Protective effect of bilingualism on aging, MCI, and dementia: A community‐based study

Abstract

INTRODUCTION

Lifelong bilingualism is associated with a delayed age at onset of dementia, but evidence from community‐based studies is limited. We investigated the relationship between bilingualism and the prevalence of cognitive impairment in a linguistically diverse community.

METHODS

A door‐to‐door community study was conducted from January to December 2021 in urban Bengaluru, India. 1234 individuals aged ≥60 years participated in the study. Participants were diagnosed with no cognitive impairment (NCI), mild cognitive impairment (MCI), or dementia using established diagnostic criteria.

RESULTS

Dementia prevalence was higher in monolinguals (4.9%) than bilinguals (0.4%) (P = .001). The prevalence of MCI was also higher in monolinguals (8.5%) than bilinguals (5.3%) (P = .001). The study also revealed better cognitive function in bilinguals than monolinguals with NCI, after controlling for confounding variables.

DISCUSSION

The current study provides significant support for the protective effect of bilingualism on cognitive impairment in an urban community with extensive bilingual interactional contexts in everyday life.

Highlights

• Bilingualism has been demonstrated to protect against dementia and mild cognitive impairment in a linguistically diverse community with extensive code‐switching contexts.

• Bilingual older individuals had superior baseline cognitive performance compared to monolingual older individuals.

• Bilingualism was found to have an independent effect on general cognition after adjusting for major social determinants of health in the group without cognitive impairment.

1. BACKGROUND

The burden of dementia is estimated to be substantial in low‐ and middle‐income countries due to the rapidly aging population and high frequency of risk factors. ¹ , ² Using a life‐course perspective, researchers have investigated potential risk and protective factors for dementia. ¹ Evidence indicates that education, physical activity, and social interactions are life‐course factors that influence late‐life cognition by providing cognitive reserve. ³ In addition, emerging research suggests that bilingualism across the lifespan also provides cognitive reserve. ⁴ Lifelong switching between languages is considered to result in an improved inhibitory control mechanism in bilinguals, with benefits in executive functions including task switching and working memory. ⁵ , ⁶ , ⁷ The cognitive advantage in healthy older bilinguals is attributed to the enhanced structural integrity of the brain, improved functional connectivity, and more efficient use of neural resources resulting from the lifelong use of two or more languages. ⁸ , ⁹ , ¹⁰ The bilingual advantage in neural and cognitive mechanisms persists in individuals with dementia as well. ¹¹ , ¹² , ¹³

A large body of literature has focused on studying the protective effects of bilingualism against cognitive decline and dementia in clinical settings. Bilinguals with dementia demonstrated a delay of 4–5 years to manifest the symptoms compared to monolinguals. ¹¹ , ¹⁴ , ¹⁵ Moreover, active bilingualism also delayed the cognitive symptoms of mild cognitive impairment (MCI), a phase between normal aging and dementia. ¹⁶ , ¹⁷ , ¹⁸ This cognitive advantage in bilingual dementia is the result of brain reserve and neural compensation due to the lifelong usage of two or more languages. ¹⁹ Comparable cognitive performance was noted in bilinguals with dementia despite greater neuronal degeneration in the brain than in monolinguals with dementia. ²⁰ The modulation in functional connectivity in specific brain regions in bilingual older adults helps to cope with the pathology due to neurodegeneration. ¹³ , ²¹ Only a few community studies have investigated the role of bilingualism on the prevalence, incidence, and progression of dementia and MCI. ⁴ , ²² , ²³ The risk of developing cognitive impairment was reduced with an increase in the number of languages spoken by older individuals. ²² The potential advantage of bilingualism on late‐life cognition and protection against age‐related cognitive decline was found to be independent of childhood intelligence. ⁴ Further, the incidence and the risk of progression to MCI from healthy aging were lower in individuals who learned an additional language in early life. ²³

