Individuals with Opioid Use Disorder (OUD) often struggle to maintain sobriety, with many experiencing relapse within the first year. While medication-assisted treatment (MAT) is among the most effective approaches, access to intensive care is often limited by financial barriers. Mobile health (mHealth) technologies offer a promising, cost-effective alternative by enabling continuous monitoring and timely intervention through tools such as ecological momentary assessments (EMAs), wearable sensors, and smartphone data. In this study, we explore the feasibility of using mHealth data to predict emotions that align with cravings in OUD patients undergoing MAT. Using data collected from EMAs, wearables, smartphone tracking, and surveys, we demonstrate that machine learning models can accurately predict emotional states associated with cravings. These findings highlight the potential of mHealth systems to support individuals with OUD through timely and scalable interventions.
Electrodermal activity (EDA) provides a direct indicator of sympathetic nervous system arousal through changes in skin conductance. However, wearable EDA sensing poses challenges such as inconsistent skin contact, electrode impedance variability, motion artifacts, and power constraints. To address these issues, this study presents mobile EDA (mEDA), a compact device driven by a stabilized direct-current source. A validation study was conducted on ten healthy adult participants in a time-synchronized protocol to collect data from BIOPAC and mEDA concurrently. mEDA recordings employed gel electrodes for P1-P5 and dry (textile) electrodes for P6-P10, while the BIOPAC MP160 system used gel electrodes for all participants. Participants underwent a 30-minute protocol of resting, deep breathing, and three cognitive tasks. The preprocessing pipeline consisted of low-pass filter and artifact (sharp peaks and flat line) removal. Cleaned signals were converted into frequency domain components for decomposition into low and high frequency components, skin conductance level (SCL), and skin conductance response (SCR) respectively. SCL and SCR were converted back to the time domain to analyze performance metrics between both devices. Pearson correlation, coherence, and Dynamic Time Warping (DTW) were computed on SCL, while zero-crossing peaks were counted for SCR analysis. With gel electrodes, the average Pearson correlation was 0.92 and the SCR peak count difference was 38. For textile electrodes, the correlation was 0.88 with a peak count difference of 119. Both configurations achieved coherence above 0.95 and DTW below 0.5 for most participants. These results demonstrate mEDA's reliable performance in capturing both tonic and phasic EDA across electrode configurations.
Excessive ultraviolet (UV) exposure is the principal driver of melanoma, yet at-risk individuals seldom receive timely, context-aware cues to apply protection. Existing wrist-worn UV monitors often struggle to recognize timely outdoor exposure because UV readings vary sharply with wrist orientation and sensor angle. To address this gap, we developed a wrist-watch form-factor device that embeds an AS7331 UV photodiode beneath a hemispherical polytetrafluoroethylene (PTFE) dome, which diffuses incident light and stabilizes the sensor's angular response. Ten participants wore the device during routine daily activities, generating more than 93k datapoints of annotated indoor-outdoor data. We implemented an on-device logistic-regression classifier, integrating UVA, UVB, and inertial features to distinguish indoor from outdoor contexts. Under leave-one-participant-out cross-validation, the PTFE-dome watch achieved 94% accuracy and a weighted F1 score of 0.95, outperforming both a flat-window GUVA-S12SD sensor (71% accuracy, F1 = 0.72) and a cylindrical-PTFE enclosure (84% accuracy, F1 = 0.85). By coupling a compact PTFE dome with on-device machine learning (ML), our system delivers reliable, on-wrist UV context sensing and paves the way for unobtrusive, personalized interventions to reduce cumulative UV exposure.
