An Automated, Online Feasibility Randomized Controlled Trial of a Just-In-Time Adaptive Intervention for Smoking Cessation (Quit Sense).

INTRODUCTION: Learned smoking cues from a smoker's environment are a major cause of lapse and relapse. Quit Sense, a theory-guided Just-In-Time Adaptive Intervention smartphone app, aims to help smokers learn about their situational smoking cues and provide in-the-moment support to help manage these when quitting. METHODS: A two-arm feasibility randomized controlled trial (N = 209) to estimate parameters to inform a definitive evaluation. Smoker's willing to make a quit attempt were recruited using online paid-for adverts and randomized to "usual care" (text message referral to NHS SmokeFree website) or "usual care" plus a text message invitation to install Quit Sense. Procedures, excluding manual follow-up for nonresponders, were automated. Follow-up at 6 weeks and 6 months included feasibility, intervention engagement, smoking-related, and economic outcomes. Abstinence was verified using cotinine assessment from posted saliva samples. RESULTS: Self-reported smoking outcome completion rates at 6 months were 77% (95% CI 71%, 82%), viable saliva sample return rate was 39% (95% CI 24%, 54%), and health economic data 70% (95% CI 64%, 77%). Among Quit Sense participants, 75% (95% CI 67%, 83%) installed the app and set a quit date and, of those, 51% engaged for more than one week. The 6-month biochemically verified sustained abstinence rate (anticipated primary outcome for definitive trial), was 11.5% (12/104) among Quit Sense participants and 2.9% (3/105) for usual care (adjusted odds ratio = 4.57, 95% CIs 1.23, 16.94). No evidence of between-group differences in hypothesized mechanisms of action was found. CONCLUSIONS: Evaluation feasibility was demonstrated alongside evidence supporting the effectiveness potential of Quit Sense. IMPLICATIONS: Running a primarily automated trial to initially evaluate Quit Sense was feasible, resulting in modest recruitment costs and researcher time, and high trial engagement. When invited, as part of trial participation, to install a smoking cessation app, most participants are likely to do so, and, for those using Quit Sense, an estimated one-half will engage with it for more than 1 week. Evidence that Quit Sense may increase verified abstinence at 6-month follow-up, relative to usual care, was generated, although low saliva return rates to verify smoking status contributed to considerable imprecision in the effect size estimate.

Evaluating Listening Performance for COVID-19 Detection by Clinicians and Machine Learning: Comparative Study.

BACKGROUND: To date, performance comparisons between men and machines have been carried out in many health domains. Yet machine learning (ML) models and human performance comparisons in audio-based respiratory diagnosis remain largely unexplored. OBJECTIVE: The primary objective of this study was to compare human clinicians and an ML model in predicting COVID-19 from respiratory sound recordings. METHODS: In this study, we compared human clinicians and an ML model in predicting COVID-19 from respiratory sound recordings. Prediction performance on 24 audio samples (12 tested positive) made by 36 clinicians with experience in treating COVID-19 or other respiratory illnesses was compared with predictions made by an ML model trained on 1162 samples. Each sample consisted of voice, cough, and breathing sound recordings from 1 subject, and the length of each sample was around 20 seconds. We also investigated whether combining the predictions of the model and human experts could further enhance the performance in terms of both accuracy and confidence. RESULTS: The ML model outperformed the clinicians, yielding a sensitivity of 0.75 and a specificity of 0.83, whereas the best performance achieved by the clinicians was 0.67 in terms of sensitivity and 0.75 in terms of specificity. Integrating the clinicians' and the model's predictions, however, could enhance performance further, achieving a sensitivity of 0.83 and a specificity of 0.92. CONCLUSIONS: Our findings suggest that the clinicians and the ML model could make better clinical decisions via a cooperative approach and achieve higher confidence in audio-based respiratory diagnosis.

Exploring Longitudinal Cough, Breath, and Voice Data for COVID-19 Progression Prediction via Sequential Deep Learning: Model Development and Validation.

