Evaluation of the accuracy and reliability for photoplethysmography based heart rate and beat-to-beat detection during daily activities

Introducing the Pi-CON Methodology to Overcome Usability Deficits during Remote Patient Monitoring

The adoption of telehealth has soared, and with that the acceptance of Remote Patient Monitoring (RPM) and virtual care. A review of the literature illustrates, however, that poor device usability can impact the generated data when using Patient-Generated Health Data (PGHD) devices, such as wearables or home use medical devices, when used outside a health facility. The Pi-CON methodology is introduced to overcome these challenges and guide the definition of user-friendly and intuitive devices in the future. Pi-CON stands for passive, continuous, and non-contact, and describes the ability to acquire health data, such as vital signs, continuously and passively with limited user interaction and without attaching any sensors to the patient. The paper highlights the advantages of Pi-CON by leveraging various sensors and techniques, such as radar, remote photoplethysmography, and infrared. It illustrates potential concerns and discusses future applications Pi-CON could be used for, including gait and fall monitoring by installing an omnipresent sensor based on the Pi-CON methodology. This would allow automatic data collection once a person is recognized, and could be extended with an integrated gateway so multiple cameras could be installed to enable data feeds to a cloud-based interface, allowing clinicians and family members to monitor patient health status remotely at any time.

From lab to life: Evaluating the reliability and validity of psychophysiological data from wearable devices in laboratory and ambulatory settings

Despite the increasing popularity of ambulatory assessment, the reliability and validity of psychophysiological signals from wearable devices is unproven in daily life settings. We evaluated the reliability and validity of physiological signals (electrocardiogram, ECG; photoplethysmography, PPG; electrodermal activity, EDA) collected from two wearable devices (Movisens EcgMove4 and Empatica E4) in the lab (N = 67) and daily life (N = 20) among adults aged 18-64 with Mindware as the laboratory gold standard. Results revealed that both wearable devices' valid data rates in daily life were lower than in the laboratory (Movisens ECG 82.94 vs. 93.10%, Empatica PPG 8.79 vs. 26.14%, and Empatica EDA 41.16 vs. 42.67%, respectively). The poor valid data rates of Empatica PPG signals in the laboratory could be partially attributed to participants' hand movements (r = - .27, p = .03). In laboratory settings, heart rate (HR) derived from both wearable devices exhibited higher concurrent validity than heart rate variability (HRV) metrics (ICCs 0.98-1.00 vs. 0.75-0.97). The number of skin conductance responses (SCRs) derived from Empatica showed higher concurrent validity than skin conductance level (SCL, ICCs 0.38 vs. 0.09). Movisens EcgMove4 provided more reliable and valid HRV measurements than Empatica E4 in both laboratory (split-half reliability: 0.95-0.99 vs. 0.85-0.98; concurrent validity: 0.95-1.00 vs. 0.75-0.98; valid data rate: 93.10 vs. 26.14%) and ambulatory settings (split-half reliability: 0.99-1.00 vs. 0.89-0.98; valid data rate: 82.94 vs. 8.79%). Although the reliability and validity of wearable devices are improving, findings suggest researchers should select devices that yield consistently robust and valid data for their measures of interest.

Undersampling and cumulative class re-decision methods to improve detection of agitation in people with dementia

Agitation is one of the most prevalent symptoms in people with dementia (PwD) that can place themselves and the caregiver's safety at risk. Developing objective agitation detection approaches is important to support health and safety of PwD living in a residential setting. In a previous study, we collected multimodal wearable sensor data from 17 participants for 600 days and developed machine learning models for detecting agitation in 1-min windows. However, there are significant limitations in the dataset, such as imbalance problem and potential imprecise labels as the occurrence of agitation is much rarer in comparison to the normal behaviours. In this paper, we first implemented different undersampling methods to eliminate the imbalance problem, and came to the conclusion that only 20% of normal behaviour data were adequate to train a competitive agitation detection model. Then, we designed a weighted undersampling method to evaluate the manual labeling mechanism given the ambiguous time interval assumption. After that, the postprocessing method of cumulative class re-decision (CCR) was proposed based on the historical sequential information and continuity characteristic of agitation, improving the decision-making performance for the potential application of agitation detection system. The results showed that a combination of undersampling and CCR improved F1-score and other metrics to varying degrees with less training time and data.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13534-023-00313-8.

