Physiological Measurement

Photoplethysmography is a key sensing technology which is used in wearable devices such as smartwatches and fitness trackers. Currently, photoplethysmography sensors are used to monitor physiological parameters including heart rate and heart rhythm, and to track activities like sleep and exercise. Yet, wearable photoplethysmography has potential to provide much more information on health and wellbeing, which could inform clinical decision making. This Roadmap outlines directions for research and development to realise the full potential of wearable photoplethysmography. Experts discuss key topics within the areas of sensor design, signal processing, clinical applications, and research directions. Their perspectives provide valuable guidance to researchers developing wearable photoplethysmography technology.

Respiratory rate (RR) is an important physiological parameter whose abnormality has been regarded as an important indicator of serious illness. In order to make RR monitoring simple to perform, reliable and accurate, many different methods have been proposed for such automatic monitoring. According to the theory of respiratory rate extraction, methods are categorized into three modalities: extracting RR from other physiological signals, RR measurement based on respiratory movements, and RR measurement based on airflow. The merits and limitations of each method are highlighted and discussed. In addition, current works are summarized to suggest key directions for the development of future RR monitoring methodologies.

Objective. Cardiovascular diseases are a major cause of mortality globally, and electrocardiograms (ECGs) are crucial for diagnosing them. Traditionally, ECGs are stored in printed formats. However, these printouts, even when scanned, are incompatible with advanced ECG diagnosis software that require time-series data. Digitizing ECG images is vital for training machine learning models in ECG diagnosis, leveraging the extensive global archives collected over decades. Deep learning models for image processing are promising in this regard, although the lack of clinical ECG archives with reference time-series data is challenging. Data augmentation techniques using realistic generative data models provide a solution. Approach. We introduce ECG-Image-Kit, an open-source toolbox for generating synthetic multi-lead ECG images with realistic artifacts from time-series data, aimed at automating the conversion of scanned ECG images to ECG data points. The tool synthesizes ECG images from real time-series data, applying distortions like text artifacts, wrinkles, and creases on a standard ECG paper background. Main results. As a case study, we used ECG-Image-Kit to create a dataset of 21 801 ECG images from the PhysioNet QT database. We developed and trained a combination of a traditional computer vision and deep neural network model on this dataset to convert synthetic images into time-series data for evaluation. We assessed digitization quality by calculating the signal-to-noise ratio and compared clinical parameters like QRS width, RR, and QT intervals recovered from this pipeline, with the ground truth extracted from ECG time-series. The results show that this deep learning pipeline accurately digitizes paper ECGs, maintaining clinical parameters, and highlights a generative approach to digitization. Significance. The toolbox has broad applications, including model development for ECG image digitization and classification. The toolbox currently supports data augmentation for the 2024 PhysioNet Challenge, focusing on digitizing and classifying paper ECG images.

Objective: To validate the accuracy of the Oura ring in the quantification of resting heart rate (HR) and heart rate variability (HRV). Background: Wearable devices have become comfortable, lightweight, and technologically advanced for assessing health behavior. As an example, the novel Oura ring integrates daily physical activity and nocturnal cardiovascular measurements. Ring users can follow their autonomic nervous system responses to their daily behavior based on nightly changes in HR and HRV, and adjust their behavior accordingly after self-reflection. As wearable photoplethysmogram (PPG) can be disrupted by several confounding influences, it is crucial to demonstrate the accuracy of ring measurements. Approach: Nocturnal HR and HRV were assessed in 49 adults with simultaneous measurements from the Oura ring and the gold standard ECG measurement. Female and male participants with a wide age range (15–72 years) and physical activity status were included. Regression analysis between ECG and the ring outcomes was performed. Main results: Very high agreement between the ring and ECG was observed for nightly average HR and HRV (r² = 0.996 and 0.980, respectively) with a mean bias of −0.63 bpm and −1.2 ms. High agreement was also observed across 5 min segments within individual nights in (r² = 0.869 ± 0.098 and 0.765 ± 0.178 in HR and HRV, respectively). Significance: Present findings indicate high validity of the Oura ring in the assessment of nocturnal HR and HRV in healthy adults. The results show the utility of this miniaturised device as a lifestyle management tool in long-term settings. High quality PPG signal results prompt future studies utilizing ring PPG towards clinically relevant health outcomes.

Objective. Photoplethysmography is a non-invasive optical technique that measures changes in blood volume within tissues. It is commonly and being increasingly used for a variety of research and clinical applications to assess vascular dynamics and physiological parameters. Yet, contrary to heart rate variability measures, a field which has seen the development of stable standards and advanced toolboxes and software, no such standards and limited open tools exist for continuous photoplethysmogram (PPG) analysis. Consequently, the primary objective of this research was to identify, standardize, implement and validate key digital PPG biomarkers. Approach. This work describes the creation of a standard Python toolbox, denoted pyPPG, for long-term continuous PPG time-series analysis and demonstrates the detection and computation of a high number of fiducial points and digital biomarkers using a standard fingerbased transmission pulse oximeter. Main results. The improved PPG peak detector had an F1-score of 88.19% for the state-of-the-art benchmark when evaluated on 2054 adult polysomnography recordings totaling over 91 million reference beats. The algorithm outperformed the open-source original Matlab implementation by ∼5% when benchmarked on a subset of 100 randomly selected MESA recordings. More than 3000 fiducial points were manually annotated by two annotators in order to validate the fiducial points detector. The detector consistently demonstrated high performance, with a mean absolute error of less than 10 ms for all fiducial points. Significance. Based on these fiducial points, pyPPG engineered a set of 74 PPG biomarkers. Studying PPG time-series variability using pyPPG can enhance our understanding of the manifestations and etiology of diseases. This toolbox can also be used for biomarker engineering in training data-driven models. pyPPG is available on https://physiozoo.com/.

The results of many medical measurements are directly or indirectly influenced by the autonomic nervous system (ANS). For example pupil size or heart rate may demonstrate striking moment-to-moment variability. This review intends to elucidate the physiology behind this seemingly unpredictable system.

