Abstract
INTRODUCTION
The relationship between biological sex, considered a risk factor for Alzheimer's disease (AD), and bilingualism, a resilience factor, is unclear. We assessed this relationship in 335 individuals with mild cognitive impairment (MCI) in a Canadian cohort.
METHODS
We used univariate analysis and structural equation modelling to study the relationship between female sex and bilingualism. We created a resilience index (RI) for each participant using the residual approach. Logistic and linear regressions predicted cognitive and brain health in relation to RI.
RESULTS
Overall, bilingual males had increased RI. Higher RI was associated with less risk of AD and less neuropathology and glial activation as indexed by plasma p‐tau181, neurofilament light, and glial fibrillary acidic protein.
DISCUSSION
In MCI, the combination of elevated estradiol levels due to aromatization and bilingualism may provide synergistic protection for verbal memory, making old bilingual males more resilient.
Highlights
Sex steroids influence verbal memory
In a structural equation modeling (SEM) model, verbal memory mediates cognitive decline
Elevated estradiol from aromatization makes old bilinguals more resilient
Keywords: bilingualism, GFAP, mild cognitive impairment, NfL, p‐tau181, risk and resilience, sex differences
1. BACKGROUND
In Canada, approximately 500,000 persons are living with Alzheimer's disease (AD), with two‐thirds being female. 1 Given that beyond aging, female sex is the largest risk factor for AD, 2 it is surprising that resilience factors specific to sex have not been studied more extensively. 3 To address this, it is essential to examine sex differences in prodromal AD stages, like mild cognitive impairment (MCI). In mixed sex studies (i.e., female and male, only controlling for sex), MCI is classified into two subtypes depending on the presence of memory problems: amnestic MCI (aMCI), and non‐amnestic MCI (naMCI); those with aMCI are at heightened risk of progressing to AD. 4 aMCI can be further divided into single‐domain (SD‐aMCI; only memory affected), and multiple‐domain (MD‐aMCI; memory plus other cognitive functions affected). 4 Virtually nothing is known about resilience factors specific to males or females with any form of MCI.
Important components of resilience are cognitive and brain reserve 5 ‐ concepts that in mixed sex studies are used to explain why some people maintain cognitive function in the face of higher AD pathology. 6 When sex is considered, verbal memory is linked with increased resilience in females 7 who maintain it in the face of more AD pathology than males. 7 , 8 With MCI, females have also shown better performance in the Rey Auditory Verbal Learning Task (RAVLT) than males—an index of verbal memory. 8 This sex difference has been replicated in females having equivalent levels of brain hypometabolism as measured by glucose metabolic rates (TLGluMR) than males, 9 and in females with higher levels of plasma phosphorylated tau (p‐tau) than males. 10
The female verbal advantage is clinically relevant because verbal memory tests are key in the diagnosis of dementia and may explain why females are diagnosed later and with greater pathology than males. 11 Conversely, the combination of lower pathology and worse verbal memory performance in men indicates a different resilience profile that warrants further investigation. Indeed, the distinction between sexes in cognitive reserve (CR) suggests that the relationship between verbal memory and CR might be influenced by other factors. Specifically, the observed differences in CR profiles linked to verbal memory suggest that exploring bilingualism may provide valuable insights 12 —verbal memory performance can vary across individuals with knowledge of different languages. 13 , 14 For instance, in studies including both males and females but only controlling for sex, multilingualism has been shown to be a resilience factor; bilingual older adults at baseline show superior verbal episodic memory (recalled more verbally presented items) than monolinguals. 15 Better verbal episodic memory has also been shown in bilinguals with dementia neuropathology, outperforming monolinguals on the RAVLT. 16 Other mixed sex studies show that bilingualism may enhance CR 17 , 18 ; bilinguals also present with the first symptoms of AD approximately 4 years later than monolinguals, despite comparable clinical profiles. This pattern has been replicated in multilingual cohorts, 19 and multiple international cohorts. 20 , 21 , 22 This finding also extends to patients with MCI, 23 , 24 though one study shows that bilingualism only benefits patients with SD‐aMCI and not multiple domain. 25 Other studies not controlling for MCI subtype demonstrate bilinguals converting faster from MCI to AD 26 and showing greater gray matter loss than monolinguals despite having similar clinical profiles. 27 Note that the fact that bilinguals maintain comparable cognitive performance despite greater brain atrophy aligns with the well‐documented memory and attention benefits observed in bilinguals.
An important gap in the literature is understanding how bilingualism interacts with biological sex to influence AD pathology and CR. To approach this gap, we asked: (i) is there an interaction between bilingualism and resilience in females and males; (ii) does bilingualism offer greater resilience benefits to females or males, and (iii) if bilingualism enhances resilience more for one sex, what might explain this difference?
To answer these questions, we accessed a large and well‐characterized cohort of Canadian participants, the Comprehensive Assessment of Neurodegeneration and Dementia (COMPASS‐ND) of the Canadian Consortium on Neurodegeneration and Aging (CCNA).
RESEARCH IN CONTEXT
Systematic review: We conducted an extensive literature review using traditional sources (e.g., PubMed). This is the first study to analyze the relationship among biological sex, sex steroids, and bilingualism in patients with mild cognitive impairment (MCI).
Interpretation: We show evidence of sex‐and language‐specific resilience profiles in MCI. In females, declines in estradiol and testosterone are linked with cognitive decline, but a verbal advantage may act as a form of cognitive reserve (CR). In males, higher estradiol from aromatization and bilingualism may enhance resilience. Greater resilience is linked with lower odds of Alzheimer's disease (AD), lower neuropathology, and less glial activation.
Future directions: Given that previous literature has shown effects of bilingualism as a function of MCI subtype, future research should explore sex‐and language‐specific resilience profiles by subtype, using longitudinal approaches to track resilience across AD progression.
2. METHODS
2.1. Cohort
This study included male and female participants with MCI drawn from the COMPASS‐ND cohort, data release 7. 28 , 29 The cohort for the current study included 335 participants from different sites across Canada: Québec City, Sherbrooke, Montréal, Ottawa, Peterborough, Toronto, Kitchener/Waterloo, London, Saskatoon, Edmonton, Calgary, Vancouver, and Victoria. Data collection sites include memory clinics, stroke clinics, movement disorder clinics, and behavioral neurology clinics, as well as both academic and private research groups. 28 All participants were community‐dwelling adults aged 60+ and older. The COMPASS‐ND study was approved by the Jewish General Hospital Research Ethics Board 29 and followed the ethical code for research with humans as stated by the Declaration of Helsinki. All participants provided written informed consent to participate in COMPASS‐ND. Details on the design, recruitment of the COMPASS‐ND cohort from the CCNA, and data management have already been published. 29 , 30
See Cohort S1, for details about bilingualism, other demographic variables, clinical assessment, neuropsychological assessment, sex steroids, and dementia biomarkers.
