A semi-quantitative method to denote generic physical activity phenotypes from long-term accelerometer data--the...

Area of Science:

• Biomedical Engineering
• Public Health
• Wearable Technology

Background:

• Objective physical activity measurement using accelerometers is crucial for large cohort studies.
• Existing methods often lack comprehensive analysis of activity patterns beyond duration and intensity.
• There is a need for methods to characterize activity regularity and its relation to health outcomes.

Purpose of the Study:

• To introduce the ATLAS index (Activity Types from Long-term Accelerometric Sensor data) for characterizing physical activity phenotypes.
• To develop a 3D graphical representation for visualizing activity patterns.
• To validate the ATLAS index using long-term accelerometer data from diverse cohorts.

Main Methods:

• The ATLAS index quantifies physical activity based on three dimensions: regularity (reg), duration (dur), and intensity (int).
• Activity events are identified from long-term accelerometer data.
• Data from 12 exemplary datasets across three cohort studies were used for validation.

Main Results:

• Five archetypical activity types were defined: insufficiently active, busy bee, cardio-active, endurance athlete, and weekend warrior.
• These types exhibit distinct combinations of regularity, intensity, and duration.
• A common 3D graph visually distinguished these activity patterns.

Conclusions:

• The ATLAS index uniquely incorporates activity regularity alongside duration and intensity.
• The 3D representation facilitates visual and computational comparison of activity types in populations.
• Further validation is needed to establish normative values and group centroids for the ATLAS index.