Objectives To estimate the relationship between moderate alcohol consumption and brain health, determining the threshold intake for harm and identifying whether population subgroups are at differential risk.

Design Observational cohort study. Alcohol consumption was determined at baseline assessment visit using touchscreen questionnaire (2016-10). Multi-modal MRI brain and cognitive testing were performed subsequently (2014-20). Clinical data was extracted from linked Hospital Episode Statistics.

Setting UK Biobank study. Brain imaging was performed on identical scanners with identical protocols at three UK centres (2014-20).

Participants 25,378 participants (mean age 54.9±7.4 years).

Main outcome measures Brain health as defined by structural and functional MRI brain measures.

Results Alcohol consumption was negatively linearly associated with global brain grey matter volume (beta= -0.1, 95%CI= -0.11 to -0.09, p<2×10−16). The association with alcohol was stronger than other modifiable factor tested and…