The review is split up into: 1. The peripheral ANS, parallel innervation by the sympathetic and parasympathetic branches, their transmitters and co-transmitters. It treats questions like the supposed sympatho/vagal balance, organization in plexuses and the ‘little brains’ that are active like in the enteric system or around the heart. Part 2 treats ANS-function in some (example-) organs in more detail: the eye, the heart, blood vessels, lungs, respiration and cardiorespiratory coupling. Part 3 poses the question of who is directing what? Is the ANS a strictly top-down directed system or is its organization bottom-up? Finally, it is concluded that the ‘noisy numbers’ in medical measurements, caused by ANS variability, are part and parcel of how the system works. This topical review is a one-man’s undertaking and may possibly give a biased view. The author has explicitly indicated in the text where his views are not (yet) supported by facts, hoping to provoke discussion and instigate new research.

Objective. Heart failure (HF) is considered a global pandemic because of increasing prevalence, high mortality rate, frequent hospitalization, and associated economic burden. This study explores a noninvasive method that may help in managing HF patients by predicting HF readmission. Methods. Seismocardiogram (SCG) signal is the low-frequency chest vibration produced by the mechanical activity of the heart. SCG signal was acquired from 101 patients with HF, including those readmitted to the hospital during the study period. SCG signals were segmented into heartbeats and clustered based on respiration phases. Features were extracted from each cluster. Several conventional machine learning (ML) models were developed using selected SCG and heart rate variability features. Furthermore, SCG signals were transformed into images using a time–frequency distribution method. Images were used to train a deep learning model. The models were able to predict the readmission status of HF patients. Results. ML algorithms achieved higher accuracy than the deep learning model in classifying the readmitted and non-readmitted HF patients. K-nearest neighbor achieved the highest classification accuracy (89.4% accuracy, 87.8% sensitivity, 90.1% specificity, 78.2% precision, and 82.7% F1-score). A detailed discussion of the extracted features was provided, correlating them with HF conditions. Conclusions. The study results suggest that SCG signals may be useful for readmission prediction of HF patients.

Objective. The present study investigated how breathing stimuli affect both non-linear and linear metrics of the autonomic nervous system (ANS). Approach. The analysed dataset consisted of 70 young, healthy volunteers, in whom arterial blood pressure (ABP) was measured noninvasively during 5 min sessions of controlled breathing at three different frequencies: 6, 10 and 15 breaths min⁻¹. CO₂ concentration and respiratory rate were continuously monitored throughout the controlled breathing sessions. The ANS was characterized using non-linear methods, including phase-rectified signal averaging (PRSA) for estimating heart acceleration and deceleration capacity (AC, DC), multiscale entropy, approximate entropy, sample entropy, and fuzzy entropy, as well as time and frequency-domain measures (low frequency, LF; high-frequency, HF; total power, TP) of heart rate variability (HRV). Main results. Higher breathing rates resulted in a significant decrease in end-tidal CO₂ concentration (p < 0.001), accompanied by increases in both ABP (p < 0.001) and heart rate (HR, p < 0.001). A strong, linear decline in AC and DC (p < 0.001 for both) was observed with increasing breathing rate. All entropy metrics increased with breathing frequency (p < 0.001). In the time-domain, HRV metrics significantly decreased with breathing frequency (p < 0.01 for all). In the frequency-domain, HRV LF and HRV HF decreased (p = 0.038 and p = 0.040, respectively), although these changes were modest. There was no significant change in HRV TP with breathing frequencies. Significance. Alterations in CO₂ levels, a potent chemoreceptor trigger, and changes in HR most likely modulate ANS metrics. Non-linear PRSA and entropy appear to be more sensitive to breathing stimuli compared to frequency-dependent HRV metrics. Further research involving a larger cohort of healthy subjects is needed to validate our observations.

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. Electrocardiography and blood pressure (BP) measurement are two widely used tools for diagnosis and monitoring cardiovascular diseases. While the electrocardiogram (ECG) and BP have been considered complementary modalities, there are also systematic relationships between them. Therefore, advancements in portable and wearable ECG devices, along with promising results in cuff-less BP measurement using a combination of ECG and other bio-signals have led researchers to hypothesize the possibility of estimating BP and classifying BP categories (e.g. normal vs. hypertensive) using only ECG. However, the literature is divided on this topic: some studies support this hypothesis, while others reject it. Approach. In this study, regression and classification machine learning (ML) models were developed to explore the feasibility of estimating BP and predicting BP categories (normal vs. hypertensive) from 30 s ECGs using an extensive dataset from AliveCor Inc. which includes 124 427 records from 7412 subjects. The ECG and BP recordings were asynchronous with variable counts and time lags. Therefore, a 3.5 min time window before and after each ECG recording was used to calculate the mean BP measurement. Sex-aware ML models were trained using a comprehensive feature vector comprising 280 features: 128 explainable ECG features developed by the research team and 150 ECG features extracted by the Black Swan team, one of the top-performing teams in the PhysioNet Challenge 2017. Additionally, the average time gap between each ECG and the corresponding BP measurement, along with the subject’s age, were included as two supplementary features. Main results. Our best regression ML models achieved a mean absolute error of 12.59 mmHg for estimating systolic BP and 7.43 mmHg for diastolic BP, with correlation coefficients of 0.35 and 0.38 between the predicted and actual values, respectively. The best BP normal-hypertensive classification model achieved an area under the receiver operating characteristic curve of 0.655. Significance. Using a large dataset of ECG and BP recordings, this study found that ML models did not achieve acceptable performance in predicting BP values or classifying BP categories, indicating that BP cannot be reliably estimated from the ECG.

Objective. 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. Approach. 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. Main results. We apply this approach to a dataset comprising approximately 1.6 million hours of continuous ECG and PPG data from over 10 000 unique patients in a pediatric intensive care unit. 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. Significance. It is our hope that this work will spur discussion around the standardization of PAT measurement using routinely collected signals in a clinical environment.

Objective. Peripheral edema is a common issue among elderly individuals with chronic conditions such as heart failure (HF). Continuous, non-invasive monitoring may enable earlier intervention, reduced hospital readmissions, and improved quality of life. This systematic review aims to evaluate the use of bioimpedance (BI) as a method for monitoring peripheral edema, with a particular focus on portable and wearable applications for remote health management. Approach. A systematic search was conducted across PubMed, IEEE Xplore, and Web of Science to identify studies utilizing BI for the detection or monitoring of lower limb edema with potential for portability or wearability. Main results. Fourteen studies met the inclusion criteria. Five studies focused on HF patients, while nine involved other populations, such as healthy individuals, patients with limb injuries, or those on hemodialysis. Ten studies featured or proposed portable BI devices, whereas four remained at the proof-of-concept stage without portable implementations. There was significant variability in device design, measurement protocols, and target populations. While existing results show promise, few studies evaluated systems in real-world or long-term monitoring scenarios. Significance. BI is a promising, non-invasive approach for the continuous monitoring of peripheral edema, particularly in remote and home-based settings. However, current research is limited by small sample sizes, lack of standardization, and minimal validation in diverse, real-world environments. Further development of wearable systems and robust clinical validation is essential to support broader clinical adoption.

