With the increasing popularity and feasibility of implementing Ecological Momentary Assessment (EMA), research on affective dynamics has expanded considerably (1–3). In parallel, advances in artificial intelligence (AI) have enabled scalable extraction of rich multimodal features (e.g., facial, vocal, and linguistic) from video data, which may serve as adjunct complementary behavioural indicators to subjective ratings of affective states typically captured in EMAs. Emerging research has demonstrated the promising utility of these features in predicting the diagnostic status of mental health problems and the presence of transdiagnostic symptomology (4–6). However, how these behavioural indicators relate to momentary affect in naturalistic settings, and whether they explain unique variance in mental health symptoms beyond self-reported mood, remains unexplored. Given the recent rise in perinatal depression and anxiety (7), exploring these questions in the perinatal period will inform the potential clinical value of implementing naturalistic video screeners of postpartum mental health symptoms in perinatal care provision. The present study therefore has two primary aims: 1) to examine the extent to which multimodal behavioural indicators map onto momentary affect ratings across a one-week EMA protocol (3 pings daily) in birthing parents who are 1-6 months postpartum; 2) to test whether behavioural indicators explain variance in mental health symptoms above and beyond self-reported momentary mood, thereby evaluating their incremental validity as markers of emotional functioning.