Search Results (41)

Objective: To evaluate the feasibility and effectiveness of wrist-worn wearable devices combined with machine learning (ML) approaches for detecting a diverse array of seizure types beyond generalized tonic–clonic (GTC), including focal, generalized, and subclinical seizures. Materials and Methods: Twenty-eight patients undergoing inpatient video-EEG monitoring at Mayo Clinic were concurrently monitored using Empatica E4 wrist-worn devices. These devices captured accelerometry, blood volume pulse, electrodermal activity, skin temperature, and heart rate. Seizures were annotated by neurologists. The data were preprocessed to experiment with various segment lengths (10 s and 60 s) and multiple feature sets. Three ML strategies, XGBoost, deep learning models (LSTM, CNN, Transformer), and ROCKET, were evaluated using leave-one-patient-out cross-validation. Performance was assessed using area under the receiver operating characteristic curve (AUROC), seizure-wise recall (SW-Recall), and false alarms per hour (FA/h). Results: Detection performance varied by seizure type and model. GTC seizures were detected most reliably (AUROC = 0.86, SW-Recall = 0.81, FA/h = 3.03). Hyperkinetic and tonic seizures showed high SW-Recall but also high FA/h. Subclinical and aware-dyscognitive seizures exhibited the lowest SW-Recall and highest FA/h. MultiROCKET and XGBoost performed best overall, though no single model was optimal for all seizure types. Longer segments (60 s) generally reduced FA/h. Feature set effectiveness varied, with multi-biosignal sets improving performance across seizure types. Conclusions: Wrist-worn wearables combined with ML can extend seizure detection beyond GTC seizures, though performance remains limited for non-motor types. Optimizing model selection, feature sets, and segment lengths, and minimizing false alarms, are key to clinical utility and real-world adoption. Full article

Remote photoplethysmography (rPPG) offers a promising solution for non-contact driver monitoring by detecting subtle blood flow-induced facial color changes from video. However, motion artifacts in dynamic driving environments remain key challenges. This study presents an rPPG framework that combines signal processing techniques before and after applying Eulerian Video Magnification (EVM) for pulse rate (PR) estimation in driving simulators. While not novel, the approach offers insights into the efficiency of the EVM method and its time complexity. We compare results of the proposed rPPG approach against reference Empatica E4 data and also compare it with existing achievements from the literature. Additionally, the possible bias of the Empatica E4 is further assessed using an independent dataset with both the Empatica E4 and the Faros 360 measurements. EVM slightly improves PR estimation, reducing the mean absolute error (MAE) from 6.48 bpm to 5.04 bpm (the lowest MAE (~2 bpm) was achieved under strict conditions) with an additional time required for EVM of about 20 s for 30 s sequence. Furthermore, statistically significant differences are identified between younger and older drivers in both reference and rPPG data. Our findings demonstrate the feasibility of using rPPG-based PR monitoring, encouraging further research in driving simulations. Full article

This study exploited the skin temperature signal derived from a wrist-worn wearable device to define potential digital biomarkers for glycemia levels. Characterization of the skin temperature signal measured through the Empatica E4 device was obtained in 16 subjects (data taken from a dataset freely available on PhysioNet) by deriving standard metrics and a set of novel metrics describing both the current and the retrospective behavior of the signal. For each subject and for each metric, values that correspond to when glycemia was inside the tight range (70–140 mg/dL) were compared through the Wilcoxon rank-sum test against those above or below the range. For hypoglycemia characterization (below range), retrospective behavior of skin temperature described by the metric CV_{T SD} (standard deviation of the series of coefficient of variation) proved to be the most effective both in daytime and nighttime (100% and 50% of the analyzed subjects, respectively). On the other side, for hyperglycemia characterization (above range), differences were observed between daytime and nighttime, with current behavior of skin temperature, described by M2_T (deviation from the reference value of 32 °C), being the most informative during daytime, whereas retrospective behavior, described by SD_{T hhmm} (standard deviation of the series of means), showed the highest effectiveness during nighttime. Proposed variability features outperformed standard metrics, and in future studies, their integration with other digital biomarkers of glycemia could improve the performance of applications devoted to non-invasive detection of glycemic events. Full article