2.2. Statistical analysis
All statistical analyses were conducted in R version 4.4.2. 31 See Statistics S2 for details about the descriptive statistics, structural equation modeling (SEM), sensitivity analysis, and magnetic resonance imaging (MRI) procedure.
3. RESULTS
3.1. Cohort characteristics
The cohort characteristics are shown below in Table 1. There were group differences in second language proficiency, education, immigration, sex steroids, and logical memory (see Supporting Information S.3.).
TABLE 1.
Demographic and clinical characteristics for MCI patients.
| Female | Male | |||
|---|---|---|---|---|
| Characteristic | Bilingual (n = 69) | Monolingual (n = 72) | Bilingual (n = 111) | Monolingual (n = 83) |
| Age | 72.7 (8.7) | 72.5 (5.2) | 73.8 (0.2) | 73.1 (0.4) |
| Education | 15.1 (5.6) | 14.1 (5.6) | 15.7 (2.8) | 14.3 (0) |
| Immigration a | 34 (47.8%) | 10 (13.8%) | 56 (50.4%) | 15 (18.0%) |
| Bilingualism score b | 6.3 (2.2) | 0.0 (0) | 6.4 (1.0) | 0.0 (0) |
| E2 (pmol/L) | 40.6 (12.7) | 50.6 (33.2) | 103.6 (11.3) | 107.2 (0.7) |
| TFCI (pmol/L) | 21.5 (59.3) | 15.0 (4.2) | 275.7 (60.1) | 264.7 (36.0) |
| MoCA | 23.4 (2.8) | 23.3(2.8) | 23.2 (2.8) | 23.1 (0.7) |
| CERAD | 18.4 (1.4) | 18.0 (3.5) | 18.3 (0.7) | 18.1(0.0) |
| Logical memory immediate recall | 10.02 (2.12) | 8.17 (7.77) | 9.42 (0.70) | 8.15 (3.53) |
| Logical memory delayed recall | 7.90 (4.24) | 5.84 (7.07) | 7.36 (1.41) | 6.09 (2.82) |
| CDR | 0.33 (0.3) | 0.42 (0) | 0.36 (0.3) | 0.41 (0.3) |
| GDS | 7.0 (0.7) | 7.7 (10.6) | 6.1 (2.1) | 6.5 (6.3) |
| LAWTON | 22.2 (2.5) | 21.8 (2.1) | 21.0 (0.7) | 21.4 (0) |
Note: Mean (SD) is reported for continuous variables and number (%) for categorical variables.
Abbreviations: CDR, Clinical Dementia Rating; CERAD, Consortium to Establish a Registry for Alzheimer's Disease; E2, estradiol; GDS, Geriatric Depression Scale; LAWTON, Lawton instrumental activities of daily living; MoCA, Montreal Cognitive Assessment; TFCI, total free testosterone concentration index.
Immigration includes number of people not born in Canada.
Bilingual score was based on participants’ rating of their bilingual proficiency in reading, writing, speaking, and listening. A bilingualism score was calculated with the averages of these subcomponents.
3.2. Is there an interaction between biological sex and bilingualism in cognitive or brain health?
The aim of this analysis was to investigate whether bilingualism, a potential CR factor, interacts with biological sex in relation to cognitive and brain health.
3.2.1. Cognitive performance
The analysis only revealed simple main effects of sex for the Delis–Kaplan Executive Function System (D‐KEFS) color–word interference, switching time (F 1,331 = 4.164; p = 0.04), the RAVLT (F 1,331 = 18.22; p < 0.001), and associative memory in both immediate (F 1,331 = 8.630; p = 0.003) and delayed recall (F 1,331 = 10.549; p = 0.001). For D‐KEFS color–word interference (switching time), males had lower scores (M = 81.95) than females (M = 86.72). Females outperformed males on the RAVLT (M = 6.5 and M = 4.8, respectively). For associative memory immediate recall, females had higher scores (M = 3.58) than males (M = 2.99). For associative memory delayed recall, females also had higher scores (M = 3.02) than males (M = 2.28).
See Table 2 below for more details and Table S1 for all p‐values and non‐significant effects.
TABLE 2.
Means and standard deviations for cognitive and brain health.
