Infants’ Everyday Exposure to Caregiver Facial Expressions: Evidence from Naturalistic Head-Mounted Camera Data

Infants’ early emotional experiences are shaped by the facial expressions they observe in daily life. However, most research has relied on isolated and posed expressions presented in laboratory settings, which may not reflect the expressive environment infants are exposed to. To capture the real-world social environments of infants, we analysed audiovisual data from the longitudinal SAYCam dataset, which recorded infants’ (N = 3) naturalistic experiences via head-mounted cameras. Infants wore head-mounted cameras twice a week for at least one hour, from 6 months and 32 months of age. Facial expressions displayed toward infants were identified using automatic facial expression recognition software (iMotions AFFDEX), applied to 280 hours of infant-perspective video. Using mixed effects models, we examined how the intensity, complexity, and frequency of expressions varied with infant age, expresser gender, and in contexts of infant distress. Across approximately 13 hours of detected emotional expressions, we found that neutral faces were most frequently observed by infants, while joyful expressions were expressed with the greatest intensity. Expression intensity for some categories (particularly joy, surprise, and disgust) decreased with age, and overall expression complexity also declined. We also found evidence for three hypothesized infant-directed expressions: love/concern, mock surprise, and exaggerated joy. Mothers were more likely to display love/concern toward their infant, and fathers more likely to display mock surprise and exaggerated joy. The use of these expressions declined with age, reflecting changes in parental expressivity toward infants as they age. Finally, when examining scenarios of infant distress, identified by analyzing audio pitch linked to crying, we found that infant crying was associated with changes in adult facial expressivity, including less positive valence, and shifts in expression type. By combining automated facial expression recognition with naturalistic head-mounted camera video data, these findings provide a detailed account of the expressive inputs that infants encounter in daily life, and how this changes with age and caregiver. More broadly, they show that studying infants in naturalistic contexts can deepen our understanding of the social environments that shape emotional development.

Recommended citation: Palmer, K., Van Der Zant, T., Holdsworth, I., Rayner, C., Tuza, A., Nelson, N., "Infants’ Everyday Exposure to Caregiver Facial Expressions: Evidence from Naturalistic Head-Mounted Camera Data." Available at SSRN: https://ssrn.com/abstract=6175290 (2026)
Download Paper