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

Both general fluid intelligence (g(f)) and performance on some ‘frontal tests’ of cognition decline with age. Both types of ability are at least partially dependent on the integrity of the frontal lobes, which also deteriorate with age. Overlap between these two methods of assessing complex cognition in older age remains unclear. Such overlap could be investigated using inter-test correlations alone, as in previous studies, but this would be enhanced by ascertaining whether frontal test performance and g(f) share neurobiological variance. To this end, we examined relationships between g(f) and 6 frontal tests (Tower, Self-Ordered Pointing, Simon, Moral Dilemmas, Reversal Learning and Faux Pas tests) in 90 healthy males, aged ~ 73 years. We interpreted their correlational structure using principal component analysis, and in relation to MRI-derived regional frontal lobe volumes (relative to maximal healthy brain size). g(f) correlated significantly and positively (.24 ≤ r ≤ .53) with the majority of frontal test scores. Some frontal test scores also exhibited shared variance after controlling for g(f). Principal component analysis of test scores identified units of g(f)-common and g(f)-independent variance. The former was associated with variance in the left dorsolateral (DL) and anterior cingulate (AC) regions, and the latter with variance in the right DL and AC regions. Thus, we identify two biologically-meaningful components of variance in complex cognitive performance in older age and suggest that age-related changes to DL and AC have the greatest cognitive impact.