Background
The continuous use of two languages has been shown to be among the different social, physical, mental, and leisure activities [
1] that can promote cognitive reserve (CR) [
2‐
4]. CR refers to individual differences in clinical resilience to brain pathology as a result of differences in neural efficiency and/or the use of a compensatory neural network [
5,
6]. This relationship between CR and bilingualism has often been related to the supposed cognitive benefits of having to manage two languages—a phenomenon, however, that is currently under dispute [
7‐
10].
Most of the evidence supporting the potential effect of bilingualism on CR comes from studies with brain-damaged patients with cognitive decline or dementia (Alzheimer’s disease (AD), single-domain amnestic mild cognitive impairment (MCI) patients, or other patients with dementias, such as the vascular type). In this regard, the most consistent finding is that the onset of the clinical symptoms associated with the disease is significantly delayed in bilinguals, compared to monolinguals—a delay of about 4–5 years [
3,
9,
11‐
16]. Moreover, protective effects of bilingualism against age-related cognitive decline have been found independently of baseline cognitive ability (childhood intelligence), therefore dismissing reverse causality—the fact that childhood differences such as intelligence could lead to bilingualism instead of bilingualism leading to cognitive differences [
17]. However, these results have not always been replicated [
4,
18].
The delay in the onset of dementia in bilinguals has been proposed as evidence of the contribution of bilingualism to CR [
9]. Two brain mechanisms are suggested to explain the increased CR: neural reserve and neural compensation. Neural reserve addresses the idea that CR could be associated with individual differences in the resilience of pre-existing cognitive networks [
9,
19]. In this regard, evidence from neuroanatomic studies with healthy older participants suggests that bilingualism promotes gray matter volume and white matter integrity as a result of using two languages [
20‐
22]. Neural compensation appears when cognitive function is maintained in the presence of brain atrophy due to better utilization of alternative networks [
9,
19,
23]. It has been suggested that this compensation promotes brain reserve; that is, it increases brain size in specific areas that allow more plasticity to overcome pathology and neurological insult [
23].
This study takes this perspective and investigates brain atrophy in bilingual and monolingual individuals with MCI. Therefore, if bilingualism promotes neural reserve and compensation, bilinguals might be expected to tolerate greater amounts of neuropathology or atrophy once the disease is manifested. This prediction is based on neuroimaging studies showing that, at the same cognitive level, bilinguals have more brain pathology than monolinguals. The first study to investigate the CR effects of bilingualism showed that bilingual patients suffering from AD exhibited much greater brain atrophy than monolinguals in regions associated with the pathology, such as the left middle temporal lobes [
24]. Because the two language groups showed the same degree of cognitive decline, the interpretation is that bilinguals would tolerate greater amounts of neuropathology or atrophy before the disease is manifested. Consistent with this interpretation, bilinguals with AD also showed greater levels of hypometabolism in the left parietal, temporal, and frontal areas than monolinguals [
16]. However, it is worth noting that the opposite results have been reported in MCI patients [
25].
The second objective of the study is to prospectively investigate the atrophy rate in MCI participants. Although the number of studies is limited, there is evidence suggesting that bilingualism promotes neural compensation. On the one hand, studies in healthy older individuals and patients with Alzheimer’s disease have shown increased functional connectivity in bilinguals compared to monolinguals [
16,
26]. On the other hand, evidence from studies investigating structural differences in healthy older participants suggests that bilingualism contributes to brain reserve, by showing that monolinguals present more extended age-related brain atrophy than bilinguals in diverse areas of the frontal, parietal, and temporal lobules and that these differences are associated with cognitive performance on different kinds of tasks [
27‐
30]. Together, these results support the view that bilingualism contributes to neural compensation and brain reserve during all the stages of neurodegeneration. However, this interpretation is limited by the fact that all these findings stem from cross-sectional designs that could be biased by cohort effects [
31].
In the present study, we adopt both a cross-sectional and a longitudinal perspective to determine the role of bilingualism in brain atrophy and cognitive decline. Based on the reviewed literature, we hypothesize the following: (1) MCI bilinguals matched with MCI monolinguals on a cognitive level, and sociodemographic factors would transversally present greater brain atrophy; (2) MCI bilinguals would show less atrophy and cognitive decline than monolinguals longitudinally. By combining the cross-sectional and longitudinal data, we attempt to provide a plausible explanation for the nature and origin of the bilingual delay in the onset of dementia.
