Multichannel ballistocardiography: A comparative analysis of heartbeat detection across different body locations.

The paper presents a validation of novel multichannel ballistocardiography (BCG) measuring system, enabling heartbeat detection from information about movements during myocardial contraction and dilatation of arteries due to blood expulsion. The proposed methology includes novel sensory system and signal processing procedure based on Wavelet transform and Hilbert transform. Because there are no existing recommendations for BCG sensor placement, the study focuses on investigation of BCG signal quality measured from eight different locations within the subject's body. The analysis of BCG signals is primarily based on heart rate (HR) calculation, for which a J-wave detection based on decision-making processes was used. Evaluation of the proposed system was made by comparing with electrocardiography (ECG) as a gold standard, when the averaged signal from all sensors reached HR detection sensitivity higher than 95% and two sensors showed a significant difference from ECG measurement.

Advancing the integration of biosignal-based automated pain assessment methods into a comprehensive model for addressing cancer pain.

BACKGROUND: Tailoring effective strategies for cancer pain management requires a careful analysis of multiple factors that influence pain phenomena and, ultimately, guide the therapy. While there is a wealth of research on automatic pain assessment (APA), its integration with clinical data remains inadequately explored. This study aimed to address the potential correlations between subjective and APA-derived objectives variables in a cohort of cancer patients. METHODS: A multidimensional statistical approach was employed. Demographic, clinical, and pain-related variables were examined. Objective measures included electrodermal activity (EDA) and electrocardiogram (ECG) signals. Sensitivity analysis, multiple factorial analysis (MFA), hierarchical clustering on principal components (HCPC), and multivariable regression were used for data analysis. RESULTS: The study analyzed data from 64 cancer patients. MFA revealed correlations between pain intensity, type, Eastern Cooperative Oncology Group Performance status (ECOG), opioids, and metastases. Clustering identified three distinct patient groups based on pain characteristics, treatments, and ECOG. Multivariable regression analysis showed associations between pain intensity, ECOG, type of breakthrough cancer pain, and opioid dosages. The analyses failed to find a correlation between subjective and objective pain variables. CONCLUSIONS: The reported pain perception is unrelated to the objective variables of APA. An in-depth investigation of APA is required to understand the variables to be studied, the operational modalities, and above all, strategies for appropriate integration with data obtained from self-reporting. TRIAL REGISTRATION: This study is registered with ClinicalTrials.gov, number (NCT04726228), registered 27 January 2021, https://classic. CLINICALTRIALS: gov/ct2/show/NCT04726228?term=nct04726228&draw=2&rank=1.

Continuous sepsis trajectory prediction using tensor-reduced physiological signals.

The quick Sequential Organ Failure Assessment (qSOFA) system identifies an individual's risk to progress to poor sepsis-related outcomes using minimal variables. We used Support Vector Machine, Learning Using Concave and Convex Kernels, and Random Forest to predict an increase in qSOFA score using electronic health record (EHR) data, electrocardiograms (ECG), and arterial line signals. We structured physiological signals data in a tensor format and used Canonical Polyadic/Parallel Factors (CP) decomposition for feature reduction. Random Forests trained on ECG data show improved performance after tensor decomposition for predictions in a 6-h time frame (AUROC 0.67 ± 0.06 compared to 0.57 ± 0.08, p = 0.01 ). Adding arterial line features can also improve performance (AUROC 0.69 ± 0.07, p < 0.01 ), and benefit from tensor decomposition (AUROC 0.71 ± 0.07, p = 0.01 ). Adding EHR data features to a tensor-reduced signal model further improves performance (AUROC 0.77 ± 0.06, p < 0.01 ). Despite reduction in performance going from an EHR data-informed model to a tensor-reduced waveform data model, the signals-informed model offers distinct advantages. The first is that predictions can be made on a continuous basis in real-time, and second is that these predictions are not limited by the availability of EHR data. Additionally, structuring the waveform features as a tensor conserves structural and temporal information that would otherwise be lost if the data were presented as flat vectors.

Compressed Deep Learning Models for Wearable Atrial Fibrillation Detection through Attention.