| Cognitive performance | Female bilingual (n = 69) | Female monolingual (n = 72) | Male bilingual (n = 111) | Male monolingual (n = 83) |
|---|---|---|---|---|
| TMTA_time | 43.5 (21.1) | 41.7 (20.1) | 41.8 (15.2) | 41.9 (17.8) |
| TMTB_time | 120.08 (7.78) | 120.58 (106.07) | 121.04 (16.97) | 113.71 (2.83) |
| D‐KEFS_color word interference_inhibition time | 79.75 (7.07) | 77.51 (4.95) | 72.81 (0) | 82.43 (9.19) |
| D‐KEFS_color word switching_time | 87.57 (30.41) | 86.17 (28.99) | 79.31 (30.4) | 83.31 (11.31) |
| Face name association immediate recall | 3.73 (2.12) | 3.43 (2.83) | 2.87 (4.24) | 3.11 (0.70) |
| Face name association delayed recall | 3.13 (2.83) | 2.91 (3.54) | 2.23 (2.13) | 2.33 (1.41) |
| Digit symbol | 51.54 (17) | 51.61 (17.6) | 51.27 (2.8) | 51.18 (6.3) |
| D‐KEFS_category fluency_total raw score | 35.24 (12.02) | 35.23 (4.94) | 34.03 (6.36) | 35.58 (2.82) |
| D‐KEFS letter fluency_total raw score | 37.62 (6.3) | 35.37 (2.1) | 34.83 (12) | 34.60 (8.4) |
| Brief visuospatial memory total_recall | 14.7 (5.16) | 14.6 (12.2) | 13.4 (2.1) | 14.6 (16.2) |
| Brief visuospatial memory_delayed recall | 5.8 (2.1) | 5.5 (5.6) | 5.7(1.41) | 5.8 (7.07) |
| WAIS III Vocabulary Test | 48.19 (20.5) | 48.39 (12.02) | 46.9 (12.02) | 46.7 (5.6) |
| RAVLT_A7 | 7.1 (3.5) | 6.0 (2.1) | 5.2 (3.5) | 4.5 (1.4) |
| Brain markers | ||||
| Aβ 40 (pg/mL) | 109.4 (18.3) | 114.1 (29.8) | 110.1 (26.8) | 116.3 (26.6) |
| Aβ 42(pg/mL) | 6.7 (2.2) | 7.0 (2) | 7 (1.7) | 7.1 (1.7) |
| Aβ 42/40 (pg/mL) | 0.06 (0) | 0.06 (0) | 0.06 (0) | 0.06 (0) |
| p‐tau181 (pg/mL) | 2.91 (3.0) | 3.36 (3.2) | 2.82 (1.52) | 3.23 (1.98) |
| Neurofilament light (pg/mL) | 26.8 (3.06) | 24.8 (11.1) | 26.9 (25.1) | 29.5 (41.1) |
| GFAP (pg/mL) | 164 (73.1) | 177.5 (100) | 144.5 (117) | 152.2 (83.5) |
| Total GM volume (mm3) | 1254602 (21335) | 1251295 (168087) | 1418403 (36465) | 1468237 (233268) |
| Temporal left GM volume (mm3) | 95478 (13432) | 95693 (12860) | 102148 (14124) | 107088 (12613) |
| Temporal right GM volume (mm3) | 95407 (14024) | 95485 (12478) | 102423 (15388) | 106899 (7912) |
Abbreviations: D‐KEFS, Delis–Kaplan Executive Function System; GFAP, glial fibrillary acidic protein; GM, gray matter; RAVLT_A7, Rey's Auditory Verbal Learning Test‐trial A7; TMTA, Trail Making Test Part A; TMTB, Trail Making Test Part B; WAIS III, Wechsler Adult Intelligence Scale®, Third Edition.
3.2.2. Brain markers
There were simple main effects of bilingualism in gray matter volume in both the left (F 1,331 = 5.56; p = 0.01) and right temporal lobes (F 1,331 = 4.38; p = 0.03). Bilinguals had less gray matter volume in the left (M = 98813 mm3) and right (M = 98915 mm3) temporal lobes than monolinguals (M = 101391 mm3, M = 101192 mm3; respectively).
There were simple main effects of sex for total gray matter volume (F 1,331 = 173; p < 0.001), gray matter volume in the left (F 1,331 = 54.7; p < 0.001) and right temporal lobe (F 1,331 = 58.33; p < 0.001), and glial fibrillary acidic protein (GFAP) (F 1,331 = 5.30; p = 0.03). Females had less total gray matter volume (M = 1252948 mm3) than males (M = 1443320 mm3). They also had less gray matter volume in the left (M = 95586 mm3) and right (M = 95446 mm3) temporal lobes than males (M = 104618 mm3, M = 104661 mm3; respectively). Females (M = 170.5 pg/mL) also had more neuropathology in the form of GFAP than males (M = 148.3 pg/mL).
See Table 2 below for more details and Table S1 for non‐significant effects.
Given the lack of interaction between biological sex and bilingualism using univariate models, the question of whether bilingual females or males demonstrate greater resilience remained unresolved and warranted further investigation using multivariate techniques to disentangle the effects of other related variables.
3.3. Does bilingualism offer greater resilience benefits to females or males?
To answer this question, we first constructed an SEM (model 1) to see the association between biological sex, bilingualism, and other important variables: estradiol (E2), total free testosterone concentration index (TFCI), immigration, age, education, and Cognition. Model 1 fit the data well (Comparative Fit Index [CFI] = 0.90; Tucker–Lewis Index [TLI] = 0.88; root mean square error of approximation [RMSEA] = 0.05, standardized root mean square residual [SRMR] = 0.05) and is presented in S.4; Figure S1; Table S2.
3.3.1. Resilience index
We used the path coefficients of model 1 to calculate a resilience index (RI) (for details see S.5).
As shown in Table 3, bilingual males had greater RI than the rest of the groups. This was followed by bilingual females, monolingual males, and monolingual females.
TABLE 3.
Resilience index.
| Group | RI | Group comparison | Bootstrap 95%, CI for group differences | Significance |
|---|---|---|---|---|
| Bilingual male | 0.55 | Bilingual male vs. bilingual female | [0.07, 0.16] | Yes |
| Bilingual female | 0.50 | Bilingual male vs. monolingual male | [−0.09, 0.02] | No |
| Monolingual male | 0.44 | Bilingual male vs. monolingual female | [0.11, 0.20] | Yes |
| Monolingual female | 0.39 | Bilingual female vs. monolingual male | [−0.21, −0.09] | Yes |
| – | – |
Bilingual female vs. monolingual female |
[−0.01, 0.09] | No |
| – | – |
Monolingual male vs. monolingual female |
[0.13, 0.25] | Yes |
Abbreviations: CI, confidence interval; RI, resilience index.
We used bootstrap methods to assess the differences between groups (number of bootstraps = 1000, see S.5.1 for details). The analysis revealed that bilingual males had a significantly higher RI than bilingual females and monolingual females. Moreover, bilingual females had significantly greater RI than monolingual males, and bilingual males had higher RI than monolingual females (see Table 3 below).
3.4. Why might bilingualism benefit males more than females?
To address this question, we developed a second SEM model (Model 2), which incorporated a verbal memory measure (RAVLT) as a mediator. Model 2 was structurally similar to model 1, with the key difference being the addition of a mediator to assess direct, indirect, and total effects on Cognition (see S.6; Figure S2 for conceptual model).
3.4.1. Model 2
Model 2 fit the data well (CFI = 0.97; TLI = 0.95; RMSEA = 0.04; SRMR = 0.03), and is presented below in Table 4, and Figure 1.
TABLE 4.