Several life‐course factors that contribute to cognitive reserve can potentially confound the effects of bilingualism. Education, socioeconomic status, immigration, and occupational skills are the major factors that have been implicated. ¹¹ , ¹² , ¹⁴ , ²⁴ , ²⁵ , ²⁶ However, bilingual effects on cognitive aging have been demonstrated as being independent of these confounding variables. ⁴ , ¹¹ , ¹⁸ , ²⁷ , ²⁸ , ²⁹ The consequences of bilingualism on cognitive and neural systems are also influenced by language‐related factors and bilingual experience such as language proficiency, contextual use, number of languages known, and language switching. ³⁰ , ³¹ , ³² Recently, sociolinguistic factors, including linguistic diversity and social context of bilingualism, are also recognized as affecting the cognitive effects of bilingualism. ³³ , ³⁴ , ³⁵

India offers a unique opportunity to study bilingualism, cognition, and dementia since the population is non‐immigrant and linguistically diverse, speaking 122 major languages, 22 official languages, and 1599 other languages. ³⁶ The linguistic communities in India speak more than one language, ³⁷ with extensive code‐switching and code‐mixing, and bilingualism is regarded as a societal norm. ³⁸ Bilingualism is shown to protect against dementia in India, as evidenced by clinic‐based studies, ¹¹ , ¹⁸ but this hypothesis needs to be tested in community settings as well. Hence, the aim of the current study was to investigate the potential relationship between bilingualism and the prevalence of cognitive impairment in a community‐dwelling population from urban Bengaluru, which is one of the most linguistically diverse areas in India.

2. METHODS

2.1. Study sample

This study was part of a baseline cognitive assessment of an ongoing epidemiological study to evaluate the burden, risk, and protective factors of dementia and MCI in the Indian context. ³⁹

A cross‐sectional, door‐to‐door community survey was conducted in an urban cohort residing in Jayanagar, South Bengaluru, to estimate the prevalence of dementia and MCI. Jayanagar is one of the zones in the Greater Bengaluru Municipal Corporation, which has seven wards consisting of 420 Census Enumeration Blocks (the most elementary census unit) with a total population of approximately 231,085. The study lasted from January 2021 to December 2021, during which 55 Census Enumeration Blocks were surveyed. Before the survey, the geographical boundaries of the study area were established based on the administrative population census records, ⁴⁰ , ⁴¹ and initiatives for awareness and community engagement activities were carried out for better community response and participation. ⁴² The study team consisted of social workers, clinical psychologists, neuropsychologists, speech‐language pathologists (SLPs), general physicians, epidemiologists, and neurologists who were trained in conducting field surveys and cognitive evaluations. Household enumerations were done by social workers to identify individuals who were ≥60 years old, had resided in the study area for at least six continuous calendar months, and were therefore eligible for the study. Those who met the eligibility criteria were invited to participate, and written consent for participation was obtained from both participants and their caregivers. Detailed sociodemographic information, including education, occupation, socioeconomic status, and language background, was obtained. The occupation was classified based on the National Classification of Occupation (NCO), and a modified Kuppuswamy Socioeconomic Scale ⁴³ was used for classifying socioeconomic status. Information about whether the participants were native residents of Bengaluru or migrants from other districts or states was obtained. A clinical examination was performed by a general physician, following which a set of neuropsychological tests was administered by a clinical psychologist and SLPs for dementia diagnosis. Older individuals with untreated hearing loss, severe visual impairment, intellectual impairment since childhood, major psychiatric illness, a history of head injury, and substance use disorders were excluded from the study. All examinations, which lasted for approximately 1 hour, were done at the participants’ homes. Any missing relevant data, inconsistencies, or incomplete examinations were resolved by revisiting the participants and, if necessary, conducting re‐examinations.

