Search

The rapidly increasing prevalence of debilitating breathing disorders, such as chronic obstructive pulmonary disease (COPD), calls for a meaningful integration of artificial intelligence (AI) into respiratory healthcare. Deep learning techniques are “data hungry” whilst patient-based data is invariably expensive and time consuming to record. To this end, we introduce a novel COPD-simulator, a physical apparatus with an easy to replicate design which enables rapid and effective generation of a wide range of COPD-like data from healthy subjects, for enhanced training of deep learning frameworks. To ensure the faithfulness of our domain-aware COPD surrogates, the generated waveforms are examined through both flow waveforms and photoplethysmography (PPG) waveforms (as a proxy for intrathoracic pressure) in terms of duty cycle, sample entropy, FEV1/FVC ratios and flow-volume loops. The proposed simulator operates on healthy subjects and is able to generate FEV1/FVC obstruction ratios ranging from greater than 0.8 to less than 0.2, mirroring values that can observed in the full spectrum of real-world COPD. As a final stage of verification, a simple convolutional neural network is trained on surrogate data alone, and is used to accurately detect COPD in real-world patients. When training solely on surrogate data, and testing on real-world data, a comparison of true positive rate against false positive rate yields an area under the curve of 0.75, compared with 0.63 when training solely on real-world data.

Abstract Background Insomnia, eating disorders, heart problems and even strokes are just some of the illnesses that reveal the negative impact of stress overload on health and well-being. Early detection of stress is therefore of utmost importance. Whereas the gold-standard for detecting stress is by means of questionnaires, more recent work uses wearable sensors to find continuous and qualitative physical markers of stress. As some physiological stress responses, e.g. increased heart rate or sweating and chills, might also occur when doing sports, a more profound approach is needed for stress detection than purely considering physiological data. Methods In this paper, we analyse the added value of context information during stress detection from wearable data. We do so by comparing the performance of models trained purely on physiological data and models trained on physiological and context data. We consider the user’s activity and hours of sleep as context information, where we compare the influence of user-given context versus machine learning derived context. Results Context-aware models reach higher accuracy and lower standard deviations in comparison to the baseline (physiological) models. We also observe higher accuracy and improved weighted F1 score when incorporating machine learning predicted, instead of user-given, activities as context information. Conclusions In this paper we show that considering context information when performing stress detection from wearables leads to better performance. We also show that it is possible to move away from human labeling and rely only on the wearables for both physiology and context.

Given that the prevalence of smartwatches has allowed them to become a hallmark in health monitoring, they are primed to provide accessible body composition estimations. The purpose of this study was to evaluate the precision and agreement of smartwatch-based bioimpedance analysis (SW-BIA) and multifrequency bioimpedance analysis (MFBIA) against a 4-compartment (4C) model criterion. A total of 186 participants (114 females) underwent body composition assessments necessary for a 4C model and SW-BIA and MFBIA. Values of total body water (TBW) from each device were compared with those obtained from bioimpedance spectroscopy. Precision was adequate though slightly lower for the smartwatch compared with other methods. No device demonstrated equivalence with the 4C model. Specifically, the SW-BIA overestimated and MFBIA underestimated body fat. MFBIA, but not SW-BIA, demonstrated equivalence for TBW. Overall error was higher for males using the smartwatch compared with females. While these findings do not invalidate the use of smartwatch-based estimates, clinicians should consider that there may be large errors relative to clinical measures. If this wearable device is intended to be used to monitor body composition change over time, these findings demonstrate the need for future research to evaluate its accuracy during follow-up testing. [ABSTRACT FROM AUTHOR]

