Associations of dietary patterns with brain morphology in children: Results from a prospective population-based study

doi:10.21203/rs.3.rs-1622356/v1

Download PDF

Research Article

Associations of dietary patterns with brain morphology in children: Results from a prospective population-based study

https://doi.org/10.21203/rs.3.rs-1622356/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 08 May, 2023

Read the published version in European Journal of Epidemiology →

You are reading this latest preprint version

Dietary patterns in childhood have been associated with child neurodevelopment and cognitive performance, while the underlying neurobiological pathway is unclear. We aimed to examine associations of dietary patterns in infancy and mid-childhood with pre-adolescent brain morphology. We included 1888 and 2326 children with dietary data at age one or eight years, respectively, and structural neuroimaging at age 10 years in the Generation R Study. Measures of brain morphology were obtained using magnetic resonance imaging. Dietary intake was assessed using food-frequency questionnaires, from which we calculated diet quality scores based on dietary guidelines and from which derived dietary patterns using principal component analyses. Associations were examined using regression models adjusted for potential confounding and controlled for multiple testing. At age one year, children with higher adherence to dietary patterns labeled as ‘Snack, processed foods and sugar’ had smaller cerebral white matter volume (B = -4.30, 95%CI -6.92, -1.68). At age eight years, higher adherence to diet patterns labeled as ‘Whole grains, soft fats and dairy’ was associated with a larger total brain (B = 8.93, 95%CI 4.54, 13.32), and larger cerebral gray matter volumes (B = 5.17, 95%CI 2.86, 7.48). Children with higher diet quality and better adherence to a ‘Whole grains, soft fats and dairy’ dietary pattern at age eight years showed greater brain gyrification and larger surface area, clustered primarily in the dorsolateral prefrontal cortex. In conclusion, dietary patterns in early- and mid-childhood are associated with differences in brain morphology. These findings may explain part of the previously reported relation between dietary patterns and neurodevelopment in children.

dietary patterns

childhood

MRI

brain volume

hippocampus

amygdala

cortical thickness

Diet is an important modifiable factor that modulates brain development by supplying energy and nutrients. In particular, differences in dietary intake in childhood may be crucial for optimal brain development, as the structural and functional complexity of the brain develops rapidly throughout childhood and adolescence [1]. Previous studies have revealed associations between specific nutrients and neurodevelopment in children [2–4], however, these studies do not account for the process by which individual nutrients interact with each other in a synergistic and complex manner in the human body. Therefore, investigating overall diet by looking at dietary patterns can complement the understanding of the association of nutrition with brain development.

Overall dietary patterns, which are often quantified by predefined indices (e.g., diet quality score) or empirically derived from dietary data, are closely related to eating behaviors in the real world setting by considering how foods and nutrients are combined [5]. Examples that are often studied are: prudent dietary patterns, often featured as high whole grains, fruits and vegetables intake or western-like dietary patterns, characterized by high intakes of refined grains, saturated fat, and sugar. Previous studies have shown that dietary patterns are established in early childhood and remain relatively stable into adulthood, with less healthy patterns showing the strongest stability [6–8]. With ongoing global nutrition transitioning towards unhealthy diets [9], more than half of children have poor-quality diets in low-, middle-, and high-income countries [10–13].

Dietary patterns in childhood are likely to have long-term effects on health, including brain development. A growing body of research has identified temporal associations between dietary patterns and children’s cognitive development. For instance, a better diet quality [14] or a prudent dietary pattern at different ages through childhood is linked to higher IQ scores of children and adolescents [15], even when accounting for socioeconomic status. In contrast, children and adolescents with higher adherence to a western-like dietary pattern showed diminished cognitive functioning and lower IQ scores [16, 17]. Neuroimaging studies have identified positive links between IQ and alterations in both global and regional brain volumetric measures in children measured using structural magnetic resonance imaging (MRI) [18, 19], although the influence of nutrition on the neurobiological underpinnings of cognitive performance has been less well studied. Neurodevelopmental disorders are also found to be associated with undesirable dietary patterns. Recent studies have shown that children with autistic traits and attention deficit hyperactivity disorder symptoms had a lower diet quality [20, 21]. Likewise, these neurodevelopmental disorders have been associated with brain morphology [22–24]. Taken together, those associations suggest a potential neurobiological pathway underlying the associations of dietary patterns with neurodevelopmental outcomes.

Dietary effects on neurodevelopment may be global or specific to certain brain regions. The hippocampus and amygdala may be particularly sensitive to the effects of specific dietary patterns, such as high-fat diet diets. The hippocampus is involved in integrating multimodal information and is critical for memory, learning, and is also involved in appetitive, digestive and learned eating behaviors [25]. The amygdala has been implicated in feeding, the reward systems associated with eating, and modulating food consumption [26]. Animal studies have found that a western diet has adverse effects on hippocampal dependent memory function and neurogenesis. Additionally, western diets also induce inflammation in the hippocampus [27, 28], and enhancing amygdala-dependent, cue-based memory in juvenile mice [29]. While these animal studies provide important information related to differences in dietary patterns to brain development, it is unclear whether different dietary patterns are associated with the volumetric differences of hippocampus and amygdala in children. Only one recent study in children investigating the effect of a western diet on hippocampal and amygdala volume did not observe an association. However, they found that increased fat consumption was related to a smaller hippocampal volume [30]. This study has a small sample size and the western diet was calculated by combining percentage of daily calories from fat and sugar, which may limit the generalizability of the findings.

Within this context, we aimed to examine the prospective association of dietary patterns in infancy and mid-childhood with brain morphology assessed at the age of 10 years in a large population-based cohort. Based on evidence from animal studies, we examined the relationship between dietary patterns and two regional brain volumes (hippocampus and amygdala). Due to the paucity of information on the association of overall diet with brain morphology in the literature, we used an exploratory approach that involves global, subcortical volumetric measures (total brain, cerebral white matter, cerebral gray matter, hippocampus and amygdala) and surface-based brain measures (gyrification index, surface area and cortical thickness). We used two approaches to define dietary patterns: a priori diet quality score based on national dietary guidelines, and a posteriori dietary patterns derived as principal components in the population. For global and subcortical volumetric measures, we hypothesized that higher adherence to a western-like dietary pattern and lower diet quality score would be associated with smaller total brain volume at 10 years of age, and would be associated with smaller hippocampal and amygdala volume. We expected that positive associations would be found for a prudent dietary pattern or a better diet quality. For surface-based measures, we did not have prior hypotheses. The findings of this study will facilitate future hypotheses on overall diet and brain development, as this has yet to be studied in children.

Study design and population

This study was embedded in The Generation R Study, an ongoing population-based prospective cohort from early fetal life onward in Rotterdam, the Netherlands. The detailed information on the general design of the Generation R Study has been described in detail elsewhere [31]. Of 9749 eligible pregnant women with a delivery date between April 2002 to January 2006, 7893 mother-child dyads consented for the postnatal follow-up. The study was approved by the Medical Ethics Committee of Erasmus Medical Center, Rotterdam. Written informed consent was obtained from all participating children and their parents.

Dietary information was obtained using validated food frequency questionnaires (FFQs), which were completed, generally by the mother, for 3629 children at one year-of-age (response rate: 71.3%) [32] and for 4733 children at eight years-of-age (response rate: 61.8%) [12]. Structural T₁-weighted images were obtained when children were 10 years [33]. We excluded children who did not visit the neuroimaging research center, or did not provide consent for the MRI scan. For those with valid dietary data at age one year, images were available for 2303 children, and for those with dietary data at eight years of age, images were available for 2813 children. Further, children were excluded if images could not be reconstructed, had poor quality, if major incidental findings were found, or if the gyrification index could not be calculated (Fig.1). The final study population comprised 1888 children for analysis of diet at age one year and 2326 children for analysis of diet at age eight years.

Dietary assessment

Dietary intake was assessed when children were approximately one year-of-age (median 12.9 months, IQR: 12.7, 13.9) and eight years of age (median 8.2 years, IQR: 8.0, 8.2) using two age-specific validated semi-quantitative FFQs. The FFQs included food items based on foods that were frequently consumed by 9-18 months-old and 2-to-12 year-old children in national food consumption surveys in the Netherlands [34, 35], resulting in 221 items for age one year and 71 items for age eight years. Each FFQ contains questions on frequency of consumption, portion size, the types of food items and corresponding food preparation methods over the past 4 weeks. The information was converted into grams of individual food items per day based on standardized portion sizes using SAS VoVris (Vovris V2.4, TNO, 1999-2006). Total energy and nutrient intake were calculated using reference data from the Dutch Food Composition Tables (NEVO 2001). The FFQ at age 1 year was evaluated for nutrient intakes against three 24-hour recalls in a representative sample of 32 Dutch children aged 14 months (r ranging from 0.36 to 0.74) [32]. The FFQ at age 8 years was validated for energy intake (r = 0.62) using the doubly labeled water method among 4 to 6 years Dutch children (n = 30) [36].