2.2. Data collection procedure and instruments used

2.2.1. Language status

Bengaluru is one of the most linguistically diverse urban areas in India, with at least 107 spoken languages. ³⁶ Kannada is the predominantly spoken language in Bengaluru, followed by English, Tamil, Telugu, Urdu, Hindi, Malayalam, and Marathi. The language combinations in the study cohort are presented in Table S1. It is also recognized that bilingual speakers are common, and bilingualism is considered the norm in Bengaluru. ⁴⁴ In the current study, bilinguals were defined as “those who meet the communicative demands of the self and society in their normal functioning in two or more languages in their interaction with other speakers of any or all of these languages.” ¹¹ , ⁴⁵ Language status was later confirmed on the modified version of the language use questionnaire (LUQ), which was developed and systematically validated in the Indian context. ³⁷ , ⁴⁶ , ⁴⁷ The LUQ was used to rate language proficiency in each of the known languages in response to 10 questions on a 5‐point scale (1 = cannot do; 2 = do it with great difficulty; 3 = do it with moderate difficulty; 4 = can do it, but not very well; 5 = do it very easily and very well). This included a rating of basic world knowledge (including days of the week, months, naming fruits/vegetables, counting numbers) and reading, writing, conversation, narration, metalinguistic skills, and translation skills in particular languages. The percentage of proficiency was estimated based on the ratings obtained on LUQ. The testing language chosen for cognitive assessment was the language in which the participant achieved ≥70% scores on the LUQ and was also reported as the most comfortable language by the participant. Those who achieved scores above 70% in only one language were categorized as monolinguals, and those who achieved scores above 70% in more than one language were categorized as bilinguals. ³⁷ , ⁴⁶ , ⁴⁷

RESEARCH IN CONTEXT

1. Systematic review: A review of studies suggests that using more than one language throughout life provides cognitive reserve, thereby delaying the age at onset of dementia. However, firm conclusions cannot be drawn about this assumption because of the differences in the sociolinguistic and interactional contexts of the study populations and limited community‐based evidence.

2. Interpretation: We found that bilingualism lowered the prevalence of dementia and mild cognitive impairment in older individuals. Furthermore, bilinguals had higher scores on the test of general cognitive function. The study supports previous reports on the independent protective effects of bilingualism against dementia in the Indian context.

3. Future directions: The relationship between bilingualism and the risk of cognitive impairment and dementia should be addressed further in the future, using larger longitudinal studies and populations with diverse sociolinguistic backgrounds.

2.3. Neuropsychological tests

All participants underwent interviews and cognitive assessments using tests that were standardized and validated for the local population. Addenbrooke's Cognitive Examination (ACE‐III) ⁴⁸ was administered to the participants to assess general cognitive function. Caregiver interviews included the administration of the Clinical Dementia Rating (CDR) ⁴⁹ scale to determine the severity of cognitive impairment and the Instrumental Activities of Daily Living‐Elderly (IADL‐E) ⁵⁰ to determine daily functionality.

2.3.1. Diagnosis of dementia and MCI

Sociodemographic details, clinical features, and performance on neuropsychological tests were reviewed by neurologists, SLPs, and neuropsychologists experienced in diagnosing cognitive impairment. Participants were labeled as having dementia, MCI, or no cognitive impairment (NCI) based on the scores achieved on ACE‐III, CDR, and IADL‐E. The dementia diagnosis was in accordance with the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM‐IV) criteria, and the MCI diagnosis was based on revised Petersen's criteria. ⁵¹ Participants with significant cognitive decline that interfered with daily activities were diagnosed as having dementia, ⁵² , ⁵³ and those with significant cognitive decline that did not interfere with independent functioning in daily life were diagnosed as having MCI. ⁵⁴ Participants who did not have a cognitive decline in the presence of independent functioning in daily life were considered to have NCI.

2.4. Statistical analysis

Quantitative variables with a normal distribution were expressed as mean and standard deviation (SD) and categorical variables as frequency and percentage. The clinical and demographic profiles of monolinguals and bilinguals were compared, including age, education, sex, occupation, socioeconomic status, and ACE‐III score. Comparisons between groups were performed using an independent sample t‐test/one‐way analysis of variance or Mann–Whitney U test for continuous variables, followed by Bonferroni adjustment and chi‐squared test or Fisher's exact test for categorical variables. A univariate general linear model (GLM) was applied to identify the independent effect of bilingualism on general cognitive function (as measured using ACE‐III total score). In the model, bilingualism, sex, socioeconomic status, occupational status, and diagnosis were added as fixed factors, while age and years of education were covariates. The same analysis was conducted for NCI and MCI groups separately. The interaction effect of bilingualism and other covariables were also calculated using univariate GLM. A P value of < .05 was considered statistically significant. All data were analyzed using the Statistical Package for the Social Sciences software version 16.0 (SPSS, Chicago, IL, USA).