Discussion
We investigated the neural bases of the putative protective effect of bilingualism against dementia by comparing the brain atrophy of bilinguals and monolinguals suffering from MCI. To this end, we selected two groups of monolingual and bilingual MCI patients with similar sociodemographic characteristics and education levels, living in the same area in the city of Valencia. The cross-sectional analysis showed that MCI bilinguals showed a greater amount of brain atrophy than MCI monolinguals, but no differences in global cognitive level or age. In the present study, we also took prospective longitudinal measures to shed light on atrophy rates in both groups. In agreement with our hypotheses, monolinguals showed higher brain atrophy rates and more cognitive decline than bilinguals in a 7-month period. Specifically, in this period, monolinguals, but not bilinguals, showed significant brain atrophy and cognitive decline. Together, our results suggest that the active use of two languages throughout life not only promotes CR, but also brain reserve, providing a neural-based framework that could explain why bilinguals, compared to monolinguals, show a delay in the onset of dementia.
The results of the present study are consistent with previous cross-sectional studies showing that bilinguals require a greater amount of neuropathology in the brain to manifest the same cognitive status level. Specifically, one study using positron emission tomography (PET) in patients with AD showed that bilinguals had lowered hypometabolism, especially in the temporo-parietal cortex [
16], whereas another study using computed tomography showed more GM atrophy in bilingual AD patients than in monolingual patients [
24]. Our cross-sectional analysis is consistent with all these results, but it presents some additional features that should be specifically discussed. First, our study is the first to demonstrate this neuroprotection in MCI patients. This condition is a preclinical form of dementia that preserves the ability to perform daily life activities. Our clinical sample was mostly composed of multi-domain amnestic MCI patients, and the results showed no age differences between the language groups because bilinguals were only 7 months older than monolinguals. These results were consistent with a previous report showing age differences in single-domain, but not multi-domain, MCI patients [
14]. In that study, the authors proposed the possible coexistence of vascular risk factors as a possible explanation for the lack of age effects in this group, but this is unlikely in our sample because patients with vascular problems identified in the MRI were removed from the sample, and because bilingualism also protects against deterioration in stroke patients [
42]. Thus, our data are the first to show increased CR in MCI, that is, higher atrophy in MCI bilinguals at the same cognitive level as MCI monolinguals. Second, we have demonstrated the effect of bilingualism on brain atrophy in the absence of relevant between-group differences in factors such as education, age, or the environment (i.e., the same city of residence), as obtained in previous studies [
16,
24]. It is important that the two groups did not differ in their cognitive status, as assessed by a neuropsychological profile. Thus, the results obtained in the present study strengthen the interpretation of a bilingual advantage because we controlled for the potential confounding factors. According to recent proposals [
9], differences may arise from a better capacity of bilinguals to functionally compensate for the greater loss of brain parenchyma. This increased neural compensation would arise from the continuous use of two languages, which entails a stronger use of certain brain areas involved in language control and executive functions [
16]. The fact that our monolingual individuals could be better categorized as
passive bilinguals favors this interpretation, given that the observed differences would not arise from the knowledge of the second language per se, but rather from the frequency of use.
RBM analysis revealed that the between-group significant differences shown in our cross-sectional study were mainly located in the lingual and supramarginal gyrus. These two regions have been reported to be affected by AD in functional and structural meta-analyses [
43,
44]. Specific alterations in MCI individuals have been found in the lingual gyrus during tasks involving episodic memory [
45]. Moreover, differences in the supramarginal gyrus have been shown when comparing anatomical likelihood estimate maps of MCI converters and non-converters in a meta-analysis integrating results from different neuroimaging modalities [
46]. The right supramarginal area was also involved in previous morphometric studies in bilingualism. For instance, increased GM volume in this region has been related to better proficiency in the second language and the number of non-native languages spoken [
47]. Furthermore, healthy older bilinguals showed higher white matter integrity than monolinguals in the superior longitudinal fasciculus [
21,
22], that is, the white matter tract that connects the supramarginal gyrus to the frontal and temporal regions. All these results may suggest that the brain areas related to the use of language would increase their efficiency in bilinguals and, in turn, compensate for the effects of AD neuropathology.
