A large body of research supports links between nutrition and neurodevelopment [1
]. Adequate concentrations of vitamins, minerals, essential elements, protein, and fatty acids have been shown to have beneficial effects on the developing brain, while deficiencies of these nutrients can be harmful [7
]. Although humans generally ingest nutrients in combination, the majority of studies on nutrition and neurodevelopment have either examined single nutrients or nutrient ratios [1
]. This approach assumes that each nutrient acts independently, but there may be interactions that explain much more of the variance. Finally, the current paradigm does not reflect the real-life scenario of food consumption, in which nutrients are always consumed as mixtures. Therefore, we propose that research on the joint effect of nutrient mixtures on health is needed. In this study, we focus on neurodevelopmental health and early-life nutrition.
A large body of literature has linked multiple food components to health. For example, various vitamins and minerals have independently been demonstrated to play important roles in brain development [1
]. Deficiencies in vitamins A and D have been implicated in learning and memory development, as well as language and psychomotor development, respectively [4
]. Deficiencies in B vitamins, particularly B9 (folate), B1 (thiamine), and B12 (cobalamin), have been associated with neural tube defects, delayed language development, and impaired motor functioning, respectively [2
]. Additionally, adequate levels of nutrient minerals such as iron, manganese, selenium, and zinc have been shown to be neuroprotective. Nonetheless, while we require sufficient levels of many nutrients, nutrients can be toxic when consumed in excess. Thus, both deficiency and excess consumption can adversely affect neurodevelopment [4
Protein and polyunsaturated fatty acids are essential macronutrients for brain growth and neuron differentiation. Protein is fundamental for neurotransmitter production as well as synapse and dendrite formation [7
], and polyunsaturated fats contribute to neural cellular membrane volume and fluidity, protein enzyme activation, and neuron signal transduction [11
]. Expectedly, adequate consumption of both macronutrients during early development has been shown to contribute to more favorable neurodevelopmental outcomes among preterm or low-birth-weight infants [12
Some studies have examined nutrition profiles among children with neurodevelopmental disorders. They show that poor nutrition intake (according to the recommended dietary allowance and estimated average requirement guidelines) and multiple, rather than single, nutrient deficiencies are more common among these children and, in some cases, their mothers [14
]. While not in and of itself proof of interactions among nutrient mixtures, this finding may be consistent with a nutrient mixture’s effect on neurodevelopment. In addition, while these studies did not examine prenatal nutrition profiles, the findings could potentially point to a role of general poor nutrition in the etiology and/or course of neurodevelopmental disorders. Prospective studies are needed to address this issue, as developmental disorders themselves may alter diet, through altered child behavior, parental nonmedical interventions, pica behaviors, or changes in appetite.
Therefore, we examined the impact of prenatal and childhood nutrient mixtures on child neurodevelopment. We hypothesized that prenatal and childhood diets rich in beneficial nutrients would positively predict child neurodevelopmental outcomes, while diets lower in these nutrients and/or higher in sugar, sodium, and/or saturated fat would negatively predict child neurodevelopmental outcomes.
All diets are a mixture of nutrients in various relative doses. This complex covariance structure has long been an unmeasured source of variance in the study of health effects and individual nutrients. To our knowledge, this is the first study to examine the relationship between maternal and child nutrient mixtures and a well-known health effect sensitive to nutrition: child neurodevelopment. We estimated the overall mixture effect of nutrients and the relative contribution of each nutrient to the children’s McCarthy Scales of Children’s Abilities (MSCA) scores, using data from empirically validated food frequency questionnaires. Results for both good and poor nutrition indices support the validity of this approach as we found significant associations in the expected directions.
As hypothesized, mothers who consumed more nutritious diets during pregnancy tended to have children with more favorable neurodevelopmental outcomes, while mothers who consumed less nutritious diets and/or higher levels of sodium, saturated fat, and/or sugar during pregnancy tended to have children with poorer neurodevelopmental outcomes. This suggests that the consumption of more comprehensively nutritious prenatal diets favorably affects child neurodevelopment, while the consumption of less comprehensively nutritious prenatal diets may hinder it. Our finding that certain prenatal micronutrients contributed more to relationships with specific skills and abilities than others shows consistency with prior literature. For example, research has linked infant thiamine intake with verbal development [3
]. Additionally, vitamin B6, which contributed strongly to the prenatal mixture association with motor abilities, has previously been implicated in infant gross motor development [29
]. Conversely, certain macronutrients may have more global neurodevelopmental effects. The finding that monounsaturated fats were important for the mixture association with memory, quantitative, and perceptual development is consistent with their role in comprising the structural and functional foundation of the developing brain [30
]. Although fiber, which contributed strongly to the mixture association for various neurodevelopmental abilities, has not been directly examined in this regard, it positively contributes to overall microbiome function [31
], and thus may indirectly facilitate neurodevelopment by enhancing the absorption of other nutrients that are important for it or by altering the microbiome to a more favorable developmental profile.