3. RESULTS

3.1. Demographic profile of the whole cohort

The study cohort consisted of 1234 participants (Figure 1). Out of the 1234 participants included in the study, 1147 (93%) were native residents of Bengaluru, and 87 (7%) were migrants residing in Bengaluru. The mean age of the cohort was 71.18 years (SD 8.24 years; range 60 to 104 years), and the mean years of education was 11.46 years (SD 5.54 years; range 0 to 22 years). Of the total 1234 participants, 53.9% (n = 665) were women. A total of 687 (55.67%) had received more than 10 years of education. According to NCO (2004), 44.1% of the total participants were skilled workers, 19% were semiskilled workers, and 36.9% were homemakers. The majority of the total participants were from middle‐income (51.3%) and lower‐income (40.1%) families, and only 8.6% were from high‐income families. Of the overall cohort, 24 (1.9%) participants were diagnosed with dementia, and 80 (6.5%) were diagnosed with MCI.

FIGURE 1.

Study flow diagram.

More than half of the participants in our cohort were bilingual (n = 803, 65%). The overall sociodemographic details of monolingual and bilingual speakers are presented in Table 1. The bilingual cohort had more men, higher levels of education, and occupations with higher skill levels, and the majority of them were from upper‐ or upper‐middle‐class families compared to monolinguals. Bilinguals with MCI (mean age: 75.49 years; SD: 8.43 years) were older than monolinguals by 3.5 years (mean age: 72.00 years; SD: 8.36) with marginal significance (P = .068). The same trend in age was observed in bilingual (mean age: 80.67 years; SD: 2.08 years) and monolingual (mean age: 75.80; SD: 8.88 years) dementia participants (Table 1). Statistical comparison of clinical and sociodemographic characteristics between monolingual and bilingual dementia was not possible due to small numbers.

TABLE 1.

Sociodemographic characteristics of monolingual and bilingual participants with dementia, mild cognitive impairment (MCI), and no cognitive impairment (NCI).

