Domain-general inhibitory control in bilingual language switching: item-specific vs whole-language inhibition

This thesis examines the role of domain-general inhibitory control in bilingual speech production. It has been suggested that correct language selection in bilingual production relies on inhibitory control, with recent evidence pointing towards the involvement of domain-general mechanisms. In particular, functional neuroimaging studies report activation of the pre-supplementary motor area (pre-SMA), a brain area responsible for domain-general inhibitory control, during language switching tasks. However, it remains unclear whether the pre-SMA plays an essential role as part of the language control network or simply co-activates with it. In this thesis, I investigate the causal relationship between neuronal activity in the pre-SMA and behavioural performance in language switching, by transiently disrupting this brain area using transcranial magnetic stimulation (TMS). It has been proposed that there may be two levels of inhibition in bilingual control: item-specific inhibition and whole-language inhibition. I start with the hypothesis that these are both at work in language switching and one or both may rely on domain-general inhibitory control.Two experiments were carried out to test this hypothesis. The first was a behavioural experiment which established the presence of these two levels of inhibition in language switching. Mandarin-English bilinguals performed a picture-naming task involving univalent items (always named in a particular language) and bivalent items (same picture requiring responses in different languages on different trials). In this design, the effect of whole-language inhibition was reflected in the performance decrement on switch trials compared to stay trials, while the effect of item-specific inhibition was reflected in the reduced performance for bivalent items compared to univalent items. In the second experiment, I investigated the causal involvement of the pre-SMA in language control. A repetitive TMS protocol was used to achieve transient disruption of this brain region while the same picture-naming task was performed. The impact on behavioural performance was assessed with respect to the two levels of inhibition. Disruption of the pre-SMA was found to modulate item-specific but not whole-language inhibition, suggesting that only item-specific control recruits this domain-general inhibitory mechanism.