BACKGROUND: Recent work has shown the potential of using audio data (eg, cough, breathing, and voice) in the screening for COVID-19. However, these approaches only focus on one-off detection and detect the infection, given the current audio sample, but do not monitor disease progression in COVID-19. Limited exploration has been put forward to continuously monitor COVID-19 progression, especially recovery, through longitudinal audio data. Tracking disease progression characteristics and patterns of recovery could bring insights and lead to more timely treatment or treatment adjustment, as well as better resource management in health care systems. OBJECTIVE: The primary objective of this study is to explore the potential of longitudinal audio samples over time for COVID-19 progression prediction and, especially, recovery trend prediction using sequential deep learning techniques. METHODS: Crowdsourced respiratory audio data, including breathing, cough, and voice samples, from 212 individuals over 5-385 days were analyzed, alongside their self-reported COVID-19 test results. We developed and validated a deep learning-enabled tracking tool using gated recurrent units (GRUs) to detect COVID-19 progression by exploring the audio dynamics of the individuals' historical audio biomarkers. The investigation comprised 2 parts: (1) COVID-19 detection in terms of positive and negative (healthy) tests using sequential audio signals, which was primarily assessed in terms of the area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity, with 95% CIs, and (2) longitudinal disease progression prediction over time in terms of probability of positive tests, which was evaluated using the correlation between the predicted probability trajectory and self-reported labels. RESULTS: We first explored the benefits of capturing longitudinal dynamics of audio biomarkers for COVID-19 detection. The strong performance, yielding an AUROC of 0.79, a sensitivity of 0.75, and a specificity of 0.71 supported the effectiveness of the approach compared to methods that do not leverage longitudinal dynamics. We further examined the predicted disease progression trajectory, which displayed high consistency with longitudinal test results with a correlation of 0.75 in the test cohort and 0.86 in a subset of the test cohort with 12 (57.1%) of 21 COVID-19-positive participants who reported disease recovery. Our findings suggest that monitoring COVID-19 evolution via longitudinal audio data has potential in the tracking of individuals' disease progression and recovery. CONCLUSIONS: An audio-based COVID-19 progression monitoring system was developed using deep learning techniques, with strong performance showing high consistency between the predicted trajectory and the test results over time, especially for recovery trend predictions. This has good potential in the postpeak and postpandemic era that can help guide medical treatment and optimize hospital resource allocations. The changes in longitudinal audio samples, referred to as audio dynamics, are associated with COVID-19 progression; thus, modeling the audio dynamics can potentially capture the underlying disease progression process and further aid COVID-19 progression prediction. This framework provides a flexible, affordable, and timely tool for COVID-19 tracking, and more importantly, it also provides a proof of concept of how telemonitoring could be applicable to respiratory diseases monitoring, in general.

Uncertainty-aware Health Diagnostics via Class-balanced Evidential Deep Learning.

Uncertainty quantification is critical for ensuring the safety of deep learning-enabled health diagnostics, as it helps the model account for unknown factors and reduces the risk of misdiagnosis. However, existing uncertainty quantification studies often overlook the significant issue of class imbalance, which is common in medical data. In this paper, we propose a class-balanced evidential deep learning framework to achieve fair and reliable uncertainty estimates for health diagnostic models. This framework advances the state-of-the-art uncertainty quantification method of evidential deep learning with two novel mechanisms to address the challenges posed by class imbalance. Specifically, we introduce a pooling loss that enables the model to learn less biased evidence among classes and a learnable prior to regularize the posterior distribution that accounts for the quality of uncertainty estimates. Extensive experiments using benchmark data with varying degrees of imbalance and various naturally imbalanced health data demonstrate the effectiveness and superiority of our method. Our work pushes the envelope of uncertainty quantification from theoretical studies to realistic healthcare application scenarios. By enhancing uncertainty estimation for class-imbalanced data, we contribute to the development of more reliable and practical deep learning-enabled health diagnostic systems1.

A smoking cessation smartphone app that delivers real-time 'context aware' behavioural support: the Quit Sense feasibility RCT.