Dietary patterns

Both a priori and a posteriori methods were used to evaluate dietary patterns. A priori Diet-Quality Scores (DQS) at ages one and eight years were quantified on the basis of Dutch dietary guidelines [12, 32]. Ten food groups were considered at each age, with intake cut-offs based on age-specific dietary recommendations. Specifically, DQS at one year-of-age (DQS-1y) includes scores for vegetables (≥100 g/d); fruit (≥150 g/d); bread and cereals (≥70 g/d); rice, pasta, potatoes, and legumes (≥70 g/d); dairy (≥350 g/d); meat, poultry, eggs, and meat substitutes (≥35 g/d); fish (≥15 g/d); oils and fats (≥25 g/d); candy and snacks (≤20 g/d); and sugar-sweetened beverages (≤100 g/d) [32]. DQS for children at eight years-of-age (DQS-8y) includes fruit (≥150g/d); vegetables (≥150g/d); whole grains (≥90g/d); fish (≥60g/wk); legumes (≥84g/wk); nuts (≥15g/d); dairy (≥300g/d); oils and soft or liquid margarines (≥30g/d); sugar-containing beverages (≤150g/d); and high-fat and processed meat (≤250g/wk) [12]. To determine the DQS for each participant, the ratio of reported and recommended intake for each food group was first calculated. Then a summary score was calculated by summing the individual food group scores. The overall diet quality scores range from 0 to 10, with higher scores reflecting a higher adherence to the dietary guidelines. Both DQS-1y and DQS-8y were validated for intake of macronutrients and micronutrients among children participating in the Generation R Study [12, 32].

A posteriori dietary patterns were derived using principal components analysis (PCA). Food items from the FFQ at age one year were classified into 27 food groups (Supplemental Table 2) in line with previous studies [37, 38], and food items from the FFQ at age eight years were grouped into 26 food groups (Supplemental Table 3). The varimax rotation method was used to obtain the orthogonal components with the aim of minimizing the covariance between the components. The choice of numbers of dietary patterns to extract was based on the scree plot and Kaiser criterion test. Consequently, the dietary patterns at ages one and eight years with eigenvalues of 1.3 or larger were considered as the most common dietary patterns in the study population. Factor loadings were obtained for each food group, indicating the correlation between the food group and each dietary pattern. Dietary pattern scores for each participant were calculated by summing the food group intake weighted by the corresponding factor loading. The scores were standardized with higher scores reflecting higher similarity to the extracted dietary patterns.

Brain morphometry assessments

At the age of 10 years (median 9.9, IQR 9.7, 10.1), the children were invited to the MRI center. Prior to the MRI procedure, the children participated in a mock MRI session. MR images were acquired on a 3.0 Tesla GE Discovery MR750w MRI system (General Electric Healthcare, Milwaukee, WI, USA) scanner using an 8-channel head coil. The high-resolution T₁-weighted sequence were obtained using a 3D coronal inversion recovery fast spoiled gradient recalled (IR-FSPGR, BRAVO) sequence (TR = 8.77 ms, TE = 3.4 ms, TI = 600 ms, NEX = 1, flip angle = 10°, field of view = 220 × 220 mm, number of slices = 230, in-plane resolution = 1.0 mm³) [33].

Volumetric segmentation and cortical reconstruction were processed using FreeSurfer version 6.0 analysis suite (https://surfer.nmr.mgh.harvard.edu/) with standard processing procedure which have been described previously [39]. Global and regional volumes, including total brain, cerebral white and gray matter, hippocampus and amygdala volumes were measured. For surface-based analyses, cortical thickness, surface area and gyrification were quantified. The quality of cortical reconstructions was visually inspected, and images were removed in case of insufficient quality [39].

Covariates

Several child and maternal characteristics were considered as potential confounders based on prior literature [12, 40-43]. Child sex and date of birth were derived from medical records filled in by obstetricians and community midwives. Self-reported questionnaires during pregnancy were used to collect information on maternal education and household income, maternal psychopathology symptoms and lifestyle-related factors during pregnancy, including smoking status, alcohol use, and folic acid supplement use. Maternal highest education level was dichotomized into low (ranging from no education up to lower vocational training) and high (higher vocational training/university) educational level. Household income was categorized into < 1200€, 1200 - 2200€ and > 2200€ per month. Smoking during pregnancy was categorized into never smoking, smoking until pregnancy was known, and continued smoking. Alcohol use during pregnancy was categorized into never drinking, drinking until pregnancy was known and continued drinking. Folic acid use was categorized into no use during embryogenesis, started the first 10 week and started periconceptional. Maternal psychopathology symptoms were assessed at the third trimester by the Brief Symptom Inventory [44], from which the global severity index was calculated. Maternal dietary intake during pregnancy was measured by food frequency questionnaires, from which diet quality scores were calculated based on dietary guidelines [45]. Child ethnic background was determined based on the country of birth of the parents, and grouped into Dutch, non-Dutch Western, and non-Dutch non-Western, in line with previous work in the same cohort [46]. Child energy intake was estimated from food frequency questionnaires at ages one and eight years old. Height and weight were measured by trained staff at the research center when children were 10 years old. Child’s sex- and age-specific BMI (kg/m²) standard deviation (SD) scores were calculated using Dutch reference growth curves. Intracranial volumes were extracted from FreeSurfer metrics.

Statistical analysis

Characteristics of the study population were described as mean (SD) for continuous variables with normal distribution, median (IQR) for continuous variables with a skewed distribution, or percentages for categorical variables.

We examined individual associations of a priori and a posteriori dietary patterns with brain morphology using multiple linear regression models considering analysis of global, subcortical brain morphometry and surface-based morphometry. For analysis of global and subcortical brain morphometry, the outcomes of interest were total brain, cerebral white matter, cerebral gray matter, hippocampal and amygdala volumes. Next, the associations of a priori and significant a posteriori dietary patterns with surface-based morphometries, including cortical thickness, surface area and gyrification, were examined. We used two models to examine the associations. Model 1 included child sex and age at neuroimaging assessment. Model 2 was further adjusted for maternal education, household income, child ethnic background, child energy intake, child BMI at the age of 10 years, maternal diet quality, smoking, alcohol use, folic acid use, maternal psychopathology symptoms during pregnancy. For analyses including the hippocampus and amygdala, intracranial volumes were additionally included in model 2 to test whether associations are independent of a global effect.

To reduce potential bias due to missing values on covariates, we applied multiple imputations by generating five independent datasets with 50 iterations using the MICE package in R. The results of pooled analyses are presented. The analyses of surface-based morphometry were performed using QDECR package [47] in R version 3.6.3 and all other statistical analyses were carried out using R version 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria). Two-sided α < 0.05 was considered statistical significance. In the analyses of global and regional brain morphometry, correction for multiple testing was performed using the Benjamini-Hochberg approach [48] for 9 dietary patterns (4 for age one year; 5 for age eight years) per 5 primary outcomes (45 tests in total) with a false discovery rate (FDR) of 0.05. In analyses of surface-based morphometry, the Gaussian Monte Carlo Simulations with a cluster-wise correction were used to correct for multiple testing. The cluster-forming threshold was set to p = 0.001, which corresponds to a false-positive rate of 0.05 [49]. Bonferroni correction were further applied for each brain hemisphere (p < 0.025 cluster-wise).

We performed additional analyses to test the robustness of the results. First, information on maternal and child characteristics between respondents and non-respondents with FFQ and neuroimaging data was compared. Second, we repeated multiple regression analyses restricted to children with a Dutch ethnic background, as the FFQs were developed and validated for Dutch children.

Population characteristics

The characteristics of study participants is reported in Table 1. The majority of mothers was highly educated, had a high household income, and had a healthy lifestyle during pregnancy. The majority of children had a Dutch ethnic background. The information of missing values of covariates is listed in Supplemental Table 1.

Dietary patterns

The characteristics of diet quality scores and dietary patterns among children at one and eight years-of-age are shown in Table 2 and Table 3, respectively.

The mean DQS-1y was 4.3 (±1.4) (Table 1). With PCA, three dietary patterns were identified, as shown in Table 2, which were named after the three food groups with the highest factor loadings: (1) ‘Vegetables, potatoes and grains’ dietary pattern; (2) ‘Snacks, processed foods and sugar’ dietary pattern; (3) ‘Butter and margarines, whole grains and dairy’ dietary pattern. Together, these three dietary patterns explained 28.1% of the variation in food intake of one-year-old children.

The mean DQS-8y was 4.5 (±1.2) (Table 1). Four dietary patterns were identified by PCA, as shown in Table 3, which were named as: (1) ‘Snacks, potatoes and processed foods’ dietary pattern; (2) ‘Fish, vegetables and fruit’ dietary pattern; (3) ‘Whole grains, soft fats and dairy’ dietary pattern; (4) ‘Meat replacement, legumes and nuts’ dietary pattern. In total, the four dietary patterns explained 28.8% of the variation in food intake of eight-year-old children.