Objective. Electrocardiography and blood pressure (BP) measurement are two widely used tools for diagnosis and monitoring cardiovascular diseases. While the electrocardiogram (ECG) and BP have been considered complementary modalities, there are also systematic relationships between them. Therefore, advancements in portable and wearable ECG devices, along with promising results in cuff-less BP measurement using a combination of ECG and other bio-signals have led researchers to hypothesize the possibility of estimating BP and classifying BP categories (e.g. normal vs. hypertensive) using only ECG. However, the literature is divided on this topic: some studies support this hypothesis, while others reject it. Approach. In this study, regression and classification machine learning (ML) models were developed to explore the feasibility of estimating BP and predicting BP categories (normal vs. hypertensive) from 30 s ECGs using an extensive dataset from AliveCor Inc. which includes 124 427 records from 7412 subjects. The ECG and BP recordings were asynchronous with variable counts and time lags. Therefore, a 3.5 min time window before and after each ECG recording was used to calculate the mean BP measurement. Sex-aware ML models were trained using a comprehensive feature vector comprising 280 features: 128 explainable ECG features developed by the research team and 150 ECG features extracted by the Black Swan team, one of the top-performing teams in the PhysioNet Challenge 2017. Additionally, the average time gap between each ECG and the corresponding BP measurement, along with the subject’s age, were included as two supplementary features. Main results. Our best regression ML models achieved a mean absolute error of 12.59 mmHg for estimating systolic BP and 7.43 mmHg for diastolic BP, with correlation coefficients of 0.35 and 0.38 between the predicted and actual values, respectively. The best BP normal-hypertensive classification model achieved an area under the receiver operating characteristic curve of 0.655. Significance. Using a large dataset of ECG and BP recordings, this study found that ML models did not achieve acceptable performance in predicting BP values or classifying BP categories, indicating that BP cannot be reliably estimated from the ECG.

Objective. Remote photoplethysmography (rPPG) offers a promising method for contactless pulse rate (PR) monitoring, which is particularly valuable for remote patient care. However, signal noise—caused by factors such as motion and lighting—can significantly impact measurement accuracy. Approach. We present a hybrid algorithm that combines frequency-domain analysis to estimate initial PR and a time-domain approach to refine this estimate, improving robustness in challenging conditions. Main results. The combined method increases accuracy and success rate compared to time-domain methods alone. Evaluated against time-aligned electrocardiogram, it achieved a root mean square error (RMSE) as low as 2.0 bpm and an r² of 0.96. On a larger outpatient dataset, the RMSE was 3.2 bpm with an r² of 0.93. Importantly, no significant performance difference was observed across varying skin tones. Significance. These results demonstrate that the proposed PR algorithm enables reliable, contactless pulse monitoring in real-world conditions, supporting broader adoption of rPPG for inclusive and scalable remote health monitoring.

Objective. Heart failure (HF) is considered a global pandemic because of increasing prevalence, high mortality rate, frequent hospitalization, and associated economic burden. This study explores a noninvasive method that may help in managing HF patients by predicting HF readmission. Methods. Seismocardiogram (SCG) signal is the low-frequency chest vibration produced by the mechanical activity of the heart. SCG signal was acquired from 101 patients with HF, including those readmitted to the hospital during the study period. SCG signals were segmented into heartbeats and clustered based on respiration phases. Features were extracted from each cluster. Several conventional machine learning (ML) models were developed using selected SCG and heart rate variability features. Furthermore, SCG signals were transformed into images using a time–frequency distribution method. Images were used to train a deep learning model. The models were able to predict the readmission status of HF patients. Results. ML algorithms achieved higher accuracy than the deep learning model in classifying the readmitted and non-readmitted HF patients. K-nearest neighbor achieved the highest classification accuracy (89.4% accuracy, 87.8% sensitivity, 90.1% specificity, 78.2% precision, and 82.7% F1-score). A detailed discussion of the extracted features was provided, correlating them with HF conditions. Conclusions. The study results suggest that SCG signals may be useful for readmission prediction of HF patients.

Objective. Peripheral edema is a common issue among elderly individuals with chronic conditions such as heart failure (HF). Continuous, non-invasive monitoring may enable earlier intervention, reduced hospital readmissions, and improved quality of life. This systematic review aims to evaluate the use of bioimpedance (BI) as a method for monitoring peripheral edema, with a particular focus on portable and wearable applications for remote health management. Approach. A systematic search was conducted across PubMed, IEEE Xplore, and Web of Science to identify studies utilizing BI for the detection or monitoring of lower limb edema with potential for portability or wearability. Main results. Fourteen studies met the inclusion criteria. Five studies focused on HF patients, while nine involved other populations, such as healthy individuals, patients with limb injuries, or those on hemodialysis. Ten studies featured or proposed portable BI devices, whereas four remained at the proof-of-concept stage without portable implementations. There was significant variability in device design, measurement protocols, and target populations. While existing results show promise, few studies evaluated systems in real-world or long-term monitoring scenarios. Significance. BI is a promising, non-invasive approach for the continuous monitoring of peripheral edema, particularly in remote and home-based settings. However, current research is limited by small sample sizes, lack of standardization, and minimal validation in diverse, real-world environments. Further development of wearable systems and robust clinical validation is essential to support broader clinical adoption.

