Wrist-Worn Accelerometer-Brand Independent Posture Classification.

Area of Science:

• Biomedical Engineering
• Physical Activity Measurement
• Wearable Technology

Background:

• Accurate posture classification is crucial for understanding sedentary behavior and its health implications.
• Wrist-worn accelerometers offer a convenient method for monitoring physical activity and posture.
• Standardizing data analysis across different accelerometer brands is essential for comparability.

Purpose of the Study:

• To evaluate the accuracy of the Sedentary Sphere algorithm for classifying posture.
• To assess the performance of the Sedentary Sphere using data from two popular wrist-worn triaxial accelerometers: GENEActiv and ActiGraph GT3X+.

Main Methods:

• Laboratory-based study involving 34 adults performing various lying, sitting, and upright activities.
• Free-living study where participants wore GENEActiv and ActiGraph accelerometers on their wrist, and an activPAL3 on their thigh for 2 days.
• Data analysis using the Sedentary Sphere with 15-second epochs.

Main Results:

• In the laboratory, the Sedentary Sphere achieved 74-75% accuracy for sedentary posture and 90-91% for upright posture with both accelerometers.
• Free-living estimates of total sedentary time showed minimal bias compared to the thigh-worn activPAL3.
• High intraclass correlations (ICC) were observed between the wrist-worn accelerometers and the thigh-worn device, indicating strong agreement.

Conclusions:

• The Sedentary Sphere is effective for classifying adult posture using wrist-worn accelerometer data.
• The algorithm demonstrates equal validity with data from both GENEActiv and ActiGraph accelerometers.
• This approach facilitates consistent and comparable posture and sedentary behavior assessments across studies using different devices.