Correlations between a priori and a posteriori dietary patterns were calculated by the Pearson correlation coefficient and reported in Supplemental Table 4. DQS-1y is highly correlated with the ‘Vegetables, potatoes and grains’ dietary pattern at the same age (r = 0.69). DQS-8y is moderately associated with the ‘Whole grains, soft fats and dairy’ dietary pattern at age eight years (r = 0.56).

Associations of a priori and a posteriori dietary patterns with global and subcortical brain morphometry

Table 4 reports the associations of dietary patterns with global brain volumes. A higher adherence to the ‘Snacks, processed foods and sugar’ dietary pattern at the age of one year was negatively associated with total brain (-7.14 cm³/SD; 95% CI: -12.79, -1.49) and cerebral white matter volumes (-4.30 cm³/SD; 95% CI: -6.92, -1.68) at age 10 years, while a higher adherence to the ‘Butter and margarines, cereals and dairy’ dietary pattern at age one year was positively associated with cerebral gray matter volume at age 10 years (2.57 cm³/SD; 95% CI: 0.20, 4.94). Only the association of cerebral white matter volume remained significant after correction for multiple testing.

A higher adherence to the ‘Snacks, potatoes and processed foods’ dietary pattern at the age of 8 years was negatively associated with total brain (-7.63 cm³/SD; 95% CI: -13.19, -2.06) and cerebral gray matter volumes (-4.07 cm³/SD; 95% CI: -7.00, -1.14) at age 10. Conversely, a higher adherence to the ‘Whole grains, soft fats and dairy’ dietary pattern was positively associated with total brain (8.93 cm³/SD; 95% CI: 4.54, 13.32), cerebral white matter (2.95 cm³/SD; 95% CI: 0.91, 4.99) and cerebral gray matter volumes (5.17 cm³/SD; 95% CI: 2.86, 7.48). A higher DQS-8y was positively associated with cerebral gray matter volume (2.30 cm³/SD; 95% CI: 0.15, 4.45). The association of the ‘Whole grains, soft fats and dairy’ dietary pattern with total brain and cerebral gray matter volume remained present after multiple testing correction.

No associations were observed between all dietary patterns at age one or eight years and hippocampal or amygdala volume at age 10 years (Table 5). In order to examine if associations observed for cerebral white and cerebral gray matter volumes were independent of a global effect, we additionally adjusted for intracranial volume in model 2. Only the association of the ‘Snacks, processed foods and sugar’ dietary pattern at one year with cerebral white matter volume remained significant (-1.44 cm³/SD; 95% CI: -2.84, -0.04).

Associations of a priori and a posteriori dietary patterns with surface-based brain morphometry

The cortical regions that were associate with DQS-8y and a ‘Whole grains, soft fats and dairy’ dietary pattern at eight years of age were illustrated in Fig.2 and Fig.3. Regions in model 2 were considered statistically significant. Higher DQS-8y was positively associated with gyrification in regions of the frontal, parietal and temporal lobes, including the rostral middle frontal, superior frontal, caudal middle frontal, precentral, postcentral, supramarginal, superior temporal, and transverse temporal gyrus in the right hemisphere, and caudal middle frontal, rostral middle frontal, parsopercularis, superior parietal, supramarginal, post central, paracentral, precuneus and posterior cingulate gyrus in the left hemisphere (Fig.2A). A higher adherence to the ‘Whole grains, soft fats and dairy’ dietary pattern at age eight years was positively associated with gyrification in small regions of fusiform, parahippocampal, inferior temporal, rostral middle frontal, superior frontal gyrus in the right hemisphere (Fig.2B). The positive association of higher DQS-8y with surface area was found in small clusters in rostral middle frontal and superior frontal regions in the left hemisphere (Fig.3A). Likewise, positive associations of a higher adherence to the ‘Whole grains, soft fats and dairy’ dietary pattern at age eight years with surface area was observed in small clusters in prefrontal and occipital lobe, including superior frontal, superior parietal and lateral occipital regions in the right hemisphere, and superior temporal, middle temporal and superior frontal regions in the left hemisphere (Fig.3B). No associations were found between cortical thickness with DQS-8y or the ‘Whole grains, soft fats and dairy’ dietary pattern at eight years. In addition, we did not find significant associations of diet quality at the age of one year with either of the three surface-based cortical measures. Specific information on the associated brain regions and their p values are reported in the Supplemental Table 5.

Additional analyses

The comparison of characteristics between respondents and non-respondents to FFQ and neuroimaging showed that respondents had a higher socioeconomic status, a higher percentage of folic acid usage before or during pregnancy, a better diet quality during pregnancy, and a higher percentage of children with a Dutch ethnic background (Supplemental Table 6). In general, the analyses including only children with a Dutch ethnic background showed stronger effect estimates and found more statistically significant associations, compared to the whole study population. Specifically, a higher adherence to the ‘Snacks, potatoes and processed foods’ at eight years was associated with larger total brain and cerebral gray matter volume; a higher adherence to the ‘Whole grains, soft fats and dairy’ at eight years was associated with a larger total brain, cerebral white matter and cerebral gray matter volumes. While the effect size of association between the ‘Snacks, processed foods and sugar’ dietary pattern at age one year and cerebral white matter in the additional analysis was comparable to the main analysis, it was no longer significant (Supplemental Table 7 and 8).

This large population-based prospective cohort study explored associations of different dietary patterns in infancy and in mid-childhood with global, regional brain volumes and surface-based brain morphometry when children were 10 years of age. Diet quality at one and eight years-of-age as assessed by adherence to dietary guidelines was low to moderate. Global brain volumes at age 10 were negatively associated with dietary patterns characterized by high intake of snack and processed foods at age one and age eight years. Interestingly, global brain volumes were positively associated with a dietary pattern characterized by whole grains, soft fats and dairy at age eight years. No volumetric differences were found in the hippocampus or amygdala for dietary patterns at one and eight years-of-age. Higher DQS-8y or higher adherence to the aforementioned pattern with high intake in whole grains, soft fats and dairy at age eight years was associated with a greater gyrification and larger surface area in widespread areas of brain, with significant areas primarily clustered in association cortices, notably in the prefrontal cortex. In the following discussion, we compared the dietary patterns high in snacks and processed foods intake with western-like dietary patterns, while the dietary patterns high in whole grains intake were compared with prudent dietary patterns due to their high factor loadings for those food groups. Overall, our findings suggest that having a prudent dietary pattern in school age, specifically one rich in whole grains, soft fat and dairy, is linked to larger global brain volumes, whereas consuming a western-like dietary pattern in infancy and school age is associated with lower global brain volume. Our study provides novel information that extend earlier studies examining the association of human overall diet with brain health in childhood.

Our findings of the association between brain morphology and children’s dietary patterns in early- and mid-childhood, especially those related to a prudent dietary pattern and a western-like dietary pattern, complement previous studies showing that overall dietary patterns are associated with school attainment [50, 51] and cognitive performance [15–17, 52, 53] in children. Considering that total brain volume [54], cerebral gray matter volume [55] and white matter microstructure [56] have been associated with cognitive ability of children, our results suggest that neuroanatomical correlates may underlie the association between dietary patterns and cognitive development in children. Furthermore, our results indicated that a prudent or better diet quality was associated with brain regions functionally implicated in appetite regulation. Among those regions, most clustered in the dorsolateral prefrontal cortex (DLPFC). The prefrontal cortex is a complex region, but is notably involved in dopamine-mediated executive function, regulation of reward, and the inhibition of impulsive behaviors. Specifically, the DLPFC is associated with satiety, food craving, and executive functioning [57, 58]. A randomized controlled trail (RCT) in healthy young men with normal BMI found increased neuronal activity in right DLPFC reduced overall caloric intake and diminished self-reported appetite scores [59], which may suggest DLPFC involvement in food intake-related control mechanisms. However, it is unclear to what extent the functional connectivity translates into differences in brain morphology, although theories of gyrification support a relationship between folding patterns of the brain and increased connectivity [60]. Caution is needed regarding the translation of functional connectivity studies with the observed association between dietary patterns and brain morphology.

Among all significant associations, a prudent dietary pattern at age eight years was positively associated and a western-like dietary pattern at age one year was negatively associated with global brain volumes at age 10 years after multiple testing correction. The association of ‘Snacks, processed foods and sugar’ dietary pattern at one year with cerebral white matter volumes persisted after controlling for intracranial volume, suggesting that cerebral white matter development may be specifically susceptible to a western-like diet specifically in infancy. A possible mechanistic interpretation is that overall diet influences brain development if adherence to a dietary pattern with a higher risk of nutrient inadequacy occurs during a period of high need, considering the rapid rate of brain development within early life [2, 61]. The majority of white matter myelination occurs during the first 2 years of life [62]. Meanwhile, research has shown that children with a western-like dietary pattern were exposed to excess fat and sugar intake or inadequate nutrient intake, such as omega-3 polyunsaturated fatty acids [63], which serves as an important nutrient for nerve cell myelination and has shown beneficial effect on cognitive impairment [64, 65].

