Internet use and cognitive trajectories in middle-aged and older people: A longitudinal mixed effects analysis of two...

Cognitive decline is a major public health concern in aging populations. Emerging evidence suggests that Internet engagement may provide mental stimulation and slow cognitive aging, but longitudinal cross-national evidence remains limited.

This study aimed to examine how first time Internet use relates to cognitive trajectories in middle-aged and older people, comparing cognitive performance before and after Internet adoption in two cohorts.

A longitudinal analysis was conducted using data from two cohort studies: the English Longitudinal Study of Aging (ELSA) and the China Health and Retirement Longitudinal Study (CHARLS).

Participants were aged ≥45 years and had repeated cognitive assessments using the Mini-Mental State Examination. Global cognition and episodic memory were the primary outcomes, orientation, attention and calculation were also measured. First time Internet use was a time varying exposure. Linear mixed effects models were used to compare cognitive slopes before versus after Internet adoption within individuals, adjusting for age, education, and other covariates. Beta coefficients (β) with 95% confidence intervals (CI) were reported.

At baseline, Internet users had higher cognitive scores than non-users, particularly for episodic memory in ELSA (β = 0.607, 95% CI 0.528-0.686) and global cognition in CHARLS (β = 0.402, 95% CI 0.365-0.439). After Internet adoption, cognitive trajectories improved mainly in global cognition and episodic memory. In ELSA, post-adoption slopes were more favorable for global cognition (β = 0.110, 95% CI 0.061-0.159) and episodic memory (β = 0.410, 95% CI 0.378-0.442). In CHARLS, episodic memory also improved after adoption (β = 0.044, 95% CI 0.025-0.063), while the post-adoption change in global cognition was not statistically significant (β = 0.013, 95% CI -0.001-0.027). Orientation, attention and calculation showed no consistent post-adoption benefits across cohorts. The protective effect of Internet use was greater in ELSA than in CHARLS, and it varied by age group. Younger participants benefited more in ELSA, whereas in CHARLS the effect was greatest in those aged ≥65 years.

Initiating Internet use in mid-to-later life was associated with a slower decline in cognitive function, especially memory. The effects varied by cognitive domain and age, suggesting that digital engagement may support brain health in aging. Encouraging Internet use among older people could help maintain cognitive function in later life.