Background: During a quit attempt, cues from a smoker's environment are a major cause of brief smoking lapses, which increase the risk of relapse. Quit Sense is a theory-guided Just-In-Time Adaptive Intervention smartphone app, providing smokers with the means to learn about their environmental smoking cues and provides 'in the moment' support to help them manage these during a quit attempt. Objective: To undertake a feasibility randomised controlled trial to estimate key parameters to inform a definitive randomised controlled trial of Quit Sense. Design: A parallel, two-arm randomised controlled trial with a qualitative process evaluation and a 'Study Within A Trial' evaluating incentives on attrition. The research team were blind to allocation except for the study statistician, database developers and lead researcher. Participants were not blind to allocation. Setting: Online with recruitment, enrolment, randomisation and data collection (excluding manual telephone follow-up) automated through the study website. Participants: Smokers (323 screened, 297 eligible, 209 enrolled) recruited via online adverts on Google search, Facebook and Instagram. Interventions: Participants were allocated to 'usual care' arm (n = 105; text message referral to the National Health Service SmokeFree website) or 'usual care' plus Quit Sense (n = 104), via a text message invitation to install the Quit Sense app. Main outcome measures: Follow-up at 6 weeks and 6 months post enrolment was undertaken by automated text messages with an online questionnaire link and, for non-responders, by telephone. Definitive trial progression criteria were met if a priori thresholds were included in or lower than the 95% confidence interval of the estimate. Measures included health economic and outcome data completion rates (progression criterion #1 threshold: ≥ 70%), including biochemical validation rates (progression criterion #2 threshold: ≥ 70%), recruitment costs, app installation (progression criterion #3 threshold: ≥ 70%) and engagement rates (progression criterion #4 threshold: ≥ 60%), biochemically verified 6-month abstinence and hypothesised mechanisms of action and participant views of the app (qualitative). Results: Self-reported smoking outcome completion rates were 77% (95% confidence interval 71% to 82%) and health economic data (resource use and quality of life) 70% (95% CI 64% to 77%) at 6 months. Return rate of viable saliva samples for abstinence verification was 39% (95% CI 24% to 54%). The per-participant recruitment cost was £19.20, which included advert (£5.82) and running costs (£13.38). In the Quit Sense arm, 75% (95% CI 67% to 83%; 78/104) installed the app and, of these, 100% set a quit date within the app and 51% engaged with it for more than 1 week. The rate of 6-month biochemically verified sustained abstinence, which we anticipated would be used as a primary outcome in a future study, was 11.5% (12/104) in the Quit Sense arm and 2.9% (3/105) in the usual care arm (estimated effect size: adjusted odds ratio = 4.57, 95% CIs 1.23 to 16.94). There was no evidence of between-arm differences in hypothesised mechanisms of action. Three out of four progression criteria were met. The Study Within A Trial analysis found a £20 versus £10 incentive did not significantly increase follow-up rates though reduced the need for manual follow-up and increased response speed. The process evaluation identified several potential pathways to abstinence for Quit Sense, factors which led to disengagement with the app, and app improvement suggestions. Limitations: Biochemical validation rates were lower than anticipated and imbalanced between arms. COVID-19-related restrictions likely limited opportunities for Quit Sense to provide location tailored support. Conclusions: The trial design and procedures demonstrated feasibility and evidence was generated supporting the efficacy potential of Quit Sense. Future work: Progression to a definitive trial is warranted providing improved biochemical validation rates. Trial registration: This trial is registered as ISRCTN12326962. Funding: This award was funded by the National Institute for Health and Care Research (NIHR) Public Health Research programme (NIHR award ref: 17/92/31) and is published in full in Public Health Research; Vol. 12, No. 4. See the NIHR Funding and Awards website for further award information.