Deep learning (DL) models have shown promise for the accurate detection of atrial fibrillation (AF) from electrocardiogram/photoplethysmography (ECG/PPG) data, yet deploying these on resource-constrained wearable devices remains challenging. This study proposes integrating a customized channel attention mechanism to compress DL neural networks for AF detection, allowing the model to focus only on the most salient time-series features. The results demonstrate that applying compression through channel attention significantly reduces the total number of model parameters and file size while minimizing loss in detection accuracy. Notably, after compression, performance increases for certain model variants in key AF databases (ADB and C2017DB). Moreover, analyzing the learned channel attention distributions after training enhances the explainability of the AF detection models by highlighting the salient temporal ECG/PPG features most important for its diagnosis. Overall, this research establishes that integrating attention mechanisms is an effective strategy for compressing large DL models, making them deployable on low-power wearable devices. We show that this approach yields compressed, accurate, and explainable AF detectors ideal for wearables. Incorporating channel attention enables simpler yet more accurate algorithms that have the potential to provide clinicians with valuable insights into the salient temporal biomarkers of AF. Our findings highlight that the use of attention is an important direction for the future development of efficient, high-performing, and interpretable AF screening tools for wearable technology.

Atrial Fibrillation Prediction Based on Recurrence Plot and ResNet.

Atrial fibrillation (AF) is the most prevalent form of arrhythmia, with a rising incidence and prevalence worldwide, posing significant implications for public health. In this paper, we introduce an approach that combines the Recurrence Plot (RP) technique and the ResNet architecture to predict AF. Our method involves three main steps: using wavelet filtering to remove noise interference; generating RPs through phase space reconstruction; and employing a multi-level chained residual network for AF prediction. To validate our approach, we established a comprehensive database consisting of electrocardiogram (ECG) recordings from 1008 AF patients and 48,292 Non-AF patients, with a total of 2067 and 93,129 ECGs, respectively. The experimental results demonstrated high levels of prediction precision (90.5%), recall (89.1%), F1 score (89.8%), accuracy (93.4%), and AUC (96%) on our dataset. Moreover, when tested on a publicly available AF dataset (AFPDB), our method achieved even higher prediction precision (94.8%), recall (99.4%), F1 score (97.0%), accuracy (97.0%), and AUC (99.7%). These findings suggest that our proposed method can effectively extract subtle information from ECG signals, leading to highly accurate AF predictions.

Rare Presentation of Wide QRS Tachycardia in a Patient in Their 40s.

This article describes the case of a 40-year-old individual who presented with fulminant myocarditis. Initial ECG displayed sinus tachycardia with a heart rate of 117 bpm, QS complexes in leads V1-V3, ST-segment depression in leads II, III, aVF, V5-V6, and ST-segment elevation >0.2 mV in leads V1 through V3. The initial clinical assessment suggested an acute anteroseptal myocardial infarction. However, subsequent diagnostic evaluation through coronary angiography disclosed that the coronary arteries were normal. Therefore, clinicians should carefully consider the differential diagnosis between these conditions, as their management strategies differ markedly. Two hours after admission, the patient unexpectedly developed syncope. The ECG findings were consistent with the typical characteristics of bidirectional ventricular tachycardia. Our report described the appearance and morphology as well as mechanism of bidirectional ventricular tachycardia in detail. Additionally, we delineate differential diagnoses for disease that can cause bidirectional ventricular tachycardia, such as aconite poisoning, digoxin overdose, immune checkpoint inhibitor (ICI), myocardial ischemia, and hereditary channelopathies, such as catecholaminergic polymorphic ventricular tachycardia (CPVT) and Andersen-Tawil syndrome. Therefore, clinicians should recognize this ECG finding immediately and initiate appropriate treatment promptly as these measures may be vital in saving the patient's life.

Seismocardiography and echocardiography: the correlation in the systolic complex.

Aim. This study aimed to investigate the correlation between seismocardiographic and echocardiographic systolic variables and whether a decrease in preload could be detected by the seismocardiography (SCG).Methods. This study included a total of 34 subjects. SCG and electrocardiography were recorded simultaneously followed by echocardiography (echo) in both supine and 30◦head-up tilted position. The SCG signals was segmented into individual heartbeats and systolic fiducial points were defined using a detection algorithm. Statistical analysis included correlation coefficient calculations and paired sample tests.Results. SCG was able to measure a decrease in preload by almost all of the examined systolic SCG variables. It was possible to correlate certain echo variables to SCG time intervals, amplitudes, and peak to peak intervals. Also, changes between supineand tilted position of some SCG variables were possible to correlate to changes in echo variables. LVET, IVCT, S', strain, SR, SV, and LVEF were significantly correlated to relevant SCG variables.Conclusion. This study showed a moderate correlation, between systolic echo and systolic SCG variables. Additionally, systolic SCG variables were able to detect a decrease in preload.

