, which were exhibited during the 24th International Conference on Biomedical Applications of EIT in Hangzhou, China. This article systematically reviewed the applications of thoracic EIT in clinical research and routine use in Chinese hospitals over the past five years.

Objective.The mean linear intercept (MLI) is often used in lung morphometry; however, its assessment is labor-intensive, time-consuming, and prone to systematic biases when using the conventional indirect method. This study examines the inherent systematic biases in the indirect method, and explores the differences between the two methods, including how methodological parameters, such as the number of accepted field-of-view (FOV) images and guideline length, affect the measurement.Approach.We developed an automated MLI measurement system that uses a multiresolution semantic segmentation model. The system enables both indirect and direct MLI methods and allows for controlled variation of measurement parameters. The number of accepted FOVs was varied from 10 to 1000, and the guideline length from 39 to 702 pixels (19.4-349.5µm).Main results.The indirect method consistently overestimated MLI due to Septa Bias and Partial Chord Bias. The standard error of MLI decreases with more accepted FOV images, and the direct method consistently yielded a lower standard error than the indirect method. Short guideline lengths (<135.9µm) have a large impact on the indirect method, whereas the direct method is relatively insensitive to this parameter.Significance.The automated MLI measurement system improves the efficiency over human raters and enables higher precision by leveraging the advantages of the direct method (e.g. lower standard error, low sensitivity to guideline length) and the analysis of a larger number of FOV images. Moreover, the segmentation model used in the system is demonstrated to be accurate, which can facilitate the development of advanced morphometry techniques.

Objective. There is growing interest in the use of physiological signals beyond electrocardiography (ECG), particularly photoplethysmography-based noninvasive arterial blood pressure (nABP), to assess autonomic nervous system (ANS) activity with minimal recording durations. This study compared heart rate variability (HRV) and pulse rate variability (PRV) derived from ECG and nABP, respectively. We investigated how signal shortening and calculation window size affect time-domain, frequency-domain, and nonlinear ANS metrics.Approach. Photoplethysmography was used to measure nABP, whereas ECG was recorded with a 3-lead device in healthy individuals (18-31 years). The HRV and PRV were analyzed using time- and frequency-domain metrics, and nonlinear indices, including entropy metrics and Poincaré plots (SD1, SD2). Agreement between signal lengths of 3 min and 5 min was assessed in 86 nABP and 70 ECG participants using intraclass correlation coefficients (ICCs). To evaluate the effect of window size, 15 min recordings from 16 participants were segmented into windows of 3 min, 5 min, and 15 min. HRV-PRV agreement was evaluated using Bland-Altman analysis.Main results. The time-domain metrics demonstrated excellent reproducibility when the signal length (ICCs ⩾ 0.96) and window size (ICCs ⩾ 0.98) were shortened, but moderate agreement between HRV and PRV. Entropy metrics were most affected by signal shortening (e.g. HRV multiscale entropy ICC (95%CI]): 0.67 (0.47-0.80); PRV approximate entropy: 0.45 (0.15-0.64)). Shorter window sizes affected selected ANS metrics, including reduced SD2 (p= 0.003 for HRV,p< 0.001 for PRV) and increased frequency-domain values (p< 0.001 for HRV and PRV).Significance. Time-domain metrics are more robust to reductions in signal length and calculation window size but demonstrate lower interchangeability between HRV and PRV. Both signal length and window size influence selected ANS metrics and should be considered, particularly when employing entropy-based indices in wearable, remote, and short-duration physiological monitoring.

Objective. To develop and evaluate method pipelines combining superimposition template subtraction (STS) and independent component analysis (ICA) for the most temporally accurate fetal electrocardiogram (fECG) signals extraction from abdominal recordings.Approach. Four method pipelines were developed by combining versions of STS and ICA algorithms to leverage their complementary strengths while mitigating their individual weaknesses. These pipelines were designed to adapt to various signal characteristics and were tested using recordings from the 2013 PhysioNet challenge and abdominal and direct fetal ECG database.Main results. Over the whole dataset, the best performing method pipeline achieved an average F1 score of 95.2% for fetal heart rate detection using a small error window of only 10 ms, demonstrating effective maternal signal suppression and accurate fetal signal extraction.Significance. Noninvasive monitoring of fetal health through electrocardiography could enable early detection of distress, but is challenged by the presence of overlapping maternal and fetal signals. This work demonstrates that strategically combining STS and ICA techniques can overcome these challenges and provide highly accurate fECG extraction.

