Objectively measured peri-vaccination sleep does not predict COVID-19 breakthrough infection.

Prior studies have shown that sleep duration peri-vaccination influences an individual's antibody response. However, whether peri-vaccination sleep affects real-world vaccine effectiveness is unknown. Here, we tested whether objectively measured sleep around COVID-19 vaccination affected breakthrough infection rates. DETECT is a study of digitally recruited participants who report COVID-19-related information, including vaccination and illness data. Objective sleep data are also recorded through activity trackers. We compared the impact of sleep duration, sleep efficiency, and frequency of awakenings on reported breakthrough infection after the 2nd vaccination and 1st COVID-19 booster. Logistic regression models were created to examine if sleep metrics predicted COVID-19 breakthrough infection independent of age and gender. Self-reported breakthrough COVID-19 infection following 2nd COVID-19 vaccination and 1st booster. 256 out of 5265 individuals reported a breakthrough infection after the 2nd vaccine, and 581 out of 2583 individuals reported a breakthrough after the 1st booster. There was no difference in sleep duration between those with and without breakthrough infection. Increased awakening frequency was associated with breakthrough infection after the 1st booster with 3.01 ± 0.65 awakenings/hour in the breakthrough group compared to 2.82 ± 0.65 awakenings/hour in those without breakthrough (P < 0.001). Cox proportional hazards modeling showed that age < 60 years (hazard ratio 2.15, P < 0.001) and frequency of awakenings (hazard ratio 1.17, P = 0.019) were associated with breakthrough infection after the 1st booster. Sleep duration was not associated with breakthrough infection after COVID vaccination. While increased awakening frequency during sleep was associated with breakthrough infection beyond traditional risk factors, the clinical implications of this finding are unclear.

AI-Enhanced Reconstruction of the 12-Lead Electrocardiogram via 3-Leads with Accurate Clinical Assessment.

The 12-lead electrocardiogram (ECG) is an integral component to the diagnosis of a multitude of cardiovascular conditions. It is performed using a complex set of skin surface electrodes, limiting its use outside traditional clinical settings. We developed an artificial intelligence algorithm, trained over 600,000 clinically acquired ECGs, to explore whether fewer leads as input are sufficient to reconstruct a full 12-lead ECG. Two limb leads (I and II) and one precordial lead (V3) were required to generate a reconstructed synthetic 12-lead ECG highly correlated with the original ECG. An automatic algorithm for detection of acute myocardial infarction (MI) performed similarly for original and reconstructed ECGs (AUC=0.94). When interpreted by cardiologists, reconstructed ECGs achieved an accuracy of 81.4±5.0% in identifying ST elevation MI, comparable with the original 12-lead ECGs (accuracy 84.6±4.6%). These results will impact development efforts to innovate ECG acquisition methods with simplified tools in non-specialized settings.

Prediction of atrial fibrillation from at-home single-lead ECG signals without arrhythmias.

Early identification of atrial fibrillation (AF) can reduce the risk of stroke, heart failure, and other serious cardiovascular outcomes. However, paroxysmal AF may not be detected even after a two-week continuous monitoring period. We developed a model to quantify the risk of near-term AF in a two-week period, based on AF-free ECG intervals of up to 24 h from 459,889 patch-based ambulatory single-lead ECG (modified lead II) recordings of up to 14 days. A deep learning model was used to integrate ECG morphology data with demographic and heart rhythm features toward AF prediction. Observing a 1-day AF-free ECG recording, the model with deep learning features produced the most accurate prediction of near-term AF with an area under the curve AUC = 0.80 (95% confidence interval, CI = 0.79-0.81), significantly improving discrimination compared to demographic metrics alone (AUC 0.67; CI = 0.66-0.68). Our model was able to predict incident AF over a two-week time frame with high discrimination, based on AF-free single-lead ECG recordings of various lengths. Application of the model may enable a digital strategy for improving diagnostic capture of AF by risk stratifying individuals with AF-negative ambulatory monitoring for prolonged or recurrent monitoring, potentially leading to more rapid initiation of treatment.

Sensor-based surveillance for digitising real-time COVID-19 tracking in the USA (DETECT): a multivariable, population-based, modelling study.