Timely decision making is critical to the effectiveness of mobile health (mHealth) interventions. At predefined timepoints called "decision points," intelligent mHealth systems such as just-in-time adaptive interventions (JITAIs) estimate an individual's biobehavioral context from sensor or survey data and determine whether and how to intervene. For interventions targeting habitual behavior (e.g., oral hygiene), effectiveness often hinges on delivering support shortly before the target behavior is likely to occur. Current practice schedules decision points at a fixed interval (e.g., one hour) before user-provided behavior times, and the fixed interval is kept the same for all individuals. However, this one-size-fits-all approach performs poorly for individuals with irregular routines, often scheduling decision points after the target behavior has already occurred, rendering interventions ineffective. In this paper, we propose SigmaScheduling, a method to dynamically schedule decision points based on uncertainty in predicted behavior times. When behavior timing is more predictable, SigmaScheduling schedules decision points closer to the predicted behavior time; when timing is less certain, SigmaScheduling schedules decision points earlier, increasing the likelihood of timely intervention. We evaluated SigmaScheduling using real-world data from 68 participants in a 10-week trial of Oralytics, a JITAI designed to improve daily toothbrushing. SigmaScheduling increased the likelihood that decision points preceded brushing events in at least 70% of cases, preserving opportunities to intervene and impact behavior. Our results indicate that SigmaScheduling can advance precision mHealth, particularly for JITAIs targeting time-sensitive, habitual behaviors such as oral hygiene or dietary habits.
Mental health disorders are often comorbid, highlighting the need for predictive models that can address multiple outcomes simultaneously. Multi-task learning (MTL) provides a principled approach to jointly model related conditions, enabling shared representations that improve robustness and reduce reliance on large disorder-specific datasets. In this work, we present a tri-modal speech-based framework that integrates text transcriptions, acoustic landmarks, and vocal biomarkers within a large language model (LLM)-driven architecture. Beyond static assessments, we introduce a longitudinal modeling strategy that captures temporal dynamics across repeated clinical interactions, offering deeper insights into symptom progression and relapse risk. Our MTL design simultaneously predicts depression relapse, suicidal ideation, and sleep disturbances, reflecting the comorbid nature of adolescent mental health. Evaluated on the Depression Early Warning (DEW) dataset, the proposed longitudinal trimodal MTL model achieves a balanced accuracy of 70.8%, outperforming unimodal, single-task, and non-longitudinal baselines. These results demonstrate the promise of combining MTL with longitudinal monitoring for scalable, noninvasive prediction of adolescent mental health outcomes.
Prone positioning improves oxygenation in patients with acute respiratory distress syndrome (ARDS), but traditional methods are labor-intensive and complicates continuous monitoring capabilities. Building on a previously developed mechanically assisted proning vest (V/Q vest), this work focuses on integrating and validating a multi-modal sensor network for non-invasive respiratory monitoring. The system incorporates thoracic bioimpedance electrodes and microphones to derive respiratory parameters, including tidal volume (TV), respiratory rate (RR), phase timing and lung sounds. Data from ten healthy adults were collected during controlled breathing tasks in seated and supine postures. Both seated and supine positions showed strong agreement between IP-derived and spirometer-derived TV, with $\mathbf{R}^{\mathbf{2}} \boldsymbol{=} \mathbf{0. 9 0}$, MAE $\boldsymbol{=} \mathbf{0. 1 6} \boldsymbol{\pm} \mathbf{0. 0 7} \mathbf{L}$ in the seated posture and $\mathbf{R}^{\mathbf{2}} =0.94$, $\text{MAE}=0.12 \pm 0.05 \mathrm{L}$ in the supine posture. A flow-based correction method was applied to decouple lung sound intensity from airflow. In the seated posture, mean repeated-measures correlation between uncorrected sound intensity and flow was $r= 0.72$, dropping to $\mathbf{r}=0.01$ after correction. In the supine posture, $\mathbf{r}=0.71$ before correction and $\mathbf{r}=-0.07$ after correction, demonstrating successful flow-decoupling. These results support the feasibility of integrating real-time impedance and acoustic sensing into a therapeutic proning vest, laying the foundation for localized lung monitoring and data-driven respiratory management in critical care.