Contrary to the animal studies, our findings did not support our hypothesis that overall diet is associated with hippocampal and amygdala volumes. We are aware of only one study investigating the effects of a western diet on hippocampal and amygdala volume in five- to nine-year-old-children. Stadterman et al [30] reported a link between percentage of daily calories from fat, but not western diet, with an isolated decreased left hippocampal volume, but no relationship was found with the right hippocampal or amygdala volume. Their study calculated western diet as a summed percentage of daily calories from fat and sugar and the sample size was small (n = 21), however, it represents a promising step towards understanding the impact of western-like diet on hippocampus and amygdala development in children. We build upon their study investigating the association between a western-like diet and the left hippocampus but found no association. Our large sample size and the population-based sample, coupled with our approach to derive western-like dietary patterns, likely better resembles actual eating habits, although we did not observe evidence for any effects. Future research with a focused hypothesis on the hippocampus and the amygdala is needed to ascertain their association with western-like diet in children. Moreover, considering the evidence from population-based studies in adults, which suggest that the left hippocampus may be more vulnerable to western-like diets [66, 67], further investigation into individual effects on hippocampus in each hemisphere is warranted. Such studies preferably follow children into adolescence and adulthood, so that potential long-term effects can be observed.

Although the etiology of the relationships between dietary patterns and neurodevelopment remains unclear, there have been potential mechanisms proposed. First, the effect of dietary patterns on brain morphology may be mediated by epigenetic mechanisms. Diet, as an epigenetic regulator, affects multiple genes expression at levels of transcription, translation and post-translational modification. Subsequently, the variation in gene expression regulated by nutrition influences several neurobiological processes, including neurogenesis, synaptic plasticity and neuronal connectivity [68, 69], which may lead to rearrangement of brain structure. Second, differences in dietary patterns can induce differences in metabolic changes underlying brain morphology. For example, brain-derived neurotrophic factor (BDNF) is related to the morphological variation of the hippocampus and prefrontal cortex [70], and can be influenced by diet. High-fat diets have been found to increase oxidative stress and further interfere with the level of BDNF [71]. Regarding healthy dietary patterns, a randomized control trial study in adults found a higher plasma BDNF levels in an experimental group with a Mediterranean diet intervention compared to a control group [72]. Furthermore, decreased BDNF has been linked to cognitive decline [71]. While these studies were performed in adults, there may be expectations of even greater differences in children, considering the high energy consumption associated with neurodevelopment. Further, the evidence showing dietary patterns were associated with cognitive performance in children, suggests shared brain metabolic pathways that underlie brain morphological changes in children and adults. Last, diet could affect the gut microbiota which in return alter the host’s physiological responses and associated structural adaption. Short-chain fatty acids (SCFAs) are the most studied metabolites produced by microbes. Foods high in dietary fiber, such as fruits, vegetables and whole grains, increase the levels of SCFAs through gut microbial fermentation [73]. Results from animal studies indicated that SCFAs might influence gut-brain communication and brain function directly or indirectly through neurochemical pathways [74].

By reporting associations of dietary patterns in early and mid-childhood with global and regional brain volumes, our findings underline the need for nutrition research related to child development, notably in middle childhood. Nutrition has been hypothesized as an aspect of the experience-dependent environment that can influence neurodevelopment [75]. However, most studies on nutrition and brain development have focused on early life nutrition, while few studies have examined children in the ‘forgotten years.’ In fact, dietary patterns change drastically from milk-based during infancy to omnivore patterns in childhood. Previous work from our group showed that diet quality at one year-of-age changes considerably by eight years-of-age [12], indicating that relative stability of dietary patterns may occur after the age of one year. Another longitudinal study including five European countries suggests that dietary patterns are established between one and two years and remain stable to eight years of age [8]. In addition, brain metabolism associated with neurodevelopment is high, with a considerable need of energy. This changes at around the age of four to five years, coincident with the slower rate of growth during that age. Overall, changes in dietary patterns across childhood, highlight the importance of investigating the effect of diet not only in infancy, but also in mid-childhood.

Most dietary interventions for brain development to date have targeted micronutrient supplementation, showing positive effects of nutritional supplementation in certain cognitive domains in nutrient-deficient children [76]. Some interventions which used food supplementation (e.g., fortified foods) in early childhood [2] or in children with atypical neurodevelopment [77] have shown short-term effects on cognitive and motor development. Evidence from those RCTs is intriguing and denotes a potential causal link between nutrition and brain development. However, some RCT studies that provide supplements of a single nutrient or food in certain groups of children failed to detect changes in cognitive performance [78, 79]. It is noteworthy that none of RCTs considers overall diet in their study design, which may partly explain the inconsistency in findings. This could be explained by an adapted hypothetical scenario in which the effects of nutrient deficiency and poor-quality diet may show an interacting effect on children’s cognitive development [2], thus reducing the effect of nutrient supplementation. Moreover, it is largely unknown whether the effect of nutrition on brain development could be detected by objective anatomical measurement, like brain volumetric alteration. We filled in the gap of lack of evidence in a healthy pediatric population with a long follow-up period.

The strengths of our study are the population-based prospective design, availability of multiple covariates, and large-scale neuroimaging in children. In addition, we defined dietary patterns using a priori diet quality score and a posteriori dietary patterns based on food groups, which captured unrelated eating patterns in the population. Therefore, the results of our study provide a public health message on the relationships of dietary patterns and brain development in children.

The main limitation is the design of the study. Although longitudinal, the design is in essence cross-sectional, as we lack repeated measures of diet and brain imaging. Thus, our results cannot infer causality between dietary patterns and brain morphometry. Future studies with repeated measurements of dietary patterns and brain morphology, preferably those with an intervention component, are necessary to better understand the direction of the association. Another limitation involves assumptions associated with calculating the a priori and a posteriori dietary patterns analysis, which can influence the interpretation and limit the comparability of the study with others. Such assumptions include how food items were clustered into food groups, the number of principal components to be retained. Additionally, dietary patterns are likely to be a part of lifestyle behaviors, and thus there is the potential for residual confounding. However, after adjusting for BMI at the age of 10 years in the model, dietary patterns were still significantly associated with brain morphology. Several other limitations should also be acknowledged. First, the use of FFQs to quantify dietary intake is subject to measurement error [80]. However, the FFQs used in our study were validated against three 24h recalls in Dutch children at age one year and against the doubly labelled water method in Dutch children at age eight years, and had sufficient capacity of ranking participants with regard to energy intake. Moreover, we adjusted our models for energy intake to mitigate the effect of measurement error in the FFQs that individuals tend to misreport their intake of food and beverages in the same direction. Fourth, although we adjusted for several covariates including socioeconomic status, maternal psychopathological symptoms, lifestyle and child BMI, we cannot rule out genetic and unmeasured environmental confounding because of the observational nature of the study. Lastly, the non-response analyses suggested a possible selection bias, which may limit the generalizability of the study.

We found a relationship between early-life and mid-childhood dietary pattern and later brain morphological using a prospective population-based study of child development, which suggests that healthy diet may benefit neurodevelopment, which has an important public health message. The findings of our study suggest that a prudent dietary pattern in mid-childhood is positively associated with global and widespread differences in pre-adolescent brain morphometry. Adherence to a western-like dietary pattern in infancy and school age is differentially associated with cerebral white matter, cerebral gray matter and total brain volume, suggesting that the timing of neurodevelopmental processes may play an important role in understanding the effect of overall diet. With the observed temporal associations, our findings highlight the potential of overall dietary patterns to be associated with brain morphology in children, which may explain part of the previously demonstrated relation between dietary patterns and neurodevelopment and cognitive performance in children. However, further research with repeated measurements of dietary patterns and brain morphology is needed to better understand the direction of the association.

Acknowledgements

The Generation R Study is conducted by Erasmus Medical Center in close collaboration with the School of Law and the Faculty of Social Sciences at the Erasmus University, Rotterdam; the Municipal Health Service, Rotterdam area; and the Stichting Trombosedienst & Artsenlaboratorium Rijnmond (Star-MDC), Rotterdam. We gratefully acknowledge the contribution of participating mothers, general practitioners, hospitals, midwives, and pharmacies in Rotterdam, the Netherlands.

Funding

The general design of the Generation R Study is made possible by finical support from the Erasmus MC, University Medical Center, Rotterdam, Erasmus University Rotterdam, the Netherlands Organization for Health Research and Development (ZonMw), Netherlands Organization for Scientific Research (NWO), Ministry of Health, Welfare and Sport and Ministry of Youth and Families. Yuchan Mou is supported by China Scholarship Council (CSC) PhD Fellowship for her PhD study in Erasmus Medical Center, Rotterdam, the Netherlands. The scholarship file number is 201806240125, CSC URL: http://www.csc.edu.cn/. Pauline Jansen is supported by a grant from ZonMw (Mental Health Care Research Program - Fellowship 636320005). The neuroimaging data collection and image processing is supported by ZonMw TOP Grant 91211021 to Tonya White. Supercomputing resources were supported by the NWO Physical Sciences Division (Exacte Wetenschappen) and SURFsara (Cartesius compute cluster, www.surfsara.nl). The funders had no role in the study design, data collection, management, analysis and interpretation of data, and preparation or writing the manuscript.