Objective. The advancement of contactless methods of measuring the respiratory rate (RR) using RGB cameras demonstrates a significant potential for improving patient care in various environments. As these methods offer reliable and discreet monitoring, they can prevent severe health complications and improve outcomes for patients facing challenges accessing traditional healthcare facilities. Approach. This systematic review explores recent advancements in RR estimation using RGB cameras, focusing on assessing publicly available datasets and effective signal preprocessing methods. We also conducted a comprehensive analysis by comparing RGB camera-based approaches with other sensor modalities and discussed potential future research directions and indicated the necessity of developing new approaches that would mitigate existing challenges and would enhance the accuracy and reliability of non-contact RR measurement methods. Main results. We analyzed existing public datasets, assessing their diversity in lighting, skin tone, and motion, alongside the camera hardware configurations, including frame rate and resolution, utilizing different filter and feature-based techniques. While deep learning and hybrid models achieved lower errors under ideal indoor lighting and minimal motion, performance significantly declined in low light, high motion, or complex uncontrolled environments. In contrast, other sensor modalities, such as thermal and infrared sensors, achieved high accuracy across a wide range of conditions, but at greater hardware cost and system complexity, while RGB cameras remained the most cost-effective option, trading off precision for accessibility. Significance. RGB camera-based RR monitoring systems have the potential for robust applicability in clinical and nonclinical settings such as telemedicine platforms for monitoring patients breathing rates (BRs) in real time. This review highlights existing research gaps, such as insufficient real-world datasets and sensitivity to environmental variance, and emphasizes on the importance of acquiring datasets based on complex real-world scenarios, standardized benchmarks, multi-sensor fusion for addressing current limitations, and deep neural network architecture implementation for reliable non-contact RR estimation for real-world applications.

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.

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. Wearable technology like the Apple Watch is increasingly important for monitoring health metrics. Accurate measurement is crucial, as inaccuracies can impact health outcomes. Despite extensive research, findings on the Apple Watch’s accuracy vary across different conditions. While previous reviews have summarized findings, few have utilized a meta-analytic approach. This study aims to quantitatively evaluate the accuracy of the Apple Watch in measuring health metrics. The accuracy of the Apple Watch was assessed in measuring energy expenditure (EE), heart rate (HR), and step counts (steps). Approach. We searched Embase, PubMed, Scopus, and SPORTDiscus for studies on adults using the Apple Watch compared to reference measures. The Bland–Altman framework was applied to assess mean bias and limits of agreement (LoA), with robust variance estimation to address within-study correlations. Heterogeneity was assessed across variables such as age, health status, device series, activity intensity, and activity type. Additionally, the mean absolute percentage error (MAPE) reported in the included studies was summarized by subgroups. Main results. This review included 56 studies, comprising 270 effect sizes on EE (71), HR (148), and steps (51). The meta-analysis showed a mean bias of 0.30 (LoA: −2.09–2.69) for EE (kcal min⁻¹), −0.12 (LoA: −11.06–10.81) for HR (beats min⁻¹), −1.83 (LoA: −9.08–5.41) for steps (steps min⁻¹). The forest plots showed variability in LoA across subgroups. For MAPE, all subgroups for EE exceeded the 10% validity threshold, while none of the subgroups for HR exceeded this threshold. For steps, some subgroups exceeded 10%, highlighting variability in accuracy based on different conditions. Significance. This study demonstrates that while the Apple Watch generally provides accurate HR and step measurements, its accuracy for EE is limited. Although HR and step measurements showed acceptable accuracy, variability was observed across different user characteristics and measurement conditions. These findings highlight the importance of considering such factors when evaluating validity.

Objective&#xD;Sleep is hypothesized to restore near-critical dynamics in large-scale brain networks, whereas insomnia may disrupt this self-organizing process. This study aimed to determine whether insomnia alters neural avalanche dynamics and criticality-based EEG metrics, and whether these metrics enhance prediction of sleep fragmentation compared with conventional spectral measures.&#xD;Approach&#xD;Overnight high-density electroencephalography was recorded from 50 participants aged 16–69 years, including healthy sleepers and individuals with insomnia. Neural avalanches were detected as clusters of significant amplitude excursions. The branching parameter (σ) quantified temporal propagation within avalanches, while the deviation-from-criticality coefficient (DCC) indexed the system’s distance from the critical state. These criticality features were contrasted with spectral power measures in predictive models of NREM sleep fragmentation.&#xD;Main Results&#xD;Participants with insomnia exhibited reduced avalanche density and diminished slow-wave activity, accompanied by significant deviations of σ from the critical value and elevated DCC across the night. Criticality-based metrics captured fragmentation dynamics more sensitively than spectral features. In predictive modeling, criticality measures significantly outperformed spectral power in forecasting NREM fragmentation (F1-score = 0.69 vs. 0.62), with the strongest gains in mild and severe insomnia subgroups.&#xD;Significance&#xD;Insomnia is characterized by a persistent deviation from near-critical neural dynamics, reflecting compromised stability and recovery during sleep. Criticality-based EEG features provide a more mechanistic and predictive framework for identifying sleep fragmentation and may offer novel biomarkers for quantifying disrupted sleep physiology in clinical insomnia.&#xD;

Objective. Large vessel occlusion (LVO) stroke presents a major challenge in clinical practice due to the potential for poor outcomes with delayed treatment. Treatment for LVO involves highly specialized care, in particular endovascular thrombectomy, and is available only at certain hospitals. Therefore, prehospital identification of LVO by emergency ambulance services, can be critical for triaging LVO stroke patients directly to a hospital with access to endovascular therapy. Clinical scores exist to help distinguish LVO from less severe strokes, but they are based on a series of examinations that can take minutes and may be impractical for patients with dementia or those who cannot follow commands due to their stroke. There is a need for a fast and reliable method to aid in the early identification of LVO. In this study, our objective was to assess the feasibility of using 30-second photoplethysmography (PPG) recording to assist in recognizing LVO stroke. Method. A total of 88 patients, including 25 with LVO, 27 with stroke mimic (SM), and 36 non-LVO stroke patients (NL), were recorded at the Liverpool Hospital emergency department in Sydney, Australia. Demographics (age, sex), as well as morphological features and beating rate variability measures, were extracted from the PPG. A binary classification approach was employed to differentiate between LVO stroke and NL+SM (NL.SM). A 2:1 train-test split was stratified and repeated randomly across 100 iterations. Results. The best model achieved a median test set area under the receiver operating characteristic curve (AUROC) of 0.77 (0.71–0.82). Conclusion. Our study demonstrates the potential of utilizing a 30-second PPG recording for identifying LVO stroke.