Continuous glucose monitoring (CGM) devices provide critical real-time data but remain minimally invasive and require frequent replacement. This study presents a novel, personalized machine learning approach for non-invasive glucose monitoring using radiofrequency (RF) spectroscopy to address these limitations. To simulate real-world usage and ensure clinical relevance, we developed a model for a single individual using data collected during standardized meals. The model was trained and tested on data collected on separate days, ensuring that the training and test sets are drawn from distinct, non-overlapping time periods. A comprehensive machine learning pipeline was validated using 3,101 spectral features (400-3500 MHz) combined with contextual data to predict glucose levels. Our best-performing model, multi-layer perception regressor (MLP), achieved a Mean Absolute Relative Difference (MARD) of 11.6%. These findings demonstrate that a personalized machine learning model holds potential to predict glucose non-invasively. This highlights a promising path toward a more user-friendly and sustainable solution for continuous glucose management
Incorporating mobile health (mHealth) into cancer care holds growing promise for improving quality and access, especially in rural and disadvantaged populations, who continue to face disparities in distress prevalence and healthcare engagement. Regardless, sustained engagement with mHealth solutions remains challenging due to motivational, technical, and usability barriers. We developed and piloted a HIPAA-compliant mHealth app, Assuage, in a 45-day longitudinal study with seven cancer patients, the majority from rural backgrounds. Participants alternated between daily and every-other-day symptom reporting. Our findings reflect adherence trends across reporting frequencies, stable and favorable system usability measures before and after the study, and qualitative insights about participants' willingness to use mHealth for symptom self-reporting.
This paper presents the first reported measurements of impedance plethysmography (IPG) signals directly from the human fingertip for the purpose of evaluating peripheral hemodynamics. The human fingertip serves as a vital non-invasive access point for cardiovascular monitoring, with photoplethysmography (PPG) widely adopted for its assessment of superficial blood flow. However, the optical sensing depth of PPG limits its capability to fully characterize deeper peripheral hemodynamics. Here, we developed a novel four-electrode fingertip IPG prototype and measurement protocol, enabling the consistent capture of pulsatile impedance waveforms reflecting blood volume changes within the digit. Preliminary data from two participants illustrates clear signal reproducibility and enables the derivation of key pulse wave metrics, including pulse arrival time (PAT). Given IPG's electrical sensing principle, distinct from PPG's optical approach, our findings suggest that fingertip IPG offers a complementary “new window” into peripheral circulation. This work establishes fingertip IPG as a significant advancement for non-invasive physiological monitoring, holding promise for both clinical applications and wearable health sensing.
Inspired by the progress of motion synthesis models, we leverage cross-modality transfer to generate realistic synthetic Inertial Measurement Unit (IMU) data from textual descriptions, hence Text2IMU. We use an established motion synthesis model and textual descriptions to generate sequences of 3D human activities. To obtain realistic and diverse sensor readings, we created multiple body surface models with different body morphologies. With the text prompts, we let the surface models perform activities and synthesise acceleration and gyroscope data for multiple virtual IMU positions. We show that synthetic data, generated by Text2IMU, can be used to classify activities across three public benchmark datasets. We demonstrate that our Text2IMU synthesis approach does not require measured data of the target domain. Text2IMU yields an average Human Activity Recognition (HAR) accuracy of $\text{7 9. 2 \%}$ for correctly synthesised activities, which doubles the performance of synthetic sensor data obtained from baseline models. We demonstrate that synthetic HAR model training can replace empirical data acquisition when the prompted activities can be successfully generated.
Electrocardiogram (ECG) recording systems are increasingly being integrated into consumer wearable systems such as smartwatches, providing users with access to clinically-relevant information about their heart activity anytime, anywhere. The increasing adoption of in-ear wearables, known as earables, as well as their stable position on the body, makes them an attractive prospect for ECG integration. However, this comes with several challenges. Other biosignals, including those from the brain and surrounding muscles, are detectable at the ear in the same frequency bands with much higher amplitudes. This means that the ECG signal-to-noise ratio (SNR) can be extremely low at this location. The few existing denoising approaches mostly rely on autoencoders. In some cases they fail to recover the ECG morphology, and their black-box nature does not allow for explainability or understanding of limitations. To address these issues, we introduce a novel system to record and denoise ear-ECG signals, leveraging open-source hardware and the Extended Kalman Filter. In-ear audio recording of heart sounds is used to accurately determine timings of cardiac cycles. From these timings, a short-term ensemble average ECG signal is calculated, which is used to fit the parameters of a dynamical ECG model to an individual user. The Kalman filter is then applied to the full time series ECG for denoising, using the dynamical model for its state prediction steps, and heart sounds as phase measurements. We have evaluated the system with data collected from 18 participants. The results report a mean SNR of 6.4 dB, mean absolute QT interval error of 54 ms, and heart rate error of 3 BPM, demonstrating the system's potential for continuous, non-invasive, user-friendly ECG monitoring.