Competing interests

The authors have no competing interests to declare that are relevant to the content of this article.

Author contributions

Yuchan Mou, Tonya White and Trudy Voortman designed research; Pauline W. Jansen, Tonya White, Trudy Voortman provided essential materials (databases); Yuchan Mou performed statistical analysis; Yuchan Mou wrote the paper; Elisabet Blok, Monica Barroso, Pauline W. Jansen, Trudy Voortman and Tonya White provided consultation regarding the analyses and interpretation of data; Yuchan Mou and Trudy Voortman had primary responsibility for final content. All authors read and approved the final manuscript.

Ethics approval

The study was approved by the Medical Ethics Committee of Erasmus Medical Center, Rotterdam.

Consent to participate

Written informed consent was obtained from all participating children and their parents.

Casey BJ, Getz S, Galvan A. The adolescent brain. Dev Rev. 2008;28(1):62–77. doi:10.1016/j.dr.2007.08.003.
Prado EL, Dewey KG. Nutrition and brain development in early life. Nutr Rev. 2014;72(4):267–84. doi:https://doi.org/10.1111/nure.12102.
Georgieff MK. Nutrition and the developing brain: nutrient priorities and measurement. Am J Clin Nutr. 2007;85(2):614S-20S. doi:10.1093/ajcn/85.2.614S.
Richardson AJ. Omega-3 fatty acids in ADHD and related neurodevelopmental disorders. Int Rev Psychiatry. 2006;18(2):155–72. doi:10.1080/09540260600583031.
Hu FB. Dietary pattern analysis: a new direction in nutritional epidemiology. Curr Opin Lipidol. 2002;13(1):3–9.
Mikkilä V, Räsänen L, Raitakari OT, Pietinen P, Viikari J. Consistent dietary patterns identified from childhood to adulthood: The Cardiovascular Risk in Young Finns Study. Br J Nutr. 2005;93(6):923–31. doi:10.1079/BJN20051418.
Group EMCCS. Dietary Patterns Track from Infancy to Preschool Age: Cross-Sectional and Longitudinal Perspectives. J Nutr. 2015;145(4):775–82. doi:10.3945/jn.114.201988.
Luque V, Escribano J, Closa-Monasterolo R, et al. Unhealthy Dietary Patterns Established in Infancy Track to Mid-Childhood: The EU Childhood Obesity Project. J Nutr. 2018;148(5):752–9. doi:10.1093/jn/nxy025.
Popkin BM, Adair LS, Ng SW. Global nutrition transition and the pandemic of obesity in developing countries. Nutr Rev. 2012;70(1):3–21. doi:10.1111/j.1753-4887.2011.00456.x.
Carlson A, Lino M, Gerrior S, Basiotis PP. Insight. September 2001: report card on the diet quality of children ages 2 to 9. Fam Econ Nutr Rev. 2003;25(2):52–5. 15(.
Jennings A, Welch A, van Sluijs EM, Griffin SJ, Cassidy A. Diet quality is independently associated with weight status in children aged 9–10 years. J Nutr. 2011;141(3):453–9.
van der Velde LA, Nguyen AN, Schoufour JD, et al. Diet quality in childhood: the Generation R Study. Eur J Nutr. 2019;58(3):1259–69. doi:10.1007/s00394-018-1651-z.
Liu J, Rehm CD, Onopa J, Mozaffarian D. Trends in Diet Quality Among Youth in the United States, 1999–2016. JAMA. 2020;323(12):1161–74. doi:10.1001/jama.2020.0878.
Nyaradi A, Oddy WH, Hickling S, Li J, Foster JK. The relationship between nutrition in infancy and cognitive performance during adolescence. Front Nutr. 2015;2:2.
Nyaradi A, Li J, Hickling S, Foster J, Oddy WH. The role of nutrition in children's neurocognitive development, from pregnancy through childhood. Front Hum Neurosci. 2013;7:97.
Nyaradi A, Foster JK, Hickling S, et al. Prospective associations between dietary patterns and cognitive performance during adolescence. J Child Psychol Psychiatry. 2014;55(9):1017–24.
Northstone K, Joinson C, Emmett P, Ness A, Paus T. Are dietary patterns in childhood associated with IQ at 8 years of age? A population-based cohort study. J Epidemiol Community Health. 2012;66(7):624–8.
Pietschnig J, Penke L, Wicherts JM, Zeiler M, Voracek M. Meta-analysis of associations between human brain volume and intelligence differences: How strong are they and what do they mean? Neurosci Biobehavioral Reviews. 2015;57:411–32.
Lange N, Froimowitz MP, Bigler ED, Lainhart JE, Brain Development Cooperative G. Associations between IQ, total and regional brain volumes, and demography in a large normative sample of healthy children and adolescents. Dev Neuropsychol. 2010;35(3):296–317.
Harris HA, Mou Y, Dieleman GC, Voortman T, Jansen PW. Child Autistic Traits, Food Selectivity and Diet Quality: A Population-Based Study. J Nutr. 2021. doi:10.1093/jn/nxab413.
Mian A, Jansen PW, Nguyen AN, Bowling A, Renders CM, Voortman T. Children's Attention-Deficit/Hyperactivity Disorder Symptoms Predict Lower Diet Quality but Not Vice Versa: Results from Bidirectional Analyses in a Population-Based Cohort. J Nutr. 2019;149(4):642–8. doi:10.1093/jn/nxy273.
Itahashi T, Yamada T, Nakamura M, et al. Linked alterations in gray and white matter morphology in adults with high-functioning autism spectrum disorder: A multimodal brain imaging study. NeuroImage: Clin. 2015;7:155–69. doi:https://doi.org/10.1016/j.nicl.2014.11.019.
Samea F, Soluki S, Nejati V, et al. Brain alterations in children/adolescents with ADHD revisited: A neuroimaging meta-analysis of 96 structural and functional studies. Neurosci Biobehavioral Reviews. 2019;100:1–8.
Hoogman M, Bralten J, Hibar DP, et al. Subcortical brain volume differences in participants with attention deficit hyperactivity disorder in children and adults: a cross-sectional mega-analysis. The Lancet Psychiatry. 2017;4(4):310–9.
Kanoski SE, Grill HJ. Hippocampus Contributions to Food Intake Control: Mnemonic, Neuroanatomical, and Endocrine Mechanisms. Biol Psychiatry. 2017;81(9):748–56. doi:https://doi.org/10.1016/j.biopsych.2015.09.011.
Douglass AM, Kucukdereli H, Ponserre M, et al. Central amygdala circuits modulate food consumption through a positive-valence mechanism. Nat Neurosci. 2017;20(10):1384–94. doi:10.1038/nn.4623.
Nakandakari SCBR, Munoz VR, Kuga GK, et al. Short-term high-fat diet modulates several inflammatory, ER stress, and apoptosis markers in the hippocampus of young mice. Brain Behav Immun. 2019;79:284–93.
Cigliano L, Spagnuolo MS, Crescenzo R, et al. Short-term fructose feeding induces inflammation and oxidative stress in the hippocampus of young and adult rats. Mol Neurobiol. 2018;55(4):2869–83.
Boitard C, Maroun M, Tantot F, et al. Juvenile obesity enhances emotional memory and amygdala plasticity through glucocorticoids. J Neurosci. 2015;35(9):4092–103.
Stadterman J, Belthoff K, Han Y, Kadesh AD, Yoncheva Y, Roy AK. A Preliminary Investigation of the Effects of a Western Diet on Hippocampal Volume in Children. Frontiers in pediatrics. 2020;8.
Kooijman MN, Kruithof CJ, van Duijn CM, et al. The Generation R Study: design and cohort update 2017. Eur J Epidemiol. 2016;31(12):1243–64. doi:10.1007/s10654-016-0224-9.
Voortman T, Kiefte-de Jong JC, Geelen A, et al. The development of a diet quality score for preschool children and its validation and determinants in the Generation R Study. J Nutr. 2015;145(2):306–14. doi:10.3945/jn.114.199349.
White T, Muetzel RL, El Marroun H, et al. Paediatric population neuroimaging and the Generation R Study: the second wave. Eur J Epidemiol. 2018;33(1):99–125. doi:10.1007/s10654-017-0319-y.
Breedveld BC, Hulshof KFAM. Zo eten jonge peuters in Nederland 2002: Resultaten van het Voedingsstoffen Inname Onderzoek 20022002.
Netherlands Nutrition Centre [Dutch. Voedingscentrum]. Results of the Dutch Food Consumption Survey 1997–1998 [Dutch: Zo eet Nederland: resultaten van de Voedselconsumptiepeiling1997–1998]. The Hague, the Netherlands1998.
Dutman AE, Stafleu A, Kruizinga A, et al. Validation of an FFQ and options for data processing using the doubly labelled water method in children. Public Health Nutr. 2011;14(3):410–7. doi:10.1017/s1368980010002119.
Barroso M, Beth SA, Voortman T, et al. Dietary Patterns After the Weaning and Lactation Period Are Associated With Celiac Disease Autoimmunity in Children. Gastroenterology. 2018;154(8):2087–96.e7. doi:10.1053/j.gastro.2018.02.024.
Voortman T, Leermakers ETM, Franco OH, et al. A priori and a posteriori dietary patterns at the age of 1 year and body composition at the age of 6 years: the Generation R Study. Eur J Epidemiol. 2016;31(8):775–83. doi:10.1007/s10654-016-0179-x.