Going Remote: A Revised Study Protocol for a Pilot Randomized Controlled Trial for Biofeedback Treatment of Anxiety Associated With Chronic Spinal Cord Injury

Background

The incidence of anxiety in adults with spinal cord injury/disorder (SCI/D) exceeds that of the general population. Heart rate variability (HRV) biofeedback training is a potential treatment associated with a reduction in stress and anxiety, however HRV training has not been explored in the SCI/D population.

Objectives

To describe a modified protocol piloting HRV training to reduce anxiety associated with SCI/D and detail the COVID-19-related modifications.

Methods

To test the feasibility of the biofeedback treatment, 30 adults with SCI/D will complete this pilot randomized controlled trial. Enrollment started in January 2020, halted in March 2020 due to the COVID-19 pandemic, and resumed in March 2021 with a modified protocol. Protocol modifications are documented using the Framework for Reporting Adaptations and Modifications (FRAME). Participants are allocated to the treatment or control arm and undergo eight sessions of physiological monitoring at home using a commercially available HRV sensor and mobile application, which also delivers biofeedback training for those in the treatment arm. Surveys are administered following each session to capture self-reported stress, anxiety, and mood. The study is approved by the HCA-HealthONE institutional review board and is registered with clinicaltrials.gov (NCT# 03975075).

Conclusion

COVID-19 has changed the research landscape, forcing scientists to rethink their study designs to address patient and staff safety in this new context. Our modified protocol accomplished this by moving the treatment setting and delivery out of the clinic and into the home. In doing so, we address patient and staff safety, increase external validity, and reduce participant burden.

Heart Rate Modeling and Prediction Using Autoregressive Models and Deep Learning

Physiological time series are affected by many factors, making them highly nonlinear and nonstationary. As a consequence, heart rate time series are often considered difficult to predict and handle. However, heart rate behavior can indicate underlying cardiovascular and respiratory diseases as well as mood disorders. Given the importance of accurate modeling and reliable predictions of heart rate fluctuations for the prevention and control of certain diseases, it is paramount to identify models with the best performance in such tasks. The objectives of this study were to compare the results of three different forecasting models (Autoregressive Model, Long Short-Term Memory Network, and Convolutional Long Short-Term Memory Network) trained and tested on heart rate beats per minute data obtained from twelve heterogeneous participants and to identify the architecture with the best performance in terms of modeling and forecasting heart rate behavior. Heart rate beats per minute data were collected using a wearable device over a period of 10 days from twelve different participants who were heterogeneous in age, sex, medical history, and lifestyle behaviors. The goodness of the results produced by the models was measured using both the mean absolute error and the root mean square error as error metrics. Despite the three models showing similar performance, the Autoregressive Model gave the best results in all settings examined. For example, considering one of the participants, the Autoregressive Model gave a mean absolute error of 2.069 (compared to 2.173 of the Long Short-Term Memory Network and 2.138 of the Convolutional Long Short-Term Memory Network), achieving an improvement of 5.027% and 3.335%, respectively. Similar results can be observed for the other participants. The findings of the study suggest that regardless of an individual's age, sex, and lifestyle behaviors, their heart rate largely depends on the pattern observed in the previous few minutes, suggesting that heart rate can be reasonably regarded as an autoregressive process. The findings also suggest that minute-by-minute heart rate prediction can be accurately performed using a linear model, at least in individuals without pathologies that cause heartbeat irregularities. The findings also suggest many possible applications for the Autoregressive Model, in principle in any context where minute-by-minute heart rate prediction is required (arrhythmia detection and analysis of the response to training, among others).