Regression weights from the paths for model 2.
| Model 2 | |||||
|---|---|---|---|---|---|
| Paths | β | SE | Z | 95% CI | p value |
| Latent Variable | |||||
| Cognition ← Visuospatial Memory | 1.000 | 0.000 | NA | NA | 1.000 |
| Cognition ← Digit Symbol | 1.800 | 0.270 | 6.665 | 1.000 to 1.271 | 0.000 |
| Cognition ← TMTA | −1.668 | 0.255 | −6.544 | 2.330 to −2.168 | 0.000 |
| Cognition ← TMTB | −1.846 | 0.276 | −6.690 | −1.169 to −2.387 | 0.000 |
| Regressions | |||||
| RAVLT_A7 → E2 | −0.026 | 0.048 | 0.539 | −1.305 to −0.120 | 0.590 |
| RAVLT_A7 →TFCI | 0.118 | 0.065 | 1.815 | 0.245 to −0.009 | 0.070 |
| Female Sex → Cognition | 0.000 | 0.025 | −0.008 | 0.049 to −0.050 | 0.994 |
| Bilingualism → Cognition | 0.050 | 0.026 | 1.904 | 0.102 to −0.001 | 0.057 |
| Direct Effects on Verbal Ability | |||||
| Female Sex → RAVLT_A7 | 0.321 | 0.065 | 4.931 | 0.448 to 0.193 | 0.000 |
| Bilingualism → RAVLT_A7 | 0.127 | 0.053 | 2.394 | 0.230 to 0.023 | 0.017 |
| Female Sex and Bilingualism via Verbal memory on Cognition | |||||
| Indirect effect | 0.321 | 0.065 | 4.931 | 0.448 to 0.193 | 0.000 |
| Bootstrap estimation indirect effect for female sex | 0.298 | 0.062 | 4.831 | 0.00 to 0.177 | 0.000 |
| Total effect | 0.320 | 0.075 | 4.261 | 0.468 to 0.173 | 0.000 |
| Covariances | |||||
| Visuospatial Memory ↔ RAVLT | 0.481 | 0.055 | 8.757 | 0.588 to 0.373 | 0.000 |
| Visuospatial Memory ↔ Bilingualism | 0.133 | 0.052 | 2.571 | 0.234 to 0.032 | 0.010 |
Note: * p < 0.05, ** p < 0.01, *** p < 0.001.
1) Arrows in the table represent the relationship between variables, and each type of arrow has a specific meaning. The latent variable is expressed with a single arrow going from right to left indicating the neuropsychological measure influencing the latent variable ‘Cognition’ (e.g., Cognition ← Visuospatial Memory). Single‐headed arrows going from left to right (→) indicate direct effects. For example, X → Y means that X influences Y. Double‐headed arrows (↔) indicate covariances or correlations between variables (mutual relationship or shared variance).
2) In SEM, the first indicator of a latent variable is typically fixed to a value of 1 to establish the scale of the latent variable, thus these values are listed as NAs in Table 4. This is typically done to avoid identification issues or to prevent the model from being over‐ or under‐identified.
Abbreviations: E2, estradiol; MoCA, Montreal Cognitive assessment; RAVLT_A7, Rey's Auditory Verbal Learning Test‐trial A7; TFCI, Free testosterone.
FIGURE 1.
SEM showing a mediation effect of verbal memory for female sex and bilingualism on general cognition. Notes: (1) Numbers indicate estimation of path coefficients (* p < 0.05, ** p < 0.01, *** p < 0.001). fit statistics: χ 25 = <0 .001; Comparative Fit Index = 0.90; Tucker–Lewis Index = 0.88; RMSEA = 0.05; standardized root mean squared residual = 0.05. (2) Arrows in the model diagram represent the relationship between variables, and each type of arrow has a specific meaning. Single‐headed arrows (→) are used to indicate direct effects. For example, X→Y means that X influences Y. The indirect effect is not represented by a single arrow but is instead the product of the two direct paths. In our case, female sex and bilingualism→RAVLT→Cognition. So, while the diagram does not show an explicit arrow representing the indirect effect, this is implied by the relationship between female sex, bilingualism, RAVLT, and Cognition. It is the combined effect of female sex and bilingualism, influencing RAVLT, which in turn influences Cognition. Given that this model is a feedback loop, the relationship between RAVLT and cognition is expressed as a covariate. Single‐headed arrows are also used for latent variables (variables that are not directly observed but are inferred from multiple indicators), the single‐headed arrows, in this case, are used to indicate how the indicators (observed variables) are related to the latent variable. Double‐headed arrows (↔) indicate covariances or correlations between variables (mutual relationship or shared variance). (3) In SEM, the first indicator of a latent variable is typically fixed to a value of 1 to establish the scale of the latent variable. This is typically done to avoid identification issues or to prevent the model from being over‐ or under‐identified. (4) Note that female sex is a categorical variable (female = 1, male = 0). For binary categorical variables like sex, the path coefficient reflects the difference in the mean outcome (in this case, Cognition) between the two groups. A positive coefficient would mean that being female (coded as 1) increases Cognition, while a negative coefficient would suggest an increase for males. E2, estradiol; RAVLT_A7, Rey's Auditory Verbal Learning Test‐trial A7; RMSEA, root mean square error of approximation; SEM, structural equation modeling; TFCI, total free testosterone concentration index.
Direct effects
Female sex had a negative but not significant effect on Cognition (β = −0.000, SE = 0.025, p = 0.994). Bilingualism had a positive and marginally significant effect on Cognition (β = 0.050, SE = 0.026, p = 0.057).
For the verbal memory pathway, female sex (β = 0.321, SE = 0.065, p < 0.001) had a significant positive effect on RAVLT, as did Bilingualism (β = 0.127, SE = 0.053, p = 0.017). E2 had a negative but not significant association with RAVLT (β = −0.026, SE = 0.048, p = 0.590). TFCI had a positive and marginally significant association with RAVLT (β = 0.118, SE = 0.065, p = 0.070).
Indirect effect: Female sex and bilingualism's effects on cognition via verbal memory
The indirect effect for both female sex and bilingualism on cognition through verbal memory (RAVLT) was significant (β = 0.321, SE = 0.065, p < 0.001), as was the total effect (β = 0.320, SE = 0.075, p < 0.001). Note that the direct effect of female sex on verbal memory is numerically identical to the indirect effect on cognition (likely due to the feedback loop in our model). Thus, the indirect effect for female sex was recalculated and, following bootstrapping estimation (5000 samples), was found to be β = 0.015, with a 95% confidence interval (CI) of [0.003–0.027].
Covariances
Visuospatial memory had a positive and significant association with RAVLT (β = 0.481, SE = 0.055, p < 0.000) and bilingualism (β = 0.133, SE = 0.052, p = 0.010).
