Temporal fluctuations instead of stable traits: State-level approach bias predicts goal-incongruent intake under low top-down control

Food approach bias is a behavioral tendency to approach food stimuli faster than to avoid them and is thought to predispose individuals to food intake. However, relationships between bias measures and intake have proven difficult to demonstrate empirically, possibly because approach bias has typically been treated as a stable characteristic of individuals, ignoring possible bias fluctuations over time within individuals. Here, we investigated (1) the temporal variability of food approach bias, its association with (2) food craving and (3) intake, and (4) the role of top-down control in this bias-intake association. 76 participants (67 women, 9 men, average age: 29.57 years, mean BMI: 24.41) completed an impulsivity questionnaire (UPPS-scale) and were asked to conduct a smartphone-based AAT on nine days (average completed sessions: 7.5). Additionally, they reported daily craving, intake, and consumption intentions for 12 personalized foods they wanted to eat less or more often. Analyses using multilevel models showed that approach bias (1) varies considerably within individuals, (2) correlates within and between individuals with craving and (3) with intake, but only (4) for individuals with generally weak intentions or high impulsivity. Results indicate the need to re-conceptualize approach bias as having both fluctuation and more enduring components, and that top-down process should be considered in its relationship with intake. This implies that single-session approach bias measures may not predict distal outcomes such as body weight. Interventions targeting approach bias may thus aim to intervene in close temporal proximity to high-risk situations.