Depression and cognition in the elderly : neuroimaging perspective

Abstract

This thesis examines the relationship between depression and brain structure in the elderly with (Study I, III) and without (Study II, IV) cognitive impairment (Alzheimer’s disease and mild cognitive impairment). Individuals from four independent cohorts were included. Participants had either a depressive episode (Study II, III) or depressive symptoms, as measured with different depression scales (Study I, IV).

Studies I and II have cross-sectional design, and studies III and IV are longitudinal. Main outcomes were cortical thickness of the brain and volumes of different structures (hippocampus, ventral diencephalon, including hypothalamus and corpus callosum), or atrophy rate of the thickness and volumes (Study IV). We found in all the cohorts that depressive symptoms were associated with cortical thinning in the same region – the left temporoparietal junction. Depression-related thinning was observed in three cohorts (Studies I, IV) in superior temporal cortex and temporal pole. In two non-demented cohorts (Studies II, IV) angular cortex was also involved in depression. Longitudinal analysis revealed that thinning in these regions is secondary to depressive symptoms (study IV). In two cohorts (Study I, II) fusiform cortex was involved in depression. In study IV, we also were able to assess thinning which developed in parallel with depressive symptoms. It covered medial superior frontal cortex and lingual cortex. The number of depressive episodes was associated with cortical thinning in the left temporal pole in women (Study II) and reduced volume of the right ventral diencephalon in both – men and women (Study III). We have found moderating effect of gender on the relationship between cortical thickness and depression onset. Women with late-onset depression (>65 years) but not men had the widespread thinning in the prefrontal cortex compared to early-onset depressed. The volume of the right hippocampus and thickness of the superior frontal cortex were positively associated with a level of global cognition measured with the mini-mental state examination (MMSE) This effect was more pronounced in the subgroup of late-onset depressed (Study II). The volume of the right ventral diencephalon was associated with cognitive decline (MCI or dementia diagnosis) one year later in the elderly with a depressive episode (study III). Adding baseline MMSE to the classifier increased its accuracy. Total and phosphorylated tau were associated with cortical thinning in the cluster covering right posterior cingulate cortex and precuneus and cluster covering right parahippocampal and fusiform gyri in the AD patients with depressive symptoms from the KI cohort (Study I). No association has been found in non-depressed AD patients. Higher baseline saliva cortisol levels in non-demented individuals (Study IV) were associated with widespread cortical atrophy in temporal, prefrontal and parietal cortex bilaterally and the right hippocampus, independently of age and MMSE.

To sum-up, depression was associated with thinning (Studies I, II) and subsequent atrophy (Study IV) in the superior temporal, supramarginal, temporal pole, lingual, fusiform and parahippocampal cortex. Cortical thinning in the superior frontal and lingual regions developed in parallel or prior to the depressive symptoms. The afore-mentioned regions are involved in social perception (processing of the information about others, experience positive emotions related to other people and building an integrative picture of another person), and are among the first to be impaired in Alzheimer’s disease. Elevated cortisol explained atrophy in these and a number of other regions, including the hippocampus, suggesting that depression and Alzheimer’s disease may be connected via cortisol-related brain damage. Depression-related atrophy in the ventral diencephalon leads to impaired cognitive performance. Assessment of cognitive function during the depressive episode, combined with brain structural measurements may have a prognostic value. Future studies should evaluate if a detailed neurocognitive assessment of elderly patients during the depressive episode would help to identify those at high risk of dementia. It is also important to test if stress-reduction interventions in individuals at-risk of Alzheimer’s disease would be effective in its prevention.