[Characteristics of electrocardiogram in fulminant myocarditis].

Objective: To investugate the unique electrocardiogram (ECG) characteristics of fulminant myocarditis (FM) patients and provide important clues for the diagnosis of FM. Methods: This was a retrospective study. Patients diagnosed with acute myocarditis at Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology from February 2017 to April 2022 were enrolled and divided into fulminant myocarditis group (FM) and non-fulminant myocarditis group (NFM) according to clinical diagnosis. A total of 246 healthy people who underwent physical examination in the Health examination Center of Tongji Hospital at the same period were selected as the control group. The clinical data and ECG characteristics of the above 3 groups were analyzed and compared. Logistic regression model was used to analyze the influence of ECG parameters on left ventricular ejection fraction in FM patients. Receiver operating curves were constructed to evaluate the predictive value of different ECG parameters for FM. Results: A total of 180 patients were included in this study (FM group: n=123; NFM group: n=57), with an age of (35.0±16.2) years and 106 males (58.89%). Compared with NFM group, ECG was significantly abnormal in FM group, with a higher incidence of sinus tachycardia, ventricular tachycardia or ventricular fibrillation, escape rhythm, right bundle branch block, third degree atrioventricular block, ST-segment elevation, low voltage, prolonged QTc interval, and widened QRS wave in the FM group (all P<0.05). The ECG parameters showed that the amplitude of the full lead QRS wave in FM group was lower than that in NFM group (P<0.01). The average heart rate and QTc interval of FM group were significantly higher than those of NFM and control groups (all P<0.05). Although ST-segment elevation had a higher incidence in the FM group, ECG parameters showed that except for leads Ⅲ and aVF, the ST segment levels in all leads in the FM group were lower than those in the control group (all P<0.05). There was a statistically significant difference in some ST segment changes between FM and NFM groups, while there was no statistical difference between the NFM and control groups. Multivariate regression analysis showed that widened QRS wave and increased heart rate were the influencing factors for left ventricular systolic dysfunction in FM patients (OR=16.914, 95%CI: 1.367-209.224, P=0.028; OR=1.026, 95%CI: 1.010-1.042, P=0.001). Receiver operating curve analysis showed that heart rate>86.90 beat/min, QTc>431.50 ms, and RV5+SV1<1.72 mV had certain predictive value for FM diagnosis. Conclusions: FM patients displayed marked and severe ECG abnormalities, and characteristic changes in ECG can provide important first clues for the diagnosis of FM.

Applying masked autoencoder-based self-supervised learning for high-capability vision transformers of electrocardiographies.

The generalization of deep neural network algorithms to a broader population is an important challenge in the medical field. We aimed to apply self-supervised learning using masked autoencoders (MAEs) to improve the performance of the 12-lead electrocardiography (ECG) analysis model using limited ECG data. We pretrained Vision Transformer (ViT) models by reconstructing the masked ECG data with MAE. We fine-tuned this MAE-based ECG pretrained model on ECG-echocardiography data from The University of Tokyo Hospital (UTokyo) for the detection of left ventricular systolic dysfunction (LVSD), and then evaluated it using multi-center external validation data from seven institutions, employing the area under the receiver operating characteristic curve (AUROC) for assessment. We included 38,245 ECG-echocardiography pairs from UTokyo and 229,439 pairs from all institutions. The performances of MAE-based ECG models pretrained using ECG data from UTokyo were significantly higher than that of other Deep Neural Network models across all external validation cohorts (AUROC, 0.913-0.962 for LVSD, p < 0.001). Moreover, we also found improvements for the MAE-based ECG analysis model depending on the model capacity and the amount of training data. Additionally, the MAE-based ECG analysis model maintained high performance even on the ECG benchmark dataset (PTB-XL). Our proposed method developed high performance MAE-based ECG analysis models using limited ECG data.

Design, Fabrication, and Evaluation of 3D Biopotential Electrodes and Intelligent Garment System for Sports Monitoring.

