Grey matter morphological changes in late life using magnetic resonance imaging technique

This PhD thesis examines grey matter (GM) morphological changes in an elderly non‐demented population, and is divided into three broad topic areas.While a few structural MRI studies have examined the relationship between cognitive function and brain morphology using regions of interest methodology, whole brain analyses have been generally lacking. The first topic is to examine the voxel‐wise correlation between GM volume and four cognitive domains. The findings are largely consistent with the conventional understanding of neuroanatomical basis of cognition. However, the lack of hemispheric predominance in these GM correlates, and the extensively positive correlation between GM volume and cognitive performance, perhaps reflects the characteristics of the ageing brain.The basal forebrain area (BFA) is closely connected to the hippocampus by virtue of cholinergic neuronal projections. Structural neuroimaging studies have shown reduced volumes of both structures in mild cognitive impairment (MCI); however, of interest is the understanding of how the two structures work together to influence the diagnosis of MCI and its two subtypes. The results suggest that atrophy of both brain regions is associated with increased likelihood of MCI and its two subtypes. However, when using them together as predictors, only hippocampal atrophy remains significant. The findings raise the possibility that BFA atrophy has a causal effect on MCI, which is mediated via its influence on hippocampal atrophy.MCI is a heterogeneous neurocognitive disorder that can be classified into various subtypes. However, most of these neuroanatomical studies on GM changes in MCI have focused on the amnestic MCI subtype. The third research topic aims to examine the grey matter (GM) atrophy patterns of MCI subtypes in comparison with a cognitively healthy group. The neuroanatomical hallmark of amnestic MCI, medial temporal lobe atrophy, is confirmed in this study. Diffuse GM loss was also found in the GM atrophy patterns for non‐amnestic MCI subtypes. The severity of GM loss in multi‐domain impaired subtypes was greater than in the single‐domain impaired subtypes. The findings provide further evidence for the neuroanatomical biomarkers of MCI subtype, and could potentially assist clinicians to improve diagnosis of MCI