Abstract
BACKGROUND: Late-life depression (LLD) predisposes individuals to cognitive decline, often leading to misdiagnoses as mild cognitive impairment (MCI). Voxel-based morphometry (VBM) can distinguish the profiles of these disorders according to gray matter (GM) volumes. We integrated findings from previous VBM studies for comparative analysis and extended the research into molecular profiles to facilitate inspection and intervention. METHODS: We comprehensively searched PubMed and Web of Science for VBM studies that compared LLD and MCI cases with matched healthy controls (HCs) from inception to 31st December 2023. We included 13 studies on LLD (414 LLDs, 350 HCs) and 50 on MCI (1878 MCIs, 2046 HCs). Seed-based d Mapping with Permutation of Subject Images (SDM-PSI) was used for voxel-based meta-analysis to assess GM atrophy, spatially correlated with neuropsychological profiles. We then used multimodal and linear-model analysis to assess the similarities and differences in GM volumetric changing patterns. Partial least squares (PLS) regression and gene enrichment were employed for transcription-neuroimaging associations. RESULTS: GM volumes in the left hippocampus and right parahippocampal gyrus are more affected in MCI, along with memory impairment. MCI was spatially correlated with a more extensive reduction in the levels of neurotransmitters and a severe downregulation of genes related to cellular potassium ion transport and metal ion transmembrane transporter activity. CONCLUSION: Compared to LLD, MCI exhibited more GM atrophy in the hippocampus and parahippocampal gyrus and lower gene expression of ion transmembrane transport. Our findings provided imaging-transcriptomic-genetic integrative profiles for differential diagnosis and precise intervention between LLD and MCI.