Technological advances continually reduce the effort to digitally transform health-related activities such as rehabilitation and training. Exemplary systems use tracking and vital sign monitoring to assess physical condition and training progress. This paper presents a system for body stability training and coordination evaluation, using cost-efficient tracking and monitoring solutions. It implements the use case of app-guided back posture tracking on the ICAROS Pro training device via SteamVR Tracking 2.0, with pulse and respiration rate monitoring via Zephyr BioHarness 3.0. A longitudinal study on training effects with 20 subjects was conducted, involving a representative procedure created with a sports manager. Posture errors served as the main progress indicator, and pulse and respiration rates as co-indicators. Outcomes suggest the system’s capabilities to foster comprehension of effects and steering of exercises. Further, a secondary study presents a self-developed VR-based exergame demo for future system expansion. The Empatica EmbracePlus smartwatch was used as an alternative for vital sign acquisition. The user experiences of five subjects gathered via a survey highlight its motivating and entertaining character. For both the main and secondary studies, a thorough discussion elaborates on potentials and current limitations. The developed training system can serve as template and be adjusted for further use cases, and the exergame’s reception revealed prospective extension directions. Software components are available via GitHub. Full article

Cognitive rumination, a transdiagnostic symptom across mental health disorders, has traditionally been assessed through self-report measures. However, these measures are limited by their temporal nature and subjective bias. The rise in wearable technologies offers the potential for continuous, real-time monitoring of physiological indicators associated with rumination. This scoping review investigates the current state of research on using wearable technology to detect cognitive rumination. Specifically, we examine the sensors and wearable devices used, physiological biomarkers measured, standard measures of rumination used, and the comparative validity of specific biomarkers in identifying cognitive rumination. The review was performed according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines on IEEE, Scopus, PubMed, and PsycInfo databases. Studies that used wearable devices to measure rumination-related physiological responses and biomarkers were included (n = 9); seven studies assessed one biomarker, and two studies assessed two biomarkers. Electrodermal Activity (EDA) sensors capturing skin conductance activity emerged as both the most prevalent sensor (n = 5) and the most comparatively valid biomarker for detecting cognitive rumination via wearable devices. Other commonly investigated biomarkers included electrical brain activity measured through Electroencephalogram (EEG) sensors (n = 2), Heart Rate Variability (HRV) measured using Electrocardiogram (ECG) sensors and heart rate fitness monitors (n = 2), muscle response measured through Electromyography (EMG) sensors (n = 1) and movement measured through an accelerometer (n = 1). The Empatica E4 and Empatica Embrace 2 wrist-worn devices were the most frequently used wearable (n = 3). The Rumination Response Scale (RRS), was the most widely used standard scale for assessing rumination. Experimental induction protocols, often adapted from Nolen-Hoeksema and Morrow’s 1993 rumination induction paradigm, were also widely used. In conclusion, the findings suggest that wearable technology offers promise in capturing real-time physiological responses associated with rumination. However, the field is still developing, and further research is needed to validate these findings and explore the impact of individual traits and contextual factors on the accuracy of rumination detection. Full article

The Empatica EmbracePlus is a recent innovation in medical-grade wristband wearable sensors that enable unobtrusive continuous measurement of pulse rate, electrodermal activity, skin temperature, and various accelerometry-based actigraphy measures using a minimalistic smartwatch design. The advantage of this lightweight wearable is the potential for holistic longitudinal recording and monitoring of physiological processes that index a suite of autonomic functions, as well as to provide ecologically valid insights into human behaviour, health, physical activity, and psychophysiological processes. Given the longitudinal nature of wearable recordings, EmbracePlus data collection is managed by storing raw timeseries in short ‘chunks’ in avro file format organised by universal standard time. This is memory-efficient but requires programming expertise to compile the raw data into continuous file formats that can be processed using standard techniques. Currently, there are no accessible tools available to compile and analyse raw EmbracePlus data over user-defined time periods. To address that, we introduce the LOTUS toolkit, an open-source graphical user interface that allows users to reconstitute and process EmbracePlus datasets over select time intervals. LOTUS is available on GitHub, and currently allows users to compile raw EmbracePlus data into unified timeseries stored in more familiar Excel or Matlab file formats to facilitate signal processing and analysis. Future work will expand the toolkit to process Empatica E4 and other wearable signal data, while also integrating more sophisticated functions for feature extraction and analysis. Full article

