Healthy Dietary Patterns Associated with Lower Risk of Subjective Cognitive Decline

A large prospective cohort study revealed that a higher adherence to each of six widely studied healthy dietary patterns was independently associated with a lower risk of subjective cognitive decline (SCD). Among the six patterns evaluated, the Dietary Approaches to Stop Hypertension (DASH) diet demonstrated the most pronounced association, with a 41% lower risk of SCD at the 90th versus 10th percentile of adherence.¹

Published February 23, 2026, in JAMA Neurology, the convergence of inverse associations across all 6 indices — spanning patterns with distinct nutritional rationales — suggested that broad dietary quality, rather than any single nutrient or framework, may be the operative variable in preserving cognitive health during midlife and beyond.¹

Study Overview

Senior author , assistant professor of Epidemiology and Nutrition in the Departments of Epidemiology and Nutrition at the Harvard T.H. Chan School of Public Health, and his colleagues conducted a prospective cohort study pooling data from the Nurses' Health Study (NHS, 1986–2014), NHSII (1991–2017), and the Health Professionals Follow-Up Study (HPFS, 1986–2012). The analytic sample comprised 159,347 adults (mean age 44.3 [SD 9.3] years; 82.6% female), with data analyses performed from September 2024 to November 2025.¹

Dietary patterns were evaluated using validated food-frequency questionnaires administered repeatedly over follow-up. All told, 6 indices were scored: the Alternate Healthy Eating Index 2010 (AHEI-2010), the DASH diet score, the Healthful Plant-Based Diet Index (hPDI), the Planetary Health Diet Index (PHDI), and reversed empirical dietary indices for hyperinsulinemia (rEDIH) and inflammatory potential (rEDIP).

Subjective cognitive decline was assessed using seven validated questions measuring self-perceived changes in memory and cognitive function. While SCD lacks diagnostic specificity for any single neurodegenerative etiology, prior work in the NHS cohorts had demonstrated that SCD items correlate with objective neuropsychological performance and apolipoprotein E ε4 genotype status, supporting its utility as an early-stage observational endpoint. Objective cognitive function was additionally measured by telephone-administered neuropsychological battery in the NHS subsample.

Study Findings

The full range of SCD associations across all six patterns is presented here in descending magnitude: DASH diet (risk ratio [RR] comparing 90th vs 10th percentile adherence: 0.59; 95% CI, 0.57–0.62), hPDI (RR, 0.76; 95% CI, 0.65–0.85), rEDIH (RR, 0.76; 95% CI, 0.73–0.80), PHDI (RR, 0.80; 95% CI, 0.75–0.86), AHEI-2010 (RR, 0.84; 95% CI, 0.80–0.89), and rEDIP (RR, 0.89; 95% CI, 0.85–0.93).

On objective cognitive assessment, the DASH diet again showed the strongest association, with a mean global cognition z-score difference of 0.05 (95% CI, 0.02–0.09) comparing the 90th to 10th adherence percentile. This effect size, while statistically significant, is modest in absolute terms and should be interpreted within the context of a long-duration cohort measuring population-level trends rather than individual clinical outcomes.

Across patterns, an age-sensitive signal emerged: adherence assessed during ages 45–54 years showed the strongest association with subsequent SCD risk reduction, suggesting midadulthood as a potentially important window for dietary influence on cognitive aging trajectories.¹ At the food-group level, higher vegetable and fish intake were the components most consistently associated with better outcomes, while higher consumption of red and processed meats was associated with worse outcomes — findings shared across several of the dietary frameworks.

Although these patterns reflected different nutritional priorities, the AHEI-2010 broadly captured dietary quality, hPDI emphasizes plant-forward eating, PHDI incorporates planetary sustainability, and the empirical indices specifically target insulin and inflammatory pathways. Ultimately, all six demonstrated statistically significant, dose-dependent inverse associations with SCD risk.¹

Expert Perspective: Interpreting the Evidence

The cross-pattern consistency of inverse associations arguably warrants as much clinical attention as the DASH diet's relative superiority. A finding that holds across plant-based, anti-inflammatory, anti-hyperinsulinemic, cardiovascular, and sustainability-oriented dietary models implies that the cognitive benefits of healthy eating are not contingent on adherence to any specific framework, which may have practical implications for dietary counseling in heterogeneous patient populations.

External commentators have noted that this observational evidence, however robust in scale and duration, does not establish causality. Individuals who maintain high dietary quality over decades may differ systematically in physical activity, sleep quality, educational attainment, and social engagement­–confounders that are difficult to fully disentangle even with multivariable adjustment. The study population, comprising predominantly White health professionals, also limits direct extrapolation to broader demographic groups.

Limitations and Unanswered Questions

Several important methodological constraints bear consideration. The cohort was composed primarily of health professionals — nurses and male health professionals — who likely have higher health literacy, socioeconomic resources, and health-seeking behaviors than the general population, potentially overstating real-world effect sizes. More than 80% of participants were female, limiting extrapolation to men.

Dietary assessment relied on self-reported food frequency questionnaires subject to recall bias and measurement error; while repeated assessment over decades is a strength, systematic misclassification cannot be excluded. SCD is a subjective, non-specific endpoint with limited individual-level diagnostic utility, and its correlation with eventual dementia diagnosis varies across populations. The objectively measured cognition endpoint was telephone-administered and available only in the NHS subsample, limiting statistical power for that outcome.

Additionally, the study cannot determine whether the observed associations reflect the influence of specific dietary constituents, overall lifestyle patterns correlated with healthy eating, or both. Randomized dietary intervention trials using incident dementia, neuroimaging, or cerebrospinal fluid biomarker endpoints will be necessary to test whether changing dietary patterns in midlife causally modifies cognitive aging trajectories. The null primary endpoint in the 2023 MIND RCT remains a salient reference point for managing expectations in this space.

Clinical Context: The Growing Dementia Burden

The 2024 Lancet Commission on Dementia Prevention identified 14 potentially modifiable risk factors — including physical inactivity, obesity, diabetes, hypertension, and dietary quality — as collectively accounting for approximately 45% of dementia cases globally.3 The current study addresses a recognized gap: while individual dietary patterns have been examined separately in prior literature, few studies have evaluated multiple frameworks head-to-head in the same cohort using the same methodology and endpoints.

Subjective cognitive decline has gained traction as a research construct for identifying individuals in the preclinical phase of neurodegenerative disease. Population-level studies suggest that sustained self-perceived cognitive decline correlates with subsequent objective impairment and dementia risk, making it a pragmatic endpoint for observational studies spanning decades — particularly where randomized controlled trials with dementia incidence endpoints would require prohibitively long follow-up.²

REFERENCES
1. Chen H, Cortese M, Flores-Torres MH, et al. Dietary Patterns and Indicators of Cognitive Function. JAMA Neurol. Published online February 23, 2026. doi:10.1001/jamaneurol.2026.0062
2. Livingston G, Huntley J, Liu KY, et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. Lancet. 2024;404(10452):572-628. doi:10.1016/S0140-6736(24)01296-0