A5076145211- Statistical Methods and Inference
- Bayesian Methods and Mixture Models
- Statistical Methods and Bayesian Inference
- Physical Activity and Health
- Sleep and related disorders
- Gene expression and cancer classification
- Sleep and Wakefulness Research
- Mental Health Research Topics
- Advanced Statistical Methods and Models
- Machine Learning in Healthcare
- Health disparities and outcomes
- Heart Rate Variability and Autonomic Control
- Balance, Gait, and Falls Prevention
- Statistical Distribution Estimation and Applications
- Dementia and Cognitive Impairment Research
- Obstructive Sleep Apnea Research
- Obesity, Physical Activity, Diet
- Diabetes, Cardiovascular Risks, and Lipoproteins
- Nutritional Studies and Diet
- Health, Environment, Cognitive Aging
- Cognitive Science and Mapping
- Ergonomics and Musculoskeletal Disorders
- Air Quality and Health Impacts
- Innovation Diffusion and Forecasting
- Statistical Methods in Clinical Trials
University of South Carolina
2022-2025
Harvard University Press
2023
Johns Hopkins University
2021-2022
North Carolina State University
2018-2020
The purpose of this study was to evaluate the reliability and validity raw accelerometry output from research-grade consumer wearable devices compared accelerations produced by a mechanical shaker table. Raw data total 40 (i.e., n = 10 ActiGraph wGT3X-BT, Apple Watch Series 7, Garmin Vivoactive 4S, Fitbit Sense) were reference an orbital table at speeds ranging 0.6 Hz (4.4 milligravity-mg) 3.2 (124.7mg). Two-way random effects absolute intraclass correlation coefficients (ICC) tested...
Abstract Introduction Few studies have explored whether gait measured continuously within a community setting can identify individuals with Alzheimer's disease (AD). This study tests the feasibility of this method to at earliest stage AD. Methods Mild AD (n = 38) and cognitively normal control (CNC; n 48) participants from University Kansas Disease Center Registry wore GT3x+ accelerometer for 7 days assess gait. Penalized logistic regression repeated five‐fold cross‐validation followed by...
Summary With the advent of continuous health monitoring with wearable devices, users now generate their unique streams data such as minute-level step counts or heartbeats. Summarizing these via scalar summaries often ignores distributional nature and almost unavoidably leads to loss critical information. We propose capture user-specific quantile functions (QF) use QFs predictors in scalar-on-quantile-function-regression (SOQFR). As an alternative approach, we also represent L-moments, robust...
The challenge of handling missing data is widespread in modern analysis, particularly during the preprocessing phase and various inferential modeling tasks. Although numerous algorithms exist for imputing data, assessment imputation quality at patient level often lacks personalized statistical approaches. Moreover, there a scarcity methods metric space based objects. aim this paper to introduce novel two-step framework that comprises: (i) objects taking values metrics spaces, (ii) criterion...
Abstract Wearable data is a rich source of information that can provide deeper understanding links between human behaviors and health. Existing modelling approaches use wearable summarized at subject level via scalar summaries in regression, temporal (time-of-day) curves functional analysis (FDA), distributions distributional (DDA). We propose to capture temporally local using subject-specific time-by-distribution (TD) objects. Specifically, we develop on regression (SOTDR) model...
Summary We propose a novel method for variable selection in functional linear concurrent regression. Our research is motivated by fisheries footprint study where the goal to identify important time-varying sociostructural drivers influencing patterns of seafood consumption, and hence footprint, over time, as well estimating their dynamic effects. develop variable-selection regression extending classically used scalar-on-scalar methods like lasso, smoothly clipped absolute deviation (SCAD)...
Abstract Study Objectives Evaluate wrist-placed accelerometry predicted heartrate compared to electrocardiogram (ECG) in children during sleep. Methods Children (n = 82, 61% male, 43.9% black) wore a Apple Watch Series 7 (AWS7) and ActiGraph GT9X polysomnogram. Three-Axis data was extracted from AWS7 the GT9X. Accelerometry estimates were derived jerk (the rate of acceleration change), computed using peak magnitude frequency short time Fourier Transforms Hilbert transformed magnitude....
