Investigating neural correlates of late life depression using large datasets

Poster Session C - Sunday, April 14, 2024, 5:00 – 7:00 pm EDT, Sheraton Hall ABC

Kassandra Hamilton¹, Janine Bijsterbosch¹; ¹Washington University in St. Louis

Recent longitudinal population studies have revealed a relationship between late life depression (LLD) and early visual and somatosensory cortices. Although evidence for early visual and somatosensory involvement in LLD can sporadically be found in the literature, the majority of LLD research focuses on association cortices such as frontal, cingulate, and default mode regions. As such, the role of visual and somatosensory cortices in depression is poorly understood. To address this gap in knowledge, we utilized population data from the UK Biobank (https://www.ukbiobank.ac.uk/) to summarize the structural brain correlates of depression across large-scale brain networks. Using data from 20,843 participants ages 40-69, we conducted linear regression analysis between imaging derived phenotypes (IDPs) from T1 structural MRI data and depression score data collected via questionnaire. We then mapped the results onto a low-dimensional brain parcellation (‘Yeo-7’) that distinguishes between unimodal cortices (i.e. sensorimotor and visual areas) and association cortices (i.e. default mode and frontal networks). Significant neural correlates were found in visual, somatomotor, ventral attention, limbic, and default networks. Brain regions most associated with state depression were found in the left supramarginal gyrus, left superiortemporal area, and left inferiorparietal area. These results suggest the existence of whole-brain neural correlates of LLD. The sensorimotor-association axis may provide a foundation for understanding the spatial distribution of LLD correlates across unimodal and association cortex.

Topic Area: METHODS: Neuroimaging