Background: Insomnia, eating disorders, heart problems and even strokes are just some of the illnesses that reveal the negative impact of stress overload on health and well-being. Early detection of stress is therefore of utmost importance. Whereas the gold-standard for detecting stress is by means of questionnaires, more recent work uses wearable sensors to find continuous and qualitative physical markers of stress. As some physiological stress responses, e.g. increased heart rate or sweating and chills, might also occur when doing sports, a more profound approach is needed for stress detection than purely considering physiological data.Methods: In this paper, we analyse the added value of context information during stress detection from wearable data. We do so by comparing the performance of models trained purely on physiological data and models trained on physiological and context data. We consider the user's activity and hours of sleep as context information, where we compare the influence of user-given context versus machine learning derived context.Results: Context-aware models reach higher accuracy and lower standard deviations in comparison to the baseline (physiological) models. We also observe higher accuracy and improved weighted F1 score when incorporating machine learning predicted, instead of user-given, activities as context information.Conclusions: In this paper we show that considering context information when performing stress detection from wearables leads to better performance. We also show that it is possible to move away from human labeling and rely only on the wearables for both physiology and context. [ABSTRACT FROM AUTHOR]

The objective of this study was to investigate the efficacy of pulse transit time (PTT)-pulse wave analysis (PWA) fusion in cuff-less blood pressure (BP) trend tracking based on the wearable wrist ballistocardiogram (BCG). We constructed PTT and 36 candidate BCG PWA features based on the BCG and photoplethysmogram (PPG) signals acquired at the wrist. We performed a model-based analysis to select 6 candidate BCG PWA features most sensitive to BP. We aggregated the 6 BCG PWA features into novel predictors of diastolic, systolic, and pulse BP (DP, SP, and PP) orthogonal to PTT by covariance-maximizing dimensionality reduction. Then, we evaluated and compared the efficacy of BCG PTT and BCG PTT-PWA fusion in tracking the trend of DP, SP, and PP. Unique innovations of this study, generalizable to a range of physiological signals beyond the BCG, are (i) the systematic selection of PWA features based on model-based analysis and (ii) the development of PWA-based BP predictors orthogonal to PTT. Using the experimental wrist BCG and PPG signals collected from 23 human subjects, we demonstrated that (i) BCG PTT-PWA fusion may significantly outperform BCG PTT; (ii) PTT may play the primary role in tracking BP while BCG PWA features may still have an independent and complementary value (especially in tracking PP); and (iii) model-based selection of BCG PWA features can enhance the robustness of BP trend tracking based on BCG PTT-PWA fusion.

Recently there has been great interest in photoplethysmogram signal processing. However, its minimally necessary sampling frequency for accurate heart rate variability parameters is ambiguous. In the present paper frequency-modulated 1.067 Hz cosine wave modelled the variable PPG in silico. The five-minute-long, 1 ms resolution master-signals were decimated (D) at 2-500 ms, then cubic spline interpolated (I) back to 1 ms resolution. The mean pulse rate, standard deviation, root mean square of successive pulse rate differences (RMSSD), and spectral components were computed by Varian 2.3 and compared to the master-series via relative accuracy error. Also Poincaré-plot morphology was assessed. Mean pulse rate is accurate down to 303 ms (D) and 400 ms (I). In low-variability series standard deviation required at least 5 ms (D) and 100 ms (I). RMSSD needed 10 ms (D), and 303 ms (I) in normal, whereas 2 ms (D) and 100 ms (I) in low- variability series. In the frequency domain 5 ms (D) and 100 ms (I) are required. 2 ms (D) and 100 ms (I) preserved the Poincaré-plot morphology. The minimal sampling frequency of PPG for accurate HRV analysis is higher than expected from the signal bandwidth and sampling theorem. Interpolation improves accuracy. The ratio of sampling error and expected variability should be considered besides the inherent sensitivity of the given parameter, the interpolation technique, and the pulse rate detection method.