3.5. Sensitivity analysis
A sensitivity analysis was conducted to test the robustness of our findings. Specifically, we tested our RI in the progression of clinical dementia and pathology. Thus, for this analysis, we included AD patients from the COMPASS‐ND cohort (n = 170) and calculated the odds of getting clinical AD and higher pathology for our MCI patients in relation to the RI. In these models, RI served as the dependent variable, and AD status was treated as the categorical predictor variable (possible outcome). The total number of participants for this analysis was 505 (335 MCI and 170 AD patients, see Table 5 for the clinical profile of AD patients).
TABLE 5.
Clinical profile of AD patients.
| Characteristic | Female | Male | ||
|---|---|---|---|---|
| Bilingual (n = 40) | Monolingual (n = 32) | Bilingual (n = 62) | Monolingual (n = 36) | |
| Age | 75 (7.9) | 76.7 (6.7) | 75.7 (7.6) | 74.2 (7.5) |
| Education | 15 (3.3) | 14.4 (2.9) | 15.5 (3.5) | 16.6 (4.5) |
| MoCA | 17.7 (4.2) | 17 (3.2) | 19 (4) | 18.3 (3.3) |
| CERAD | 15 (3.1) | 14.8 (2.8) | 16 (2.5) | 16 (2.8) |
| Logical memory_delayed recall | 2.6 (3.6) | 0.6 (1.5) | 2.2 (3.2) | 2.2 (3.2) |
| CDR | 0.5 (0) | 0.7 (0.2) | 0.6 (0.2) | 0.8 (0.2) |
| GDS | 5.5 (5.3) | 5.8 (3.9) | 6.2 (4.1) | 5.5 (4) |
| LAWTON | 22.2 (0.8) | 18.5 (4.12) | 18.8 (3.9) | 17 (4) |
Note: Mean (SD) is reported for continuous variables and number (%) for categorical variables.
Abbreviations: CDR, Clinical Dementia Rating; CERAD, Consortium to Establish a Registry for Alzheimer's Disease; GDS, Geriatric Depression Scale; LAWTON, Lawton instrumental activities of daily living; MoCA, Montreal Cognitive Assessment.
First, a univariate logistic regression revealed that the RI was associated with a decrease in the odds of developing AD for our MCI participants (odds ratio [OR] = 0.04, 95% CI [0.003–0.49], p = 0.01). Thus, for each unit increase in the RI, the odds of the AD decreased by a factor of 0.0436 (or 96.4% reduction in the odds). Note that the fact that the interval range does not include 1 suggests that the effect of the index is significantly different from no effect (i.e., not equal to 1). However, the fact that the range is quite wide indicates uncertainty in the exact magnitude of the effect; the lower bound being so small still points to a strong negative association.
Then, linear regressions showed every unit increase in the RI was associated with 7.0 increase in Montreal Cognitive Assessment (MoCA) scores (β = 7.06, t (409) = 2.71, p = 0.00694); 4.9 increase in the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) scores (β = 4.94, t (409) = 3.05, p = 0.00246); and 5.8 increase in logical memory (β = 5.83, t (409) = 2.05, p = 0.0411). There were no significant associations between the index and any of the other diagnostic tools (e.g., CDR).
As regards pathology, linear models showed a significant negative association between the RI and the concentrations of p‐tau181, GFAP, and plasma neurofilament light (NfL). A higher index was associated with lower levels of p‐tau181 (β = −2.84, p = 0.0069), NfL (β = −43.03, p = 0.0003), and GFAP (β = −197.62, p = 0.0006, indicating that the index may reflect a protective or mitigating effect related to neurodegeneration and glial activation. There was no significant association between the index and Aβ40 (β = −0.087, p = 0.995), Aβ42 (β = 0.342, p = 0.715), or the Aβ42/40 ratio (β = 0.0036, p = 0.504), suggesting that the index does not appear to influence amyloid‐related biomarkers in this cohort.
Finally, a logistic regression with an interaction term (sex*bilingualism) revealed that monolingualism was associated with higher odds of dementia (β = 7.867, p = 0.01). The interaction between RI and bilingualism showed that the effect of resilience on dementia might be stronger for bilinguals than monolinguals (β = −20.8, p = 0.01). The interaction between sex and bilingualism showed that being bilingual and male was associated with lower odds of dementia (β = −10.21, p = 0.004).
4. DISCUSSION
4.1. Summary of findings
This is the first study to examine resilience by sex in bilingual persons.
Our analyses revealed a sex difference in how bilingualism contributes to resilience, with bilingual males showing a higher RI, which is associated with decreased odds of developing AD. Further, sex steroids significantly influenced verbal memory for persons with MCI, and SEM model 2 suggested that verbal memory may mediate cognitive decline. In females, a decline in E2 and TFCI may negatively impact cognition, though women seem to have a verbal advantage, potentially serving as a form of CR. In men, aromatase activity influenced verbal memory, with a stronger effect observed in bilingual individuals. These novel findings suggest that sex steroids and bilingualism interact to shape resilience profiles, offering new insights into dementia progression.
4.2. Interaction between bilingualism, a possible CR factor, and the general risk associated with female sex
Irrespective of bilingualism, females with MCI outperformed males in cognitive switching, verbal episodic memory, and associative memory; they also showed greater gray matter loss in right and left temporal lobes and astrocytopathy than males. This pattern aligns with the CR theory. However, very few mechanisms for CR in women have been suggested in the AD literature. 3 One proposed mechanism is a verbal advantage. 7 Our findings support this; females with MCI outperformed men in the RAVLT despite having greater biomarker pathology.
As regards bilingualism, bilinguals also showed a pattern of CR, and our findings replicate what previous studies have shown. 17 , 18 , 26 , 27 When sex is excluded as a factor, bilinguals with MCI had greater brain atrophy in the temporal lobe despite a similar cognitive profile to monolinguals. However, this type of CR seems to be different. Bilinguals have more structural damage than monolinguals but maintain comparable cognitive performance, suggesting slower functional decline despite progressive atrophy.
Taken together, these first findings indicate that both females and bilinguals show reserve, but likely through different underlying mechanisms. This has important clinical implications; personalized strategies may be needed depending on the type of reserve on which a person relies. 32
4.3. Bilingualism offers greater resilience benefits to males
SEM model 1 emphasized the need to consider sex steroids and other related variables to analyze the relationship between bilingualism and biological sex. Both E2 and TFCI were negatively associated with female sex, likely due to the decline in sex steroids in old adult females. Thus, the model identified the importance of factoring in sex steroids, rather than just biological sex, when studying cognitive impairment.