This study presents the development and evaluation of an innovative intelligent garment system, incorporating 3D knitted silver biopotential electrodes, designed for long-term sports monitoring. By integrating advanced textile engineering with wearable monitoring technologies, we introduce a novel approach to real-time physiological signal acquisition, focusing on enhancing athletic performance analysis and fatigue detection. Utilizing low-resistance silver fibers, our electrodes demonstrate significantly reduced skin-to-electrode impedance, facilitating improved signal quality and reliability, especially during physical activities. The garment system, embedded with these electrodes, offers a non-invasive, comfortable solution for continuous ECG and EMG monitoring, addressing the limitations of traditional Ag/AgCl electrodes, such as skin irritation and signal degradation over time. Through various experimentation, including impedance measurements and biosignal acquisition during cycling activities, we validate the system's effectiveness in capturing high-quality physiological data. Our findings illustrate the electrodes' superior performance in both dry and wet conditions. This study not only advances the field of intelligent garments and biopotential monitoring, but also provides valuable insights for the application of intelligent sports wearables in the future.

Higher-Order Spectral Analysis Combined with a Convolution Neural Network for Atrial Fibrillation Detection-Preliminary Study.

The global burden of atrial fibrillation (AFIB) is constantly increasing, and its early detection is still a challenge for public health and motivates researchers to improve methods for automatic AFIB prediction and management. This work proposes higher-order spectra analysis, especially the bispectrum of electrocardiogram (ECG) signals combined with the convolution neural network (CNN) for AFIB detection. Like other biomedical signals, ECG is non-stationary, non-linear, and non-Gaussian in nature, so the spectra of higher-order cumulants, in this case, bispectra, preserve valuable features. The two-dimensional (2D) bispectrum images were applied as input for the two CNN architectures with the output AFIB vs. no-AFIB: the pre-trained modified GoogLeNet and the proposed CNN called AFIB-NET. The MIT-BIH Atrial Fibrillation Database (AFDB) was used to evaluate the performance of the proposed methodology. AFIB-NET detected atrial fibrillation with a sensitivity of 95.3%, a specificity of 93.7%, and an area under the receiver operating characteristic (ROC) of 98.3%, while for GoogLeNet results for sensitivity and specificity were equal to 96.7%, 82%, respectively, and the area under ROC was equal to 96.7%. According to preliminary studies, bispectrum images as input to 2D CNN can be successfully used for AFIB rhythm detection.

Wearable ECG Device and Machine Learning for Heart Monitoring.

With cardiovascular diseases (CVD) remaining a leading cause of mortality, wearable devices for monitoring cardiac activity have gained significant, renewed interest among the medical community. This paper introduces an innovative ECG monitoring system based on a single-lead ECG machine, enhanced using machine learning methods. The system only processes and analyzes ECG data, but it can also be used to predict potential heart disease at an early stage. The wearable device was built on the ADS1298 and a microcontroller STM32L151xD. A server module based on the architecture style of the REST API was designed to facilitate interaction with the web-based segment of the system. The module is responsible for receiving data in real time from the microcontroller and delivering this data to the web-based segment of the module. Algorithms for analyzing ECG signals have been developed, including band filter artifact removal, K-means clustering for signal segmentation, and PQRST analysis. Machine learning methods, such as isolation forests, have been employed for ECG anomaly detection. Moreover, a comparative analysis with various machine learning methods, including logistic regression, random forest, SVM, XGBoost, decision forest, and CNNs, was conducted to predict the incidence of cardiovascular diseases. Convoluted neural networks (CNN) showed an accuracy of 0.926, proving their high effectiveness for ECG data processing.

IoT-Based Heartbeat Rate-Monitoring Device Powered by Harvested Kinetic Energy.

