We investigate the use of audio-inspired features to characterize ECG-derived skin sympathetic nerve activity (SKNA). Building on concepts from audio signal processing, we designed a set of SKNA-optimized spectral and energy-related features tailored to the transient, burst-like nature of sympathetic activation. High-frequency ECG signals were recorded from 20 participants during rest and during cognitive stress induced by the Stroop Color and Word Test. Applied to this two-phase protocol, the proposed features captured consistent differences between baseline and stress, achieving robust classification performance using support vector machines with leave-one-subject-out cross-validation. Results demonstrate that audio-inspired features, when physiologically constrained, provide a powerful and interpretable representation of SKNA dynamics, highlighting their promise for noninvasive assessment of autonomic function.

Most deep learning models of multiclass arrhythmia classification are tested on fingertip photoplethysmographic (PPG) data, which has higher signal-to-noise ratios compared to smartwatch-derived PPG, and the best reported sensitivity value for premature atrial/ventricular contraction (PAC/PVC) detection is only 75%. To improve upon PAC/PVC detection sensitivity while maintaining high AF detection, we use multi-modal data which incorporates 1D PPG, accelerometers, and heart rate data as the inputs to a computationally efficient 1D bi-directional Gated Recurrent Unit (1D-Bi-GRU) model to detect three arrhythmia classes. We used motion-artifact prone smartwatch PPG data from the NIH-funded Pulsewatch clinical trial. Our multimodal model tested on 72 subjects achieved an unprecedented 83% sensitivity for PAC/PVC detection while maintaining a high accuracy of 97.31% for AF detection. These results outperformed the best state-of-the-art model by 20.81% for PAC/PVC and 2.55% for AF detection even while our model was computationally more efficient (14 times lighter and 2.7 faster).

Accurate estimation of sleepiness levels is crucial for managing sleep-related health risks and preventing cognitive degradation that can lead to accidents in the workplace. However, using machine learning (ML) to estimate these levels from speech remains challenging, with reported weak correlations (<0.40) with ground-truth sleepiness levels. Developing effective ML models requires high-quality sleepiness data from noticeably sleepy individuals, but collecting such data through prolonged sleep deprivation is both risky and costly. We propose that acted (feigned) sleepy speech can effectively represent realistic sleepiness and be used to train ML models for estimating sleepiness levels and detecting sleepiness-associated cognitive performance degradation. Our study demonstrates that: (1) human listeners perceive acted sleepy speech as sleepier than both non-sleepy and genuinely sleepy speech, and (2) ML models trained on acted speech can accurately estimate sleepiness levels in individuals who have been awake for 25 hours, achieving a correlation of 0.54 with ground-truth sleepiness levels while using fewer samples. Furthermore, the acted sleepy speech-based ML model detects cognitive performance degradation (F1 score = 0.80) in sleep-deprived individuals, outperforming models trained on real sleepy speech (F1 score = 0.32). Our approach provides efficient, effective, and scalable solutions to not only update benchmarks but also enhance the capabilities of real-world AI applications, from voice assistants to AI agents, ultimately supporting human health, workplace safety, and daily tasks.

This study addresses the challenges of arrhythmia detection, including atrial fibrillation (AF), using continuously collected smartwatch photoplethysmography (PPG) data. We propose a novel application of the densely connected convolutional neural network (DenseNet) for AF detection using smartwatch PPG data with effective differentiation from premature atrial contractions (PACs) and premature ventricular contractions (PVCs). While smartwatches offer a convenient platform for continuous monitoring, current research often relies on controlled settings using fingertip PPG and struggles to differentiate AF from other arrhythmias, particularly PAC/PVC$s$. Unlike prior controlled-setting studies, our data collection includes 14 days of continuous smartwatch PPG data in real-world settings, which truly challenges the performance of arrhythmia classification algorithms. Notably, the proposed DenseNet model excelled at distinguishing AF (sensitivity: 0.950, specificity: 0.949), demonstrating promise for real-world continuous AF monitoring. The proposed DenseNet model also achieved high specificity (0.958) for the classification of PAC/PVCs despite the inherent difficulty in their identification from PPG data. These findings demonstrate the efficacy of DenseNet in accurately identifying AF and distinguishing it from other arrhythmias, highlighting its potential for widespread clinical application.