Objective: Hemodynamic monitoring is essential in preclinical research. Currently available techniques are either invasive or complex to implement. Inductive plethysmography (IP) provides an alternative for estimating stroke volume and cardiac output, as the IP signal includes ventilatory and cardiogenic oscillations (COS). COS monitoring, also defined as thoracocardiography (TCG), has been validated in humans and large laboratory animals. A recent study demonstrated proof of concept in COS extraction from the TCG signal recorded during respiratory pauses in mechanically-ventilated laboratory rats using a high-resolution IP device. The present study aims to develop an ensemble averaging (EA) algorithm, triggered by the electrocardiogram (ECG) R-peak, to extract COS from TCG signals in rats and continuously estimate stroke volume and cardiac output. Approach: After an evaluation of the IP device using the EA technique on a mechanical test bench, the applicability of the EA technique was tested in anesthetized rats without ventilatory support during a pharmacological challenge. The ability of the algorithm to track stroke volume and cardiac output changes during the hemodynamic test was also evaluated. Main Results: Metrological evaluation of the IP device using the EA technique demonstrated linearity across the physiological operating range and resolution sufficient to detect volume changes of less than 10% of typical physiological values. Although the assumptions underlying the use of EA cannot be fully satisfied for COS extraction—due to quasi-synchrony with the ECG R-peak and signal non-stationarities—the method enabled extraction of satisfactory average COS waveforms, from which the system reliably captured positive and negative inotropic effects consistent with reference measurements during the pharmacological protocol. Significance: The evaluated algorithm demonstrates advancement over previous studies by enabling hemodynamic monitoring under usage conditions. Further studies are needed to extend its applicability to complex and physiologically relevant scenarios, positioning this technology as a potential non-invasive tool for preclinical research.

Objective. To develop a comprehensive physiological dataset for assessing internal and external stress and to propose robust automated stress recognition methods based on photoplethysmographic (PPG) signals.&#xD;Approach. We established the Internal and External Stress Dataset (IESD), comprising PPG signals from 107 participants subjected to four distinct stress-inducing paradigms. Exploratory analyses revealed significant differences in heart rate variability (HRV) across these paradigms, underscoring the necessity for advanced methods capable of differentiating various stress types. To address this, we introduced a transfer learning-based inter-paradigm stress recognition model utilizing a Domain Adversarial Neural Network (DANN) combined with Maximum Mean Discrepancy (MMD) for robust feature extraction.&#xD;Main results. Analysis identified significant differences between internal and external stress, as well as among different external paradigms. Our proposed model demonstrated superior accuracy in recognizing homologous stress compared to heterologous stress within the same target domain, achieving accuracies of 73.86\% (TSST to ST) and 60.41\% (TSST to VWT). Moreover, the deep feature extraction significantly improved recognition performance and robustness across both intra- and inter-paradigm contexts.&#xD;Significance. This study provides a valuable dataset and advanced methodology to enhance automated stress detection capabilities, effectively differentiating internal and external stress. The application of deep learning significantly improves recognition accuracy, offering promising prospects for future research and practical applications in stress monitoring.

Objective. Calf blood pressure plays an important role in various clinical applications, such as determining the appropriate cuff pressure for compression therapy and diagnosis of lower extremity vascular diseases, which necessitates the integration of built-in measurement methods within the device. This study aimed to investigate the differences in human calf blood pressure assessments resulting from the application of various in vivo external compression strategies. Approach. An experimental procedure incorporating different compression strategies, specially the low-pressure-sustained mode, the slow-inflation mode and the slow-deflation mode, was conducted to capture the dynamic responses of the photoplethysmography (PPG) signal for calf blood pressure evaluation. Nineteen subjects, including 13 males and 6 females, participated in this experimental study. Feature points related to calf blood pressure were extracted from dynamic responses of the PPG signal and subjected to statistical analysis. A lumped parameter model of the human lower extremity was developed to assist in analyzing the biomechanical factors underlying these differences. Main results. The experimental results indicated that the dynamic behaviors of the PPG signal followed clear and consistent patterns, and the calf blood pressure value derived from the PPG signal in the slow-inflation mode was significantly higher than that obtained in the slow-deflation mode. Model simulation results showed that the differences in the evaluated values of the calf blood pressure between the slow-inflation and slow-deflation modes were caused by the different collapse processes of calf arteries and veins. Significance. This study could help understand the forced collapse process of blood vessels in the calf and inspire new ideas on calf blood pressure evaluation and personalized compression therapy.

Objective. 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. Approach. 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. Main results. We apply this approach to a dataset comprising approximately 1.6 million hours of continuous ECG and PPG data from over 10 000 unique patients in a pediatric intensive care unit. 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. Significance. It is our hope that this work will spur discussion around the standardization of PAT measurement using routinely collected signals in a clinical environment.

Objective. Large vessel occlusion (LVO) stroke presents a major challenge in clinical practice due to the potential for poor outcomes with delayed treatment. Treatment for LVO involves highly specialized care, in particular endovascular thrombectomy, and is available only at certain hospitals. Therefore, prehospital identification of LVO by emergency ambulance services, can be critical for triaging LVO stroke patients directly to a hospital with access to endovascular therapy. Clinical scores exist to help distinguish LVO from less severe strokes, but they are based on a series of examinations that can take minutes and may be impractical for patients with dementia or those who cannot follow commands due to their stroke. There is a need for a fast and reliable method to aid in the early identification of LVO. In this study, our objective was to assess the feasibility of using 30-second photoplethysmography (PPG) recording to assist in recognizing LVO stroke. Method. A total of 88 patients, including 25 with LVO, 27 with stroke mimic (SM), and 36 non-LVO stroke patients (NL), were recorded at the Liverpool Hospital emergency department in Sydney, Australia. Demographics (age, sex), as well as morphological features and beating rate variability measures, were extracted from the PPG. A binary classification approach was employed to differentiate between LVO stroke and NL+SM (NL.SM). A 2:1 train-test split was stratified and repeated randomly across 100 iterations. Results. The best model achieved a median test set area under the receiver operating characteristic curve (AUROC) of 0.77 (0.71–0.82). Conclusion. Our study demonstrates the potential of utilizing a 30-second PPG recording for identifying LVO stroke.