Remote patient-monitoring systems are helpful since they can provide timely and effective healthcare facilities. Such online telemedicine is usually achieved with the help of sophisticated and advanced wearable sensor technologies. The modern type of wearable connected devices enable the monitoring of vital sign parameters such as: heart rate variability (HRV) also known as electrocardiogram (ECG), blood pressure (BLP), Respiratory rate and body temperature, blood pressure (BLP), respiratory rate, and body temperature. The ubiquitous problem of wearable devices is their power demand for signal transmission; such devices require frequent battery charging, which causes serious limitations to the continuous monitoring of vital data. To overcome this, the current study provides a primary report on collecting kinetic energy from daily human activities for monitoring vital human signs. The harvested energy is used to sustain the battery autonomy of wearable devices, which allows for a longer monitoring time of vital data. This study proposes a novel type of stress- or exercise-monitoring ECG device based on a microcontroller (PIC18F4550) and a Wi-Fi device (ESP8266), which is cost-effective and enables real-time monitoring of heart rate in the cloud during normal daily activities. In order to achieve both portability and maximum power, the harvester has a small structure and low friction. Neodymium magnets were chosen for their high magnetic strength, versatility, and compact size. Due to the non-linear magnetic force interaction of the magnets, the non-linear part of the dynamic equation has an inverse quadratic form. Electromechanical damping is considered in this study, and the quadratic non-linearity is approximated using MacLaurin expansion, which enables us to find the law of motion for general case studies using classical methods for dynamic equations and the suitable parameters for the harvester. The oscillations are enabled by applying an initial force, and there is a loss of energy due to the electromechanical damping. A typical numerical application is computed with Matlab 2015 software, and an ODE45 solver is used to verify the accuracy of the method.

Driver Fatigue Detection Using Heart Rate Variability Features from 2-Minute Electrocardiogram Signals While Accounting for Sex Differences.

Traffic accidents due to fatigue account for a large proportion of road fatalities. Based on simulated driving experiments with drivers recruited from college students, this paper investigates the use of heart rate variability (HRV) features to detect driver fatigue while considering sex differences. Sex-independent and sex-specific differences in HRV features between alert and fatigued states derived from 2 min electrocardiogram (ECG) signals were determined. Then, decision trees were used for driver fatigue detection using the HRV features of either all subjects or those of only males or females. Nineteen, eighteen, and thirteen HRV features were significantly different (Mann-Whitney U test, p < 0.01) between the two mental states for all subjects, males, and females, respectively. The fatigue detection models for all subjects, males, and females achieved classification accuracies of 86.3%, 94.8%, and 92.0%, respectively. In conclusion, sex differences in HRV features between drivers' mental states were found according to both the statistical analysis and classification results. By considering sex differences, precise HRV feature-based driver fatigue detection systems can be developed. Moreover, in contrast to conventional methods using HRV features from 5 min ECG signals, our method uses HRV features from 2 min ECG signals, thus enabling more rapid driver fatigue detection.

A Novel Workflow for Electrophysiology Studies in Patients with Brugada Syndrome.

Brugada Syndrome (BrS) is a primary electrical epicardial disease characterized by ST-segment elevation followed by a negative T-wave in the right precordial leads on the surface electrocardiogram (ECG), also known as the 'type 1' ECG pattern. The risk stratification of asymptomatic individuals with spontaneous type 1 ECG pattern remains challenging. Clinical and electrocardiographic prognostic markers are known. As none of these predictors alone is highly reliable in terms of arrhythmic prognosis, several multi-factor risk scores have been proposed for this purpose. This article presents a new workflow for processing endocardial signals acquired with high-density RV electro-anatomical mapping (HDEAM) from BrS patients. The workflow, which relies solely on Matlab software, calculates various electrical parameters and creates multi-parametric maps of the right ventricle. The workflow, but it has already been employed in several research studies involving patients carried out by our group, showing its potential positive impact in clinical studies. Here, we will provide a technical description of its functionalities, along with the results obtained on a BrS patient who underwent an endocardial HDEAM.

A Novel Instruction Driven 1-D CNN Processor for ECG Classification.

Electrocardiography (ECG) has emerged as a ubiquitous diagnostic tool for the identification and characterization of diverse cardiovascular pathologies. Wearable health monitoring devices, equipped with on-device biomedical artificial intelligence (AI) processors, have revolutionized the acquisition, analysis, and interpretation of ECG data. However, these systems necessitate AI processors that exhibit flexible configuration, facilitate portability, and demonstrate optimal performance in terms of power consumption and latency for the realization of various functionalities. To address these challenges, this study proposes an instruction-driven convolutional neural network (CNN) processor. This processor incorporates three key features: (1) An instruction-driven CNN processor to support versatile ECG-based application. (2) A Processing element (PE) array design that simultaneously considers parallelism and data reuse. (3) An activation unit based on the CORDIC algorithm, supporting both Tanh and Sigmoid computations. The design has been implemented using 110 nm CMOS process technology, occupying a die area of 1.35 mm2 with 12.94 µW power consumption. It has been demonstrated with two typical ECG AI applications, including two-class (i.e., normal/abnormal) classification and five-class classification. The proposed 1-D CNN algorithm performs with a 97.95% accuracy for the two-class classification and 97.9% for the five-class classification, respectively.