BACKGROUND: Traditional viral illness surveillance relies on in-person clinical or laboratory data, paper-based data collection, and outdated technology for data transfer and aggregation. We aimed to assess whether continuous sensor data can provide an early warning signal for COVID-19 activity as individual physiological and behavioural changes might precede symptom onset, care seeking, and diagnostic testing. METHODS: This multivariable, population-based, modelling study recruited adult (aged ≥18 years) participants living in the USA who had a smartwatch or fitness tracker on any device that connected to Apple HealthKit or Google Fit and had joined the DETECT study by downloading the MyDataHelps app. In the model development cohort, we included people who had participated in DETECT between April 1, 2020, and Jan 14, 2022. In the validation cohort, we included individuals who had participated between Jan 15 and Feb 15, 2022. When a participant joins DETECT, they fill out an intake survey of demographic information, including their ZIP code (postal code), and surveys on symptoms, symptom onset, and viral illness test dates and results, if they become unwell. When a participant connects their device, historical sensor data are collected, if available. Sensor data continue to be collected unless a participant withdraws from the study. Using sensor data, we collected each participant's daily resting heart rate and step count during the entire study period and identified anomalous sensor days, in which resting heart rate was higher than, and step count was lower than, a specified threshold calculated for each individual by use of their baseline data. The proportion of users with anomalous data each day was used to create a 7-day moving average. For the main cohort, a negative binomial model predicting 7-day moving averages for COVID-19 case counts, as reported by the Centers for Disease Control and Prevention (CDC), in real time, 6 days in the future, and 12 days in the future in the USA and California was fitted with CDC-reported data from 3 days before alone (H0) or in combination with anomalous sensor data (H1). We compared the predictions with Pearson correlation. We then validated the model in the validation cohort. FINDINGS: Between April 1, 2020, and Jan 14, 2022, 35 842 participants enrolled in DETECT, of whom 4006 in California and 28 527 in the USA were included in our main cohort. The H1 model significantly outperformed the H0 model in predicting the 7-day moving average COVID-19 case counts in California and the USA. For example, Pearson correlation coefficients for predictions 12 days in the future increased by 32·9% in California (from 0·70 [95% CI 0·65-0·73] to 0·93 [0·92-0·94]) and by 12·2% (from 0·82 [0·79-0·84] to 0·92 [0·91-0·93]) in the USA from the H0 model to the H1 model. Our validation model also showed significant correlations for predictions in real time, 6 days in the future, and 12 days in the future. INTERPRETATION: Our study showed that passively collected sensor data from consenting participants can provide real-time disease tracking and forecasting. With a growing population of wearable technology users, these sensor data could be integrated into viral surveillance programmes. FUNDING: The National Center for Advancing Translational Sciences of the US National Institutes of Health, The Rockefeller Foundation, and Amazon Web Services.

Inter-individual variation in objective measure of reactogenicity following COVID-19 vaccination via smartwatches and fitness bands.

The ability to identify who does or does not experience the intended immune response following vaccination could be of great value in not only managing the global trajectory of COVID-19 but also helping guide future vaccine development. Vaccine reactogenicity can potentially lead to detectable physiologic changes, thus we postulated that we could detect an individual's initial physiologic response to a vaccine by tracking changes relative to their pre-vaccine baseline using consumer wearable devices. We explored this possibility using a smartphone app-based research platform that enabled volunteers (39,701 individuals) to share their smartwatch data, as well as self-report, when appropriate, any symptoms, COVID-19 test results, and vaccination information. Of 7728 individuals who reported at least one vaccination dose, 7298 received an mRNA vaccine, and 5674 provided adequate data from the peri-vaccine period for analysis. We found that in most individuals, resting heart rate (RHR) increased with respect to their individual baseline after vaccination, peaked on day 2, and returned to normal by day 6. This increase in RHR was greater than one standard deviation above individuals' normal daily pattern in 47% of participants after their second vaccine dose. Consistent with other reports of subjective reactogenicity following vaccination, we measured a significantly stronger effect after the second dose relative to the first, except those who previously tested positive to COVID-19, and a more pronounced increase for individuals who received the Moderna vaccine. Females, after the first dose only, and those aged <40 years, also experienced a greater objective response after adjusting for possible confounding factors. These early findings show that it is possible to detect subtle, but important changes from an individual's normal as objective evidence of reactogenicity, which, with further work, could prove useful as a surrogate for vaccine-induced immune response.

Passive detection of COVID-19 with wearable sensors and explainable machine learning algorithms.