Wearable Devices Suitable for Monitoring Twenty Four Hour Heart Rate Variability in Military Populations

Heart rate variability (HRV) measurements provide information on the autonomic nervous system and the balance between parasympathetic and sympathetic activity. A high HRV can be advantageous, reflecting the ability of the autonomic nervous system to adapt, whereas a low HRV can be indicative of fatigue, overtraining or health issues. There has been a surge in wearable devices that claim to measure HRV. Some of these include spot measurements, whilst others only record during periods of rest and/or sleep. Few are capable of continuously measuring HRV (≥24 h). We undertook a narrative review of the literature with the aim to determine which currently available wearable devices are capable of measuring continuous, precise HRV measures. The review also aims to evaluate which devices would be suitable in a field setting specific to military populations. The Polar H10 appears to be the most accurate wearable device when compared to criterion measures and even appears to supersede traditional methods during exercise. However, currently, the H10 must be paired with a watch to enable the raw data to be extracted for HRV analysis if users need to avoid using an app (for security or data ownership reasons) which incurs additional cost.

It is All in the Wrist: Wearable Sleep Staging in a Clinical Population versus Reference Polysomnography

PURPOSE

There is great interest in unobtrusive long-term sleep measurements using wearable devices based on reflective photoplethysmography (PPG). Unfortunately, consumer devices are not validated in patient populations and therefore not suitable for clinical use. Several sleep staging algorithms have been developed and validated based on ECG-signals. However, translation from these techniques to data derived by wearable PPG is not trivial, and requires the differences between sensing modalities to be integrated in the algorithm, or having the model trained directly with data obtained with the target sensor. Either way, validation of PPG-based sleep staging algorithms requires a large dataset containing both gold standard measurements and PPG-sensor in the applicable clinical population. Here, we take these important steps towards unobtrusive, long-term sleep monitoring.

METHODS

We developed and trained an algorithm based on wrist-worn PPG and accelerometry. The method was validated against reference polysomnography in an independent clinical population comprising 244 adults and 48 children (age: 3 to 82 years) with a wide variety of sleep disorders.

RESULTS

The classifier achieved substantial agreement on four-class sleep staging with an average Cohen's kappa of 0.62 and accuracy of 76.4%. For children/adolescents, it achieved even higher agreement with an average kappa of 0.66 and accuracy of 77.9%. Performance was significantly higher in non-REM parasomnias (kappa = 0.69, accuracy = 80.1%) and significantly lower in REM parasomnias (kappa = 0.55, accuracy = 72.3%). A weak correlation was found between age and kappa (ρ = -0.30, p<0.001) and age and accuracy (ρ = -0.22, p<0.001).

CONCLUSION

This study shows the feasibility of automatic wearable sleep staging in patients with a broad variety of sleep disorders and a wide age range. Results demonstrate the potential for ambulatory long-term monitoring of clinical populations, which may improve diagnosis, estimation of severity and follow up in both sleep medicine and research.

Direct application of an ECG-based sleep staging algorithm on reflective photoplethysmography data decreases performance

OBJECTIVE

The maturation of neural network-based techniques in combination with the availability of large sleep datasets has increased the interest in alternative methods of sleep monitoring. For unobtrusive sleep staging, the most promising algorithms are based on heart rate variability computed from inter-beat intervals (IBIs) derived from ECG-data. The practical application of these algorithms is even more promising when alternative ways of obtaining IBIs, such as wrist-worn photoplethysmography (PPG) can be used. However, studies validating sleep staging algorithms directly on PPG-based data are limited.

RESULTS

We applied an automatic sleep staging algorithm trained and validated on ECG-data directly on inter-beat intervals derived from a wrist-worn PPG sensor, in 389 polysomnographic recordings of patients with a variety of sleep disorders. While the algorithm reached moderate agreement with gold standard polysomnography, the performance was significantly lower when applied on PPG- versus ECG-derived heart rate variability data (kappa 0.56 versus 0.60, p < 0.001; accuracy 73.0% versus 75.9% p < 0.001). These results show that direct application of an algorithm on a different source of data may negatively affect performance. Algorithms need to be validated using each data source and re-training should be considered whenever possible.