Higher prenatal calcium also strongly predicted more favorable neurodevelopment, and there are several mechanisms by which it may contribute. First, calcium supplements have been shown to reduce blood lead among pregnant women in Mexico [32
], which may indirectly improve the neurodevelopment of offspring. Second, calcium may positively impact fetal brain development directly since it facilitates the formation of neural synaptic connections and neurotransmission [33
]. Third, adequate calcium is important for thyroid function [35
], and normal maternal thyroid function during pregnancy is important for healthy infant neurodevelopment [36
]. Thus, maternal calcium could directly and indirectly contribute to infant neurodevelopment via these mechanisms.
Findings from the poor prenatal nutrition index were consistent with the good prenatal nutrition index. Lower monounsaturated fat contributed strongly to the association of the mixture with poorer quantitative and memory abilities, and lower thiamine intake contributed strongly to the association of the mixture with poorer verbal abilities. This provides further support for the importance of these nutrients for these abilities. In addition, higher saturated fat consumption within the context of the mixture predicted poorer neurodevelopment, particularly in the quantitative realm. Reasons for this association are not entirely clear; however, it could be that higher saturated fat intake contributes to maternal health problems that negatively affect fetal brain development. Still, there is evidence that saturated fats may in fact be essential to neurodevelopment or a proxy for other nutrients, such as cholesterol, that are critical to it. The APOE4 gene variant that is linked to neurodegeneration, for example, is predictive of better performance in cognitive skills in children [37
] as well as higher serum levels of cholesterol and saturated fats [38
]. Some research also suggests that medium-chain triglycerides found in certain foods high in saturated fat, particularly coconut oil, may have brain health benefits [39
]. Therefore, more research is needed to better understand this finding. Lower fiber intake also appeared to contribute relatively strongly within the mixture to associations with various poorer neurodevelopmental outcomes. Although increased prevalence of fiber deficiency (along with other nutrient deficiencies) has been reported among children with neurodevelopmental disorders [15
], future studies would be needed to determine whether this, or even lower intake within the recommended intake range, could be a contributing factor.
Consistent with our hypothesis, children with more nutritious diets tended to have more favorable perceptual performance, while children with less nutritious diets and/or higher levels of sodium, saturated fat, and/or sugar consumption tended to perform more poorly on memory, perceptual, and quantitative tasks. These associations were weaker than in prenatal models, which could suggest that prenatal nutrition plays a greater role in neurodevelopment than childhood nutrition; however, future studies are needed to further explore this finding. Consistent with prenatal analyses, within the nutrient mixture, higher calcium and monounsaturated fat intakes appeared to be particularly important for favorable neurodevelopment, especially perception, while lower protein intake, essential for brain growth and neurotransmitter synthesis [7
], contributed relatively strongly to the association with poorer memory. Higher sodium and saturated fat intake also contributed relatively strongly within the mixture to associations with poorer performance in various domains. It is possible that consuming snack foods high in sodium and saturated fat that also contain other additives and preservatives known to interfere with attentional processes [40
] may have accounted for this association, or that consuming such foods is a marker of other poor health behaviors that can impact neurodevelopment.
Our study is limited by several factors. First, given that the FFQs measured eating habits over the course of one month for women and one week for children, they may not have captured typical eating habits for people whose eating habits were uncharacteristic during that time. In addition, there may have been changes in nutrition patterns between birth and age of 4 years that could have affected neurodevelopment, but were not captured by the childhood FFQ. Questionnaires have recall error that can affect results, although we believe that this is likely to be nondifferential with respect to neurodevelopment measured at 4–5 years following the prenatal FFQ. In addition, dietary intake does not perfectly correlate with the absorption of nutrients, which is another source of nondifferential error. For example, most women consumed excess levels of iron; however, in Mexico, iron deficiency is still prevalent because the bioavailability of supplemental iron is low and phytic acid consumption is high. We did not include dietary folate intake in our models because folate concentrations from food were not distinguished from folic acid concentrations from supplements at the time of study. Lastly, the prenatal sample size was small, which may limit the generalizability of these results. Future studies examining the consumption of combinations of nutrients and neurodevelopment with larger sample sizes are needed to provide further validation for our findings.