Bilingualism was positively and significantly associated with immigration; the model also identified a non‐significant effect between immigration and cognition. Interestingly, immigration is a risk factor in the US population, 33 but this might not be the case in Canada, where immigration can be viewed as a protective factor, as in the healthy immigrant effect, 34 and while the trauma of immigration might be similar, settlement practices differ. 35 Our results showed that while bilingualism and immigration are positively related, immigration may not influence cognition in our cohort.
Additionally, age had a negative association with cognition, and this was expected as age is the highest risk factor for AD. 36 Education had a positive effect on cognition, and this was also expected as education has been suggested as a form of CR. 6
When all these variables were considered, and a RI was created, bilingual males had the highest RI. Bilingual females had a lower RI than bilingual males but higher than monolingual males and females. This underscores the significance of bilingualism as a key factor in resilience.
Importantly, the sensitivity analysis showed that a higher RI was associated with decreased odds of developing AD, and increased values in the MoCA, CERAD, and logical memory. A higher RI was also associated with decreased concentrations of p‐tau 181, GFAP, and NfL. The aggregation of tau and the neurotoxic effects of the protein are crucial in the development of AD, 37 and GFAP is a biomarker for glia and an indicator of the severity of AD. 38 Similarly, NfL has been suggested as a promising biomarker, as patients with AD typically show high concentrations of NfL. 39 As regards p‐tau, females in our cohort had higher p‐tau 181 than males, a finding that is consistent with recent studies showing that the decline in E2 after spontaneous menopause is associated with increased tau in females. 40 , 41 Our study adds to these results by also showing increased GFAP and NfL in menopausal females with MCI.
Our study is also the first to show decreased p‐tau 181, GFAP, and NfL in bilinguals than in monolinguals. The bilinguals in our cohort had less neuropathology and glial activation. This pattern indicates that individuals with higher resilience, as indicated by lower biomarkers, may be better equipped to maintain brain health. Indeed, our results for bilinguals are consistent with a brain maintenance hypothesis stating that individuals with higher resilience can better maintain brain health than those with lower resilience. In this respect, our results are in concert with a recent study using the COMPASS‐ND cohort, which found bilingualism to be a predictor for brain maintenance using cortical thickness measures. 42
4.4. Why does bilingualism offer greater resilience benefits to males than females?
SEM model 2 identified verbal memory as a crucial variable mediating cognition for both females and bilinguals. The model also showed that males had higher levels of both E2 and TFCI, that both E2 and TFCI were positively associated, and that TFCI had a positive association with verbal memory. There was also a covariance between bilingualism and visuospatial memory.
It is well established in the bilingualism literature that bilinguals outperform monolinguals on non‐verbal cognitive tasks. 43 While this benefit may hold broadly, in older adult males who are undergoing aromatization of testosterone (conversion of TFCI to E2), sex steroids might affect verbal memory, making older adult bilingual males more resistant to verbal memory decline.
The enzyme aromatase found in various tissues (fat tissue, brain, and testis) converts testosterone into E2. As men age, testosterone levels decline, and aromatase activity increases, which results in men having a large proportion of testosterone converted into E2. E2 plays a crucial role in verbal memory 44 ; and several studies have shown that men with higher E2 (either naturally or due to increased aromatization) may show better performance in verbal memory. 45 Moreover, testosterone may alter visuospatial memory in males through aromatization to E2. 46 Thus, visuospatial memory may be altered in bilingual males, and they may rely more on verbal memory to compensate for cognitive decline. Bilingualism itself is a form of resilience, and the interplay between E2 aromatization and bilingualism may provide bilingual men with an additional layer of resilience.
4.5. Limitations
Our study has limitations. While we developed a RI based on the residual approach, our SEM models only included variables that we know are relevant for our cohort and are based on existing bilingualism and dementia research in Canada (e.g., immigration status, bilingual proficiency). Other cohorts may need to include other variables to study resilience in their bilingual populations. For instance, due to data constraints, our analysis focused on bilingual proficiency as the primary measure of bilingualism. 47 , 48 Other factors, such as age of acquisition, occupational language demands, social interaction, and community engagement, may also influence bilingual language use and cognitive outcomes. While these factors could not be included in the current study, they should be explored in future work. Recent evidence, however, suggests that multilingualism itself, rather than individual factors, may have a protective effect on cognitive aging. A recent large‐scale study conducted across 27 European countries (n = 86,149) found that bilingualism was associated with delayed cognitive aging, even after adjusting for a wide range of linguistic, physical, social, and sociopolitical factors. 49
Also, previous literature has shown effects of bilingualism as a function of MCI subtype, bilingualism seems to impact the aMCI subtype but not the multiple domains 25 ; thus, future research should pursue the possibility that there are also different potential interactions with sex. Additionally, future studies should explore how environmental health and other sociocultural factors may interact with sex, as these variables could also influence biomarker levels and resilience in the context of AD.
5. CONCLUSIONS
Overall, we found different resilience profiles that may be associated with the progression from MCI to AD, but the direction and causality of this relationship remain uncertain given the cross‐sectional nature of our study. Importantly, our study identified sex and language differences in the application of diagnostic tools such as logical memory. All MCI participants (irrespective of sex or language) were diagnosed with MCI and had similar MoCA and CERAD scores. However, bilinguals with MCI outperformed monolinguals in logical memory, a verbal episodic memory task. This may be attributed to both biological and social factors. While lifelong bilingualism provides CR benefits in executive functioning (non‐verbal tasks), females (irrespective of language) have a verbal advantage in verbal episodic memory, and thus bilingual females may have a different resilience profile than monolingual women. Moreover, aromatization to E2 may alter the CR mechanism for bilingual males who, in our study, show greater resilience than bilingual females and monolingual males. Thus, it is important to consider different resilience profiles when interpreting MCI diagnosis. Regarding the distinct mechanisms of resilience, we propose that females may rely more on a verbal memory advantage, while bilinguals show resilience through slower functional decline despite greater brain atrophy. For sex‐specific interventions, it may be beneficial to focus on enhancing verbal memory and supporting hormone balance in females. These findings highlight the need for further research into sex and language‐specific diagnostic criteria, as well as the refinement of cognitive testing tools to ensure that they are equally sensitive across females, males, bilinguals, and monolinguals.
CONFLICT OF INTEREST STATEMENT
The authors have no financial or non‐financial interests to disclose. Author disclosures are available in the Supporting Information.
CONSENT STATEMENT
Written informed consent was obtained from all participants.
