Answers are grounded in this paper and cite their sources.
Background Dietary intake and emotional symptoms are associated in later life, but the direction of their longitudinal associations may vary. We examined food-intake frequency and four emotional symptoms across two successive observation windows in older Chinese adults. Methods This longitudinal cohort study used the 2008, 2011 and 2014 waves of the Chinese Longitudinal Healthy Longevity Survey. Cross-lagged panel networks included 12 food-intake items and four emotional-symptom items. Regularized regressions were fitted for 2008–2011 (N = 6,490) and 2011–2014 (N = 3,968). Paired bootstrap comparisons used 3,796 participants with complete data across all three waves, with false discovery rate correction across 256 paths. We examined network stability, age strata, missing data, complete-case weighting, age/sex adjustment, alternative model specifications and exclusion of the unhappiness item. Results In 2008–2011, more frequent fruit, nut and meat intake was associated with lower subsequent loneliness or uselessness. In 2011–2014, uselessness was associated with less frequent subsequent nut intake (β=−0.051, 95% confidence interval − 0.092 to − 0.025), and uselessness and loneliness had more outgoing connections within the emotional network. Unadjusted paired comparisons suggested changes in nuts→uselessness (Δβ = 0.051), uselessness→nuts (Δβ=−0.059) and loneliness→anxiety (Δβ = 0.056). These cross-lagged differences remained exploratory after correction (q = 0.099–0.228). Eight autoregressive differences survived correction at q < 0.05. The later-window pattern had the same direction in both age strata. Six of 480 age-stratified path-by-window comparisons were nominally significant (approximate P < 0.05), all six in 2008–2011; none survived false discovery rate (FDR) correction (smallest q = 0.83). Core cross-domain path directions were consistent under multiple imputation, inverse probability weighting and exclusion of the unhappiness item. Conclusions Diet–emotion associations showed possible changes across observation windows, alongside stronger autoregressive persistence. Cross-lagged differences were exploratory after multiple-testing correction. Independent samples and designs that distinguish within-person change from between-person differences are needed to test whether these patterns reflect stages of ageing and have implications for intervention. Trial registration Clinical trial number: not applicable.
No takes yet. Share an insight, caveat, or question.
Wen et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: