Potential effect of skull thickening on the associations between cognition and brain atrophy in ageing

Background: intracranial volume (ICV) is commonly used as a marker of premorbid brain size in neuroimaging studies as it is thought to remain fixed throughout adulthood. However, inner skull table thickening would encroach on ICV and could mask actual brain atrophy. Objective: we investigated the effect that thickening might have on the associations between brain atrophy and cognition. Methods: the sample comprised 57 non-demented older adults who underwent structural brain MRI at mean age 72.7 +/- 0.7 years and were assessed on cognitive ability at mean age 11 and 73 years. Principal component analysis was used to derive factors of general cognitive ability (g), information processing speed and memory from the recorded cognitive ability data. The total brain tissue volume and ICV with (estimated original ICV) and without (current ICV) adjusting for the effects of inner table skull thickening were measured. General linear modelling was used to test for associations. Results: all cognitive ability variables were significantly (P < 0.01) associated with percentage total brain volume in ICV measured without adjusting for skull thickening (g: eta(2) = 0.177, speed: eta(2) = 0.264 and memory: eta(2) = 0.132). After accounting for skull thickening, only speed was significantly associated with percentage total brain volume in ICV (eta(2) = 0.085, P = 0.034), not g or memory. Conclusions: not accounting for skull thickening when computing ICV can distort the association between brain atrophy and cognitive ability in old age. Larger samples are required to determine the true effect.