Objective. Diabetes mellitus presents a significant global health burden, with patients demonstrating high prevalence of lower extremity atherosclerotic disease (LEAD) and poor prognosis. Despite the crucial need for early screening, primary healthcare lacks accessible LEAD screening protocols for people with diabetes. This study proposed a photoplethysmography (PPG)-based approach to enhance detection sensitivity for this high-risk population.Approach. This study collected toe PPG signals from 104 participants with diabetes, including 54 participants with LEAD. PPG signals underwent preprocessing followed by extraction of 162 features from 7 dimensions. Through a hybrid feature selection framework integrating feature extraction rate filtering and embedded random forest (RF) algorithms, 6 key PPG features were identified for RF classification model construction. The model was evaluated using metrics including sensitivity, specificity, accuracy,F1 score and Kappa coefficient, with DUS results serving as the reference standard.Results.The model achieved 85% sensitivity and 79% specificity, with 82% accuracy andF1-score, indicating good overall performance. The model's Kappa coefficient was 0.63, indicating good agreement with the DUS.Significance. This work demonstrates the feasibility of PPG-based method for screening LEAD in people with diabetes.

Objective. Hypertension increasingly affects younger populations alongside rising obesity rates, and impaired baroreflex (BR) function could contribute to its development. This study investigated changes in BR control of the cardiac chronotropic (ccBR) and vascular resistance (vrBR) arms in young normotensive patients with obesity and explored associations with sex- and age-independent anthropometric measures (body mass index (iso-BMI) and waist to hip ratio (OSS of WHR)), insulin resistance (HOMA_IR), and arterial stiffness index CAVI₀.Approach.Twenty-three normotensive adolescents and young adults with obesity (17 females, median age: 17.1 years) and twenty-two sex- and age-matched healthy lean participants (16 females, median age: 17.4 years) were examined during four phases: supine rest, head-up-tilt (HUT), supine recovery, and mental arithmetics task (MA). The causal coupling and gain in the frequency-domain of the ccBR and vrBR arms were assessed non-invasively from the spontaneous variability series of arterial pressure, heart period, and peripheral vascular resistance using a partial spectral decomposition method in the low frequency band (0.04-0.15 Hz).Main results.Patients with obesity showed lower ccBR gain during HUT and persistently lower vrBR gain during supine rest and HUT. No significant associations were found between iso-BMI, OSS of WHR, HOMA_IR, CAVI₀, and spectral parameters during supine rest, except for a significant negative correlation between iso-BMI and changes in ccBR spectral gain as a response to MA.Significance.Advanced non-invasive methods accounting for causality in evaluating two BR arms revealed early BR impairment in young participants with obesity, affecting both the ccBR arm and the less-explored vrBR arm.

Objective. Sleep apnea is a common sleep disorder associated with severe health risks, necessitating accurate and efficient detection methods.Approach. This study proposes ModelS4Apnea, a deep learning framework for sleep apnea detection from electrocardiogram (ECG) spectrograms, integrating structured state space models (S4) for temporal modeling. The framework consists of a convolutional neural network module for local feature extraction, an S4 module for capturing long-range dependencies, and a classification module for final predictions.Main results. The model was trained and evaluated on the Apnea-ECG dataset, achieving an accuracy of 0.933, anF1-score of 0.912, a sensitivity of 0.916, and a specificity of 0.944, outperforming most prior studies while maintaining computational efficiency.Significance. Compared to existing methods, ModelS4Apnea provides high classification performance with significantly fewer trainable parameters than long short-term memory-based models, reducing training time and memory consumption. The model's ability to aggregate segment-level predictions enabled perfect per-recording classification, demonstrating its robustness in diagnosing sleep apnea across entire recordings. Moreover, its low memory footprint and fast inference speed make it well-suited for wearable devices, home-based monitoring, and clinical applications, offering a scalable and efficient solution for automated sleep apnea detection. Future work may explore multi-modal data integration, real-world deployment, and further optimizations to enhance its clinical applicability and reliability.