This essay aims to highlight how the current state of medical biometrics produces a transitional state of being between measurement and data recollection in which there no longer is an analogon for the self, how this is affecting our traditional image of the narrative self, and consequently how it directly impacts our very belief in a life-story. [ABSTRACT FROM AUTHOR]

The rapidly increasing prevalence of debilitating breathing disorders, such as chronic obstructive pulmonary disease (COPD), calls for a meaningful integration of artificial intelligence (AI) into respiratory healthcare. Deep learning techniques are "data hungry" whilst patient-based data is invariably expensive and time consuming to record. To this end, we introduce a novel COPD-simulator, a physical apparatus with an easy to replicate design which enables rapid and effective generation of a wide range of COPD-like data from healthy subjects, for enhanced training of deep learning frameworks. To ensure the faithfulness of our domain-aware COPD surrogates, the generated waveforms are examined through both flow waveforms and photoplethysmography (PPG) waveforms (as a proxy for intrathoracic pressure) in terms of duty cycle, sample entropy, FEV 1 /FVC ratios and flow-volume loops. The proposed simulator operates on healthy subjects and is able to generate FEV 1 /FVC obstruction ratios ranging from greater than 0.8 to less than 0.2, mirroring values that can observed in the full spectrum of real-world COPD. As a final stage of verification, a simple convolutional neural network is trained on surrogate data alone, and is used to accurately detect COPD in real-world patients. When training solely on surrogate data, and testing on real-world data, a comparison of true positive rate against false positive rate yields an area under the curve of 0.75, compared with 0.63 when training solely on real-world data., (© 2024 The Authors.)

Much of the information related to breathing is contained within the photoplethysmography (PPG) signal, through changes in venous blood flow, heart rate and stroke volume. We aim to leverage this fact, by employing a novel deep learning framework which is a based on a repurposed convolutional autoencoder. Our corr-encoder model aims to encode all of the relevant respiratory information contained within photoplethysmography waveform, and decode it into a waveform that is similar to a gold standard respiratory reference — the Capnogram. The model is employed on two photoplethysmography data sets, namely Capnobase and BIDMC. We show that the model is capable of producing respiratory waveforms that approach the gold standard, while in turn producing state of the art respiratory rate estimates. We also show that when it comes to capturing more advanced respiratory waveform characteristics such as duty cycle, our model is for the most part unsuccessful. A suggested reason for this, in light of a previous study on in-ear PPG, is that the respiratory variations in finger-PPG are far weaker compared with other recording locations. Importantly, our model can perform these waveform estimates in a fraction of a millisecond, giving it the capacity to produce over 6 hours of respiratory waveforms in a single second. Moreover, we attempt to interpret the behaviour of the kernel weights within the model, showing that in part our model intuitively selects different breathing frequencies. The model proposed in this work could help to improve the usefulness of consumer PPG-based wearables for medical applications, where detailed respiratory information is required. • A novel convolutional Corr-encoder to extract respiratory information from PPG. • The low complexity model can produce hours of respiratory waveforms in one second. • The model is evaluated on two separate datasets, Capnobase and BIDMC. • Median respiratory rate error is 0.39BPM with Capnobase and 0.89BPM with BIDMC. [ABSTRACT FROM AUTHOR]

The non-invasive estimation of blood oxygen saturation (SpO2) by pulse oximetry is of vital importance clinically, from the detection of sleep apnea to the recent ambulatory monitoring of hypoxemia in the delayed post-infective phase of COVID-19. In this proof of concept study, we set out to establish the feasibility of SpO2 measurement from the ear canal as a convenient site for long term monitoring, and perform a comprehensive comparison with the right index finger—the conventional clinical measurement site. During resting blood oxygen saturation estimation, we found a root mean square difference of 1.47% between the two measurement sites, with a mean difference of 0.23% higher SpO2 in the right ear canal. Using breath holds, we observe the known phenomena of time delay between central circulation and peripheral circulation with a mean delay between the ear and finger of 12.4 s across all subjects. Furthermore, we document the lower photoplethysmogram amplitude from the ear canal and suggest ways to mitigate this issue. In conjunction with the well-known robustness to temperature induced vasoconstriction, this makes conclusive evidence for in-ear SpO2 monitoring being both convenient and superior to conventional finger measurement for continuous non-intrusive monitoring in both clinical and everyday-life settings.