Background and Objectives Hearing aids may reduce the risk of dementia among individuals with hearing loss. However, no evidence is available from randomized controlled trials (RCTs) on effectiveness use in reducing incident dementia. Using target trial emulation, we leveraged an existing longitudinal cohort study to estimate association between initiation Research Design Methods The Health Retirement Study was used emulate non-institutionalized participants aged ≥50 years self-reported...
ABSTRACT Introduction This study examined the potential of a device agnostic approach for predicting physical activity from consumer wearable accelerometry compared with research-grade accelerometry. Methods Seventy-five 5- to 12-year-olds (58% male, 63% White) participated in 60-min protocol. Children wore wrist-placed wearables (Apple Watch Series 7 and Garmin Vivoactive 4) (ActiGraph GT9X) concurrently an indirect calorimeter (COSMED K5). Activity intensities (i.e., inactive, light,...
We develop a functional proportional hazards mixture cure model with scalar and covariates measured at the baseline. The model, useful in studying populations fraction of particular event interest is extended to data. employ expectation–maximization algorithm semiparametric penalized spline-based approach estimate dynamic coefficients incidence latency part. proposed method computationally efficient simultaneously incorporates smoothness estimated via roughness penalty. Simulation studies...
We develop a new method for variable selection in nonlinear additive function-on-scalar regression (FOSR) model. Existing methods FOSR have focused on the linear effects of scalar predictors, which can be restrictive assumption presence multiple continuously measured covariates. propose computationally efficient approach existing using functional principal component scores response and extend this framework to The proposed provides unified flexible FOSR, allowing Numerical analysis...
Abstract The purpose of this study was to evaluate the reliability and validity raw accelerometry output from research-grade consumer wearable devices compared accelerations produced by a mechanical shaker table. Raw data total 40 (i.e., n=10 ActiGraph wGT3X-BT, Apple Watch Series 7, Garmin Vivoactive 4S, Fitbit Sense) were criterion an orbital table at speeds ranging 0.6 Hz (4.4 milligravity-mg) 3.2 (124.7mg). For testing, identical oscillated for 5 trials that lasted 2 minutes each. 1...
We develop a new method for variable selection in nonparametric functional concurrent regression. The commonly used linear model (FLCM) is far too restrictive assuming linearity of the covariate effects, which not necessarily true many real‐world applications. (NPFCM), on other hand, much more flexible and can capture complex dynamic relationships present between response covariates. extend classically methods, e.g., group LASSO, SCAD MCP, to perform NPFCM. show via numerical simulations...
Complex survey designs are commonly employed in many medical cohorts. In such scenarios, developing case-specific predictive risk score models that reflect the unique characteristics of study design is essential. This approach key to minimizing potential selective biases results. The objectives this paper are: (i) To propose a general framework for regression and classification using neural network (NN) modeling, which incorporates weights into estimation process; (ii) introduce an...
Mobile devices (e.g., tablets and smartphones) have been rapidly integrated into the lives of children impacted how engage with digital media. The portability these allows for sporadic, on-demand interaction, reducing accuracy self-report estimates mobile device use. Passive sensing applications objectively monitor time spent on a given but are unable to identify who is using device, significant limitation in child screen research. Behavioral biometric authentication, embedded sensors...
The advent of wearable and sensor technologies now leads to functional predictors which are intrinsically infinite dimensional. While the existing approaches for data survival outcomes lean on well-established Cox model, proportional hazard (PH) assumption might not always be suitable in real-world applications. Motivated by physiological signals encountered digital medicine, we develop a more general flexible time-transformation model estimating conditional function with both scalar...
Studies have examined the association between dual sensory impairment and late-life cognitive outcomes in U.S with inconsistent findings.
ABSTRACT Mobile health studies often collect multiple within‐day self‐reported assessments of participants' behavior and well‐being on different scales such as physical activity (continuous scale), pain levels (truncated mood states (ordinal the occurrence daily life events (binary scale). These assessments, when indexed by time day, can be treated analyzed functional data corresponding to their respective types: continuous, truncated, ordinal, binary. Motivated these examples, we develop a...