Driving is integral to many people’s daily existence, but aggressive driving behavior increases the risk of road traffic collisions. Young drivers are more prone to aggressive driving and danger perception impairments. A driver’s physiological state (e.g., fatigue, anger, or stress) can negatively affect their driving performance. This is especially true for young drivers who have limited driving experience. This research focuses on examining the connection between emotional arousal and aggressive driving behavior in young drivers, using predictive analysis based on electrodermal activity (EDA) data through neural networks. The study involved 20 participants aged 18 to 30, who completed 84 driving sessions. During these sessions, their EDA signals and driving behaviors, including acceleration and braking, were monitored using an Empatica E4 wristband and a telematics device. This study conducted two key analyses using neural networks. The first analysis used a comprehensive set of EDA features to predict emotional arousal, achieving an accuracy of 65%. The second analysis concentrated on predicting aggressive driving behaviors by leveraging the top 10 most significant EDA features identified from the arousal prediction model. Initially, the arousal prediction was performed using the complete set of EDA features, from which feature importance was assessed. The top 10 features with the highest importance were then selected to predict aggressive driving behaviors. Another aggressive driving behavior prediction with a refined set of difference features, representing the changes from baseline EDA values, was also utilized in this analysis to enhance the prediction of aggressive driving events. Despite moderate accuracy, these findings suggest that EDA data, particularly difference features, can be valuable in predicting emotional states and aggressive driving, with future research needed to incorporate additional physiological measures for enhanced predictive performance. Full article

Existing literature thoroughly reviews the detrimental consequences that adverse childhood experiences (ACEs) have physically, emotionally, neurobiologically, and financially. It is imperative to develop effective treatments that offer a sense of hope to children who have been impacted. The established relationship between high ACE scores and physiological hyperarousal due to emotional dysregulation is clear in the literature. This relationship indicates that taking psychophysiological measures may be an effective method of gauging the effectiveness of trauma treatments. This study measured the heart rate of a child who had experienced multiple ACEs, during 16 child-centered play therapy (CCPT) sessions, using the Empatica (E4) wristband. Bayesian change point analysis was conducted and multiple changes in the heart rate mean were detected and identified within each session’s time series. Additionally, changes in heart rate variability during the 16 sessions were observed and points of interest, e.g., highest and lowest observed heart rates, were noted. Results suggested the number of breakpoints in the heart rate means within each session, as well as the location, i.e., the time of each breakpoint, so that each significantly detectable change in heart rate mean as well as sessions of noted differences in heart rate variability were discussed alongside what was occurring within the video recorded sessions. Full article

(1) Background: Heart rate variability (HRV) has been investigated in the context of ST-segment elevation myocardial infarction (STEMI). This study contributes to the field by assessing short-term HRV during primary percutaneous coronary intervention (PCI) using wearable technology, providing real-time insights into autonomic function. (2) Methods: This single-center, observational cohort study included 104 STEMI patients undergoing primary percutaneous coronary intervention (PCI). HRV parameters (including SDNN, RMSSD, pNN50, HF, SD1, and SD2/SD1 ratio) were measured using a wearable device (Empatica E4 wristband, CE certified). Measurements were taken throughout the entire duration of the primary PCI, as well as specifically during the initial 5 min and the final 5 min of the procedure. The association between HRV parameters and adverse outcomes, including in-hospital mortality and in-hospital major adverse cardiovascular events (MACE), were assessed. (3) Results: HRV parameters significantly decreased after myocardial revascularization, particularly SDNN, RMSSD, pNN50, HF, SD1, and SD2/SD1 ratio. Significant associations were found between reduced SD2/SD1 ratio, approximate entropy, and adverse outcomes, including increased in-hospital mortality and in-hospital MACE (respectively, p = 0.007, p = 0.017 and p = 0.006 and p = 0.005). The SD2/SD1 ratio was significantly lower in patients who died during the hospital stay (p = 0.008) compared to survivors. Approximate entropy was also significantly lower in deceased patients (p = 0.019). (4) Conclusions: Real-time HRV monitoring using wearable technology offers valuable data regarding dynamic physiological changes during primary PCI. Further studies are required to validate these preliminary results and to explore their potential implications for clinical practice. Full article

