关键词: 老年人, 外语学习, 执行功能, 认知改善, 元分析

Abstract: Foreign language learning is increasingly regarded as an accessible, non-pharmacological form of cognitive engagement in later life. However, empirical findings concerning its cognitive benefits remain inconsistent. In particular, previous syntheses have mainly focused on the cumulative effects of lifelong bilingual experience, leaving unclear whether foreign language learning initiated in old age can itself improve executive function (EF). The present meta-analysis aimed to quantify the effects of late-life foreign language learning on EF, identify the conditions under which such effects are more likely to emerge, and clarify the mechanisms underlying the observed cognitive gains. Particular attention was given to distinguishing language-specific benefits from the more general effects of cognitive stimulation and to evaluating competing theoretical accounts of both language input and executive-processing outcomes.
To investigate the effects of late-life foreign language learning on EF, 20 studies comprising 63 effect sizes were identified through a rigorous systematic literature search conducted in accordance with the PRISMA 2020 guidelines. A three-level random-effects meta-analytic model was used to account for within-study dependence among multiple effect sizes. Hedges’ g was used as the standardized effect-size index. The included studies comprised randomized controlled trials (RCTs), non-randomized controlled trials, and single-group pre-post designs. Moderator analyses examined intervention characteristics (linguistic distance, learning duration, learning intensity, and control-group type), participant characteristics (age and cognitive status), and outcome characteristics (EF components). Subgroup analyses, linear meta-regression, natural-spline analyses, and multiple meta-regression were conducted where appropriate. In addition, the relative support for the interference inhibition hypothesis and the processing complexity hypothesis was quantified, while the Bilingual Inhibitory Control Advantage (BICA) and Bilingual Executive Processing Advantage (BEPA) hypotheses were evaluated through differences across EF components.
The results showed that late-life foreign language learning yielded a small but significant overall effect on EF (g = 0.37, p < 0.001). When the analysis was restricted to RCTs, the effect remained small but statistically significant (g = 0.29, 95% CI [0.12, 0.46], p < 0.01), suggesting that the positive overall effect was retained under more rigorous research designs. However, the estimated language-specific effect was substantially smaller when comparing language learning against active control groups (g = 0.11, 95% CI [-0.02, 0.23], p = 0.104). Further restricting this active-control comparison to RCTs yielded a non-significant effect of similar magnitude (g = 0.11, 95% CI [-0.03, 0.24], p = 0.119).
Moderation analyses showed that linguistic distance significantly moderated the effect of language learning on EF, with greater distance associated with larger cognitive gains. Exploratory analyses suggested a greater benefit for older adults with cognitive impairment than for cognitively healthy older adults, although this finding is based on only three studies and awaits further confirmation. Notably, when linguistic distance and cognitive status were simultaneously entered into a multiple meta-regression model, cognitive status remained a significant independent moderator, whereas the effect of linguistic distance was no longer significant. Their independent contributions could not be fully separated because all studies involving cognitively impaired participants used typologically distant target languages. Linear meta-regression showed no significant moderating effects of learning duration, learning intensity, or age. No significant moderating effects were found for control-group type or EF component either. Nonlinear analyses, however, revealed a significant moderating effect of learning intensity, whereas the overall nonlinear effect of learning duration alone was not significant. In the joint model, duration and intensity showed relatively independent nonlinear contributions, with the effect of learning intensity being more pronounced.
Furthermore, the processing complexity hypothesis received significantly stronger support than the interference inhibition hypothesis, suggesting a greater role for processing demands in executive engagement. At the outcome level, the absence of significant differences across EF components supported BEPA rather than BICA, indicating broad optimization of executive processing rather than selective enhancement of inhibitory control. Together with the attenuated effect observed in RCTs comparing foreign language learning with active control conditions, these findings suggest that late-life foreign language learning may function primarily as a form of comprehensive cognitive stimulation, with language-specific executive gains being relatively limited.
The study extends previous research in three main respects. First, late-life foreign language learning is distinguished from lifelong bilingual experience, allowing short-term language learning to be evaluated explicitly from an intervention perspective. Second, intervention, participant, and outcome characteristics are integrated within a three-level meta-analytic framework, with cognitive status included as a moderator and nonlinear dose-response patterns further examined. Third, competing theoretical accounts at both the language-input and executive-output levels are quantitatively evaluated, providing a more fine-grained account of the mechanisms underlying cognitive gains. Overall, the findings clarify the magnitude, boundary conditions, and possible mechanisms of EF gains associated with late-life foreign language learning, while indicating that gains specifically attributable to language training itself may be limited. Future research should employ well-matched active control groups, include a wider range of language pairs and participant profiles, and adopt longitudinal designs to determine the long-term maintenance and optimal intensity of late-life foreign language learning interventions.

Key words: older adults, foreign language learning, executive functions, cognitive enhancement, meta-analysis