Neuroimage Reports · Nov 9, 2026
Association of bilingual language use with brain age: MRI evidence from bilinguals, translators, and interpreters
By Chen M, Korenar M, Schrantee A et al.
Background Bilingualism has been linked to cognitive reserve and later onset of Mild Cognitive Impairment, but its neuroanatomical correlates remain unclear. We examined whether structural MRI-derived brain age varies across bilingual groups with different professional language experience. Objective To test whether brain age differs among non-professional bilinguals, translators, and simultaneous interpreters. Methods We included 105 Czech–English bilingual adults (44 non-professional bilinguals, 35 translators, 26 simultaneous interpreters). Brain age was estimated using six pretrained models. The primary outcome was the age-corrected brain-age gap (BAGcorr). Group differences and age × group interactions were tested across models with FDR correction. Prediction accuracy and residual age dependence were also characterised. The model showing the largest standardised BAGcorr group effect was taken forward for exploratory post hoc age-slope and regional-volume analyses. Results No group difference in BAGcorr survived FDR correction across the six models. No age × group interaction survived correction, although nominal interactions were observed for BrainAge and DeepBrainNet. Model-quality criteria ranked the models differently, with no single model consistently preferred. In BrainAge, the sex- and ICV-adjusted BAGcorr difference between interpreters and non-professional bilinguals was −1.61 years (95% CI: −3.66 to 0.44). Exploratory age-slope and regional-volume associations were model-specific and did not provide robust evidence of group differences. Conclusions Professional bilingual experience was not robustly associated with age-corrected brain-age gap across pretrained models. Brain-age inference was sensitive to model choice and age-bias correction. BrainAge-specific interpreter patterns should therefore be considered exploratory and hypothesis-generating rather than evidence of a general difference in biological brain ageing.