Dietary factors and MRI biomarkers of brain ageing in general populations: a comprehensive systematic review

This systematic review examined the associations of dietary factors such as nutrients, food intake, dietary patterns and dietary biomarkers with structural and functional brain MRI biomarkers, focusing on macrostructural, microstructural, lesion and perfusion measures, and functional activity/connectivity. Articles published in English were systematically searched in PubMed, Embase and PsycInfo up to 19 July 2024. A total of thirty-eight prospective cohort studies (twenty-three cross-sectional and fifteen longitudinal analyses) and thirteen intervention studies were included. Cross-sectional analyses revealed heterogenous associations: baked fish correlated with larger hippocampal volumes (<i>β</i> = 0·21), while oily fish, dairy products and tofu adversely related to ventricle grade. Pro-inflammatory dietary patterns were positively associated with silent infarct risk (DII Q4 <i>v</i>. Q1, OR = 1·77), whereas anti-inflammatory patterns tended to favour brain preservation. Longitudinal studies demonstrated more consistent protective associations: green tea consumption (+100 mL/d) reduced hippocampal atrophy by 0·024%/year, prudent dietary patterns preserved +203 mm³ left hippocampal volume over 4 years and higher plasma carotenoids decreased medial temporal lobe loss by 0·02 cm³/year. However, null findings were common across multiple dietary factors. Interventions showed limited structural benefits (effective in only two of six studies), while polyphenol-rich supplements more consistently improved cerebral perfusion and functional connectivity. Longitudinal and intervention studies demonstrated more consistent patterns than cross-sectional analyses; however, current evidence remains limited for clinical translation. Findings from cross-sectional analyses, despite being from prospective cohorts, require careful interpretation. Further replication across diverse populations and standardised long-term studies are needed before translating these associations into clinical practice.