Peripheral Artery Disease (PAD) is a common atherosclerotic condition that is underdiagnosed due to the lack of accessible screening options. Photoplethysmography (PPG) serves as a potentially valuable tool in accessible screening for PAD due to its ubiquitous nature and ability to be measured on a smartphone. However, the relationship between PPG and PAD is underexplored. In this paper, we seek to identify features of a PPG signal that correlate with PAD. In an analysis of 5,237 legs from $\mathrm{N}=2,362$ unique patients, we find significant correlations with multiple different features and the ankle-brachial index (ABI), which is used to diagnose PAD. Additionally, these features agree with physiological explanations of PAD and how the disease affects blood flow. These results set up the ability of future work to develop an accessible screening tool for PAD that uses physiologically relevant features of PPG morphology.
This paper introduces a novel wearable solution for continuous respiratory monitoring through electrocardiogramderived respiration (EDR) using custom-designed, dual-sided grid-patterned inkjet-printed (IJP) flexible dry electrodes and real-time smartphone-based analysis. The proposed electrode design reduces silver ink usage while maintaining signal quality and wearer comfort. We first compared ECG signal quality across gel, one-sided, and two-sided grid-patterned electrodes. Our mobile application, CardioHelp, processes ECG signals in real time to extract respiratory waveforms and continuously updates the respiration rate. EDR performance was validated against a commercial respiration belt across four activity conditions in five healthy adults. Bland-Altman and statistical analyses revealed minimal bias (mean difference $<0.5$ bpm), MAE $\leq 0.36$ bpm, and RMSE≤0.38 bpm. These results confirm robust and reliable performance. This integrated solution provides an affordable and practical approach to continuous cardiorespiratory monitoring in everyday life.
This paper presents the design and evaluation of a smart cane prototype designed to enhance independent navigation for individuals with visual impairments. The system integrates two VL53L0X Time-of-Flight (ToF) sensors and an ESP8266MOD D1 Mini microcontroller to detect obstacles at varying heights with high accuracy. Real-time data acquisition enables the generation of intuitive auditory and haptic alerts via a buzzer and vibration motor, offering multimodal feedback based on proximity. The device was tested in real-world settings and iteratively improved through user feedback collected from visually impaired individuals at a local support institution in Barranquilla, Colombia. Results demonstrate a detection precision of under 5 cm within a 1-meter range and a high user acceptance rate, with 90% of participants recommending the device. The system's low-cost architecture (USD 44) and Wi-Fi capability support future expansion, including mobile integration and cloud connectivity. This work contributes to Sustainable Development Goals 10 and 11 by promoting equitable access to mobility tools through affordable, contextually relevant assistive technology.
We present a compact, deployable heart rate (HR) estimation system using photoplethysmography (PPG) and inertial measurement unit (IMU) data, combining TimeWeaver, a conditional diffusion model for metadata-aware synthetic augmentation, with progressive structured pruning of Temporal Convolutional Networks (TCNs). Our smallest model, with 1.56k parameters, achieves a mean absolute error (MAE) of 4.92 BPM on the PPG-DaLiA dataset and supports real-time inference ($<{40 ms}$ latency) on a 64 MHz ARM Cortex-M4F microcontroller (MCU) without requiring quantization. Synthetic data conditioned on subject metadata, HR, and activity type significantly enhances model generalization, enabling pruned models to match or exceed the accuracy of larger baselines, achieving over a 23% improvement compared to training on real data alone. Our work establishes a new Pareto frontier for real-time, on-device HR monitoring using diffusion-augmented training and sub-2 k parameter models.
This paper introduces CLEAR-APG, a novel acoustic sensing approach that enables reliable heart rate monitoring in unconstrained environments using off-the-shelf active noise cancellation (ANC) headphones. By emitting ultrasonic signals into the user's ear canal via the headphone speaker and analyzing their echoes, which can detect the frequency of a pulsating vein along the canal wall. However, everyday activities such as exercising, speaking, or eating cause jaw movements that deform the ear canal, overwhelming the subtle deformation caused by blood flowing. To overcome this challenge, we employ the ANC headphone's built-in gyroscope to capture body motion and identify how various motion patterns influence the heartbeat waveform. Building on this insight, we propose a multi-modal method that effectively denoises the heartbeat waveform measurements and further accurately extracts heart rate. We implement CLEARAPG on ANC earbuds and conduct comprehensive field studies on 14 users. The results show that CLEAR-APG achieves an average heart rate error of 4.01% across seven different activities, satisfying industry-required margin of 10% heart rate error.