The rising incidence of type 2 diabetes underscores the need for technological innovations aimed at enhancing diabetes management by aiding individuals in monitoring their dietary intake. This has resulted in the development of technologies capable of tracking the timing and content of an individual’s meals. However, the ability to use non-invasive wearables to estimate or classify the carbohydrate content of the food an individual has just consumed remains a relatively unexplored area. This study investigates carbohydrate content classification using postprandial heart rate responses from non-invasive wearables. We designed and developed timeStampr, an iOS application for collecting timestamps essential for data labeling and establishing ground truth. We then conducted a pilot study in controlled, yet naturalistic settings. Data were collected from 23 participants using an Empatica E4 device worn on the upper arm, while each participant consumed either low-carbohydrate or carbohydrate-rich foods. Due to sensor irregularities with dark skin tones and non-compliance with the study’s health criteria, we excluded data from three participants. Finally, we configured and trained a Light Gradient Boosting Machine (LGBM) model for carbohydrate content classification. Our classifiers demonstrated robust performance, with the carbohydrate content classification model consistently achieving at least 84% in accuracy, precision, recall, and AUCROC within a 60 s window. The results of this study demonstrate the potential of postprandial heart rate responses from non-invasive wearables in carbohydrate content classification. Full article

This study examines the performance of various machine learning (ML) models in predicting Interstitial Glucose (IG) levels using data from wrist-worn wearable sensors. The insights from these predictions can aid in understanding metabolic syndromes and disease states. A public dataset comprising information from the Empatica E4 smart watch, the Dexcom Continuous Glucose Monitor (CGM) measuring IG, and a food log was utilized. The raw data were processed into features, which were then used to train different ML models. This study evaluates the performance of decision tree (DT), support vector machine (SVM), Random Forest (RF), Linear Discriminant Analysis (LDA), K-Nearest Neighbors (KNN), Gaussian Naïve Bayes (GNB), lasso cross-validation (LassoCV), Ridge, Elastic Net, and XGBoost models. For classification, IG labels were categorized into high, standard, and low, and the performance of the ML models was assessed using accuracy (40–78%), precision (41–78%), recall (39–77%), F1-score (0.31–0.77), and receiver operating characteristic (ROC) curves. Regression models predicting IG values were evaluated based on R-squared values (−7.84–0.84), mean absolute error (5.54–60.84 mg/dL), root mean square error (9.04–68.07 mg/dL), and visual methods like residual and QQ plots. To assess whether the differences between models were statistically significant, the Friedman test was carried out and was interpreted using the Nemenyi post hoc test. Tree-based models, particularly RF and DT, demonstrated superior accuracy for classification tasks in comparison to other models. For regression, the RF model achieved the lowest RMSE of 9.04 mg/dL with an R-squared value of 0.84, while the GNB model performed the worst, with an RMSE of 68.07 mg/dL. A SHAP analysis identified time from midnight as the most significant predictor. Partial dependence plots revealed complex feature interactions in the RF model, contrasting with the simpler interactions captured by LDA. Full article

Stress has various impacts on the health of human beings. Recent success in wearable sensor development, combined with advancements in deep learning to automatically detect features from raw data, opens several interesting applications related to detecting emotional states. Being able to accurately detect stress-related emotional arousal in an acute setting can positively impact the imminent health status of humans, i.e., through avoiding dangerous locations in an urban traffic setting. This work proposes an explainable deep learning methodology for the automatic detection of stress in physiological sensor data, recorded through a non-invasive wearable sensor device, the Empatica E4 wristband. We propose a Long-Short Term-Memory (LSTM) network, extended through a Deep Generative Ensemble of conditional GANs (LSTM DGE), to deal with the low data regime of sparsely labeled sensor measurements. As explainability is often a main concern of deep learning models, we leverage Integrated Gradients (IG) to highlight the most essential features used by the model for prediction and to compare the results to state-of-the-art expert-based stress-detection methodologies in terms of precision, recall, and interpretability. The results show that our LSTM DGE outperforms the state-of-the-art algorithm by 3 percentage points in terms of recall, and 7.18 percentage points in terms of precision. More importantly, through the use of Integrated Gradients as a layer of explainability, we show that there is a strong overlap between model-derived stress features for electrodermal activity and existing literature, which current state-of-the-art stress detection systems in medical research and psychology are based on. Full article