Effects of an Unexpected and Expected Event on Older Adults' Autonomic Arousal and Eye Fixations During Autonomous Driving

Driving cessation for some older adults can exacerbate physical, cognitive, and mental health challenges due to loss of independence and social isolation. Fully autonomous vehicles may offer an alternative transport solution, increasing social contact and encouraging independence. However, there are gaps in understanding the impact of older adults’ passive role on safe human–vehicle interaction, and on their well-being. 37 older adults (mean age ± SD = 68.35 ± 8.49 years) participated in an experiment where they experienced fully autonomous journeys consisting of a distinct stop (an unexpected event versus an expected event). The autonomous behavior of the vehicle was achieved using the Wizard of Oz approach. Subjective ratings of trust and reliability, and driver state monitoring including visual attention strategies (fixation duration and count) and physiological arousal (skin conductance and heart rate), were captured during the journeys. Results revealed that subjective trust and reliability ratings were high after journeys for both types of events. During an unexpected stop, overt visual attention was allocated toward the event, whereas during an expected stop, visual attention was directed toward the human–machine interface (HMI) and distributed across the central and peripheral driving environment. Elevated skin conductance level reflecting increased arousal persisted only after the unexpected event. These results suggest that safety-critical events occurring during passive fully automated driving may narrow visual attention and elevate arousal mechanisms. To improve in-vehicle user experience for older adults, a driver state monitoring system could examine such psychophysiological indices to evaluate functional state and well-being. This information could then be used to make informed decisions on vehicle behavior and offer reassurance during elevated arousal during unexpected events.

Validating Measures of Electrodermal Activity and Heart Rate Variability Derived From the Empatica E4 Utilized in Research Settings That Involve Interactive Dyadic States

Portable and wireless devices that collect physiological data are becoming more and more sought after in clinical and psychophysiological research as technology swiftly advances. These devices allow for data collection in interactive states, such as dyadic therapy, with reduced restraints compared to traditional laboratory devices. One such portable device is the Empatica E4 wristband (Empatica Srl, Milan, Italy) which allows quantifying cardiac interbeat intervals (IBIs), heart rate variability (HRV), and electro-dermal activity (EDA), as well as several other acceleration and temperature measures. In the current study, we aimed to assess IBI, HRV, and EDA measures, against the same data collected from the well-validated MindWare mobile impedance cardiograph device (MindWare Technology, Gahanna, OH, United States). We assessed the E4 strictly as a research instrument and not as a clinical tool. We were specifically interested in the wristbands’ performance during naturalistic interactive face-to-face conversations which inherently involve more hand movements. We collected data from 30 participants, nested in 15 dyads, which were connected to both devices simultaneously, during rest and during a social conversation. After preprocessing and analyses, we found that mean IBIs obtained by the E4 and the MindWare device, were highly similar during rest and during conversation. Medium to high correlations were found between the devices with respect to several HRV measures, with higher correlations during rest compared to conversation. The E4 failed to produce reliable EDA data. We conclude by discussing the strengths and limitations of the E4 during seated conversational states and suggest optimal ways to collect and analyze data with the E4.

Objective stress monitoring based on wearable sensors in everyday settings

Monitoring people's stress levels has become an essential part of behavioural studies for physical and mental illnesses conducted within the biopsychosocial framework. There have been several stress assessment studies in laboratory-based controlled settings. However, the results of these studies do not always translate effectively to an everyday context. The current state of wearable sensor technology allows us to develop systems measuring the physiological signals reflecting stress 24/7 while capturing the context. In this paper, we present a stress monitoring system that provides objective daily stress measurements in everyday settings based on three physiological signals: electrocardiogram (ECG), photoplethysmogram (PPG), and galvanic skin response (GSR) using Shimmer3 ECG, Shimmer3 GSR+, and Empatica E4 wearable sensors. We perform controlled stress assessment experiments on 17 participants in which we successfully detect stress with a 94.55% accuracy for 10-fold cross-validation and an 85.71% accuracy for subject-wise cross-validation. In everyday settings, the system assesses stress with an 81.82% accuracy. We also examine whether motion artefacts affect stress assessment and filter the low-confidence readings to minimise false alarms.