Sensor-Assisted Analysis of Autonomic and Cerebrovascular Dysregulation following Concussion in an Individual with a History of Ten Concussions: A Case Study.

INTRODUCTION: Concussion is known to cause transient autonomic and cerebrovascular dysregulation that generally recovers; however, few studies have focused on individuals with an extensive concussion history. METHOD: The case was a 26-year-old male with a history of 10 concussions, diagnosed for bipolar type II disorder, mild attention-deficit hyperactivity disorder, and a history of migraines/headaches. The case was medicated with Valproic Acid and Escitalopram. Sensor-based baseline data were collected within six months of his injury and on days 1-5, 10, and 14 post-injury. Symptom reporting, heart rate variability (HRV), neurovascular coupling (NVC), and dynamic cerebral autoregulation (dCA) assessments were completed using numerous biomedical devices (i.e., transcranial Doppler ultrasound, 3-lead electrocardiography, finger photoplethysmography). RESULTS: Total symptom and symptom severity scores were higher for the first-week post-injury, with physical and emotional symptoms being the most impacted. The NVC response showed lowered activation in the first three days post-injury, while autonomic (HRV) and autoregulation (dCA) were impaired across all testing visits occurring in the first 14 days following his concussion. CONCLUSIONS: Despite symptom resolution, the case demonstrated ongoing autonomic and autoregulatory dysfunction. Larger samples examining individuals with an extensive history of concussion are warranted to understand the chronic physiological changes that occur following cumulative concussions through biosensing devices.

MDDBranchNet: A Deep Learning Model for Detecting Major Depressive Disorder Using ECG Signal.

Major depressive disorder (MDD) is a chronic mental illness which affects people's well-being and is often detected at a later stage of depression with a likelihood of suicidal ideation. Early detection of MDD is thus necessary to reduce the impact, however, it requires monitoring vitals in daily living conditions. EEG is generally multi-channel and due to difficulty in signal acquisition, it is unsuitable for home-based monitoring, whereas, wearable sensors can collect single-channel ECG. Classical machine-learning based MDD detection studies commonly use various heart rate variability features. Feature generation, which requires domain knowledge, is often challenging, and requires computation power, often unsuitable for real time processing, MDDBranchNet is a proposed parallel-branch deep learning model for MDD binary classification from a single channel ECG which uses additional ECG-derived signals such as R-R signal and degree distribution time series of horizontal visibility graph. The use of derived branches was able to increase the model's accuracy by around 7%. An optimal 20-second overlapped segmentation of ECG recording was found to be beneficial with a 70% prediction threshold for maximum MDD detection with a minimum false positive rate. The proposed model evaluated MDD prediction from signal excerpts, irrespective of location (first, middle or last one-third of the recording), instead of considering the entire ECG signal with minimal performance variation stressing the idea that MDD phenomena are likely to manifest uniformly throughout the recording.

Evaluating regression and probabilistic methods for ECG-based electrolyte prediction.

Imbalances in electrolyte concentrations can have severe consequences, but accurate and accessible measurements could improve patient outcomes. The current measurement method based on blood tests is accurate but invasive and time-consuming and is often unavailable for example in remote locations or an ambulance setting. In this paper, we explore the use of deep neural networks (DNNs) for regression tasks to accurately predict continuous electrolyte concentrations from electrocardiograms (ECGs), a quick and widely adopted tool. We analyze our DNN models on a novel dataset of over 290,000 ECGs across four major electrolytes and compare their performance with traditional machine learning models. For improved understanding, we also study the full spectrum from continuous predictions to a binary classification of extreme concentration levels. Finally, we investigate probabilistic regression approaches and explore uncertainty estimates for enhanced clinical usefulness. Our results show that DNNs outperform traditional models but model performance varies significantly across different electrolytes. While discretization leads to good classification performance, it does not address the original problem of continuous concentration level prediction. Probabilistic regression has practical potential, but our uncertainty estimates are not perfectly calibrated. Our study is therefore a first step towards developing an accurate and reliable ECG-based method for electrolyte concentration level prediction-a method with high potential impact within multiple clinical scenarios.