Smokers often fail to quit because of urges to smoke triggered by their surroundings (e.g. being around smokers). We developed a smartphone app ('Quit Sense') which learns about an individual's surroundings and locations where they smoke. During a quit attempt, Quit Sense uses in-built sensors to identify when smokers are in those locations and sends 'in the moment' advice to help prevent them from smoking. We ran a feasibility study to help plan for a future large study to see if Quit Sense helps smokers to quit. This feasibility study was designed to tell us how many participants complete study measures; recruitment costs; how many participants install and use Quit Sense; and estimate whether Quit Sense may help smokers to stop and how it might do this. We recruited 209 smokers using online adverts on Google search, Facebook and Instagram, costing £19 per participant. Participants then had an equal chance of receiving a web link to the National Health Service SmokeFree website ('usual care group') or receive that same web link plus a link to the Quit Sense app ('Quit Sense group'). Three-quarters of the Quit Sense group installed the app on their phone and half of these used the app for more than 1 week. We followed up 77% of participants at 6 months to collect study data, though only 39% of quitters returned a saliva sample for abstinence verification. At 6 months, more people in the Quit Sense group had stopped smoking (12%) than the usual care group (3%). It was not clear how the app helped smokers to quit based on study measures, though interviews found that the process of training the app helped people quit through learning about what triggered their smoking behaviour. The findings support undertaking a large study to tell us whether Quit Sense really does help smokers to quit.

Uncertainty Estimation with Data Augmentation for Active Learning Tasks on Health Data.

Supervised machine learning (ML) is revolutionising healthcare, but the acquisition of reliable labels for signals harvested from medical sensors is usually challenging, manual, and costly. Active learning can assist in establishing labels on-the-fly by querying the user only for the most uncertain -and thus informative- samples. However, current approaches rely on naive data selection algorithms, which still require many iterations to achieve the desired accuracy. To this aim, we introduce a novel framework that exploits data augmentation for estimating the uncertainty introduced by sensor signals.Our experiments on classifying medical signals show that our framework selects informative samples up to 50% more diverse. Sample diversity is a key indicator of uncertainty, and our framework can capture this diversity better than previous solutions as it picks unlabelled samples with a higher average point distance during the first queries compared to the baselines, which pick samples that are closer together. Through our experiments, we show that augmentation-based uncertainty makes better decisions, as the more informative signals are labelled first and the learner is able to train on samples with more diverse features earlier on, thus enabling the potential expansion of ML in more real-life healthcare use cases.

Human-centred artificial intelligence for mobile health sensing: challenges and opportunities.

Advances in wearable sensing and mobile computing have enabled the collection of health and well-being data outside of traditional laboratory and hospital settings, paving the way for a new era of mobile health. Meanwhile, artificial intelligence (AI) has made significant strides in various domains, demonstrating its potential to revolutionize healthcare. Devices can now diagnose diseases, predict heart irregularities and unlock the full potential of human cognition. However, the application of machine learning (ML) to mobile health sensing poses unique challenges due to noisy sensor measurements, high-dimensional data, sparse and irregular time series, heterogeneity in data, privacy concerns and resource constraints. Despite the recognition of the value of mobile sensing, leveraging these datasets has lagged behind other areas of ML. Furthermore, obtaining quality annotations and ground truth for such data is often expensive or impractical. While recent large-scale longitudinal studies have shown promise in leveraging wearable sensor data for health monitoring and prediction, they also introduce new challenges for data modelling. This paper explores the challenges and opportunities of human-centred AI for mobile health, focusing on key sensing modalities such as audio, location and activity tracking. We discuss the limitations of current approaches and propose potential solutions.

Denoising Cough Sound Recordings Using Neural Networks.

Objective cough sound evaluation is useful in the diagnosis and management of respiratory diseases. However, the performance of cough sound analysis models can degrade in the presence of background noises common in everyday environments. This brings forward the need for cough sound denoising. This work utilizes a method for denoising cough sound recordings using signal processing and machine learning techniques, inspired by research in the field of speech enhancement. It uses supervised learning to find a mapping between the noisy and clean spectra of cough sound signals using a fully connected feed-forward neural network. The method is validated on a dataset of 300 manually annotated cough sound recordings corrupted with babble noise. The effect of various signal processing and neural network parameters on denoising performance is investigated. The method is shown to improve cough sound quality and intelligibility and outperform conventional denoising methods.

EarSet: A Multi-Modal Dataset for Studying the Impact of Head and Facial Movements on In-Ear PPG Signals.