Validity of Wrist-Worn photoplethysmography devices to measure heart rate: A systematic review and meta-analysis

Heart rate (HR), when combined with accelerometry, can dramatically improve estimates of energy expenditure and sleep. Advancements in technology, via the development and introduction of small, low-cost photoplethysmography devices embedded within wrist-worn consumer wearables, have made the collection of heart rate (HR) under free-living conditions more feasible. This systematic review and meta-analysis compared the validity of wrist-worn HR estimates to a criterion measure of HR (electrocardiography ECG or chest strap). Searches of PubMed/Medline, Web of Science, EBSCOhost, PsycINFO, and EMBASE resulted in a total of 44 articles representing 738 effect sizes across 15 different brands. Multi-level random effects meta-analyses resulted in a small mean difference (beats per min, bpm) of -0.40 bpm (95 confidence interval (CI) -1.64 to 0.83) during sleep, -0.01 bpm (-0.02 to 0.00) during rest, -0.51 bpm (-1.60 to 0.58) during treadmill activities (walking to running), while the mean difference was larger during resistance training (-7.26 bpm, -10.46 to -4.07) and cycling (-4.55 bpm, -7.24 to -1.87). Mean difference increased by 3 bpm (2.5 to 3.5) per 10 bpm increase of HR for resistance training. Wrist-worn devices that measure HR demonstrate acceptable validity compared to a criterion measure of HR for most common activities.

Preliminary Comparison of Zero-Gravity Chair With Tilt Table in Relation to Heart Rate Variability Measurements

Heart rate variability (HRV) measurements are performed using a tilt-table (TT) to diagnose dysfunctionality in the autonomic nervous system (ANS) and the cardiovascular system. To maintain homeostasis, the ANS adapts to body position changes through alterations in sympathetic and parasympathetic responses that can be quantified by extracting time-domain and frequency-domain parameters from the heart rate signal. When body position is changed from supine to erect, a healthy subject’s response also shows changes in ANS activity. However, TT can be unsafe or uncomfortable for elderly or overweight subjects. Furthermore, it may induce anxiety which alters the HRV measurements. This study proposes an alternative strategy to replace the TT with a zero-gravity chair (ZGC). The statistical analysis between HRV parameters from the TT and the ZGC shows that ZGC can be a feasible alternative to TT. Therefore, ZGC can be used as a more convenient, secure, stable and safer option to the traditional HRV analysis with TT.

Pulse Rate Variability in Emergency Physicians During Shifts: Pilot Cross-Sectional Study

Background

The high prevalence of physician burnout, particularly in emergency medicine, has garnered national attention in recent years. Objective means of measuring stress while at work can facilitate research into stress reduction interventions, and wearable photoplethysmography (PPG) technology has been proposed as a potential solution. However, the use of low-burden wearable biosensors to study training and clinical practice among emergency physicians (EP) remains untested.

Objective

This pilot study aimed to (1) determine the feasibility of recording on-shift photoplethysmographic data from EP, (2) assess the quality of these data, and (3) calculate standard pulse rate variability (PRV) metrics from the acquired dataset and examine patterns in these variables over the course of an academic year.

Methods

A total of 21 EP wore PPG biosensors on their wrists during clinical work in the emergency department during a 9-hour shift. Recordings were collected during the first quarter of the academic year, then again during the fourth quarter of the same year for comparison. The overall rate of usable data collection per time was computed. Standard pulse rate (PR) and PRV metrics from these two time points were calculated and entered into Student t tests.

Results

More than 400 hours of data were entered into these analyses. Interpretable data were captured during 8.54% of the total recording time overall. In the fourth quarter of the academic year compared with the first quarter, there was no significant difference in median PR (75.8 vs 76.8; P=.57), mean R-R interval (0.81 vs 0.80; P=.32), SD of R-R interval (0.11 vs 0.11; P=.93), root mean square of successive difference of R-R interval (0.81 vs 0.80; P=.96), low-frequency power (3.5×103 vs 3.4×103; P=.79), high-frequency power (8.5×103 vs 8.3×103; P=.91), or low-frequency to high-frequency ratio (0.42 vs 0.41; P=.43), respectively. Power estimates for each of these tests exceeded .90. A secondary analysis of the resident-only subgroup similarly showed no significant differences over time, despite power estimates greater than .80.