Sleepiness results in an increased susceptibility to workplace accidents due to reduced cognitive functioning. Early detection of signs of sleepiness is crucial for mitigating work-related risks. Establishing a comprehensive database of sleepiness data is fundamental for developing effective sleepiness identification models. However, obtaining data on realistic sleepiness proves challenging in practice. This study explores a more efficient approach to sleepiness simulation to address this challenge. In this paper, we hypothesize that actively inducing feelings of sleepiness can precipitate sleepiness. In our experiment, ten participants simulated sleepiness while listening to a speech delivered by a tired, sleep-deprived individual. We measured self-reported levels of sleepiness and two widely-used physiological biomarkers-electrodermal activity (EDA) and electrocardiogram (ECG) signals-during this induced state. Our results shed light on the potential of this sleepiness simulation approach, unveiling physiological markers linked to induced sleepiness. By introducing this innovative sleepiness simulation method, our research advances building large sleepy datasets, which contribute to the proactive identification of sleepiness-related risks in the workplace.

Sleep deprivation is known to impair cognitive performance and contribute to workplace accidents. Bio-signal-based sleepiness detection approaches, facilitated by mobile healthcare systems, have emerged as a promising tool for preventing such accidents. However, due to the wide-ranging effects of sleep deprivation on human physiology and the sympathetic nervous system, accurately determining sleepiness remains challenging. In this study, we hypothesize that speaking loudly could increase the sensitivity of electrothermal activity (EDA) to sleep deprivation, as EDA measures changes in sympathetic nervous system activity. To validate our hypothesis, twenty participants underwent 25 hours of prolonged wakefulness, during which they vocalized loudly and completed a cognitive performance test, 10-min Psychomotor Vigilance Test (PVT) to measure their prolonged reaction time. Two EDA measures retrieved in speaking versus PVT tasks were compared. As result, EDA obtained during the speaking activity exhibited a stronger group correlation of ~-0.96 and individual correlation of ~-0.53 with sleep deprived indicators (i.e., reaction time) compared to the EDA acquired during the PVT task. We found a feasibility of performance degradation can be predictive by capturing the activity of the sympathetic nervous system precisely during the speaking activity. Moreover, the study emphasizes the importance of speaking activity in investigating sleepiness detection approaches based on speech, which have gained significant attention in recent research.

Data visualization is critical to unraveling hidden information from complex and high-dimensional data. Interpretable visualization methods are critical, especially in the biology and medical fields, however, there are limited effective visualization methods for large genetic data. Current visualization methods are limited to lower-dimensional data and their performance suffers if there is missing data. In this study, we propose a literature-based visualization method to reduce high-dimensional data without compromising the dynamics of the single nucleotide polymorphisms (SNP) and textual interpretability. Our method is innovative because it is shown to (1) preserves both global and local structures of SNP while reducing the dimension of the data using literature text representations, and (2) enables interpretable visualizations using textual information. For performance evaluations, we examined the proposed approach to classify various classification categories including race, myocardial infarction event age groups, and sex using several machine learning models on the literature-derived SNP data. We used visualization approaches to examine clustering of data as well as quantitative performance metrics for the classification of the risk factors examined above. Our method outperformed all popular dimensionality reduction and visualization methods for both classification and visualization, and it is robust against missing and higher-dimensional data. Moreover, we found it feasible to incorporate both genetic and other risk information obtained from literature with our method.

Hateful and offensive speech on online social media platforms has been exacerbated by the turbulent and chaotic circumstances brought on by the coronavirus pandemic. A particularly contentious issue was lockdown orders issued by state governments designed to keep citizens safe by controlling the spread of the virus. To compel the government to relax these orders and restore normalcy, antilockdown protests were organized in many states. The economic, ideological, political, and health concerns related to the lockdowns and the associated protests were debated vigorously on social media platforms, many times using offensive content. Detecting such insulting and humiliating content is especially important during tumultuous times, when tensions are high, because such expressions online can quickly precipitate violence in the physical world. This paper presents an approach to detect hateful and offensive content from Twitter feeds collected after anti-lockdown protests in Lansing, Michigan. These tweets were labeled using a comprehensive definition of what constitutes offensive content based on its potential to trigger and incite people. Linguistic and auxiliary features were extracted from these labeled tweets. These features were further processed through feature selection and dimensionality reduction techniques. The preprocessed feature set was used to train machine learning models, which detect offensive content with an accuracy of around 84%. Our approach demonstrates the feasibility of identifying and tagging offensive content in politically motivated situations, even when such speech is dominated by contextual and circumstantial information. It can thus be used to mitigate the damage caused by widespread dissemination of offensive content.