Examining the neurostructural architecture of intelligence: The Lothian Birth Cohort 1936 Study

Examining underlying neurostructural correlates of specific cognitive abilities is practically and theoretically complicated by the existence of the positive manifold (all cognitive tests positively correlate): if a brain structure is associated with a cognitive task, how much of this is uniquely related to the cognitive domain, and how much is due to covariance with all other tests across domains (referred to as general cognitive functioning, general intelligence, or g)? We quantitatively address this question by examining associations between brain structural MRI measures (global tissue volumes, white matter hyperintensities, white matter diffusion measures, vertex-wise cortical volume) with g and cognitive domains (processing speed, crystallised ability, memory, visuospatial ability) using 13 cognitive tests in 697 participants from the Lothian Birth Cohort 1936 study (mean age 72.5 years; SD=0.7). There were a range of FDR-corrected cognitive domain-specific significant associations between greater scores in specific cognitive domains and global MRI brain structural measures indicating better brain health (r range 0.008 to 0.269, p<.05), and regions implicated by vertex-wise regional cortical volume associations were consistent with previous research, including a widespread number of medial and lateral areas of the frontal, temporal and parietal lobes. However, at both global and regional level, much of the domain-MRI associations were accounted for by g. Having removed g-related variance, domain-specific association magnitudes with global brain MRI measures were attenuated by 27.9-59.7% (M = 46.2%) across cognitive domains, with only processing speed retaining all significant associations. At the regional cortical level, g appeared to account for the majority (range 22.1-88.4%; M = 52.8% across cognitive domains) of regional domain-specific associations: crystallised and memory domains had almost no unique cortical correlates, whereas processing speed maintained limited associations unique from g in medial temporal and occipital regions, and anterior regions of the left lateral temporal lobe. Visuospatial ability also retained some associations in lateral temporal, parietal and anterior cingulate cortex. The greatest spatial overlaps across cognitive domains (as denoted by g) were present in the medial and lateral temporal, lateral parietal and right lateral frontal areas. The majority of brain structural correlates of domain-level cognitive functions are shared, and we characterise the neurostructural substrates common to multiple types of complex cognition. Truly domain-specific correlates were relatively weak and sparse, with the possible exception of the domains of visuospatial ability and processing speed.