Conclusions

Although the use of PPG biosensors to record real-time physiological data from EP while providing clinical care seems operationally feasible, this study fails to support the notion that such an approach can efficiently provide reliable estimates of metrics of interest. No significant differences in PR or PRV metrics were found at the end of the year compared with the beginning. Although these methods may offer useful applications to other domains, it may currently have limited utility in the contexts of physician training and wellness.

Stressing the accuracy: Wrist-worn wearable sensor validation over different conditions

Wearable sensors are promising instruments for conducting both laboratory and ambulatory research in psychophysiology. However, scholars should be aware of their measurement error and the conditions in which accuracy is achieved. This study aimed to assess the accuracy of a wearable sensor designed for research purposes, the E4 wristband (Empatica, Milan, Italy), in measuring heart rate (HR), heart rate variability (HRV), and skin conductance (SC) over five laboratory conditions widely used in stress reactivity research (seated rest, paced breathing, orthostatic, Stroop, speech task) and two ecological conditions (slow walking, keyboard typing). Forty healthy participants concurrently wore the wristband and two gold standard measurement systems (i.e., electrocardiography and finger SC sensor). The wristband accuracy was determined by evaluating the signal quality and the correlations with and the Bland-Altman plots against gold standard-derived measurements. Moreover, exploratory analyses were performed to assess predictors of measurement error. Mean HR measures showed the best accuracy over all conditions. HRV measures showed satisfactory accuracy in seated rest, paced breathing, and recovery conditions but not in dynamic conditions, including speaking. Accuracy was diminished by wrist movements, cognitive and emotional stress, nonstationarity, and larger wrist circumferences. Wrist SC measures showed neither correlation nor visual resemblance with finger SC signal, suggesting that the two sites may reflect different phenomena. Future studies are needed to assess the responsivity of wrist SC to emotional and cognitive stress. Limitations and implications for laboratory and ambulatory research are discussed.

Occupational physical activity in brewery and office workers

Active lifestyles are beneficial to health and well-being but our workplaces may not be inherently supportive of physical activity at work. With the increasing use of technology in the workplace, many jobs are becoming more sedentary. The purpose of this study was to characterize levels of occupational physical activity (OPA) among active and sedentary workers. Two types of activity trackers (Fitbit Charge HR and Hexoskin) were used to assess activity measures (steps, heart rate, and energy expenditure) among workers during one full work shift. The first objective of the study was to assess the agreement between two types of accelerometer-based activity trackers as measures of occupational physical activity. The second objective of this study was to assess differences in measures of OPA among workers in generally physically active (brewery) and sedentary (office) work environments. Occupational physical activity data were collected from 50 workers in beer-brewing tasks and 51 workers in office work tasks. The 101 subjects were from the brewing service sector, a call center, and an engineering office within a manufacturing facility. A two-factor repeated measures analysis of variance (ANOVA) was used to assess the two activity tracking devices while two-sample t-tests were used to compare the two worker groups. There were statistically significant differences in total steps and mean heart rate between the two devices. When comparing the two groups of workers there were statistically significant differences in measures of steps, mean heart rate, and energy expenditure. The results of the present study provide quantitative evidence that levels of OPA should be identified for different work groups.

Towards a wearable sensor system for continuous occupational cold stress assessment

This study investigated the usefulness of continuous sensor data for improving occupational cold stress assessment. Eleven volunteer male subjects completed a 90-120-min protocol in cold environments, consisting of rest, moderate and hard work. Biomedical data were measured using a smart jacket with integrated temperature, humidity and activity sensors, in addition to a custom-made sensor belt worn around the chest. Other relevant sensor data were measured using commercially available sensors. The study aimed to improve decision support for workers in cold climates, by taking advantage of the information provided by data from the rapidly growing market of wearable sensors. Important findings were that the subjective thermal sensation did not correspond to the measured absolute skin temperature and that large differences were observed in both metabolic energy production and skin temperatures under identical exposure conditions. Temperature, humidity, activity and heart rate were found to be relevant parameters for cold stress assessment, and the locations of the sensors in the prototype jacket were adequate. The study reveals the need for cold stress assessment and indicates that a generalised approached is not sufficient to assess the stress on an individual level.