Search

Since its emergence in late 2019, the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused a pandemic with more than 55 million reported cases and 1.3 million estimated deaths worldwide. While epidemiological and clinical characteristics of COVID-19 have been reported, risk factors underlying the transition from mild to severe disease among patients remain poorly understood. In this retrospective study, we analysed data of 879 confirmed SARS-CoV-2 positive patients admitted to a two-site NHS Trust hospital in London, England, between January 1st and May 26th, 2020, with a majority of cases occurring in March and April. We extracted anonymised demographic data, physiological clinical variables and laboratory results from electronic healthcare records (EHR) and applied multivariate logistic regression, random forest and extreme gradient boosted trees. To evaluate the potential for early risk assessment, we used data available during patients’ initial presentation at the emergency department (ED) to predict deterioration to one of three clinical endpoints in the remainder of the hospital stay: admission to intensive care, need for invasive mechanical ventilation and in-hospital mortality. Based on the trained models, we extracted the most informative clinical features in determining these patient trajectories. Considering our inclusion criteria, we have identified 129 of 879 (15%) patients that required intensive care, 62 of 878 (7%) patients needing mechanical ventilation, and 193 of 619 (31%) cases of in-hospital mortality. Our models learned successfully from early clinical data and predicted clinical endpoints with high accuracy, the best model achieving area under the receiver operating characteristic (AUC-ROC) scores of 0.76 to 0.87 (F1 scores of 0.42–0.60). Younger patient age was associated with an increased risk of receiving intensive care and ventilation, but lower risk of mortality. Clinical indicators of a patient’s oxygen supply and selected laboratory results, such as blood lactate and creatinine levels, were most predictive of COVID-19 patient trajectories. Among COVID-19 patients machine learning can aid in the early identification of those with a poor prognosis, using EHR data collected during a patient’s first presentation at ED. Patient age and measures of oxygenation status during ED stay are primary indicators of poor patient outcomes.

Affective computing aims to detect a person's affective state (e.g. emotion) based on observables. The link between affective states and biophysical data, collected in lab settings, has been established successfully. However, the number of realistic studies targeting affect detection in the wild is still limited. In this paper we present an exploratory field study, using physiological data of 11 healthy subjects. We aim to classify arousal, State-Trait Anxiety Inventory (STAI), stress, and valence self-reports, utilizing feature-based and convolutional neural network (CNN) methods. In addition, we extend the CNNs to multi-task CNNs, classifying all labels of interest simultaneously. Comparing the F1 score averaged over the different tasks and classifiers the CNNs reach an 1.8% higher score than the classical methods. However, the F1 scores barely exceed 45%. In the light of these results, we discuss pitfalls and challenges for physiology-based affective computing in the wild.

Seismocardiography (SCG) is a cardiac diagnostic method which evaluates vibrations of the chest via accelerometers. Current prototypical wearable SCG sensor patches are combined with electrocardiography (ECG) sensors which enables the estimation of cardiac timing intervals such as pre-ejection period (PEP). We envision a system with only SCG sensors to increase the battery run time and wearing comfort. Therefore, we present a method to predict the timing of ECG R-peak based on SCG data. The developed machine learning (ML) algorithms were evaluated with respect to algorithm complexity, body postures, and personalization schemes, on a dataset consisting of 10 subjects. Our findings show that R-peaks can be predicted with a mean error of 4:16 ms for supine, 9:43 ms for sitting, and 14:3 ms for standing posture. While personalization results only in minor improvements, using more complex ML classifier is beneficial for standing posture, which is more affected by motion artifacts.

Deep learning (DL) methods receive increasing attention within the field of human activity recognition (HAR) due to their success in other machine learning domains. Nonetheless, a direct transfer of these methods is often not possible due to domain specific challenges (e.g. handling of multi-modal sensor data, lack of large labeled datasets). In this paper, we address three key aspects for the future development of robust DL methods for HAR: (1) Is it beneficial to apply data specific normalization? (2) How to optimally fuse multimodal sensor data? (3) How robust are these approaches with respect to available training data? We evaluate convolutional neuronal networks (CNNs) on a new large real-world multimodal dataset (RBK) as well as the PAMAP2 dataset. Our results indicate that sensor specific normalization techniques are required. We present a novel pressure specific normalization method which increases the F1-score by ∼ 4.5 percentage points (pp) on the RBK dataset. Further, we show that late- and hybrid fusion techniques are superior compared to early fusion techniques, increasing the F1-score by up to 3.5 pp (RBK dataset). Finally, our results reveal that in particular CNNs based on a shared filter approach have a smaller dependency on the amount of available training data compared to other fusion techniques.

In extracranial robotic radiotherapy, tumour motion due to respiration is compensated based external markers. Two models are typically used to enable a real-time adaptation. A prediction model, which compensates time latencies of the treatment systems due to e.g. kinematic limitations, and a correlation model, which estimates the internal tumour position based on external markers. We present a novel approach based on multi-task Gaussian Processes (MTGP) which enables an efficient combination of both models by simultaneously learning the correlation and temporal delays between markers. The approach is evaluated using datasets acquired from porcine and human studies. We conclude that the prediction accuracy of MTGP is superior to that of existing methods and can be further increased by using multivariate input data. We investigate the dependency of the number of internal training points and the potential for using the marginal likelihood for model selection.

Gaussian process (GP) models are a flexible means of performing nonparametric Bayesian regression. However, GP models in healthcare are often only used to model a single univariate output time series, denoted as single-task GPs (STGP). Due to an increasing prevalence of sensors in healthcare settings, there is an urgent need for robust multivariate time-series tools. Here, we propose a method using multitask GPs (MTGPs) which can model multiple correlated multivariate physiological time series simultaneously. The flexible MTGP framework can learn the correlation between multiple signals even though they might be sampled at different frequencies and have training sets available for different intervals. Furthermore, prior knowledge of any relationship between the time series such as delays and temporal behavior can be easily integrated. A novel normalization is proposed to allow interpretation of the various hyperparameters used in the MTGP. We investigate MTGPs for physiological monitoring with synthetic data sets and two real-world problems from the field of patient monitoring and radiotherapy. The results are compared with standard Gaussian processes and other existing methods in the respective biomedical application areas. In both cases, we show that our framework learned the correlation between physiological time series efficiently, outperforming the existing state of the art.

Gaussian process (GP) models are a flexible means of performing non-parametric Bayesian regression. However, the majority of existing work using GP models in healthcare data is defined for univariate output time-series, denoted as single-task GPs (STGP). Here, we investigate how GPs could be used to model multiple correlated univariate physiological time-series simultaneously. The resulting multi-task GP (MTGP) framework can learn the correlation within multiple signals even though they might be sampled at different frequencies and have training sets available for different intervals. We illustrate the basic properties of MTGPs using a synthetic case-study with respiratory motion data. Finally, two real-world biomedical problems are investigated from the field of patient monitoring and motion compensation in radiotherapy. The results are compared to STGPs and other standard methods in the respective fields. In both cases, MTGPs learned the correlation between physiological time-series efficiently, which leads to improved modelling accuracy. © 2014 IEEE.