Mobile technologies for the recording of vital signs and neural signals are envisaged to underpin the operation of future health services. For practical purposes, unobtrusive devices are favoured, such as those embedded in a helmet or incorporated onto an earplug. However, these locations have so far been underexplored, as the comparably narrow neck impedes the propagation of vital signals from the torso to the head surface. To establish the principles behind electrocardiogram (ECG) recordings from head and ear locations, we first introduce a realistic three-dimensional biophysics model for the propagation of cardiac electric potentials to the head surface, which demonstrates the feasibility of head-ECG recordings. Next, the proposed biophysics propagation model is verified over comprehensive real-world experiments based on head- and in-ear-ECG measurements. It is shown both that the proposed model is an excellent match for the recordings, and that the quality of head- and ear-ECG is sufficient for a reliable identification of the timing and shape of the characteristic P-, Q-, R-, S- and T-waves within the cardiac cycle. This opens up a range of new possibilities in the identification and management of heart conditions, such as myocardial infarction and atrial fibrillation, based on 24/7 continuous in-ear measurements. The study therefore paves the way for the incorporation of the cardiac modality into future ‘hearables’, unobtrusive devices for health monitoring.

Wearable health monitoring has emerged as a promising solution to the growing need for remote health assessment and growing demand for personalized preventative care and wellness management. Vital signs can be monitored and alerts can be made when anomalies are detected, potentially improving patient outcomes. One major challenge for the use of wearable health devices is their energy efficiency and battery-lifetime, which motivates the recent efforts towards the development of self-powered wearable devices. This article proposes a method for context aware dynamic sensor selection for power optimized physiological prediction using multi-modal wearable data streams. We first cluster the data by physical activity using the accelerometer data, and then fit a group lasso model to each activity cluster. We find the optimal reduced set of groups of sensor features, in turn reducing power usage by duty cycling these and optimizing prediction accuracy. We show that using activity state-based contextual information increases accuracy while decreasing power usage. We also show that the reduced feature set can be used in other regression models increasing accuracy and decreasing energy burden. We demonstrate the potential reduction in power usage using a custom-designed multi-modal wearable system prototype.

Nurses regularly perform patient handling activities. These activities with awkward postures expose healthcare providers to a high risk of overexertion injury. The recognition of patient handling activities is the first step to reduce injury risk for caregivers. The current practice on workplace activity recognition is based on human observational approach, which is neither accurate nor projectable to a large population. In this paper, we aim at addressing these challenges. Our solution comprises a smart wearable device and a novel spatio-temporal warping (STW) pattern recognition framework. The wearable device, named Smart Insole 2.0, is equipped with a rich set of sensors and can provide an unobtrusive way to automatically capture the information of patient handling activities. The STW pattern recognition framework fully exploits the spatial and temporal characteristics of plantar pressure by calculating a novel warped spatio-temporal distance, to quantify the similarity for the purpose of activity recognition. To validate the effectiveness of our framework, we perform a pilot study with eight subjects, including eight common activities in a nursing room. The experimental results show the overall classification accuracy achieves 91.7%. Meanwhile, the qualitative profile and load level can also be classified with accuracies of 98.3% and 92.5%, respectively. [ABSTRACT FROM AUTHOR]

The non-invasive estimation of blood oxygen saturation (SpO2) by pulse oximetry is of vital importance clinically, from the detection of sleep apnea to the recent ambulatory monitoring of hypoxemia in the delayed post-infective phase of COVID-19. In this proof of concept study, we set out to establish the feasibility of SpO2 measurement from the ear canal as a convenient site for long term monitoring, and perform a comprehensive comparison with the right index finger—the conventional clinical measurement site. During resting blood oxygen saturation estimation, we found a root mean square difference of 1.47% between the two measurement sites, with a mean difference of 0.23% higher SpO2 in the right ear canal. Using breath holds, we observe the known phenomena of time delay between central circulation and peripheral circulation with a mean delay between the ear and finger of 12.4 s across all subjects. Furthermore, we document the lower photoplethysmogram amplitude from the ear canal and suggest ways to mitigate this issue. In conjunction with the well-known robustness to temperature induced vasoconstriction, this makes conclusive evidence for in-ear SpO2 monitoring being both convenient and superior to conventional finger measurement for continuous non-intrusive monitoring in both clinical and everyday-life settings. [ABSTRACT FROM AUTHOR]