Photoplethysmography (PPG) is a simple, yet powerful technique to study blood volume changes by measuring light intensity variations. However, PPG is severely affected by motion artifacts, which hinder its trustworthiness. This problem is pressing in earables since head movements and facial expressions cause skin and tissue displacements around and inside the ear. Understanding such artifacts is fundamental to the success of earables for accurate cardiovascular health monitoring. However, the lack of in-ear PPG datasets prevents the research community from tackling this challenge. In this work, we report on the design of an ear tip featuring a 3-channels PPG and a co-located 6-axis motion sensor. This, enables sensing PPG data at multiple wavelengths and the corresponding motion signature from both ears. Leveraging our device, we collected a multi-modal dataset from 30 participants while performing 16 natural motions, including both head/face and full body movements. This unique dataset will greatly support research towards making in-ear vital signs sensing more accurate and robust, thus unlocking the full potential of the next-generation PPG-equipped earables.

A summary of the ComParE COVID-19 challenges.

The COVID-19 pandemic has caused massive humanitarian and economic damage. Teams of scientists from a broad range of disciplines have searched for methods to help governments and communities combat the disease. One avenue from the machine learning field which has been explored is the prospect of a digital mass test which can detect COVID-19 from infected individuals' respiratory sounds. We present a summary of the results from the INTERSPEECH 2021 Computational Paralinguistics Challenges: COVID-19 Cough, (CCS) and COVID-19 Speech, (CSS).

Components in time-varying graphs.

Real complex systems are inherently time-varying. Thanks to new communication systems and novel technologies, today it is possible to produce and analyze social and biological networks with detailed information on the time of occurrence and duration of each link. However, standard graph metrics introduced so far in complex network theory are mainly suited for static graphs, i.e., graphs in which the links do not change over time, or graphs built from time-varying systems by aggregating all the links as if they were concurrent in time. In this paper, we extend the notion of connectedness, and the definitions of node and graph components, to the case of time-varying graphs, which are represented as time-ordered sequences of graphs defined over a fixed set of nodes. We show that the problem of finding strongly connected components in a time-varying graph can be mapped into the problem of discovering the maximal-cliques in an opportunely constructed static graph, which we name the affine graph. It is, therefore, an NP-complete problem. As a practical example, we have performed a temporal component analysis of time-varying graphs constructed from three data sets of human interactions. The results show that taking time into account in the definition of graph components allows to capture important features of real systems. In particular, we observe a large variability in the size of node temporal in- and out-components. This is due to intrinsic fluctuations in the activity patterns of individuals, which cannot be detected by static graph analysis.

Towards Adversarial Robustness with Early Exit Ensembles.

Deep learning techniques are increasingly used for decision-making in health applications, however, these can easily be manipulated by adversarial examples across different clinical domains. Their security and privacy vulnerabilities raise concerns about the practical deployment of these systems. The number and variety of the adversarial attacks grow continuously, making it difficult for mitigation approaches to provide effective solutions. Current mitigation techniques often rely on expensive re-training procedures as new attacks emerge. In this paper, we propose a novel adversarial mitigation technique for biosignal classification tasks. Our approach is based on recent findings interpreting early exit neural networks as an ensemble of weight sharing sub-networks. Our experiments on state-of-the-art deep learning models show that early exit ensembles can provide robustness generalizable to various white box and universal adversarial attacks. The approach increases the accuracy of vulnerable deep learning models up to 60 percentage points, while providing adversarial mitigation comparable to adversarial training. This is achieved without previous exposure to the adversarial perturbation or the computational burden of re-training.

Exploring machine learning for audio-based respiratory condition screening: A concise review of databases, methods, and open issues.

Auscultation plays an important role in the clinic, and the research community has been exploring machine learning (ML) to enable remote and automatic auscultation for respiratory condition screening via sounds. To give the big picture of what is going on in this field, in this narrative review, we describe publicly available audio databases that can be used for experiments, illustrate the developed ML methods proposed to date, and flag some under-considered issues which still need attention. Compared to existing surveys on the topic, we cover the latest literature, especially those audio-based COVID-19 detection studies which have gained extensive attention in the last two years. This work can help to facilitate the application of artificial intelligence in the respiratory auscultation field.