Individual smartwatch or fitness band sensor data in the setting of COVID-19 has shown promise to identify symptomatic and pre-symptomatic infection or the need for hospitalization, correlations between peripheral temperature and self-reported fever, and an association between changes in heart-rate-variability and infection. In our study, a total of 38,911 individuals (61% female, 15% over 65) have been enrolled between March 25, 2020 and April 3, 2021, with 1118 reported testing positive and 7032 negative for COVID-19 by nasopharyngeal PCR swab test. We propose an explainable gradient boosting prediction model based on decision trees for the detection of COVID-19 infection that can adapt to the absence of self-reported symptoms and to the available sensor data, and that can explain the importance of each feature and the post-test-behavior for the individuals. We tested it in a cohort of symptomatic individuals who exhibited an AUC of 0.83 [0.81-0.85], or AUC = 0.78 [0.75-0.80] when considering only data before the test date, outperforming state-of-the-art algorithm in these conditions. The analysis of all individuals (including asymptomatic and pre-symptomatic) when self-reported symptoms were excluded provided an AUC of 0.78 [0.76-0.79], or AUC of 0.70 [0.69-0.72] when considering only data before the test date. Extending the use of predictive algorithms for detection of COVID-19 infection based only on passively monitored data from any device, we showed that it is possible to scale up this platform and apply the algorithm in other settings where self-reported symptoms can not be collected.

The Physiologic Response to COVID-19 Vaccination.

Two mRNA vaccines and one adenovirus-based vaccine against SARS CoV-2 are currently being distributed at scale in the United States. Objective evidence of a specific individual's physiologic response to that vaccine are not routinely tracked but may offer insights into the acute immune response and personal and/or vaccine characteristics associated with that. We explored this possibility using a smartphone app-based research platform developed early in the pandemic that enabled volunteers (38,911 individuals between 25 March 2020 and 4 April 2021) to share their smartwatch and activity tracker data, as well as self-report, when appropriate, any symptoms, COVID-19 test results and vaccination dates and type. Of 4,110 individuals who reported at least one mRNA vaccination dose, 3,312 provided adequate resting heart rate data from the peri-vaccine period for analysis. We found changes in resting heart rate with respect to an individual baseline increased the days after vaccination, peaked on day 2, and returned to normal on day 6, with a much stronger effect after second dose with respect to first dose (average changes 1.6 versus 0.5 beats per minute). The changes were more pronounced for individuals who received the Moderna vaccine (on both doses), those who previously tested positive to COVID-19 (on dose 1), and for individuals aged <40 years, after adjusting for possible confounding factors. Taking advantage of continuous passive data from personal sensors could potentially enable the identification of a digital fingerprint of inflammation, which might prove useful as a surrogate for vaccine-induced immune response.

A Comprehensive Explanation Framework for Biomedical Time Series Classification.

In this study, we propose a post-hoc explainability framework for deep learning models applied to quasi-periodic biomedical time-series classification. As a case study, we focus on the problem of atrial fibrillation (AF) detection from electrocardiography signals, which has strong clinical relevance. Starting from a state-of-the-art pretrained model, we tackle the problem from two different perspectives: global and local explanation. With global explanation, we analyze the model behavior by looking at entire classes of data, showing which regions of the input repetitive patterns have the most influence for a specific outcome of the model. Our explanation results align with the expectations of clinical experts, showing that features crucial for AF detection contribute heavily to the final decision. These features include R-R interval regularity, absence of the P-wave or presence of electrical activity in the isoelectric period. On the other hand, with local explanation, we analyze specific input signals and model outcomes. We present a comprehensive analysis of the network facing different conditions, whether the model has correctly classified the input signal or not. This enables a deeper understanding of the network's behavior, showing the most informative regions that trigger the classification decision and highlighting possible causes of misbehavior.

Wearable sensor data and self-reported symptoms for COVID-19 detection.

Traditional screening for COVID-19 typically includes survey questions about symptoms and travel history, as well as temperature measurements. Here, we explore whether personal sensor data collected over time may help identify subtle changes indicating an infection, such as in patients with COVID-19. We have developed a smartphone app that collects smartwatch and activity tracker data, as well as self-reported symptoms and diagnostic testing results, from individuals in the United States, and have assessed whether symptom and sensor data can differentiate COVID-19 positive versus negative cases in symptomatic individuals. We enrolled 30,529 participants between 25 March and 7 June 2020, of whom 3,811 reported symptoms. Of these symptomatic individuals, 54 reported testing positive and 279 negative for COVID-19. We found that a combination of symptom and sensor data resulted in an area under the curve (AUC) of 0.80 (interquartile range (IQR): 0.73-0.86) for discriminating between symptomatic individuals who were positive or negative for COVID-19, a performance that is significantly better (P < 0.01) than a model1 that considers symptoms alone (AUC = 0.71; IQR: 0.63-0.79). Such continuous, passively captured data may be complementary to virus testing, which is generally a one-off or infrequent sampling assay.

On the Effectiveness of Deep Representation Learning: the Atrial Fibrillation Case.