Objective: Hemodynamic monitoring is essential in preclinical research. Currently available techniques are either invasive or complex to implement. Inductive plethysmography (IP) provides an alternative for estimating stroke volume and cardiac output, as the IP signal includes ventilatory and cardiogenic oscillations (COS). COS monitoring, also defined as thoracocardiography (TCG), has been validated in humans and large laboratory animals. A recent study demonstrated proof of concept in COS extraction from the TCG signal recorded during respiratory pauses in mechanically-ventilated laboratory rats using a high-resolution IP device. The present study aims to develop an ensemble averaging (EA) algorithm, triggered by the electrocardiogram (ECG) R-peak, to extract COS from TCG signals in rats and continuously estimate stroke volume and cardiac output. Approach: After an evaluation of the IP device using the EA technique on a mechanical test bench, the applicability of the EA technique was tested in anesthetized rats without ventilatory support during a pharmacological challenge. The ability of the algorithm to track stroke volume and cardiac output changes during the hemodynamic test was also evaluated. Main Results: Metrological evaluation of the IP device using the EA technique demonstrated linearity across the physiological operating range and resolution sufficient to detect volume changes of less than 10% of typical physiological values. Although the assumptions underlying the use of EA cannot be fully satisfied for COS extraction—due to quasi-synchrony with the ECG R-peak and signal non-stationarities—the method enabled extraction of satisfactory average COS waveforms, from which the system reliably captured positive and negative inotropic effects consistent with reference measurements during the pharmacological protocol. Significance: The evaluated algorithm demonstrates advancement over previous studies by enabling hemodynamic monitoring under usage conditions. Further studies are needed to extend its applicability to complex and physiologically relevant scenarios, positioning this technology as a potential non-invasive tool for preclinical research.

Objective. Peripheral edema is a common issue among elderly individuals with chronic conditions such as heart failure (HF). Continuous, non-invasive monitoring may enable earlier intervention, reduced hospital readmissions, and improved quality of life. This systematic review aims to evaluate the use of bioimpedance (BI) as a method for monitoring peripheral edema, with a particular focus on portable and wearable applications for remote health management. Approach. A systematic search was conducted across PubMed, IEEE Xplore, and Web of Science to identify studies utilizing BI for the detection or monitoring of lower limb edema with potential for portability or wearability. Main results. Fourteen studies met the inclusion criteria. Five studies focused on HF patients, while nine involved other populations, such as healthy individuals, patients with limb injuries, or those on hemodialysis. Ten studies featured or proposed portable BI devices, whereas four remained at the proof-of-concept stage without portable implementations. There was significant variability in device design, measurement protocols, and target populations. While existing results show promise, few studies evaluated systems in real-world or long-term monitoring scenarios. Significance. BI is a promising, non-invasive approach for the continuous monitoring of peripheral edema, particularly in remote and home-based settings. However, current research is limited by small sample sizes, lack of standardization, and minimal validation in diverse, real-world environments. Further development of wearable systems and robust clinical validation is essential to support broader clinical adoption.

Deep learning for continuous physiological time series, such as electrocardiography or oximetry, has achieved remarkable success in supervised learning scenarios where training and testing data are drawn from the same distribution. However, when evaluating real-world applications, models often fail to generalize due to distribution shifts between the source domain on which the model was trained and the target domain where it is deployed. A common and particularly challenging shift often encountered in reality is where the source and target domain supports do not fully overlap. In this paper, we propose a novel framework, named Deep Unsupervised Domain adaptation using variable nEighbors (DUDE), to address this challenge. We introduce a new type of contrastive loss between the source and target domains using a dynamic neighbor selection strategy, in which the number of neighbors for each sample is adaptively determined based on the density observed in the latent space. We use multiple real-world datasets as source and target domains, with target domains that included demographics, ethnicities, geographies, and comorbidities that were not present in the source domain. The experimental results demonstrate superior DUDE performance compared to baselines, with an improvement of up to 16%, as well as a set of four benchmarks.

Objective. Electrocardiography and blood pressure (BP) measurement are two widely used tools for diagnosis and monitoring cardiovascular diseases. While the electrocardiogram (ECG) and BP have been considered complementary modalities, there are also systematic relationships between them. Therefore, advancements in portable and wearable ECG devices, along with promising results in cuff-less BP measurement using a combination of ECG and other bio-signals have led researchers to hypothesize the possibility of estimating BP and classifying BP categories (e.g. normal vs. hypertensive) using only ECG. However, the literature is divided on this topic: some studies support this hypothesis, while others reject it. Approach. In this study, regression and classification machine learning (ML) models were developed to explore the feasibility of estimating BP and predicting BP categories (normal vs. hypertensive) from 30 s ECGs using an extensive dataset from AliveCor Inc. which includes 124 427 records from 7412 subjects. The ECG and BP recordings were asynchronous with variable counts and time lags. Therefore, a 3.5 min time window before and after each ECG recording was used to calculate the mean BP measurement. Sex-aware ML models were trained using a comprehensive feature vector comprising 280 features: 128 explainable ECG features developed by the research team and 150 ECG features extracted by the Black Swan team, one of the top-performing teams in the PhysioNet Challenge 2017. Additionally, the average time gap between each ECG and the corresponding BP measurement, along with the subject’s age, were included as two supplementary features. Main results. Our best regression ML models achieved a mean absolute error of 12.59 mmHg for estimating systolic BP and 7.43 mmHg for diastolic BP, with correlation coefficients of 0.35 and 0.38 between the predicted and actual values, respectively. The best BP normal-hypertensive classification model achieved an area under the receiver operating characteristic curve of 0.655. Significance. Using a large dataset of ECG and BP recordings, this study found that ML models did not achieve acceptable performance in predicting BP values or classifying BP categories, indicating that BP cannot be reliably estimated from the ECG.

Objective. Remote photoplethysmography (rPPG) offers a promising method for contactless pulse rate (PR) monitoring, which is particularly valuable for remote patient care. However, signal noise—caused by factors such as motion and lighting—can significantly impact measurement accuracy. Approach. We present a hybrid algorithm that combines frequency-domain analysis to estimate initial PR and a time-domain approach to refine this estimate, improving robustness in challenging conditions. Main results. The combined method increases accuracy and success rate compared to time-domain methods alone. Evaluated against time-aligned electrocardiogram, it achieved a root mean square error (RMSE) as low as 2.0 bpm and an r² of 0.96. On a larger outpatient dataset, the RMSE was 3.2 bpm with an r² of 0.93. Importantly, no significant performance difference was observed across varying skin tones. Significance. These results demonstrate that the proposed PR algorithm enables reliable, contactless pulse monitoring in real-world conditions, supporting broader adoption of rPPG for inclusive and scalable remote health monitoring.

