Integration of physiological and accelerometer data to improve physical activity assessment

Area of Science:

• Exercise Physiology
• Biomedical Engineering
• Public Health

Background:

• Accurate physical activity (PA) measurement is crucial for understanding health outcomes.
• Combining heart rate (HR) and accelerometer (ACC) data offers potential for improved PA assessment.
• The impact of individual calibration (IC) versus group-level calibration (GC) on HR-ACC monitoring accuracy requires further investigation.

Purpose of the Study:

• To compare the accuracy of different modeling techniques combining HR and ACC for estimating PA energy expenditure (PAEE).
• To evaluate the influence of individual calibration (IC) versus group-level calibration (GC) of HR on PAEE monitoring accuracy.
• To assess the precision of combined HR-ACC methods against single-measure ACC estimates.

Main Methods:

• 10 adults completed 6 hours of free-living activity.
• PAEE was measured using indirect calorimetry (criterion).
• PAEE was modeled using two combined HR-ACC methods (arm-leg HR+M, branched model) with IC and GC, and from hip ACC alone.

Main Results:

• Combined HR-ACC models with IC showed high accuracy (R2 = 0.81, SEE = 0.55 METs for arm-leg HR+M; R2 = 0.75, SEE = 0.61 METs for branched model).
• GC resulted in larger errors for all methods, with the branched model performing better than arm-leg HR+M.
• Both combined methods were more precise than hip ACC estimates alone (R2 = 0.41, SEE = 0.96 METs).

Conclusions:

• Combining HR and ACC data significantly enhances the accuracy of PAEE estimation.
• Individual calibration of HR is superior to group-level calibration for improving PAEE monitoring.
• These combined methods show promise for application in large-scale epidemiological studies on physical activity and health.