Supporting information
Supporting Information
Supporting Information
ACKNOWLEDGMENTS
This work was supported by the ‘Synapse Challenge’ award from the Canadian Consortium on Neurodegeneration in Aging (CCNA) to N.C., G.E., and N.P. The CCNA is supported by a grant from the Canadian Institutes of Health Research (CIHR) with funding from several partners. Part of this work was also supported by the Wilfred and Joyce Posluns Chair in Women's Brain Health and Aging from the Posluns Family Foundation, Canadian Institutes of Health Research (CIHR), Ontario Brain Institute, and Alzheimer Society of Canada to G.E. [grant WJP‐150643], the Canadian Consortium on Neurodegeneration in Aging (CCNA) Phase II to G.E. (grant CCNA 049‐04; CIHR reference number: CNA 163902), and the Jacqueline Ford Gender and Health Fund to G.E.
Calvo N, Phillips N, Bialystok E, Einstein G. Biological sex and bilingualism: Its impact on risk and resilience for dementia. Alzheimer's Dement. 2026;18:e70255. 10.1002/dad2.70255
REFERENCES
- 1. Public Health Agency of Canada . Dementia in Canada, including Alzheimer's disease: Highlights from the Canadian chronic disease surveillance system. Government of Canada. 2017.
- 2. Mielke MM. Sex and gender differences in Alzheimer's disease dementia. Psychiatr Times. 2018;35(11):14. [PMC free article] [PubMed] [Google Scholar]
- 3. Subramaniapillai S, Almey A, Rajah MN, Einstein G. Sex and gender differences in cognitive and brain reserve: implications for Alzheimer's disease in women. Front Neuroendocrinol. 2021;60:100879. [DOI] [PubMed] [Google Scholar]
- 4. Petersen RC. Mild cognitive impairment as a diagnostic entity. J Intern Med. 2004;256(3):183‐194. doi: 10.1111/j.1365-2796.2004.01388.x [DOI] [PubMed] [Google Scholar]
- 5. Stern Y, Arenaza‐Urquijo EM, Bartrés‐Faz D, et al. Whitepaper: defining and investigating cognitive reserve, brain reserve, and brain maintenance. Alzheimers Dement. 2020;16(9):1305‐1311. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Stern Y. What is cognitive reserve? theory and research application of the reserve concept. J Int Neuropsychol Soc. 2002;8(3):448‐460. [PubMed] [Google Scholar]
- 7. Sundermann EE, Biegon A, Rubin LH, et al. Better verbal memory in women than men in MCI despite similar levels of hippocampal atrophy. Neurology. 2016;86(15):1368‐1376. doi: 10.1212/WNL.0000000000002570 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Sundermann EE, Maki PM, Rubin LH, et al. Female advantage in verbal memory: evidence of sex‐specific cognitive reserve. Neurology. 2016;87(18):1916‐1924. doi: 10.1212/WNL.0000000000003288 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. for the Alzheimer's Disease Neuroimaging Initiative , Sundermann EE, Biegon A, et al. Does the female advantage in verbal memory contribute to underestimating Alzheimer's disease pathology in women versus men?. J Alzheimers Dis. 2017;56(3):947‐957. doi: 10.3233/JAD-160716 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Banks, Sarah J. , et al. Sex differences in Alzheimer's disease: do differences in tau explain the verbal memory gap?. Neurobiol Aging 107 (2021): 70‐77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Sundermann EE, Maki P, Biegon A, et al. Sex‐specific norms for verbal memory tests may improve diagnostic accuracy of amnestic MCI. Neurology. 2019;93(20):e1881‐e1889. doi: 10.1212/WNL.0000000000008467 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Calvo N, Einstein G. Steroid hormones: risk and resilience in women's Alzheimer disease. Front Aging Neurosci. 2023;15:1159435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Marian V, Spivey M. Competing activation in bilingual language processing: within‐and between‐language competition. Biling‐Lang Cogn. 2003;6(2):97‐115. [Google Scholar]
- 14. Marian V. Language‐dependent memory. In: Proceedings of the Twenty‐First Annual Conference of the Cognitive Science Society. Psychology Press; 2020:355‐360. Accessed June 20, 2025. https://www.taylorfrancis.com/chapters/edit/10.4324/9781410603494‐67/language‐dependent‐memory‐viorica‐marian [Google Scholar]
- 15. Zahodne LB, Schofield PW, Farrell MT, Stern Y, Manly JJ. Bilingualism does not alter cognitive decline or dementia risk among Spanish‐speaking immigrants. Neuropsychology. 2014;28(2):238. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Cameli L, Phillips NA, Kousaie S, Panisset M. Memory and language in bilingual Alzheimer and Parkinson patients: insights from verb inflection. In: ISB4: Proceedings of the 4th International Symposium on Bilingualism. Citeseer; 2005:452‐476. [Google Scholar]
- 17. Bialystok E, Craik FI, Freedman M. Bilingualism as a protection against the onset of symptoms of dementia. Neuropsychologia. 2007;45(2):459‐464. [DOI] [PubMed] [Google Scholar]
- 18. Bialystok E. Bilingualism: pathway to cognitive reserve. Trend Cogn Sci. 2021;25(5):355‐364. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Chertkow H, Whitehead V, Phillips N, Wolfson C, Atherton J, Bergman H. Multilingualism (but not always bilingualism) delays the onset of Alzheimer disease: evidence from a bilingual community. Alzheimer Dis Assoc Disord. 2010;24(2):118‐125. [DOI] [PubMed] [Google Scholar]
- 20. Zheng Y, Wu Q, Su F, Fang Y, Zeng J, Pei Z. The protective effect of Cantonese/Mandarin bilingualism on the onset of Alzheimer disease. Dement Geriatr Cogn Disord. 2018;45(3‐4):210‐219. [DOI] [PubMed] [Google Scholar]
- 21. Woumans, E. , Santens P, Sieben A, Versijpt J, Stevens S, Duyck W. Bilingualism delays clinical manifestation of Alzheimer's disease. Biling‐ Lang Cogn. 2015;18(3):568‐574. [Google Scholar]
- 22. Alladi S, Bak TH, Duggirala V, et al. Bilingualism delays age at onset of dementia, independent of education and immigration status. Neurology. 2013;81(22):1938‐1944. [DOI] [PubMed] [Google Scholar]
- 23. Duncan HD, Nikelski J, Pilon R, Steffener J, Chertkow H, Phillips NA. Structural brain differences between monolingual and multilingual patients with mild cognitive impairment and Alzheimer disease: evidence for cognitive reserve. Neuropsychologia. 2018;109:270‐282. [DOI] [PubMed] [Google Scholar]