Muetzel RL, Blanken LME, van der Ende J, et al. Tracking Brain Development and Dimensional Psychiatric Symptoms in Children: A Longitudinal Population-Based Neuroimaging Study. Am J Psychiatry. 2018;175(1):54–62.
Noble KG, Houston SM, Brito NH, et al. Family income, parental education and brain structure in children and adolescents. Nat Neurosci. 2015;18(5):773–8. doi:10.1038/nn.3983.
Ambrosini GL, Emmett PM, Northstone K, Jebb SA. Tracking a dietary pattern associated with increased adiposity in childhood and adolescence. Obes (Silver Spring). 2014;22(2):458–65. doi:10.1002/oby.20542.
Polanska K, Kaluzny P, Aubert AM, et al. Dietary Quality and Dietary Inflammatory Potential During Pregnancy and Offspring Emotional and Behavioral Symptoms in Childhood: An Individual Participant Data Meta-analysis of Four European Cohorts. Biol Psychiatry. 2021;89(6):550–9. doi:https://doi.org/10.1016/j.biopsych.2020.10.008.
Stein A, Pearson RM, Goodman SH, et al. Effects of perinatal mental disorders on the fetus and child. The Lancet. 2014;384(9956):1800–19.
Derogatis LR, Melisaratos N. The Brief Symptom Inventory: an introductory report. Psychol Med. 1983;13(3):595–605. doi:Doi: 10.1017/s0033291700048017.
Nguyen AN, de Barse LM, Tiemeier H, et al. Maternal history of eating disorders: Diet quality during pregnancy and infant feeding. Appetite. 2017;109:108–14. doi:https://doi.org/10.1016/j.appet.2016.11.030.
El Marroun H, Zou R, Leeuwenburg MF, et al. Association of Gestational Age at Birth With Brain Morphometry. JAMA Pediatr. 2020;174(12):1149–58.
Lamballais S, Muetzel RL. QDECR: A Flexible, Extensible Vertex-Wise Analysis Framework in R. Front Neuroinformatics. 2021;15(14). doi:10.3389/fninf.2021.561689.
Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Roy Stat Soc: Ser B (Methodol). 1995;57(1):289–300.
Greve DN, Fischl B. False positive rates in surface-based anatomical analysis. NeuroImage. 2018;171:6–14. doi:https://doi.org/10.1016/j.neuroimage.2017.12.072.
Feinstein L, Sabates R, Sorhaindo A, et al. Dietary patterns related to attainment in school: the importance of early eating patterns. J Epidemiol Commun Health. 2008;62(8):734. doi:10.1136/jech.2007.068213.
Fu M-L, Cheng L, Tu S-H, Pan W-H. Association between Unhealthful Eating Patterns and Unfavorable Overall School Performance in Children. J Am Diet Assoc. 2007;107(11):1935–43. doi:https://doi.org/10.1016/j.jada.2007.08.010.
Theodore RF, Thompson JMD, Waldie KE, et al. Dietary patterns and intelligence in early and middle childhood. Intelligence. 2009;37(5):506–13. doi:https://doi.org/10.1016/j.intell.2009.07.001.
Smithers LG, Golley RK, Mittinty MN, et al. Dietary patterns at 6, 15 and 24 months of age are associated with IQ at 8 years of age. Eur J Epidemiol. 2012;27(7):525–35. doi:10.1007/s10654-012-9715-5.
McDaniel MA. Big-brained people are smarter: A meta-analysis of the relationship between in vivo brain volume and intelligence. Intelligence. 2005;33(4):337–46. doi:https://doi.org/10.1016/j.intell.2004.11.005.
Wilke M, Sohn J-H, Byars AW, Holland SK. Bright spots: correlations of gray matter volume with IQ in a normal pediatric population. NeuroImage. 2003;20(1):202–15. doi:https://doi.org/10.1016/S1053-8119(03)00199-X.
Muetzel RL, Mous SE, van der Ende J, et al. White matter integrity and cognitive performance in school-age children: A population-based neuroimaging study. NeuroImage. 2015;119:119–28. doi:https://doi.org/10.1016/j.neuroimage.2015.06.014.
Gluck ME, Heinitz S, Alonso-Alonso M, et al. Decreased Snack Food Intake in Obesity Following Noninvasive Neuromodulation Targeted to the Left Dorsolateral Prefrontal Cortex (LDLPFC). Brain Stimulation: Basic, Translational, and Clinical Research in Neuromodulation. 2017;10(4):e37. doi:10.1016/j.brs.2017.04.063.
Gluck ME, Viswanath P, Stinson EJ. Obesity, Appetite, and the Prefrontal Cortex. Curr Obes Rep. 2017;6(4):380–8.
Jauch-Chara K, Kistenmacher A, Herzog N, Schwarz M, Schweiger U, Oltmanns KM. Repetitive electric brain stimulation reduces food intake in humans. Am J Clin Nutr. 2014;100(4):1003–9. doi:10.3945/ajcn.113.075481.
Van Essen DC. A 2020 view of tension-based cortical morphogenesis. Proceedings of the National Academy of Sciences. 2020;117(52):32868. doi:10.1073/pnas.2016830117.
White TJH. Brain Development and Stochastic Processes During Prenatal and Early Life: You Can’t Lose It if You’ve Never Had It; But It’s Better to Have It and Lose It, Than Never to Have Had It at All. J Am Acad Child Adolesc Psychiatry. 2019;58(11):1042–50. doi:https://doi.org/10.1016/j.jaac.2019.02.010.
Lebel C, Deoni S. The development of brain white matter microstructure. NeuroImage. 2018;182:207–18. doi:10.1016/j.neuroimage.2017.12.097.
Gow RV, Hibbeln JR. Omega-3 Fatty Acid and Nutrient Deficits in Adverse Neurodevelopment and Childhood Behaviors. Child and Adolescent Psychiatric Clinics. 2014;23(3):555–90. doi:10.1016/j.chc.2014.02.002.
Sinn N. Nutritional and dietary influences on attention deficit hyperactivity disorder. Nutr Rev. 2008;66(10):558–68. doi:10.1111/j.1753-4887.2008.00107.x.
Provensi G, Schmidt SD, Boehme M, et al. Preventing adolescent stress-induced cognitive and microbiome changes by diet. Proceedings of the National Academy of Sciences. 2019;116(19):9644-51. doi:10.1073/pnas.1820832116.
Jacka FN, Cherbuin N, Anstey KJ, Sachdev P, Butterworth P. Western diet is associated with a smaller hippocampus: a longitudinal investigation. BMC Med. 2015;13(1):215. doi:10.1186/s12916-015-0461-x.
Akbaraly T, Sexton C, Zsoldos E, et al. Association of Long-Term Diet Quality with Hippocampal Volume: Longitudinal Cohort Study. Am J Med. 2018;131(11):1372–81.e4. doi:10.1016/j.amjmed.2018.07.001.
Dauncey MJ. Nutrition, the brain and cognitive decline: insights from epigenetics. Eur J Clin Nutr. 2014;68(11):1179–85. doi:10.1038/ejcn.2014.173.
Dauncey MJ. Genomic and Epigenomic Insights into Nutrition and Brain Disorders. Nutrients. 2013;5(3):887–914.
Pezawas L, Verchinski BA, Mattay VS, et al. The Brain-Derived Neurotrophic Factor val66met Polymorphism and Variation in Human Cortical Morphology. J Neurosci. 2004;24(45):10099. doi:10.1523/jneurosci.2680-04.2004.
Francis H, Stevenson R. The longer-term impacts of Western diet on human cognition and the brain. Appetite. 2013;63:119–28. doi:https://doi.org/10.1016/j.appet.2012.12.018.
Sánchez-Villegas A, Galbete C, Martinez-González MA, et al. The effect of the Mediterranean diet on plasma brain-derived neurotrophic factor (BDNF) levels: the PREDIMED-NAVARRA randomized trial. Nutr Neurosci. 2011;14(5):195–201.
Fernandes J, Su W, Rahat-Rozenbloom S, Wolever TM, Comelli EM. Adiposity, gut microbiota and faecal short chain fatty acids are linked in adult humans. Nutr Diabetes. 2014;4(6):e121.
Dalile B, Van Oudenhove L, Vervliet B, Verbeke K. The role of short-chain fatty acids in microbiota–gut–brain communication. Nat Reviews Gastroenterol Hepatol. 2019;16(8):461–78. doi:10.1038/s41575-019-0157-3.
Nelson CA. A Neurobiological Perspective on Early Human Deprivation. Child Dev Perspect. 2007;1(1):13–8. doi:https://doi.org/10.1111/j.1750-8606.2007.00004.x.
Lam LF, Lawlis TR. Feeding the brain – The effects of micronutrient interventions on cognitive performance among school-aged children: A systematic review of randomized controlled trials. Clin Nutr. 2017;36(4):1007–14. doi:https://doi.org/10.1016/j.clnu.2016.06.013.
Kidd PM. Omega-3 DHA and EPA for cognition, behavior, and mood: clinical findings and structural-functional synergies with cell membrane phospholipids. Altern Med Rev. 2007;12(3):207–27.
Almaas AN, Tamnes CK, Nakstad B, et al. Long-Chain Polyunsaturated Fatty Acids and Cognition in VLBW Infants at 8 years: an RCT. Pediatrics. 2015;135(6):972–80. doi:10.1542/peds.2014-4094.
Raine A, Cheney RA, Ho R, et al. Nutritional supplementation to reduce child aggression: a randomized, stratified, single-blind, factorial trial. J Child Psychol Psychiatry. 2016;57(9):1038–46. doi:https://doi.org/10.1111/jcpp.12565.
Freedman LS, Schatzkin A, Midthune D, Kipnis V. Dealing with dietary measurement error in nutritional cohort studies. J Natl Cancer Inst. 2011;103(14):1086–92. doi:10.1093/jnci/djr189.

