BACKGROUND: While diverse dietary patterns are emerging from nutritional research, comprehensive comparisons of their effects on brain health are lacking. Furthermore, the underlying biological mechanisms through which diet influences neurological and psychiatric outcomes remain largely unclear.
METHODS: Using data from the UK Biobank, a large-scale prospective cohort (with participant enrolled between April, 2009, and June, 2012), we used Cox proportional hazards models to analyse the associations between ten dietary patterns (including the Mediterranean diet, the Dietary Approaches to Stop Hypertension (DASH) diet, and the Mediterranean-DASH Intervention for Neurodegenerative Delay (MIND) diet, the alternative healthy eating index (AHEI), the healthful plant-based dietary index (hPDI), the planetary health diet index (PHDI), the healthy low-carbohydrate diet (HLCD), the healthy low-fat diet (HLFD), the dietary index for gut microbiota (DI-GM), and the reversed energy-adjusted diet inflammatory index (rEDII)) and ultra-processed food (UPF) intake with the incident risk of five major brain disorders, including dementia, stroke, depression, anxiety, and Parkinson's disease (PD) in participants with available dietary data who were free of major brain disorders at baseline. The most neuroprotective diet identified was then validated for its associations with a range of brain health outcomes in the U.S. Health and Retirement Study (HRS, with participant data collected between 2013 and 2014) and the Chinese Longitudinal Healthy Longevity Survey (CLHLS, with participant data collected in 2018) using generalized linear models to analyze the data. Participants from these two cohorts were required to have complete data on diet, outcomes, and covariates. Finally, we employed a four-way decomposition model with multi-omics data as mediators to explore the underlying mechanisms.
FINDINGS: In the UK Biobank cohort (N = 166,916), MIND diet demonstrated the broadest neuroprotective effects. Over a median follow-up of 10·5 years, 1870 participants were diagnosed with dementia, 980 with PD, 3121 with stroke, 5113 with depression, and 5697 with anxiety. Adherence to the MIND diet was significantly associated with a reduced risk of dementia (hazard ratio [HR] = 0·87, 95% confidence interval [CI] = 0·77-0·98), stroke (HR = 0·89, 95% CI = 0·81-0·98), depression (HR = 0·77, 95% CI = 0·71-0·82), and anxiety (HR = 0·82, 95% CI = 0·76-0·88), but not PD (HR = 0·94, 95% CI = 0·80-1·11). In the HRS (n = 4496) and CLHLS (n = 9099) validation cohorts, the MIND diet remained significantly associated with lower risks of depression and anxiety, as well as better cognitive performance (in HRS: better cognitive function (β = 0·78, 95% CI = 0·47-1·10), fewer depressive symptoms (β = -0·22, 95% CI = -0·39 to -0·06), and fewer anxiety symptoms (β = -0·40, 95% CI = -0·78 to -0·02); in CLHLS: lower odds of cognitive impairment (odds ratio [OR] = 0·45, 95% CI = 0·37-0·55), depression (OR = 0·44, 95% CI = 0·38-0·52), and anxiety (OR = 0·49, 95% CI = 0·41-0·60)). In contrast, the intake of UPF was associated with an increased risk for dementia (HR = 1·40, 95% CI = 1·25-1·56), PD (HR = 1·26, 95% CI = 1·08-1·47), depression (HR = 1·42, 95% CI = 1·34-1·50), and anxiety (HR = 1·26, 95% CI = 1·19-1.33) in the UK Biobank cohort. Mechanistically, multi-omics analyses revealed that the MIND diet's protective effects were mediated through several key pathways. A favourable metabolic signature explained a substantial proportion of the reduced risk for stroke (proportion mediated = 60·63%), depression (38·97%), and anxiety (26·06%). Furthermore, slower biological ageing significantly mediated the reduced risk of dementia (19·40%). A structural equation model confirmed the overall protective pathway linking the MIND diet to better brain health via these mediators. Conversely, higher UPF intake was found to increase brain disease risk through detrimental changes in these same metabolic, and ageing pathways.
INTERPRETATION: This study supports the MIND diet as a promising intervention for promoting brain health and mitigating neuropsychiatric disorders, while highlighting UPF intake as a significant modifiable risk factor. By identifying specific metabolic, inflammatory, and ageing-related biological pathways, our findings provide potential targets for future research and targeted interventions. However, these observational associations require definitive confirmation through further experimental studies, such as randomised controlled trials.
FUNDING: National Natural Science Foundation of China and the Fujian Provincial Science and Technology Plan Project.
