Temporal Lifestyle Patterns Linked with Health Indicators
Principal Investigator
Name
Heather Eicher-Miller
Degrees
PhD
Institution
Purdue University
Position Title
Associate Professor
Email
About this CDAS Project
Study
IDATA
(Learn more about this study)
Project ID
IDATA-44
Initial CDAS Request Approval
May 27, 2021
Title
Temporal Lifestyle Patterns Linked with Health Indicators
Summary
Diet and physical activity (PA) are independent risk factors for obesity and hypertension and with a link to heart rate. Diet and PA are also inter-related, with potential synergistic effects on these health status indicators. Both dietary and PA behaviors occur in a temporal context or “temporal pattern” including daily rhythms that begin and end throughout a day. The temporal sequence of these lifestyle exposures can be used to create temporal patterns with a relationship to health status indicators. This project is proposed to develop machine learning and signal processing techniques to generate temporal lifestyle patterns including dietary, physical activity, and sedentary behavior exposures based on data from 24 hour dietary recalls, PA questionnaires, and PA monitors. To test the change of temporal lifestyle patterns over time, data from multiple dietary recalls and multiple days of PA monitors for all available assessment time points are needed. Multiple linear regression models may be used to compare clusters of temporal lifestyle patterns relationship to health status indicators. Models will be adjusted for gender, race, age, total PA counts per day, total energy estimate and multiple comparisons. The temporal lifestyle patterns are expected to be significantly associated with health status indicators including weight, BMI, blood pressure, and heart rate. This project could provide insight for early detection of lifestyle behavioral patterns prone to obesity and chronic disease.
Aims
Specific Aim #1: Develop machine learning and signal processing techniques to create temporal lifestyle patterns that allow the sample to be divided to clusters representing similar temporal lifestyle patterns.
Specific Aim #2: Apply inferential statistics to evaluate the temporal lifestyle patterns and compare to health indicators to determine the cluster relationships with health outcomes.
Collaborators
Saul Gelfand, Purdue University
Edward Delp, Purdue University
Elizabeth Richards, Purdue University
Anindya Bhadra, Purdue University
Erin Hennessy, Tufts University
Luotao Lin, Purdue University