Objective. Heart failure (HF) is considered a global pandemic because of increasing prevalence, high mortality rate, frequent hospitalization, and associated economic burden. This study explores a noninvasive method that may help in managing HF patients by predicting HF readmission. Methods. Seismocardiogram (SCG) signal is the low-frequency chest vibration produced by the mechanical activity of the heart. SCG signal was acquired from 101 patients with HF, including those readmitted to the hospital during the study period. SCG signals were segmented into heartbeats and clustered based on respiration phases. Features were extracted from each cluster. Several conventional machine learning (ML) models were developed using selected SCG and heart rate variability features. Furthermore, SCG signals were transformed into images using a time–frequency distribution method. Images were used to train a deep learning model. The models were able to predict the readmission status of HF patients. Results. ML algorithms achieved higher accuracy than the deep learning model in classifying the readmitted and non-readmitted HF patients. K-nearest neighbor achieved the highest classification accuracy (89.4% accuracy, 87.8% sensitivity, 90.1% specificity, 78.2% precision, and 82.7% F1-score). A detailed discussion of the extracted features was provided, correlating them with HF conditions. Conclusions. The study results suggest that SCG signals may be useful for readmission prediction of HF patients.

Background: Sleep staging is essential for diagnosing sleep disorders and managing sleep health. Traditional methods require time-consuming manual scoring. Recent PPG-based deep learning models perform well on local datasets but struggle with external generalization due to data drift. Methods: This study evaluates multi-source domain training for improving out-of-distribution generalization in four-class sleep staging (wake, light, deep, REM) from raw PPG time-series. The trained deep learning model is denoted SleepPPG-Net2. Additionally, we examined the impact of demographic factors, ethnicity, and obstructive sleep apnea on performance. SleepPPG-Net2 was benchmarked against two state-of-the-art models. Results: SleepPPG-Net2 outperformed benchmark models, improving generalization performance (Cohen’s kappa) by up to 21%. Performance disparities were observed in relation to age, sex, and obstructive sleep apnea severity. Conclusion: SleepPPG-Net2 enhances PPG-based sleep staging and provides insights into demographic and clinical influences on model performance.

Photoplethysmography (PPG) is a simple and low-cost optical technique that can be used to detect blood volume changes in the microvascular bed of tissue. It is often used non-invasively to make measurements at the skin surface. The PPG waveform comprises a pulsatile (‘AC’) physiological waveform attributed to cardiac synchronous changes in the blood volume with each heart beat, and is superimposed on a slowly varying (‘DC’) baseline with various lower frequency components attributed to respiration, sympathetic nervous system activity and thermoregulation. Although the origins of the components of the PPG signal are not fully understood, it is generally accepted that they can provide valuable information about the cardiovascular system. There has been a resurgence of interest in the technique in recent years, driven by the demand for low cost, simple and portable technology for the primary care and community based clinical settings, the wide availability of low cost and small semiconductor components, and the advancement of computer-based pulse wave analysis techniques. The PPG technology has been used in a wide range of commercially available medical devices for measuring oxygen saturation, blood pressure and cardiac output, assessing autonomic function and also detecting peripheral vascular disease. The introductory sections of the topical review describe the basic principle of operation and interaction of light with tissue, early and recent history of PPG, instrumentation, measurement protocol, and pulse wave analysis. The review then focuses on the applications of PPG in clinical physiological measurements, including clinical physiological monitoring, vascular assessment and autonomic function.

Near-infrared imaging (NIRI) is a neuroimaging technique which enables us to non-invasively measure hemodynamic changes in the human brain. Since the technique is very sensitive, the movement of a subject can cause movement artifacts (MAs), which affect the signal quality and results to a high degree. No general method is yet available to reduce these MAs effectively. The aim was to develop a new MA reduction method. A method based on moving standard deviation and spline interpolation was developed. It enables the semi-automatic detection and reduction of MAs in the data. It was validated using simulated and real NIRI signals. The results show that a significant reduction of MAs and an increase in signal quality are achieved. The effectiveness and usability of the method is demonstrated by the improved detection of evoked hemodynamic responses. The present method can not only be used in the postprocessing of NIRI signals but also for other kinds of data containing artifacts, for example ECG or EEG signals.

In the past few decades, analysis of heart sound signals (i.e. the phonocardiogram or PCG), especially for automated heart sound segmentation and classification, has been widely studied and has been reported to have the potential value to detect pathology accurately in clinical applications. However, comparative analyses of algorithms in the literature have been hindered by the lack of high-quality, rigorously validated, and standardized open databases of heart sound recordings. This paper describes a public heart sound database, assembled for an international competition, the PhysioNet/Computing in Cardiology (CinC) Challenge 2016. The archive comprises nine different heart sound databases sourced from multiple research groups around the world. It includes 2435 heart sound recordings in total collected from 1297 healthy subjects and patients with a variety of conditions, including heart valve disease and coronary artery disease. The recordings were collected from a variety of clinical or nonclinical (such as in-home visits) environments and equipment. The length of recording varied from several seconds to several minutes. This article reports detailed information about the subjects/patients including demographics (number, age, gender), recordings (number, location, state and time length), associated synchronously recorded signals, sampling frequency and sensor type used. We also provide a brief summary of the commonly used heart sound segmentation and classification methods, including open source code provided concurrently for the Challenge. A description of the PhysioNet/CinC Challenge 2016, including the main aims, the training and test sets, the hand corrected annotations for different heart sound states, the scoring mechanism, and associated open source code are provided. In addition, several potential benefits from the public heart sound database are discussed.

Functional near-infrared spectroscopy (fNIRS) is a powerful tool for monitoring brain functional activities. Due to its non-invasive and non-restraining nature, fNIRS has found broad applications in brain functional studies. However, for fNIRS to work well, it is important to reduce its sensitivity to motion artifacts. We propose a new wavelet-based method for removing motion artifacts from fNIRS signals. The method relies on differences between artifacts and fNIRS signal in terms of duration and amplitude and is specifically designed for spike artifacts. We assume a Gaussian distribution for the wavelet coefficients corresponding to the underlying hemodynamic signal in detail levels and identify the artifact coefficients using this distribution. An input parameter controls the intensity of artifact attenuation in trade-off with the level of distortion introduced in the signal. The method only modifies wavelet coefficients in levels adaptively selected based on the degree of contamination with motion artifact. To demonstrate the feasibility of the method, we tested it on experimental fNIRS data collected from three infant subjects. Normalized mean-square error and artifact energy attenuation were used as criteria for performance evaluation. The results show 18.29 and 16.42 dB attenuation in motion artifacts energy for 700 and 830 nm wavelength signals in a total of 29 motion events with no more than −16.7 dB distortion in terms of normalized mean-square error in the artifact-free regions of the signal.

