Timely diagnosis and treatment of heart blocks are critical for preventing fatal outcomes in patients with cardiac conduction disorders. Expert analysis of clinical 12-lead electro-cardiograms (ECG) remains the standard diagnosis apparatus. In this study, we propose a self-supervised deep learning model that can detect evidence of heart blocks in ECG signals. We build a multichannel residual neural network (KRes) and train this model in a self-supervised contrastive fashion leveraging 3.6 million ECG from patients at the Massachusetts General Hospital (MGH) to learn robust ECG representations. We evaluate the utility of such representations on a large public dataset, PTB-XL, through fine-tuning and linear probing for identifying heart blocks. To compare, we also build a baseline supervised model by adapting KRes. The training data contains a subset of 10.6 thou-sand PTB-XL ECG from patients with and without heart block patterns, and a holdout set of 1319 ECG is used for evaluation. We compare the performances of the finetuning and the linear probing of the self-supervised representations with that of the supervised model using the area under the receiver-operating curve (AUC), sensitivity, specificity, and predictive values, and found those performing equally well. Moreover, we repeat the training cycles of the three pipelines while reducing the number of training samples and demonstrate that the performance of the self-supervised representations remain steady even when labeled data is limited. This research underscores the potential of self-supervised learning in cardiac diagnostics, emphasizing generalizability and performance across diverse datasets and clinical settings, especially in data-scarce paradigms.

Myocardial Infarction (MI) is a major global health concern due to its high mortality and morbidity rates. Early detection of MI is crucial for timely medical intervention and improved patient outcomes. In this study, we investigate the feasibility of predicting MI using lead-I of electrocardiogram (ECG) data, with a focus on practical applications for wearable monitoring systems. Utilizing the PTB-XL dataset, which includes a comprehensive collection of 12-lead ECG recordings with both normal and various MI samples, we employ deep learning techniques to develop a binary classification model. For MI detection using lead-I, we achieved an AUC of 0.92 and an AUPR of 0.82 on the test set. In comparison, using 6-lead and 12-lead configurations both resulted in an AUC of 0.99. These findings demonstrate the potential for detecting MI using only lead-I, as measured by wearable devices. This advancement could significantly enhance clinical outcomes for MI patients by enabling timely detection and intervention.

Self-supervised learning (SSL) for clinical time series data has received significant attention in recent literature, since these data are highly rich and provide important information about a patient's physiological state. However, most existing SSL methods for clinical time series are limited in that they are designed for unimodal time series, such as a sequence of structured features (e.g., lab values and vitals signs) or an individual high-dimensional physiological signal (e.g., an electrocardiogram). These existing methods cannot be readily extended to model time series that exhibit multimodality, with structured features and high-dimensional data being recorded at each timestep in the sequence. In this work, we address this gap and propose a new SSL method --- Sequential Multi-Dimensional SSL --- where a SSL loss is applied both at the level of the entire sequence and at the level of the individual high-dimensional data points in the sequence in order to better capture information at both scales. Our strategy is agnostic to the specific form of loss function used at each level -- it can be contrastive, as in SimCLR, or non-contrastive, as in VICReg. We evaluate our method on two real-world clinical datasets, where the time series contains sequences of (1) high-frequency electrocardiograms and (2) structured data from lab values and vitals signs. Our experimental results indicate that pre-training with our method and then fine-tuning on downstream tasks improves performance over baselines on both datasets, and in several settings, can lead to improvements across different self-supervised loss functions.

Continuous monitoring of breathing rate (BR), minute ventilation (VE), and other respiratory parameters could transform care for and empower patients with chronic cardio-pulmonary conditions, such as asthma. However, the clinical standard for measuring respiration, namely Spirometry, is hardly suitable for continuous use. Wearables can track many physiological signals, like ECG and motion, yet respiration tracking faces many challenges. In this work, we infer respiratory parameters from wearable ECG and wrist motion signals. We propose a modular and generalizable classification-regression pipeline to utilize available context information, such as physical activity, in learning context-conditioned inference models. Novel morphological and power domain features from the wearable ECG are extracted to use with these models. Exploratory feature selection methods are incorporated in this pipeline to discover application-driven interpretable biomarkers. Using data from 15 subjects, we evaluate two implementations of the proposed inference pipeline: for BR and VE. Each implementation compares generalized linear model, random forest, support vector machine, Gaussian process regression, and neighborhood component analysis as regression models. Permutation, regularization, and relevance determination methods are used to rank the ECG features to identify robust ECG biomarkers across models and activities. This work demonstrates the potential of wearable sensors not only in continuous monitoring, but also in designing biomarker-driven preventive measures.

Research in the area of internet-of-things, cyberphysical-systems, and smart health often employ sensor systems at residences for continuous monitoring. Such research-oriented residential monitoring systems (RRMSs) usually face two major challenges, long-term reliable operation management and validation of system functionality with minimal human effort. Targeting these two challenges, this paper describes a monitor of monitoring systems with ground-truth validation capabilities, M2G. It consists of two subsystems, the Monitor2system and the Ground-truth validation system. The Monitor2system encapsulates a flexible set of general-purpose components to monitor the operation and connectivity of heterogeneous sensor devices (e.g. smart watches, smart phones, microphones, beacons, etc.), a local base-station, as well as a cloud server. It provides a user-friendly interface and supports different types of RRMSs in various contexts. The system also features a ground truth validation system to support obtaining ground truth in the field. Additionally, customized alerts can be sent to remote administrators and other personnel to report any dysfunction or inaccuracy of the system in real time. M2G is applied to three very different case studies: the M2FED system which monitors family eating dynamics [1], an in-home wireless sensing system for monitoring nighttime agitation [2], and the BESI system which monitors behavioral and environmental parameters to predict health events and to provide interventions [3]. The results indicate that M2G is a comprehensive system that (i) requires small cost in time and effort to adapt to an existing RRMS, (ii) provides reliable data collection and reduction in data loss by detecting faults in real-time, and (iii) provides a convenient and timely ground truth validation facility.