One of the main objectives of this study was to investigate how bilingualism impacts the course of dementia. For this reason, we retested a subsample of patients who did not differ at baseline on any of the cognitive or sociodemographic variables (age, education, and gender). Our longitudinal results suggest that monolinguals have a faster rate of cognitive decline and brain atrophy than bilinguals. On the neuropsychological measures, both groups showed a significant cognitive decline on the MMSE, FAQ, phonetic and semantic fluency, Boston Naming Test, clock-drawing test, and the overall measure in the 7-month period. Crucially, monolinguals showed a greater decline than bilinguals on the overall cognitive measure. This result coincides with previous findings showing a heterogeneous pattern of cognitive decline that does not focus on any specific domain [
48,
49]. In this regard, it is noteworthy that most of the patients who participated in this study were multiple-domain amnestic. Importantly, monolinguals, but not bilinguals, presented significant brain atrophy in the 7-month period. These results coincide with the findings from cross-sectional studies in healthy older individuals showing increased age-related brain atrophy in monolinguals compared to bilinguals [
27‐
30]. In this regard, the results of our study agree with these findings and suggest that the neuroprotective effect of bilingualism is also maintained during the early stages of dementia. A possible neural mechanism driving this effect was proposed by Barulli and Stern, who suggested that neural compensation may increase brain reserve by promoting neuroplasticity [
23]. Thus, the functional compensation required to maintain performance will eventually lead to changes in the brain itself. RBM analyses investigating the specific brain areas with a higher atrophy rate in monolinguals than in bilinguals did not show significant results at the pre-established threshold (
p < 0.05 FDR corrected). However, uncorrected results suggest that the main differences found in the parenchymal analyses were located in areas related to language or executive control, such as the cingulate gyrus and the striatum [
50], and in crucial areas in dementia, such as the hippocampus. Although speculative, we suggest that the continuous use of two languages in bilinguals may help to preserve the brain areas involved in controlling two languages and, potentially, executive control [
51]. This neuroprotection would have a compensatory effect on the manifestation of cognitive symptoms of MCI and dementia [
16].
This study has several limitations that should be considered. First, the sample in our study was composed only of MCI individuals. MCI cohorts are heterogeneous groups with variable rates of conversion to dementia. Therefore, cautious interpretation is required when extending the results to the protective effects of bilingualism on dementia. Second, due to participants’ drop out, only 59.6% of the initial sample performed the longitudinal recording. Furthermore, this subsample was unbalanced. Therefore, the longitudinal results were based on a relatively small sample of 32 individuals. However, the results obtained when comparing the unbalanced sample of 16 bilinguals and 43 bilinguals yielded similar results. Future studies with larger sample sizes may provide other differences not detected in our study. Third, the follow-up period could be considered short for a long-term disease such as AD. Therefore, our longitudinal analysis could be considered the first one to provide empirical evidence about short-term patterns of atrophy and cognitive decline in bilinguals and monolinguals with MCI. However, further studies investigating longitudinal changes in these populations using longer temporal windows are necessary. Fourth, the bilinguals in this study spoke Spanish and Catalan, which could be considered two similar languages. This could be a limitation in many studies of bilingualism; however, the protective influence of bilingualism on dementia has been demonstrated in different contexts (e.g., English-Spanish, English-Polish, English-Yiddish, Telugu-Hindi, Dutch-French). Furthermore, for the specific purposes of this study, this similarity might be a strength. It has been proposed that the relationship between cognitive reserve and bilingualism is due to the additional demands on the control system in bilinguals to effectively manage the two languages [
8‐
10]. The positive findings in our study suggest that the contribution of bilingualism to CR is effective even in bilinguals speaking similar languages, where the cognitive demands on the control system might be considered lower than in dissimilar languages. In fact, in our study, we compared active and passive bilinguals; therefore, our results suggest that the contribution of bilingualism to cognitive reserve is related to the active use of the two languages and not just their comprehension. The same conclusions could be drawn from other studies [
15].
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.