The automatic and unsupervised analysis of biomedical time series is of primary importance for diagnostic and preventive medicine, enabling fast and reliable data processing to reveal clinical insights without the need for human intervention. Representation learning (RL) methods perform an automatic extraction of meaningful features that can be used, e.g., for a subsequent classification of the measured data. The goal of this study is to explore and quantify the benefits of RL techniques of varying degrees of complexity, focusing on modern deep learning (DL) architectures. We focus on the automatic classification of atrial fibrillation (AF) events from noisy single-lead electrocardiographic signals (ECG) obtained from wireless sensors. This is an important task as it allows the detection of sub-clinical AF which is hard to diagnose with a short in-clinic 12-lead ECG. The effectiveness of the considered architectures is quantified and discussed in terms of classification performance, memory/data efficiency and computational complexity.

Deep Learning Techniques for Improving Digital Gait Segmentation.

Wearable technology for the automatic detection of gait events has recently gained growing interest, enabling advanced analyses that were previously limited to specialist centres and equipment (e.g., instrumented walkway). In this study, we present a novel method based on dilated convolutions for an accurate detection of gait events (initial and final foot contacts) from wearable inertial sensors. A rich dataset has been used to validate the method, featuring 71 people with Parkinson's disease (PD) and 67 healthy control subjects. Multiple sensors have been considered, one located on the fifth lumbar vertebrae and two on the ankles. The aims of this study were: (i) to apply deep learning (DL) techniques on wearable sensor data for gait segmentation and quantification in older adults and in people with PD; (ii) to validate the proposed technique for measuring gait against traditional gold standard laboratory reference and a widely used algorithm based on wavelet transforms (WT); (iii) to assess the performance of DL methods in assessing high-level gait characteristics, with focus on stride, stance and swing related features. The results showed a high reliability of the proposed approach, which achieves temporal errors considerably smaller than WT, in particular for the detection of final contacts, with an inter-quartile range below 70 ms in the worst case. This study showes encouraging results, and paves the road for further research, addressing the effectiveness and the generalization of data-driven learning systems for accurate event detection in challenging conditions.

Prediction of Adverse Glycemic Events From Continuous Glucose Monitoring Signal.

The most important objective of any diabetes therapy is to maintain the blood glucose concentration within the euglycemic range, avoiding or at least mitigating critical hypo/hyperglycemic episodes. Modern continuous glucose monitoring (CGM) devices bear the promise of providing the patients with an increased and timely awareness of glycemic conditions as these get dangerously near to hypo/hyperglycemia. The challenge is to detect, with reasonable advance, the patterns leading to risky situations, allowing the patient to make therapeutic decisions on the basis of future (predicted) glucose concentration levels. We underline that a technically sound performance comparison of the approaches proposed in recent years has yet to be done, thus it is unclear which one is preferred. The aim of this study is to fill this gap by carrying out a comparative analysis among the most common methods for glucose event prediction. Both regression and classification algorithms have been implemented and analyzed, including static and dynamic training approaches. The dataset consists of 89 CGM time series measured in diabetic subjects for 7 subsequent days. Performance metrics, specifically defined to assess and compare the event-prediction capabilities of the methods, have been introduced and analyzed. Our numerical results show that a static training approach exhibits better performance, in particular when regression methods are considered. However, classifiers show some improvement when trained for a specific event category, such as hyperglycemia, achieving performance comparable to the regressors, with the advantage of predicting the events sooner.

Toward lightweight biometric signal processing for wearable devices.

Wearable devices are becoming a natural and economic means to gather biometric data from end users. The massive amount of information that they will provide, unimaginable until a few years ago, owns an immense potential for applications such as continuous monitoring for personalized healthcare and use within fitness applications. Wearables are however heavily constrained in terms of amount of memory, transmission capability and energy reserve. This calls for dedicated, lightweight but still effective algorithms for data management. This paper is centered around lossy data compression techniques, whose aim is to minimize the amount of information that is to be stored on their onboard memory and subsequently transmitted over wireless interfaces. Specifically, we analyze selected compression techniques for biometric signals, quantifying their complexity (energy consumption) and compression performance. Hence, we propose a new class of codebook-based (CB) compression algorithms, designed to be energy efficient, online and amenable to any type of signal exhibiting recurrent patterns. Finally, the performance of the selected and the new algorithm is assessed, underlining the advantages offered by CB schemes in terms of memory savings and classification algorithms.

A method for ventricular late potentials detection using time-frequency representation and wavelet denoising.

This study proposes a method for ventricular late potentials (VLPs) detection using time-frequency representation and wavelet denoising in high-resolution electrocardiography (HRECG). The analysis is performed both with the signal averaged electrocardiography (SAECG) and in real time. A comparison between the temporal and the time-frequency analysis is also reported. In the first analysis the standard parameters QRSd, LAS40, and RMS40 were used; in the second normalized energy in time-frequency domain was calculated. The algorithm was tested adding artificial VLPs to real ECGs.