Brain Volume Reductions and Relationship With Depression and Cognitive Functioning in Suicide in Older Adults: A Cross-Sectional and Longitudinal Study Using Bayesian Multilevel Modeling

Background

In older adults, depression with cognitive impairment may be a harbinger of early dementia and a risk factor for suicidal behavior. However, the neuroanatomical correlates of these impairments are unknown, particularly when examined longitudinally.

Methods

Older adults (N = 153, mean [SD] age = 62.8 [7.7] years; 86/67 [56%/44%] female/male) with either a history of suicide attempts (n = 46), depression with no history of suicide attempts (n = 72), or no psychiatric history (n = 35) completed T1 structural magnetic resonance imaging scans. Of these participants, 51 had repeated scans (days between scans mean [SD] = 415 [252], range = 90–1091). Bayesian multilevel modeling with false discovery rate correction tested cross-sectional group differences and prospective changes in brain volumes as a function of suicide attempt history, depression severity, and cognitive functioning.

Results

Three broad categories of individual differences emerged: 1) a history of attempted suicide and impaired executive functioning related to and predicted volumes in temporal areas, neighboring parietal and occipital regions, and the hippocampus; 2) frontal and subcortical volume reductions related to depression, particularly current episode severity; and 3) cognitive impairment typical of cortical dementia predicted medial temporal volume reduction.

Conclusions

A history of suicidal behavior, depression, and dementia-related cognitive decline are accompanied by reduced brain volumes in largely non-overlapping regions that nevertheless converge on the hippocampus. The hippocampus may be a nexus where independent changes associated with depression and suicide diathesis factors co-occur with dementia-related neurodegeneration. These results support distinct neurocognitive deficits in late-life suicide in older adults. A better understanding of hippocampal structure and function in people at risk of suicide will advance both risk prediction and treatment development.