Exploring correlations between neuropsychological measures and domain-specific consistency in associations with n-3 LCPUFA status in 8-9 year-old boys and girls.

Long-chain n-3 polyunsaturated fatty acids (n-3 LCPUFA) have in some studies been associated with cognitive and socioemotional outcomes in children, but results are inconsistent possibly due to the use of different tests and potential gender-specific effects. The objective of this cross-sectional study was to explore overall patterns in neuropsychological scores as well as correlations between scores within specific domains, and to examine potential gender differences and consistency in associations with n-3 LCPUFA status. In 199 Danish 8-9 year-old children, we performed a large battery of tests and questionnaires on attention, processing speed, executive functions, memory, and socioemotional traits, and measured erythrocyte fatty acid composition. Principal component analyses (PCA) showed that most of the variation in both cognitive performance and socioemotional traits was explained by overall performance, followed by speed-accuracy trade off and externalizing vs. internalizing problems, respectively. Boys had higher speed, lower attention and higher externalizing problem scores than girls. Measures of performance within both processing speed and attention domains correlated moderately, whereas no correlations were found for measures of executive functions apart from some weak correlations for impulsivity. Parent-rated scores for both externalizing and internalizing problems correlated strongly, whereas correlations with child-rated scores were weak. Scores within specific domains did not consistently associate with n-3 LCPUFA, except for processing speed measures which all pointed to faster processing with increased n-3 LCPUFA status. Gender differences in the associations were observed for attention and impulsivity. Child- but not parent-rated internalizing and social problems tended to associate directly with n-3 LCPUFA, supported by increased internalizing problems measured by the PCA component. In conclusion, measures of speed and attention seem to represent these domains in general, whereas single measures of more complex cognitive functions should be interpreted with caution. One approach could be to use multiple tests and create multivariate scores to guide interpretations. Furthermore, the results indicate a need to consider both parent- and child-rated socioemotional scores and gender differences in neuropsychological functions e.g. in investigations of n-3 LCPUFA effects.