The ActiGraph counts processing and the assessment of vigorous activity

Area of Science:

• Physical Activity Measurement
• Biomedical Engineering
• Exercise Physiology

Background:

• ActiGraph devices are widely used to measure physical activity.
• Existing band-pass filters may introduce bias during high-speed running.
• Accurate estimation of vigorous physical activity (VPA) is crucial for health research.

Purpose of the Study:

• To evaluate the impact of novel band-pass filters on ActiGraph measurement bias during high-speed running.
• To assess the effect of these filters on estimating free-living VPA.
• To compare extended frequency ranges (0.29-4 Hz and 0.29-10 Hz) against the original (0.29-1.66 Hz).

Main Methods:

• Sixty-two participants underwent a locomotion protocol with varying speeds.
• VPA was assessed in 1121 children using three different band-pass filters.
• ActiGraph GT3X+ devices recorded raw acceleration at 30 Hz.
• Intensity cut-points were determined using linear regression and ROC analysis calibrated with indirect calorimetry.

Main Results:

• The association between ActiGraph counts and running speed was negative with the original filter (AG) but positive with the extended filters (AC4, AC10).
• Extended filters reduced measurement bias for running speeds exceeding 10 km/h.
• Estimates of time spent in free-living VPA were similar across all tested band-pass filters.

Conclusions:

• Incorporating higher frequency data in ActiGraph counts, particularly with hip/waist worn devices, mitigates measurement bias during intense running.
• Further research and device development are needed for precise capture of all VPA, including intermittent activities.