Biopotential measurement, including electrocardiograph (ECG), electroencephalograph (EEG), electromyograph (EMG), etc., is a generic tool for health monitoring, diagnosis, human-robot interface, etc. This paper presents an open-source platform for designing and fabricating dry-contact textile electrodes using programmable embroidery, which provides an accessible tool for ubiquitously embroidering textile electrodes on fabrics according to users' needs. Our method allows for precise control of geometric parameters, including stitch pattern and filling density, to optimize electrode performance. The results demonstrate that by varying the filling density of conductive threads, the electrode impedance can be systematically controlled. The experimental results showed that our embroidered electrodes achieved low skin-electrode impedance. In real-time ECG monitoring, these electrodes produced high-quality signals, which are comparable to gel electrodes. In addition, these electrodes can be quickly embroidered onto different fabric substrates such as clothes, sheets, pillows, etc., without the need for extra fabrication steps. These findings validate our platform as a feasible method for producing cost-effective, comfortable, reliable embroidered electrodes suitable for long-term wearable health monitoring. We open source the platform to make it accessible to the research community.
This paper presents a novel human-body communication technology that enables capacitive intra-body backscatter (C-IBB) communication between a batteryless ring sensor and a wrist-worn transceiver. C-IBB leverages the finite conductivity of human skin and air coupling capacitance to facilitate nearfield communication (NFC) between wearable devices. The C-IBB system features a radio frequency energy harvester connected to an impedance-matched wearable electrode, which charges a capacitor. This energy storage capacitor powers an ultra-lowpower microcontroller, enabling backscatter communication by modulating the electrode's load impedance. In this work, we developed a modular heterodyne transceiver system and intrabody channel gain emulator. These tools optimize transceiver and tag systems for realistic channel gains tailored to specific electrode configurations. We validated the system's performance on the human body, optimizing it for sensing applications in a wearable ring format. Our preliminary study reveals that the system supports a bit rate of 20.83 kbps with a bit error rate of 10−3 to 10−2 and operates effectively within a range of 23 cm.
Electrodermal activity (EDA), an electrical manifestation of the sympathetic innervation of the sweat glands, is widely used in long-term physiological monitoring, including sleep, stress, and cognitive studies. DC-source devices are more commonly used for recording EDA due to their simplicity, while AC alternatives are less adopted because of their perceived complexity. However, maintaining low noise and ensuring signal stability over extended durations remains a significant challenge. This study uses LTspice simulations to compare AC and DC constant current EDA circuits under identical conditions. The electrode–skin interface is modeled using a Randles cell, and real EDA recordings are time-compressed to modulate tissue resistance. Results show that both AC and DC circuits perform comparably well in short-term recordings; however, over time, DC signals degrade due to electrode polarization in the Randles cell model, while AC remains stable and continues to capture EDA reliably.
Radar-based fall detection systems offer significant potential to enhance the safety and quality of life for individuals with Alzheimer's Disease and Alzheimer's Disease-Related Dementias (AD/ADRD). These systems have demonstrated impressive accuracy when evaluated on standardized datasets; however, real-world deployment often reveals a marked drop in performance due to the challenges posed by environmental complexity, clutter, and variability in human behavior. This study explores two primary research questions: firstly, assessing the realism and transferability of performance metrics from standardized datasets to practical, cluttered environments; secondly, determining if and to what extent performance in realistic settings can be improved by augmenting datasets with synthetically generated falls and activities of daily life data using a U-Net diffusion model. Our findings highlight substantial performance gaps between standardized datasets and realistic conditions. Preliminary experiments demonstrate that introducing generated fall data can significantly enhance detection accuracy in practical settings, providing insights into the optimal amount of synthetic data needed to maximize detection effectiveness.