Table 1 Characteristics of study population

	Study sample for dietary patterns at one year (N = 1888)	Study sample for dietary patterns at eight years (N = 2326)
Maternal characteristics at enrollment
Educational level (Low), N (%)	548 (29.0%)	758 (32.6%)
Household income per month, N (%)
< 1200 €	187 (9.9%)	221 (9.5%)
1200 – 2200 €	372 (19.7%)	512 (22.0%)
> 2200 €	1329 (70.4%)	1593 (68.5%)
Alcohol use during pregnancy, N (%)
Never	640 (33.9%)	821 (35.3%)
Until pregnancy was known	274 (14.5%)	342 (14.7%)
Continued	974 (51.6%)	1163 (50.0%)
Smoking during pregnancy, N (%)
Never	1484 (78.6%)	1840 (79.1%)
Until pregnancy was known	187 (9.9%)	212 (9.1%)
Continued	215 (11.4%)	274 (11.8%)
Folic acid use, N (%)
No	255 (13.5%)	391 (16.8%)
Started the first 10 weeks	598 (31.7%)	730 (31.4%)
Started periconceptional	1035 (54.8%)	1205 (51.8%)
Psychopathological symptoms, median (IQR)	0.1 (0.1, 0.3)	0.1 (0.1, 0.3)
Diet quality scores during pregnancy (SD)	7.9 (1.5)	7.8 (1.5)

Child characteristics
Age at the neuroimaging assessment, median (IQR), years	9.9 (9.7, 10.0)	9.9 (9.7, 10.2)
Sex (Girls), N (%)	974 (51.6%)	1179 (50.7%)
Ethnic background
Dutch	1346 (71.3%)	1568 (67.4%)
Non-Dutch Western	151 (8.0%)	223 (9.5%)
Non-Dutch non-Western	393 (20.8%)	535 (23.0%)
Diet quality scores, mean (SD)	4.3 (1.4)	4.5 (1.2)

Values are mean (SD) for continuous variables with a normal distribution, medians (IQR) for continuous variables with a skewed distribution, or valid numbers N (%) for categorical variables. Missing data of covariates (Supplemental Table 1) were imputed with multiple imputation (M = 5 imputations).

Table 2 Characteristics of the a priori- and a posteriori-derived dietary patterns, identified among children at age one year (n = 3629) ^{1, 2}

			A posteriori-derived dietary patterns
Food groups ³	Mean intake (g/d)	DQS-1y	Vegetables, potatoes and grains	Snacks, processed foods and sugar	Butter and margarines, whole grains and dairy
Vegetable	56	+	0.79	0.13	0.04
Potatoes	35	+	0.69	0.10	0.07
Pasta, rice and other grains	27	+	0.59	0.16	0.01
Vegetable oils	1	+	0.53	0.35	0.03
Composite meals	22	Not included	0.45	0.00	0.04
Meat	10	+	0.41	-0.04	0.26
Fish and fish products	9	+	0.40	0.12	0.07
Children's meals	70	Not included	-0.38	-0.23	-0.03
Legumes	5	+	0.37	0.21	-0.05
Fruit	152	+	0.27	-0.25	0.20
Savory snacks	3	-	0.11	0.61	0.08
Refined cereals	17	Not included	-0.01	0.58	0.14
Soups and bouillons	17	Not included	0.07	0.52	-0.20
Sugar and confectionery	29	-	0.01	0.50	0.43
Eggs	2	+	0.20	0.44	-0.01
Sauces and condiments	2	Not included	0.28	0.41	0.15
Other fats	2	+	0.24	0.36	0.17
High-fat dairy products	52	+	0.05	0.25	0.09
Non-sugar-containing beverages	69	Not included	0.16	0.22	-0.07
Nuts and seeds	1	Not included	0.07	0.18	0.17
Soy products	5	Not included	0.02	0.06	0.01
Butters and margarines	8	+	0.08	0.06	0.60
Whole grains	56	+	0.16	-0.41	0.53
Formula and breastfeeding	416	+	0.03	0.02	-0.53
Meat products	14	+	0.27	0.09	0.52
Low-fat dairy products	166	+	0.05	0.06	0.51
Sugar-containing beverages	194	-	-0.07	0.23	0.50

Eigenvalues			4.0	1.9	1.6
Variance explained, % ⁴			10.9	9.1	8.1

¹ Factor loadings with and absolute value ≥ 0.20 are shown in bold.

² Total sample with FFQ data at one year.

³Included food items in the food groups are presented in Supplemental Table 2.

⁴ Total explained variance by all three a postereri derived patterns is 28.1%.

Table 3 Characteristics of the a priori- and a posteriori-derived dietary patterns, identified among children at age eight years (n = 4733)^1,2

			A posteriori-derived dietary patterns
Food groups ³	Mean intake (g/d)	DQS-8y	Snacks, potatoes and processed food	Fish, vegetables and fruit	Whole grains, soft fats and dairy	Meat replacement, legumes and nuts
Snacks and fast food	21	Not included	0.61	-0.07	-0.14	-0.02
Potatoes	30	Not included	0.60	0.08	0.14	0.11
Sauces	5	Not included	0.53	-0.06	-0.04	-0.04
Composite and ready-to-eat meals	66	Not included	0.48	0.14	0.20	0.15
Refined cereals and grains	47	Not included	0.45	0.14	-0.39	0.01
Processed meat	42	-	0.45	0.12	0.25	-0.43
Sugar and confectionery products	216	-	0.43	-0.12	0.03	-0.04
Sugar-containing beverages	186	-	0.35	-0.03	-0.15	0.01
Low- and moderate-fat fish	5	+	-0.05	0.61	-0.11	-0.04
Vegetables	92	+	0.15	0.55	0.34	0.13
Fat fish	10	+	0.08	0.50	-0.06	-0.04
Fruit	121	+	-0.14	0.45	0.28	0.02
Low-sugar-containing beverages	314	Not included	-0.03	0.41	0.02	0.06
Hard fats	1	Not included	-0.14	0.35	-0.12	0.25
White meat, unprocessed	14	Not included	0.24	0.32	-0.16	-0.32
Eggs	10	Not included	0.21	0.23	-0.18	0.15
Whole grains	98	+	-0.13	0.11	0.76	0.13
Soft fats	12	+	0.15	-0.19	0.62	-0.15
Low-fat dairy products	201	+	-0.03	0.04	0.35	-0.32
Meat replacement products	4	Not included	-0.05	0.03	0.19	0.59
Legumes	9	+	0.33	0.19	-0.02	0.37
Red meat, unprocessed	10	Not included	0.31	0.22	0.00	-0.37
Nuts and nut butters	6	+	0.08	0.01	0.20	0.36
High-fat dairy products	35	+	0.05	0.15	-0.07	0.32
Soy drinks	9	Not included	0.00	0.02	-0.06	0.23
Porridge	2	Not included	0.04	0.00	-0.09	0.14

Eigenvalues			2.5	1.9	1.7	1.4
Variance explained, % ⁴			9.2	7.0	6.9	5.8

¹ Factor loadings with and absolute value ≥ 0.20 are shown in bold.

² Total sample with FFQ data at eight years.

