Brown adipose tissue (BAT) is a highly thermogenic tissue that metabolises lipid and glucose to produce heat via uncoupling protein (UCP)1. Rodents lacking UCP1 become obese, and adult humans who are obese have less BAT(1). We have previously reported the use of infrared thermography (IRT) for the assessment of supraclavicular BAT(2), demonstrating a negative relationship between body mass index (BMI) and supraclavicular skin temperature (TSCR) in children(3). BAT activity is modulated by dietary intake in rodents, but the impact of food on BAT in children is unknown. This study aimed to explore the relationship between food exposure/preference and BAT thermogenesis in children. 36 children (8.5-11 years) were recruited after Ethics Committee approval, informed parental consent and child assent. TSCR calculated from thermograms was used as a measure of BAT thermal activity. Children’s food exposure (FFQ) and preference (FPQ) were assessed using validated questionnaires(4). FFQ consisted of an 8-point scale [‘everyday’ (1) to ‘never tried’ (8)] and FPQ consisted of a 5 point scale [l’oves'(1) to ‘hates'(5)] for 162 foods. The children’s questionnaire featured a subset of 39 foods to demonstrate agreement between the parent-child responses. Foods were classified as carbohydrate(C), dairy(D), fruit(F), protein(P), savoury snacks(SS), sweet foods(Sw) and vegetable(V). Correlation between FPQ/FFQ scores and TSCR were calculated to assess the relationships between food and TSCR, and partial correlation coefficients adjusted for known confounders of TSCR (BMI centile and ambient temperature4). Agreement was determined using Cohen’s weighted kappa, and Cronbach’s α assessed reliability of the food groupings. Groups with a reliability coefficient of <0.8 were excluded from further analysis. Values are mean±S.D. The study group consisted of 23 females, and 13 males with a weight of 33.5±6.7Kg and BMI centile of 55.6±30.8. The room temperature during IRT was 22.1±0.7°C. FFQ: The C, F, SS, Sw and V food groupings were reliable and parent-child agreement was predominantly poor. FPQ: The C, Sw and V food groupings were reliable and parent-child agreement was fair-good. A significant correlation was observed between FPQ scores for V and right-sided TSCR (Table 1). After adjustment for confounders, this relationship no longer persisted and a significant correlation between both C and Sw FPQ scores was observed (Table 1). No significant relationships were observed between FFQ scores and TSCR. For the first time we have shown that as preference for sweet foods/carbohydrate increases, BAT thermal activity decreases independent of BMI centile. The acute and chronic effects of dietary composition should now be prospectively studied to identify foods capable of BAT stimulation and obesity prevention.