Rett syndrome (RTT) is a paediatric neurodevelopmental disorder spanning four developmental stages. This multi-system disorder offers a unique window to explore genotype–phenotype relationships in a disease model. However, genetic prognosticators of RTT have limited clinical value due to the disorder’s heterogeneity on multiple levels. This case report used a precision medicine approach to better understand the clinical phenotype of RTT twins with an identical pathogenic MECP2 mutation and discordant neurodevelopmental profiles. Targeted genotyping, objective physiological monitoring of heart rate variability (HRV) parameters, and clinical severity were assessed in a RTT twin pair (5 years 7 months old) with an identical pathogenic MECP2 mutation. Longitudinal assessment of autonomic HRV parameters was conducted using the Empatica E4 wristband device, and clinical severity was assessed using the RTT-anchored Clinical Global Impression Scale (RTT-CGI) and the Multi-System Profile of Symptoms Scale (MPSS). Genotype data revealed impaired BDNF function for twin A when compared to twin B. Twin A also had poorer autonomic health than twin B, as indicated by lower autonomic metrics (autonomic inflexibility). Hospitalisation, RTT-CGI-S, and MPSS subscale scores were used as measures of clinical severity, and these were worse in twin A. Treatment using buspirone shifted twin A from an inflexible to a flexible autonomic profile. This was mirrored in the MPSS scores, which showed a reduction in autonomic and cardiac symptoms following buspirone treatment. Our findings showed that a combination of a co-occurring rs6265 BDNF polymorphism, and worse autonomic and clinical profiles led to a poorer prognosis for twin A compared to twin B. Buspirone was able to shift a rigid autonomic profile to a more flexible one for twin A and thereby prevent cardiac and autonomic symptoms from worsening. The clinical profile for twin A represents a departure from the disorder trajectory typically observed in RTT and underscores the importance of wider genotype profiling and longitudinal objective physiological monitoring alongside measures of clinical symptoms and severity when assessing genotype–phenotype relationships in RTT patients with identical pathogenic mutations. A precision medicine approach that assesses genetic and physiological risk factors can be extended to other neurodevelopmental disorders to monitor risk when genotype–phenotype relationships are not so obvious. Full article

Acute psychological stress (APS) is a complex and multifactorial phenomenon that affects metabolism, necessitating real-time detection and interventions to mitigate its effects on glycemia in people with type 1 diabetes. This study investigates the detection of APS using physiological variables measured by the Empatica E4 wristband and employs explainable machine learning to evaluate the importance of the physiological signals. The extreme gradient boosting model is developed for classification of APS and non-stress (NS) with weighted training, achieving an overall accuracy of 99.93%. The Shapley additive explanations (SHAP) technique is employed to interpret the global importance of the physiological signals, determining the order of importance for the variables from most to least as galvanic skin response (GSR), heart rate (HR), skin temperature (ST), and motion sensors (accelerometer readings). The increase in GSR and HR are positively correlated with the occurrence of APS as indicated by high positive SHAP values. The SHAP technique is also used to explain the local signal importance for particular instances of misclassified samples. The detection of APS can inform multivariable automated insulin delivery systems to intervene to counteract the APS-induced glycemic excursions in people with type 1 diabetes. Full article

Today, issues related to people’s mental health and well-being have been described as a challenge for society, globally as well as in Sweden. This calls for new approaches to mental health promotion. The aim was to evaluate the adequacy of its content and structure, describing experiences of study circles as a means of supporting participants’ self-care and self-compassion. The overall design is a descriptive QUAL + quan design, where the quantitative and qualitative results are integrated. Five participants participated in a focus group interview, of whom four completed questionnaires. One individual interview was conducted with the study circle leader. Study circles can be an arena for mental health promotion, as learning and sharing of experience contributes to a sense of coherence, as well as self-compassion and a genuine concern for one’s own and others’ well-being, but are not considered an alternative to psychiatric care for those in need of professional services. Study circles can be a possible means to support self-care and thereby promote mental health in the general population and are a valuable contribution to public health. However, in addition to modifications of the content, further research is needed on the qualifications for study circle leaders, as well as the dissemination of study circles. Full article