Physical activity as a system: A network perspective on behavioral attributes and self-efficacy.

Area of Science:

• Behavioral science
• Systems science
• Biomedical engineering

Background:

• Measuring physical activity (PA) lacks a gold standard due to its dynamic, multifactorial nature.
• Existing methods often fail to capture the complex interplay of behavioral and contextual factors influencing PA.
• A systems-based approach is needed to model PA as an interacting system.

Purpose of the Study:

• To apply network analysis to model physical activity (PA) attributes and their relationships.
• To assess gender invariance in PA networks.
• To examine the links between PA attributes and self-efficacy.

Main Methods:

• Network analysis was used to model relationships among PA attributes in 40 healthy adults over one week.
• Data were collected using self-report and smartphone accelerometry.
• Three network models were estimated: PA attribute relationships, gender comparison, and integration with self-efficacy.

Main Results:

• The network structure revealed both positive and negative connections between PA attributes (53.42% positive).
• Weekday and weekend sitting time showed the strongest positive association.
• Moderate PA, sitting time, and self-efficacy with limited facility access were central nodes, with some gender differences observed.

Conclusions:

• Network analysis is a promising framework for modeling physical activity (PA) as an interacting system.
• This approach can uncover complex relationships between PA behaviors, contextual factors, and self-efficacy.
• Findings highlight the interconnectedness of different PA components and suggest potential targets for intervention.