Breaking away from labels: The promise of self-supervised machine learning in intelligent health.

Medicine is undergoing an unprecedented digital transformation, as massive amounts of health data are being produced, gathered, and curated, ranging from in-hospital (e.g., intensive care unit [ICU]) to person-generated data (wearables). Annotating all these data for training purposes in order to feed to deep learning models for pattern recognition is impractical. Here, we discuss some exciting recent results of self-supervised learning (SSL) applications to high-resolution health signals. These examples leverage unlabeled data to learn meaningful representations that can generalize to situations where the ground truth is inadequate or simply infeasible to collect due to the high burden or associated costs. The most prominent bottleneck of deep learning today is access to labeled, carefully curated datasets, and self-supervision on health signals opens up new possibilities to eliminate data silos through general-purpose models that can transfer to low-resource environments and tasks.

Machine learning detects altered spatial navigation features in outdoor behaviour of Alzheimer's disease patients.

Impairment of navigation is one of the earliest symptoms of Alzheimer's disease (AD), but to date studies have involved proxy tests of navigation rather than studies of real life behaviour. Here we use GPS tracking to measure ecological outdoor behaviour in AD. The aim was to use data-driven machine learning approaches to explore spatial metrics within real life navigational traces that discriminate AD patients from controls. 15 AD patients and 18 controls underwent tracking of their outdoor navigation over two weeks. Three kinds of spatiotemporal features of segments were extracted, characterising the mobility domain (entropy, segment similarity, distance from home), spatial shape (total turning angle, segment complexity), and temporal characteristics (stop duration). Patients significantly differed from controls on entropy (p-value 0.008), segment similarity (p-value [Formula: see text]), and distance from home (p-value [Formula: see text]). Graph-based analyses yielded preliminary data indicating that topological features assessing the connectivity of visited locations may also differentiate patients from controls. In conclusion, our results show that specific outdoor navigation features discriminate AD patients from controls, which has significant implication for future AD diagnostics, outcome measures and interventions. Furthermore, this work illustrates how wearables-based sensing of everyday behaviour may be used to deliver ecologically-valid digital biomarkers of AD pathophysiology.

Stress Inference from Abdominal Sounds using Machine Learning.

Stress is often considered the 21st century's epidemic, affecting more than a third of the globe's population. Long-term exposure to stress has significant side effects on physical and mental health. In this work we propose a methodology for detecting stress using abdominal sounds. For this study, eight participants were either exposed to a stressful (Stroop test) or a relaxing (guided meditation) stimulus for ten days. In total, we collected 104 hours of abdominal sounds using a custom wearable device in a belt form-factor. We explored the effect of various features on the binary stress classification accuracy using traditional machine learning methods. Namely, we observed the impact of using acoustic features on their own, as well as in combination with features representing current mood state, and hand-crafted domain-specific features. After feature extraction and reduction, by utilising a multilayer perceptron classifier model we achieved 77% accuracy in detecting abdominal sounds under stress exposure. Clinical relevance- This feasibility study confirms the link between the gastrointestinal system and stress and uncovers a novel approach for stress inference via abdominal sounds using machine learning.

Longitudinal cardio-respiratory fitness prediction through wearables in free-living environments.

Cardiorespiratory fitness is an established predictor of metabolic disease and mortality. Fitness is directly measured as maximal oxygen consumption (VO2max), or indirectly assessed using heart rate responses to standard exercise tests. However, such testing is costly and burdensome because it requires specialized equipment such as treadmills and oxygen masks, limiting its utility. Modern wearables capture dynamic real-world data which could improve fitness prediction. In this work, we design algorithms and models that convert raw wearable sensor data into cardiorespiratory fitness estimates. We validate these estimates' ability to capture fitness profiles in free-living conditions using the Fenland Study (N=11,059), along with its longitudinal cohort (N = 2675), and a third external cohort using the UK Biobank Validation Study (N = 181) who underwent maximal VO2max testing, the gold standard measurement of fitness. Our results show that the combination of wearables and other biomarkers as inputs to neural networks yields a strong correlation to ground truth in a holdout sample (r = 0.82, 95CI 0.80-0.83), outperforming other approaches and models and detects fitness change over time (e.g., after 7 years). We also show how the model's latent space can be used for fitness-aware patient subtyping paving the way to scalable interventions and personalized trial recruitment. These results demonstrate the value of wearables for fitness estimation that today can be measured only with laboratory tests.

