Validation of the Activity Preference Assessment: a tool for quantifying children's implicit preferences for sedentary and physical activities.

BACKGROUND: High levels of sedentary behavior and low physical activity are associated with poor health, and the cognitive determinants of these behaviors in children and adolescents are not well understood. To address this gap, we developed a novel, non-verbal, computer-based assessment to quantify the degree to which youth prefer to be sedentary relative to physically active in their leisure time. METHODS: The Activity Preference Assessment (APA) uses a forced-choice paradigm to understand implicit decision-making processes when presented with common sedentary and physical activities. The APA bias score ranges from - 100 to + 100, with positive scores indicating a relative preference for sedentary activities, and negative scores representing a preference for physical activities. In 60 children ages 8-17 years, we assessed the validity of this behavioral task against a free-choice play observation, accelerometry-measured activity, anthropometrics and body composition, and cardiorespiratory fitness. We explored neighborhood, family, and individual-level factors that may influence implicit activity preferences. Test-retest reliability was assessed over one week. RESULTS: The majority of children (67%) preferred sedentary relative to physical activities. APA bias scores were positively associated with sedentary time during free-choice play. In girls, bias scores were negatively associated with average daily MVPA. APA bias scores were positively associated with body fat and negatively associated with cardiorespiratory fitness. These findings were independent of age, sex, and race/ethnicity. Neighborhood access to physical activity spaces, the number of people in the home, perceived physical self-competence (e.g., coordination, strength), and self-reported depressive symptoms were associated with activity preferences. The intra-class correlation for test-retest reliability was r = 0.59. CONCLUSIONS: The APA shows promise as a novel tool for quantifying children's relative preference for sedentary versus physical activities. Implicit bias scores from the APA are clinically meaningful, as shown by significant associations with adiposity and cardiorespiratory fitness. Future longitudinal studies should examine the directionality of the association between preferences and health markers, and the degree to which implicit activity preferences are modifiable. Importantly, the task only takes an average of 10 min to complete, highlighting a potential role as an efficient screening tool for the propensity to be sedentary versus physically active. TRIAL REGISTRATION: ClinicalTrials.gov NCT03624582 .

Perceptual Characterization of the Macronutrient Picture System (MaPS) for Food Image fMRI.

Food image fMRI paradigms are used widely for investigating the neural basis of ingestive behavior. However, these paradigms have not been validated in terms of ingestive behavior constructs, engagement of food-relevant neural systems, or test-retest reliability, making the generalizability of study findings unclear. Therefore, we validated the Macronutrient Picture System (MaPS) (McClernon et al., 2013), which includes food images from the six categories represented in the Geiselman Food Preference Questionnaire (FPQ) (Geiselman et al., 1998). Twenty-five healthy young adults (n = 21 female, mean age = 20.6 ± 1.1 years, mean BMI = 22.1 ± 1.9 kg/m2) rated the MaPS images in terms of visual interest, appetitive quality, nutrition, emotional valence, liking, and frequency of consumption, and completed the FPQ. In a second study, 12 individuals (n=8 female, mean age = 25.0 ± 6.5 years, mean BMI = 28.2 ± 8.7 kg/m2) viewed MaPS and control images (vegetables and non-food) during two separate 3T BOLD fMRI scans after fasting overnight. Intuitively, high fat/high sugar (HF/HS) and high fat/high complex carbohydrate (HF/HCCHO) images achieved higher liking and appetitive ratings, and lower nutrition ratings, than low fat/low complex carbohydrate/high protein (LF/LCHO/HP) images on average. Within each food category, FPQ scores correlated strongly with MaPS image liking ratings (p < 0.001). Brain activation differences between viewing images of HF/HS and vegetables, and between HF/HCCHO and vegetables, were seen in several reward-related brain regions (e.g., putamen, insula, and medial frontal gyrus). Intra-individual, inter-scan agreement in a summary measure of brain activation differences in seven reward network regions of interest was high (ICC = 0.61), and was even higher when two distinct sets of food images with matching visual ratings were shown in the two scans (ICC = 0.74). These results suggest that the MaPS provides valid representation of food categories and reliably activates food-reward-relevant neural systems.

Quantitative Comparison of 2 Dual-Energy X-ray Absorptiometry Systems in Assessing Body Composition and Bone Mineral Measurements.

Dual-energy X-ray absorptiometry (DXA) is widely used in body composition measurement and evaluation. Because of its numerous applications, the probability of instrument discrepancies has increased dramatically. This study quantitatively compares 2 different DXA systems. In this study, 96 subjects (60 female and 36 male, aged 19-82 years) were recruited and scanned using a General Electric Lunar iDXA and a Hologic Discovery scanner. Four measurements (percent fat, total mass, bone mineral density [BMD], and bone mineral content [BMC]) were quantitatively compared in the whole body and in specific anatomic regions (arms, legs, trunk, android, gynoid, head, ribs, and pelvis). A simple linear regression of each measurement was performed to examine the correlation between the 2 systems. Percent fat, total mass, BMC, and BMD were highly correlated between the 2 DXA systems, with correlation r values greater than 0.854 for both the whole body and the individual anatomic regions except for BMC and BMD in ribs. The high correlation between the 2 DXA systems with systematic differences enabled development of calibration equations for extending the multisystem measurements to advanced quantitative analyses.