Several studies in developing countries suggest that poor nutrition in infancy is linked with impaired cognitive development¹ but a number of confounding factors could explain these relations (for example, parental education, home environment, or birthweight). In developed countries most studies have focused on the importance of the type of milk and especially breastfeeding in infancy². Whether cognitive development is affected by the composition of the diet babies receive once they start the transition to solid foods is uncertain. The aim of the present study³ was to study relationships between dietary patterns in infancy and cognitive development at age four years.

The Southampton Women’s Survey (SWS)

The SWS started in 1998⁴, with the aim of identifying influences in the fetal and postnatal environment that influence growth, development and health.

The “Infant guidelines” pattern

The children’s diet was assessed at age six and twelve months using a Food Frequency Questionnaire (FFQ)⁶. Dietary patterns at each age were determined using principal components analysis. The first pattern was characterized by a high consumption of “vegetables, fruits, meat and fish, other home-prepared foods and breast milk” and a low consumption of “commercial baby foods and formula milk”. As this pattern conforms to infant feeding guidelines7, we called it the ‘infant guidelines’ pattern.

Neuropsychological assessment

At age four years, 396 mothers and children were invited to take part in a study of cognitive function. The children’s IQ was measured with the Wechsler Pre-School and Primary Scale of Intelligence; visual attention, visuomotor precision, sentence repetition and verbal fluency were measured with the NEPSY (Developmental Neuropsychological Assessment) and visual form-constancy was measured with the Test of Visual Perceptual skills. The quality of the child’s home environment was assessed using a short version of the Home Observation for Measurement of the Environment Scale (HOME-SF)⁵.

Higher intake of fruit, vegetables, meat, fish and home-prepared foods in infancy linked with higher intelligence at age 4 years 

Two hundred and forty one children had complete data on the neuropsychological tests, diet in infancy and parental characteristics. There were no differences in dietary patterns between boys and girls at either six or twelve months of age. Girls had higher scores for full-scale and verbal IQ, visual attention, visuomotor precision and sentence repetition.

In analyses adjusted for sex only, children whose diet in infancy was characterized by high consumption of fruit, vegetables and home-prepared foods (‘infant guidelines’ dietary pattern) had higher full-scale and verbal IQ and better memory performance at age four years. Further adjustment for maternal education, intelligence, social class, quality of the home environment and other potential confounding factors, weakened these associations but the relations between higher ‘infant guidelines’ diet score and full-scale and verbal IQ remained. For a standard deviation, increase in ‘infant guidelines’ diet score at six or twelve months full scale IQ rose by 0.18 (95% CI 0.04 to 0.31) of a standard deviation. For a standard deviation increase in ‘infant guidelines’ diet score at six months verbal IQ rose by 0.14 (0.01 to 0.27) of a standard deviation. There were no associations between dietary patterns in infancy and four-year performance on the other tests. Further adjustment for breastfeeding had no effect on the associations described above.

Even in well-nourished populations, dietary pattern in early life may be important for cognitive development

These results suggest that a dietary pattern characterized by high consumption of vegetables, fruits, meat and fish and other homeprepared food in infancy could have some positive effect on cognitive development. The findings persisted after adjustment for a wide range of potential confounding factors, but we cannot exclude the possibility that they could be explained by other factors in the child’s environment we are unable to measure.