³Included food items in the food groups are presented in Supplemental Table 3.

⁴ Total explained variance by all four a postereri derived patterns is 28.8%.

Table 4 Association of dietary patterns at ages one and eight years with global brain volumes at age 10 years ¹

		Model 1			Model 2
	B	95% CI	p value	B	95% CI	p value
Total brain volume (cm³)
Diet at one year
Principal components
Vegetables, potatoes and grains	0.24	-4.23, 4.71	0.9164	-2.24	-6.81, 2.33	0.3358
Snacks, processed foods and sugar	-17.03	-21.85, -12.22	0.0000	-7.14	-12.79, -1.49	0.0134
Butter and margarines, whole grains and dairy	10.37	5.92, 14.81	0.0000	4.35	-0.15, 8.84	0.0580
DQS-1y	7.26	3.04, 11.47	0.0007	1.52	-2.73, 5.77	0.4836

Diet at eight years
Principal components
Snacks, potatoes and processed foods	-11.62	-15.46, -7.77	0.0000	-7.63	-13.19, -2.06	0.0073
Fish, vegetables and fruit	0.20	-3.67, 4.07	0.9187	-1.67	-5.61, 2.27	0.4048
Whole grains, soft fats and dairy	18.13	14.29, 21.97	0.0000	8.93	4.54, 13.32	0.0001^*
Meat replacement, legumes and nuts	-3.74	-7.68, 0.21	0.0634	-3.51	-7.38, 0.37	0.0764
DQS-8y	8.75	4.97, 12.53	0.0000	3.64	-0.45, 7.74	0.0812

Cerebral white matter volume (cm³)
Diet at one year
Principal components
Vegetables, potatoes and grains	0.37	-1.67, 2.4	0.7232	-0.64	-2.76, 1.48	0.5557
Snacks, processed foods and sugar	-7.03	-9.22, -4.83	0.0000	-4.30	-6.92, -1.68	0.0013^*
Butter and margarines, whole grains and dairy	3.41	1.38, 5.44	0.001	1.22	-0.86, 3.31	0.2501
DQS-1y	3.18	1.26, 5.09	0.0012	1.10	-0.87, 3.07	0.2743

Diet at eight years
Principal components
Snacks, potatoes and processed foods	-4.15	-5.9, -2.4	0.0000	-2.53	-5.11, 0.06	0.0553
Fish, vegetables and fruit	0.47	-1.28, 2.23	0.5979	-0.14	-1.97, 1.69	0.8800
Whole grains, soft fats and dairy	6.02	4.26, 7.77	0.0000	2.95	0.91, 4.99	0.0046
Meat replacement, legumes and nuts	-1.85	-3.64, -0.06	0.0427	-1.83	-3.63, -0.03	0.0463
DQS-8y	3.02	1.3, 4.74	0.0006	1.32	-0.59, 3.22	0.1754

Cerebral gray matter volume (cm³)
Diet at one year
Principal components
Vegetables, potatoes and grains	-0.1	-2.46, 2.26	0.9351	-1.35	-3.76, 1.06	0.2706
Snacks, processed foods and sugar	-8.52	-11.06, -5.97	0.0000	-2.52	-5.50, 0.46	0.0974
Butter and margarines, whole grains and dairy	5.83	3.48, 8.17	0.0000	2.57	0.20, 4.94	0.0334
DQS-1y	3.36	1.14, 5.59	0.0031	0.27	-1.96, 2.51	0.8113

Diet at eight years
Principal components
Snacks, potatoes and processed foods	-6.05	-8.08, -4.02	0.0000	-4.07	-7.00, -1.14	0.0065
Fish, vegetables and fruit	-0.38	-2.42, 1.67	0.7164	-1.50	-3.58, 0.58	0.1569
Whole grains, soft fats and dairy	10.34	8.32, 12.36	0.0000	5.17	2.86, 7.48	0.0000^*
Meat replacement, legumes and nuts	-1.25	-3.33, 0.84	0.2404	-1.07	-3.11, 0.97	0.3040
DQS-8y	5.19	3.19, 7.19	0.0000	2.30	0.15, 4.45	0.0358

¹ The effect estimates represent the difference in cubic centimeters for brain volumes per 1 SD higher score on the dietary pattern. Model 1 was adjusted for child sex and age when brain imaging was assessed. Model 2 was additionally adjusted for maternal education, household income, child ethnic background, child energy intake, child BMI measured at the age of 10 years, maternal diet quality during pregnancy, smoking during pregnancy, alcohol use during pregnancy, folic acid use, and maternal psychopathological symptoms during pregnancy. Statistical significance (p < 0.05) in model 2 is indicated with bold.

* denotes the associations which remained statistically significant after the Benjamini Hochberg correction for multiple testing (45 tests) with a FDR ≤ 0.05.

Table 5 Association of dietary patterns at ages one and eight years with regional brain volumes at age 10 years ¹

		Model 1			Model 2
	B	95% CI	p value	B	95% CI	p value
Hippocampal volume (mm³)
Diet at one year
Principal components
Vegetables, potatoes and grains	19.86	-11.86, 51.58	0.2196	10.27	-18.63, 39.17	0.4861
Snacks, processed foods and sugar	-63.86	-98.36, -29.37	0.0003	6.82	-29.15, 42.79	0.7100
Butter and margarines, whole grains and dairy	59.37	27.77, 90.97	0.0002	21.55	-6.92, 50.02	0.1378
DQS-1y	46.69	16.78, 76.6	0.0022	9.75	-17.16, 36.66	0.4774

Diet at eight years
Principal components
Snacks, potatoes and processed foods	-56.85	-85.27, -28.44	0.0001	-14.82	-51.65, 22	0.4298
Fish, vegetables and fruit	3.92	-24.58, 32.42	0.7875	-10.46	-36.29, 15.37	0.4271
Whole grains, soft fats and dairy	69.85	41.25, 98.46	0.0000	-17.52	-46.61, 11.58	0.2378
Meat replacement, legumes and nuts	-2.66	-31.73, 26.41	0.8575	14.17	-11.3, 39.63	0.2755
DQS-8y	27.6	-0.36, 55.56	0.0530	-27.24	-54.48, 0.00	0.0500

Amygdala volume (mm³)
Diet at one year
Principal components
Vegetables, potatoes, and grains	12.13	-4.62, 28.88	0.1558	4.20	-11.08, 19.48	0.5897
Snacks, processed foods and sugar	-34.01	-52.23, -15.79	0.0003	0.68	-18.12, 19.48	0.9435
Butter and margarines, whole grains and dairy	33.1	16.41, 49.78	0.0001	7.69	-7.33, 22.7	0.3154
DQS-1y	29.6	13.82, 45.38	0.0002	8.04	-6.15, 22.22	0.2666

Diet at eight years
Principal components
Snacks, potatoes and processed foods	-27.37	-41.74, -13.01	0.0002	-11.21	-29.65, 7.23	0.2334
Fish, vegetables and fruit	2.96	-11.45, 17.36	0.6871	-4.93	-17.99, 8.13	0.4592
Whole grains, soft fats and dairy	43.46	29.04, 57.89	0.0000	2.11	-12.58, 16.79	0.7785
Meat replacement, legumes and nuts	-9.25	-23.94, 5.44	0.2170	1.07	-11.85, 13.98	0.8712
DQS-8y	16.92	2.79, 31.04	0.0189	-8.49	-22.09, 5.11	0.2210

¹ The effect estimates represent the difference in cubic centimeters for brain volumes per 1 SD higher score on the dietary pattern. Model 1 was adjusted for child sex and age when brain imaging was assessed. Model 2 was additionally adjusted for maternal education, household income, child ethnic background, child energy intake, child BMI measured at the age of 10 years, maternal diet quality during pregnancy, smoking during pregnancy, alcohol use during pregnancy, folic acid use, maternal psychopathological symptoms during pregnancy and child intracranial volume.

SupplementalMaterialsDietBrian3mayfinal.pdf

Download PDF

Journal Publication

published 08 May, 2023

Read the published version in European Journal of Epidemiology →

Reviewers agreed at journal
07 Jun, 2022
Reviewers invited by journal
05 Jun, 2022
Editor invited by journal
11 May, 2022
Editor assigned by journal
05 May, 2022
First submitted to journal
04 May, 2022

You are reading this latest preprint version

Associations of dietary patterns with brain morphology in children: Results from a prospective population-based study

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Methods

Study design and population

Dietary assessment

Dietary patterns

Brain morphometry assessments

Covariates

Statistical analysis

Results

Population characteristics

Dietary patterns

Associations of a priori and a posteriori dietary patterns with global and subcortical brain morphometry

Associations of a priori and a posteriori dietary patterns with surface-based brain morphometry

Additional analyses

Discussion

Conclusion

Declarations

Acknowledgements

Funding

Competing interests

Author contributions

Ethics approval

Consent to participate

References

Tables

Supplementary Files

Status:

Journal Publication

Version 1