Computational protocol: Correlational structure of ‘frontal’ tests and intelligence tests indicates two components with asymmetrical neurostructural correlates in old age

Similar protocols

Protocol publication

[…] Methods used to determine frontal test internal consistency are found in the Supplementary material. Pearson's product–moment correlation tests were performed to examine the relationships between test scores for variables that approximated to a normal distribution. Spearman's rank order correlation tests were used for those that were non-normally distributed. Next, the relationships between frontal tests and general fluid cognitive ability were examined. Tests for significant differences between correlations from the same sample () using the cocor package in R ( were used to test predictions that dorsal frontal tasks (i.e. Simon task, Tower task and SOPT) would correlate with gf more strongly than ventral frontal tasks (Dilemmas task, Faux Pas task and Reversal Learning). In light of the strong relationship between gf and processing speed (r (88) = .78, p < .001), we also sought to examine the differences in their correlations with frontal tasks. These analyses and the method of deriving a factor score for processing speed can be found in the Supplementary material. Though our alpha level of significance was p < .05, we acknowledged the potential for type I error amongst multiple simultaneous comparisons of inter-correlated variables by also considering interpretations at p < .01.Further analyses tested correlations between frontal test scores after partialling out gf to identify unique shared variance between frontal test scores that is not in common with gf. The correlational structure of cognitive performance was further examined using two principal component analyses (PCAs) of frontal test scores, first without and then with the inclusion of gf, followed by varimax rotation in SPSS 19. Finally, scores were derived for each participant from the rotated principal components using the PCA that included gf. These were correlated with frontal lobe regional volumes. To test whether these effects were significantly lateralised, we compared the correlation magnitudes (, as above) of principal components between homotopic frontal regions. […]

Pipeline specifications

Software tools cocor, SPSS
Application Miscellaneous