This review describes the features of modern infrared imaging technology and the standardization protocols for thermal imaging in medicine. The technique essentially uses naturally emitted infrared radiation from the skin surface. Recent studies have investigated the influence of equipment and the methods of image recording. The credibility and acceptance of thermal imaging in medicine is subject to critical use of the technology and proper understanding of thermal physiology. Finally, we review established and evolving medical applications for thermal imaging, including inflammatory diseases, complex regional pain syndrome and Raynaud's phenomenon. Recent interest in the potential applications for fever screening is described, and some other areas of medicine where some research papers have included thermal imaging as an assessment modality. In certain applications thermal imaging is shown to provide objective measurement of temperature changes that are clinically significant.

Remote photoplethysmography (rPPG) enables contact-free monitoring of the blood volume pulse using a color camera. Essentially, it detects the minute optical absorption changes caused by blood volume variations in the skin. In this paper, we show that the different absorption spectra of arterial blood and bloodless skin cause the variations to occur along a very specific vector in a normalized RGB-space. The exact vector can be determined for a given light spectrum and for given transfer characteristics of the optical filters in the camera. We show that this ‘signature’ can be used to design an rPPG algorithm with a much better motion robustness than the recent methods based on blind source separation, and even better than the chrominance-based methods we published earlier. Using six videos recorded in a gym, with four subjects exercising on a range of fitness devices, we confirm the superior motion robustness of our newly proposed rPPG methods. A simple peak detector in the frequency domain returns the correct pulse-rate for 68% of total measurements compared to 60% for the best previous method, while the SNR of the pulse-signal improves from − 5 dB to − 4 dB. For a large population of 117 stationary subjects we prove that the accuracy is comparable to the best previous method, although the SNR of the pulse-signal drops from + 8.4 dB to + 7.6 dB. We expect the improved motion robustness to significantly widen the application scope of the rPPG-technique.

The effect of an active reference in EEG recording is one of the oldest technical problems in EEG practice. In this paper, a method is proposed to approximately standardize the reference of scalp EEG recordings to a point at infinity. This method is based on the fact that the use of scalp potentials to determine the neural electrical activities or their equivalent sources does not depend on the reference, so we may approximately reconstruct the equivalent sources from scalp EEG recordings with a scalp point or average reference. Then the potentials referenced at infinity are approximately reconstructed from the equivalent sources. As a point at infinity is far from all the possible neural sources, this method may be considered as a reference electrode standardization technique (REST). The simulation studies performed with assumed neural sources included effects of electrode number, volume conductor model and noise on the performance of REST, and the significance of REST in EEG temporal analysis. The results showed that REST is potentially very effective for the most important superficial cortical region and the standardization could be especially important in recovering the temporal information of EEG recordings.

Objective: Vast 12-lead ECGs repositories provide opportunities to develop new machine learning approaches for creating accurate and automatic diagnostic systems for cardiac abnormalities. However, most 12-lead ECG classification studies are trained, tested, or developed in single, small, or relatively homogeneous datasets. In addition, most algorithms focus on identifying small numbers of cardiac arrhythmias that do not represent the complexity and difficulty of ECG interpretation. This work addresses these issues by providing a standard, multi-institutional database and a novel scoring metric through a public competition: the PhysioNet/Computing in Cardiology Challenge 2020. Approach: A total of 66 361 12-lead ECG recordings were sourced from six hospital systems from four countries across three continents; 43 101 recordings were posted publicly with a focus on 27 diagnoses. For the first time in a public competition, we required teams to publish open-source code for both training and testing their algorithms, ensuring full scientific reproducibility. Main results: A total of 217 teams submitted 1395 algorithms during the Challenge, representing a diversity of approaches for identifying cardiac abnormalities from both academia and industry. As with previous Challenges, high-performing algorithms exhibited significant drops ($\lessapprox$10%) in performance on the hidden test data. Significance: Data from diverse institutions allowed us to assess algorithmic generalizability. A novel evaluation metric considered different misclassification errors for different cardiac abnormalities, capturing the outcomes and risks of different diagnoses. Requiring both trained models and code for training models improved the generalizability of submissions, setting a new bar in reproducibility for public data science competitions.

Electrical impedance tomography (EIT) is an attractive method for clinically monitoring patients during mechanical ventilation, because it can provide a non-invasive continuous image of pulmonary impedance which indicates the distribution of ventilation. However, most clinical and physiological research in lung EIT is done using older and proprietary algorithms; this is an obstacle to interpretation of EIT images because the reconstructed images are not well characterized. To address this issue, we develop a consensus linear reconstruction algorithm for lung EIT, called GREIT (Graz consensus Reconstruction algorithm for EIT). This paper describes the unified approach to linear image reconstruction developed for GREIT. The framework for the linear reconstruction algorithm consists of (1) detailed finite element models of a representative adult and neonatal thorax, (2) consensus on the performance figures of merit for EIT image reconstruction and (3) a systematic approach to optimize a linear reconstruction matrix to desired performance measures. Consensus figures of merit, in order of importance, are (a) uniform amplitude response, (b) small and uniform position error, (c) small ringing artefacts, (d) uniform resolution, (e) limited shape deformation and (f) high resolution. Such figures of merit must be attained while maintaining small noise amplification and small sensitivity to electrode and boundary movement. This approach represents the consensus of a large and representative group of experts in EIT algorithm design and clinical applications for pulmonary monitoring. All software and data to implement and test the algorithm have been made available under an open source license which allows free research and commercial use.

Objective. Pulse oximetry is a non-invasive optical technique used to measure arterial oxygen saturation (SpO₂) in a variety of clinical settings and scenarios. Despite being one the most significant technological advances in health monitoring over the last few decades, there have been reports on its various limitations. Recently due to the Covid-19 pandemic, questions about pulse oximeter technology and its accuracy when used in people with different skin pigmentation have resurfaced, and are to be addressed. Approach. This review presents an introduction to the technique of pulse oximetry including its basic principle of operation, technology, and limitations, with a more in depth focus on skin pigmentation. Relevant literature relating to the performance and accuracy of pulse oximeters in populations with different skin pigmentation are evaluated. Main Results. The majority of the evidence suggests that the accuracy of pulse oximetry differs in subjects of different skin pigmentations to a level that requires particular attention, with decreased accuracy in patients with dark skin. Significance. Some recommendations, both from the literature and contributions from the authors, suggest how future work could address these inaccuracies to potentially improve clinical outcomes. These include the objective quantification of skin pigmentation to replace currently used qualitative methods, and computational modelling for predicting calibration algorithms based on skin colour.