Detecting sleep outside the clinic using wearable heart rate devices.

The adoption of multisensor wearables presents the opportunity of longitudinal monitoring of sleep in large populations. Personalized yet device-agnostic algorithms can sidestep laborious human annotations and objectify cross-cohort comparisons. We developed and tested a heart rate-based algorithm that captures inter- and intra-individual sleep differences in free-living conditions and does not require human input. We evaluated it on four study cohorts using different research- and consumer-grade devices for over 2000 nights. Recording periods included both 24 h free-living and conventional lab-based night-only data. We compared our optimized method against polysomnography, sleep diaries and sleep periods produced through a state-of-the-art acceleration based method. Against sleep diaries, the algorithm yielded a mean squared error of 0.04-0.06 and a total sleep time (TST) deviation of [Formula: see text]2.70 (± 5.74) and 12.80 (± 3.89) minutes, respectively. When evaluated with PSG lab studies, the MSE ranged between 0.06 and 0.11 yielding a time deviation between [Formula: see text]29.07 and [Formula: see text]55.04 minutes. These results showcase the value of this open-source, device-agnostic algorithm for the reliable inference of sleep in free-living conditions and in the absence of annotations.

Sounds of COVID-19: exploring realistic performance of audio-based digital testing.

To identify Coronavirus disease (COVID-19) cases efficiently, affordably, and at scale, recent work has shown how audio (including cough, breathing and voice) based approaches can be used for testing. However, there is a lack of exploration of how biases and methodological decisions impact these tools' performance in practice. In this paper, we explore the realistic performance of audio-based digital testing of COVID-19. To investigate this, we collected a large crowdsourced respiratory audio dataset through a mobile app, alongside symptoms and COVID-19 test results. Within the collected dataset, we selected 5240 samples from 2478 English-speaking participants and split them into participant-independent sets for model development and validation. In addition to controlling the language, we also balanced demographics for model training to avoid potential acoustic bias. We used these audio samples to construct an audio-based COVID-19 prediction model. The unbiased model took features extracted from breathing, coughs and voice signals as predictors and yielded an AUC-ROC of 0.71 (95% CI: 0.65-0.77). We further explored several scenarios with different types of unbalanced data distributions to demonstrate how biases and participant splits affect the performance. With these different, but less appropriate, evaluation strategies, the performance could be overestimated, reaching an AUC up to 0.90 (95% CI: 0.85-0.95) in some circumstances. We found that an unrealistic experimental setting can result in misleading, sometimes over-optimistic, performance. Instead, we reported complete and reliable results on crowd-sourced data, which would allow medical professionals and policy makers to accurately assess the value of this technology and facilitate its deployment.

Segmentation-free Heart Pathology Detection Using Deep Learning.

Cardiovascular (CV) diseases are the leading cause of death in the world, and auscultation is typically an essential part of a cardiovascular examination. The ability to diagnose a patient based on their heart sounds is a rather difficult skill to master. Thus, many approaches for automated heart auscultation have been explored. However, most of the previously proposed methods involve a segmentation step, the performance of which drops significantly for high pulse rates or noisy signals. In this work, we propose a novel segmentation-free heart sound classification method. Specifically, we apply discrete wavelet transform to denoise the signal, followed by feature extraction and feature reduction. Then, Support Vector Machines and Deep Neural Networks are utilised for classification. On the PASCAL heart sound dataset our approach showed superior performance compared to others, achieving 81% and 96% precision on normal and murmur classes, respectively. In addition, for the first time, the data were further explored under a user-independent setting, where the proposed method achieved 92% and 86% precision on normal and murmur, demonstrating the potential of enabling automatic murmur detection for practical use.