When investigators treat bilingualism as a dimensional, non-unitary construct, fMRIs demonstrate that differences in the duration and extent of bilingual experience effect the recruitment of different neural regions for distinct cognitive processes.
Bilingualism, though regularly treated as a unitary construct opposing monolingualism, can mean different things to different people. The extent, duration, activation, and immersion involved in speaking multiple languages shapes the experience of the speaker. While the effect of bilingualism on cognition and executive control has long been debated due to variabilities between studies or issues of replication, DeLuca, Rothman, Bialystok, and Pliatsikas argue in their 2019 paper that these effects are due to the treatment of bilingualism as a uniform concept. In order to properly analyze the effects of bilingualism on cognition, DeLuca and colleagues conducted an fMRI study in which different bilingual language experiences were identified and measured in relation to different processes of executive control.
Multiple theories work to explain how different language experiences may shape neurocognition. DeLuca and colleagues offer the “neuroemergenist” perspective, in which brain regions become more efficient when tasked with the higher cognitive demands of bilingual language use. Within the neuroemergenist perspective, the Adaptive Control Hypothesis suggests that specific contexts of language usage recruit different brain regions to handle different language commands. Further, the Bilingual Anterior to Posterior and Subcortical Shift (BAPSS) framework suggests that more second language exposure involves a corresponding shift from frontal cognitive control regions to more subcortical and posterior regions in relation to increased efficiency and automation. As tasks become more automatic and effective, they rely less on frontal control and can draw from more posterior regions instead.
DeLuca’s study investigates these theories with an fMRI task of executive control and 65 bilingual participants who speak English as their second language. In order to investigate the different “experience-based factors” (EBFs) of bilingualism, language profiles were described by the domains of “duration” and “extent” of second language (English) use. Duration offered to explain how sustained language control affected executive cognition, while extent examined whether these patterns were reinforced based on the contexts in which English was used. Duration was assessed via participants’ age of acquisition of English and length of English immersion (time in settings where predominantly English is used). Extent was evaluated by scores from the Language and Social Background Questionnaire that described the extent of English engagement at home and in the broader community. These factors were assessed independently in “Model 1” and combined in a separate analysis of “Model 2” to assess the duration of active English use via the total amount of time spent using the second language (active duration) and the length of time in immersion actively using the second language (active immersion). Active duration was calculated by multiplying the average percentage of English use in different language learning stages by the total number of years using English, while active immersion was calculated by multiplying the percentage of regular English use with time in immersive conditions.
The study utilized the Flanker task to assess executive control during an fMRI. In the Flanker task, participants are asked to identify the direction of a red arrow on a screen while white arrows pointing in either the same or opposite direction surround it. DeLuca’s task consisted of three interspersed “mixed” blocks, in which the white arrows either all pointed in the same or opposite direction as the red arrow in each different trial, one “congruent” block in which the white arrows pointed in the same direction as the red, and one “neutral” block in which the white arrows were double-sided.
Different comparisons of these blocks were assumed to demand different cognitive functions. In particular, contrasts of incongruent and congruent trials within the mixed blocks were thought to measure interference suppression, an executive control function in which irrelevant information is ignored. Contrasts of the congruent block and neutral block assessed the facilitation effect, in which additional information assists goal achievement. Finally, mixing cost, the cognitive load connected to switching between tasks, was measured in contrasts of the congruent trials from within the mixed blocks versus all of the trials from the congruent block.
Behavioral results were analyzed by differences in reaction time across blocks. Although the participants all demonstrated the expected effects of facilitation, information suppression, and mixing cost in their reaction times, these results did not differ significantly across different bilingual experience models. This lack of significance was expected, however, as differences in neurocognitive adaptations in previous studies have not transferred over to behavioral performance levels. This suggests that different bilingual and monolingual speakers may perform equally on executive control tasks, but do so in different ways with different networks.
This hypothesis is supported by DeLuca’s subsequent analysis on Model 1 variables and fMRI results. In the interference suppression contrast, lower ages of second language acquisition correlated to higher activation of regions including the angular gyrus, middle frontal gyrus, and left cerebellum. Length of immersion and second language use at home negatively correlated to activation in the cerebellum, thalamus, and precentral gyrus, meaning more English immersion and use at home resulted in less activation in these areas. The decrease in activation of these areas, which are often involved in selection and conflict monitoring, suggest the brain relies less on these regions and suppresses information more efficiently with more second language exposure. Second language use in social context positively correlated with activation in the cerebellum and anterior and posterior cingulate cortex, meaning greater use of English outside of the home resulted in greater activation. These regions have been implicated in language and cognitive control processes, indicating that greater use of a second language in social contexts increases efficiency for interference suppression in specific communicative environments, which then transfers to increased efficiency in general cognitive control.
In mixing cost contrasts, longer duration of English use (lower age of acquisition) correlated with amplified reliance on posterior regions of the brain, like the left lingual gyrus. This supports the BAPSS framework, where cognitive efficiency improves with more bilingual experience, involving a shift from frontal to posterior region activation. Longer immersion of the second language was associated with less reliance on superior parietal regions.
Facilitation contrasts demonstrated a decrease in right hemisphere regions with longer duration of second language use. This may indicate that facilitation does not have as strong of a general effect on people with prolonged bilingual experience. However, more immersion resulted in increased recruitment of left hemisphere temporal and parietal regions, demonstrating that facilitation may have more of an effect in specifically linguistic contexts for individuals with more intensive immersive experiences.
The combination of bilingual experience factors was evaluated in Model 2 to assess the duration of active English use. Activation patterns in the same contrasts proved similar to the factors of acquisition and immersion in Model 1, but to a smaller extent – not every brain region implicated in the independent duration analyses was evident in this analysis. Additionally, in the facilitation effect contrast, there were no significant activation differences. The posterior regions of activation that overlap in Model 2’s “active immersion” with Model 1’s general immersion demonstrate decreased cognitive demand within interference suppression and mixing cost tasks with higher second language immersion. This is in accordance with the idea that more automated, less cognitively demanding tasks rely less on frontal control regions and instead draw from the posterior parts of the brain. Similarly, when comparing “active duration” with Model 1’s age of acquisition, overlapping activation in the cerebellum in interference suppression tasks suggests automation increases with longer duration of second language learning.
In analyzing different bilingual experience factors independently and combined in relation to distinct cognitive processes, different neural activation trends were observed. These results support the idea of a neuroemergenist framework, in which different regions are recruited to accomplish the same task based on language experience. More generally, these results demonstrate that increased duration of second language use leads to more efficient interference suppression and language switching, and variable reliance on facilitation based on context. Additionally, increased usage of second languages at home and in the community leads to more effective management of language control and mixing cost demands. DeLuca and colleagues’ study demonstrates that different dimensions of bilingual experiences lead to different and measurable neurocognitive outcomes. In the future, DeLuca and colleagues suggest more factors of bilingualism should be theorized and tested to develop a comprehensive model of the dimensions of bilingual experiences and interactions with neural networks. For now, it is safe to conclude that the brain is built to maximize efficiency based on its inputs. If the inputs of the brain change, especially in terms of specific language experiences, there will be measurable change in how the brain adapts and learns to effectively maximize cognitive control.
DeLuca, V., Rothman, J., Bialystok, E., & Pliatsikas, C. (2019). Duration and extent of bilingual experience modulate neurocognitive outcomes. NeuroImage, 204, 116222.