46 - Maternal dietary nutrition interacts with demographic and socioeconomic information to impact offspring neurocognition by early school age

Poster Number: 46
Publication Number: 46.253

Rina Bao, Boston Children's Hospital, Cambridge, MA, United States; Yangming Ou, Boston Children's Hospital; Harvard Medical School, Boston, MA, United States

Background: Breastfeeding mother’s dietary intake of nutrition is crucial for offspring’s brain development. However, existing studies mostly investigate multiple maternal dietary nutrients in isolation rather than combinatorially. The exact impact till early school age remains less understood. Moreover, little is known on specific nutrition suggestions given individual socioeconomic variability.

Objective: Use multivariate machine learning approaches to understand how maternal dietary nutrients interact with sociodemographics and jointly impact offspring’s neurocognitive functions by early school age (4-6 years).

Design/Methods: 20 healthy term-born mother-infant dyads neonates were recruited from the well-baby nurseries at Brigham and Woman’s and Beth Israel Deaconess Medical Center during 2014-17 within the first week after delivery. Maternal dietary nutrients (73 nutrients in 6 categories: total energy, carbohydrates, fats, minerals, vitamins and antioxidants, and fibers) were quantified by food frequency questionnaires. Demographic information included mother’s age, body mass index (BMI) and infant’s birth weight, length and head circumference. Socioeconomic information included mother and father’s education levels and family income levels. Early-school age neurocognition was scored by Child Development Inventory (CDI), including 9 domain-specific scores (fine and gross motor, problem-solving, personal-social, numbers, language comprehension, etc.). We used multivariate machine learning with feature selection to identify the best combination of variables that can predict each of the 9 CDI neurocognitive scores.

Results: Univariate analysis showed that none but 1 maternal dietary nutrient were significantly associated with offspring neurocognition at early-school age (Fig 1). Therefore, multivariate analysis was needed, and it found (1) several nutrients (mainly total energy, fats, vitamins) combined with age and birth demographics, as well as parental education levels, could jointly impact offspring’s school-age neurocognition (Fig 2); and (2) CDI expressive language and letter scores were the most and least predictable neurocognitions at early school age.

Conclusion(s): Maternal dietary intake of nutrition during lactation joins sociodemographic factors to influence offspring’s early-school-age neurocognition. This is the first step toward personalized nutrition suggestion during pregnancy and lactation, given individual variability in sociodemographics. This finding needs to be further validated in larger multi-site data.