- 24. Bialystok E, Craik FIM, Binns MA, Ossher L, Freedman M. Effects of bilingualism on the age of onset and progression of MCI and AD: evidence from executive function tests. Neuropsychology. 2014;28(2):290‐304. doi: 10.1037/neu0000023 [DOI] [PubMed] [Google Scholar]
- 25. Ossher L, Bialystok E, Craik FI, Murphy KJ, Troyer AK. The effect of bilingualism on amnestic mild cognitive impairment. J Gerontol Series B Psychol Sci Soc Sci. 2013;68(1):8‐12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Berkes M, Bialystok E, Craik FI, Troyer A, Freedman M. Conversion of mild cognitive impairment to Alzheimer's disease in monolingual and bilingual patients. Alzheimer Dis Assoc Dis. 2020;34(3):225. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Calvo N, Anderson JA, Berkes M, Freedman M, Craik FI, Bialystok E. Gray matter volume as evidence for cognitive reserve in bilinguals with mild cognitive impairment. Alzheimer Dis Assoc Disord. 2023;37(1):7‐12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Phillips N A, Fogarty J, Pilon R, et al. The Comprehensive Assessment of Neurodegeneration and Dementia (COMPASS‐ND) Study neuropsychology battery of the Canadian Consortium on Neurodegeneration in Aging (CCNA): Battery Development and Description. Can J Aging. Published online 2025:1‐11. [DOI] [PubMed] [Google Scholar]
- 29. Chertkow H, Borrie M, Whitehead V, et al. The comprehensive assessment of neurodegeneration and dementia: Canadian Cohort Study. Can J Neurol Sci. 2019;46(5):499‐511. [DOI] [PubMed] [Google Scholar]
- 30. Mohaddes Z, Das S, Abou‐Haidar R, et al. National neuroinformatics framework for Canadian Consortium on Neurodegeneration in Aging (CCNA). Front Neuroinform. 2018;12:85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. R Core Team . R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. 2023. URL: https://www.R‐project.org/
- 32. Stern Y. How can cognitive reserve promote cognitive and neurobehavioral health? Arch Clin Neuropsychol. 2021;36(7):1291‐1295. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Adkins‐Jackson PB, George KM, Besser LM, et al. The structural and social determinants of Alzheimer's disease related dementias. Alzheimers Dement. 2023;19(7):3171‐3185. doi: 10.1002/alz.13027 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Vang ZM, Ng E. The impacts of COVID‐19 on immigrants and the healthy immigrant effect: reflections from Canada. Prevent Med. 2023;171:107501. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Fleras, A . Immigration and settlement in Canada: diverging paths in the U.S. and Canadian contexts. Int J Can Studies. 2019;61:1‐26. doi: 10.3138/ijcs.61.1 [DOI] [Google Scholar]
- 36. Livingston G, Huntley J, Liu KY, et al. Dementia prevention, intervention, and care: 2024 report of the Lancet Standing Commission. Lancet. 2024;404(10452):572‐628. [DOI] [PubMed] [Google Scholar]
- 37. Mattsson N, Zetterberg H, Janelidze S, et al. Plasma tau in Alzheimer disease. Neurology. 2016;87(17):1827‐1835. doi: 10.1212/WNL.0000000000003246 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Hol EM, Roelofs RF, Moraal E, et al. Neuronal expression of GFAP in patients with Alzheimer pathology and identification of novel GFAP splice forms. Mol Psychiatr. 2003;8(9):786‐796. [DOI] [PubMed] [Google Scholar]
- 39. Mattsson N, Andreasson U, Zetterberg H, Blennow K, Initiative ADN. Association of plasma neurofilament light with neurodegeneration in patients with Alzheimer disease. JAMA Neurol. 2017;74(5):557‐566. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Wood Alexander M, Honer WG, Saloner R, et al. The interplay between age at menopause and synaptic integrity on Alzheimer's disease risk in women. Sci Adv. 2025;11(10):eadt0757. doi: 10.1126/sciadv.adt0757 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Tsiknia AA, Edland SD, Sundermann EE, et al. Sex differences in plasma p‐tau181 associations with Alzheimer's disease biomarkers, cognitive decline, and clinical progression. Mol psychiatr. 2022;27(10):4314‐4322. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Coulter K, Phillips NA. Bilinguals show evidence of brain maintenance in Alzheimer's disease. Biling‐ Lang Cogn. 2024;27(5):1‐10. [Google Scholar]
- 43. Bialystok E. Bilingualism: the good, the bad, and the indifferent. Biling‐ Lang Cogn. 2009;12(1):3‐11. [Google Scholar]
- 44. Ulubaev A, Lee DM, Purandare N, Pendleton N, Wu FC. Activational effects of sex hormones on cognition in men. Clin Endocrinol. 2009;71(5): 607‐623. [DOI] [PubMed] [Google Scholar]
- 45. Cherrier MM, Matsumoto AM, Amory JK, et al. The role of aromatization in testosterone supplementation: effects on cognition in older men. Neurology. 2005;64(2):290‐296. doi: 10.1212/01.wnl.0000149639.25136.ca [DOI] [PubMed] [Google Scholar]
- 46. Beer TM, Bland LB, Bussiere JR, et al. Testosterone loss and estradiol administration modify memory in men. J Urol. 2006;175(1):130‐135. doi: 10.1016/s0022-5347(05)00049-2 [DOI] [PubMed] [Google Scholar]
- 47. DeLuca V, Rothman J, Bialystok E, Pliatsikas C. Redefining bilingualism as a spectrum of experiences that differentially affects brain structure and function. Proc Natl Acad Sci. 2019;116(15):7565‐7574. doi: 10.1073/pnas.1811513116 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Calvo N, Ibáñez A, Muñoz E, García AM. A core avenue for transcultural research on dementia: On the cross‐linguistic generalization of language‐related effects in Alzheimer's disease and Parkinson's disease. Int J Geriatr Psychiatry. 2018;33(6):814‐823. doi: 10.1002/gps.4712 [DOI] [PubMed] [Google Scholar]
- 49. Amoruso L, Hernandez H, Santamaria‐Garcia H, et al. Multilingualism protects against accelerated aging in cross‐sectional and longitudinal analyses of 27 European countries. Nat Aging. 2025;5:2340‐2354. doi: 10.1038/s43587-025-01000-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting Information
Supporting Information