Using ecological momentary assessment and wearable sensors to predict restrictive eating in anorexia nervosa

Anorexia nervosa (AN) is characterised by persistent dietary restriction (APA, 2013). Although dietary restriction is a central maintenance factor of the disorder, it fluctuates within individuals across time. Understanding when restrictive behaviours occur in daily life is crucial for developing timely interventions. Evidence suggests that negative emotions are linked to dietary restriction in daily life (Fitzsimmons-Craft et al., 2015; Haynos et al., 2017). However, most work has focused on explaining associations rather than predicting when dietary restriction is likely to occur. Advancements in ambulatory assessment allow the simultaneous measurement of self-report emotional states using ecological momentary assessment (EMA) and physiological processes using wearable sensors, here referred to as ecological physiological assessment (EPA). These approaches provide high-frequency data that may improve the prediction of eating disorder behaviours in real-world settings. Recent studies have begun to explore the prediction of dietary restriction using intensive longitudinal data and have shown promising results for models based on EMA (Levinson et al., 2023) as well as wearable sensor data (Ralph-Nearman et al., 2023). However, it remains unclear whether physiological signals add predictive value beyond self-reported emotional states, and direct comparisons between EMA-only, EPA-only, and combined models are currently lacking.