Wearable health monitoring has emerged as a promising solution to the growing need for remote health assessment and growing demand for personalized preventative care and wellness management. Vital signs can be monitored and alerts can be made when anomalies are detected, potentially improving patient outcomes. One major challenge for the use of wearable health devices is their energy efficiency and battery-lifetime, which motivates the recent efforts towards the development of self-powered wearable devices. This article proposes a method for context aware dynamic sensor selection for power optimized physiological prediction using multi-modal wearable data streams. We first cluster the data by physical activity using the accelerometer data, and then fit a group lasso model to each activity cluster. We find the optimal reduced set of groups of sensor features, in turn reducing power usage by duty cycling these and optimizing prediction accuracy. We show that using activity state-based contextual information increases accuracy while decreasing power usage. We also show that the reduced feature set can be used in other regression models increasing accuracy and decreasing energy burden. We demonstrate the potential reduction in power usage using a custom-designed multi-modal wearable system prototype. [ABSTRACT FROM AUTHOR]

With the introduction of the large number of fitness devices on the market, there are numerous possibilities for their use in managing chronic diseases in older adults. For example, monitoring people with dementia using commercially available devices that measure heart rate, breathing rate, lung volume, step count, and activity level could be used to predict episodic behavioral and psychological symptoms before they become distressing or disruptive. However, since these devices are designed primarily for fitness assessment, validation of the sensors in a controlled environment with the target cohort population is needed. In this study, we present validation results using a commercial fitness tracker, the Hexoskin sensor vest, with thirty-one participants aged 65 and older. Estimated physiological measures investigated in this study are heart rate, breathing rate, lung volume, step count, and activity level of the participants. Findings indicate that while the processed step count, heart rate, and breathing rate show strong correlations to the clinically accepted gold standard values, lung volume and activity level do not. This indicates the need to proceed cautiously when making clinical decisions using such sensors, and suggests that users should focus on the three strongly correlated parameters for further analysis, at least in the older population. The use of physiological measurement devices such as the Hexoskin may eventually become a non-intrusive way to continuously assess physiological measures in older adults with dementia who are at risk for distressing behavioral and psychological symptoms.

Mobile technologies for the recording of vital signs and neural signals are envisaged to underpin the operation of future health services. For practical purposes, unobtrusive devices are favoured, such as those embedded in a helmet or incorporated onto an earplug. However, these locations have so far been underexplored, as the comparably narrow neck impedes the propagation of vital signals from the torso to the head surface. To establish the principles behind electrocardiogram (ECG) recordings from head and ear locations, we first introduce a realistic three-dimensional biophysics model for the propagation of cardiac electric potentials to the head surface, which demonstrates the feasibility of head-ECG recordings. Next, the proposed biophysics propagation model is verified over comprehensive real-world experiments based on head- and in-ear-ECG measurements. It is shown both that the proposed model is an excellent match for the recordings, and that the quality of head- and ear-ECG is sufficient for a reliable identification of the timing and shape of the characteristic P-, Q-, R-, S- and T-waves within the cardiac cycle. This opens up a range of new possibilities in the identification and management of heart conditions, such as myocardial infarction and atrial fibrillation, based on 24/7 continuous in-ear measurements. The study therefore paves the way for the incorporation of the cardiac modality into future 'hearables', unobtrusive devices for health monitoring., Competing Interests: The authors declare no conflict of interest.