Structural neuroimaging measures and lifetime depression across levels of phenotyping in UK biobank

作者：M. Harris, S. Cox, L. de Nooij, M. Barbu, M. Adams, Xueyi Shen, I. Deary, S. Lawrie, A. McIntosh, H. Whalley · 发表于：Translational Psychiatry · 年份：2022 · DOI：10.1038/s41398-022-01926-w · 被引用次数：20 · 研究领域：Medicine

Depression is assessed in various ways in research, with large population studies often relying on minimal phenotyping. Genetic results suggest clinical diagnoses and self-report measures of depression show some core similarities, but also important differences. It is not yet clear how neuroimaging associations depend on levels of phenotyping. We studied 39,300 UK Biobank imaging participants (20,701 female; aged 44.6 to 82.3 years, M = 64.1, SD = 7.5) with structural neuroimaging and lifetime depression data. Past depression phenotypes included a single-item self-report measure, an intermediate measure of ‘probable’ lifetime depression, derived from multiple questionnaire items relevant to a history of depression, and a retrospective clinical diagnosis according to DSM-IV criteria. We tested (i) associations between brain structural measures and each depression phenotype, and (ii) effects of phenotype on these associations. Depression-brain structure associations were small (β < 0.1) for all phenotypes, but still significant after FDR correction for many regional metrics. Lifetime depression was consistently associated with reduced white matter integrity across phenotypes. Cortical thickness showed negative associations with Self-reported Depression in particular. Phenotype effects were small across most metrics, but significant for cortical thickness in most regions. We report consistent effects of lifetime depression in brain structural measures, including reduced integrit...