	Overall			Dementia		MCI			NCI
Sociodemographic variables	Monolinguals (n = 431)	Bilinguals (n = 803)	P	Monolinguals (n = 21)	Bilinguals (n = 3)	Monolinguals (n = 37)	Bilinguals (n = 43)	P	Monolinguals (n = 373)	Bilinguals (n = 757)	P
Age
Mean (SD)	70.45 (7.87)	71.50 (8.28)	.029 *	75.80 (8.88)	80.67 (2.08)	72.00 (8.36)	75.49 (8.43)	.068	70.01 (7.66)	71.27 (8.29)	.025 *
[Range]	[60–100]	[60–104]		[62–100]	[79–83]	[60–92]	[60–94]		[60–94]	[60–104]
Year of education
Mean (SD)	6.45 (4.81)	14.15 (3.76)	<.001 *	3.86 (5.05)	15.33 (0.57)	5.05 (3.88)	13.26 (4.87)	<.001 *	6.74 (4.8)	14.20 (3.67)	<.001 *
[Range]	[0–17]	[0–22]		[0–15]	[15–16]	[0–12]	[0–21]		[0–17]	[0–22]
Sex, n (%)
Male	118 (27.4%)	451 (56.2%)	<.001 *	3 (14.3%)	3 (100.0%)	14 (37.8%)	31 (72.1%)	.002 *	101 (27.1%)	417 (55.1%)	<.001 *
Female	313 (72.6%)	352 (43.8%)		18 (85.7%)	–	23 (62.2%)	12 (27.9%)		272 (72.9%)	340 (44.9%)
Education, n (%)
Low education	372 (86.3%)	175 (21.8%)	<.001 *	19 (90.5%)	–	36 (97.3%)	12 (27.9%)	<.001 *	317 (85.0%)	163 (21.5%)	<.001 *
High education	59 (13.7%)	628 (78.2%)		2 (9.5%)	3 (100%)	1 (2.7%)	31 (72.1%)		56 (15.0%)	594 (78.5%)
Current occupational status, n (%)
Legislators, senior officials, and managers	3 (0.7%)	63 (7.8%)	<.001 *	–	–	–	–	<.001 *	3 (0.8%)	63 (8.3%)	<.001 *
Professionals	13 (3.0%)	328 (40.8%)		–	3 (100%)	–	17 (39.5%)		13 (3.5%)	308 (40.7%)
Technicians and associated professionals	14 (3.2%)	73 (9.1%)		–	–	2 (5.4%)	5 (11.6%)		12 (3.2%)	68 (9.0%)
Clerks	13 (3.0%)	37 (4.6%)		1 (4.8%)	–	1 (2.7%)	3 (7.0%)		11 (2.9%)	34 (4.5%)
Service workers and shop and market workers	9 (2.1%)	42 (5.2%)		–	–	2 (5.4%)	4 (9.3%)		7 (1.9%)	38 (5.0%)
Skilled agricultural and fishery workers	14 (3.2%)	14 (1.7%)		–	–	–	–		14 (3.8%)	14 (1.8%)
Craft and related trade workers	23 (5.3%)	19 (2.4%)		1 (4.8%)	–	2 (5.4%)	1 (2.3%)		20 (5.4%)	18 (2.4%)
Plant and machine operators and assemblers	24 (5.6%)	34 (4.2%)		1 (4.8%)	–	2 (5.4%)	2 (4.7%)		21 (5.6%)	32 (4.2%)
Elementary occupations	41 (9.5%)	15 (1.9%)		4 (19.0%)	–	6 (16.2%)	3 (7.0%)		31 (8.3%)	12 (1.6%)
Workers not classified by occupations	277 (64.3%)	178 (22.2%)		14 (66.7%)	–	22 (59.5%)	8 (18.6%)		241 (64.6%)	170 (22.5%)
Socioeconomic status, n (%)
Upper (I)	0	106 (13.2%)	<.001 *	–	1 (33.3%)	–	7 (16.3%)	<.001 *	–	98 (12.9%)	<.001 *
Upper middle (II)	27 (6.3%)	403 (50.2%)		–	2 (66.7%)	1 (2.7%)	16 (37.2%)		26 (7.0%)	385 (50.9%)
Lower middle (III)	55 (12.8%)	148 (18.4%)		2 (9.5%)	–	5 (13.5%)	9 (20.9%)		48 (12.9%)	139 (18.4%)
Upper lower (IV)	243 (56.4%)	134 (16.7%)		10 (47.6%)	–	24 (64.9%)	9 (20.9%)		209 (56.0%)	125 (16.5%)
Lower (V)	106 (24.6%)	12 (1.5%)		9 (42.9%)	–	7 (18.9%)	2 (4.7%)		90 (24.1%)	10 (1.3%)

Note: The sociodemographic characteristics could not be compared between monolingual and bilingual dementia subgroups due to the small sample size. “Workers not classified by occupations” include homemakers, persons who were unemployed , and occupations not identifiable as per NCO classification.

Abbreviation: SD, standard deviation.

^*Significant, P < .05.

3.2. Comparison of prevalence of dementia and MCI between monolinguals and bilinguals

The prevalence of dementia was compared between monolingual and bilingual speakers. The results revealed that dementia prevalence was higher in monolingual speakers (21 [4.9%]) than in bilinguals (3 [0.4%]) (P = .001). The prevalence of MCI was also higher in monolinguals (37 [8.5%]) than in bilinguals (43 [5.3%]) (P = .001).

Since educational levels differed between monolingual and bilingual groups, the participants were further divided into low‐education (≤10 years of education) and high‐education (>10 years of education) subgroups (Table 2). In the low‐education group, the prevalence of dementia was higher (19 [5.1%] out of 372) in monolinguals, while none (0 out of 175) of the bilingual speakers had dementia. Similarly, in the high‐education group, monolinguals had a higher prevalence of dementia (2 [3.4%] out of 59) in comparison to bilinguals (3 [0.4%] out of 628). Bilingualism was not associated with a difference in the prevalence of MCI in either educational group (Table 2). On investigating the overall role of education in the cohort, the low‐education group had a higher prevalence of MCI (48 [8.8%] out of 547) and dementia (19 [3.5%] out of 547) compared to the high‐education group (32 [4.7%] and 5 [0.7%] out of 687 respectively). In the low‐education group, 87.8% (480 out of 547) were with no cognitive impairment (NCI) in comparison to 94.6% (650 out of 687) in the high‐ education group (Table 3). This suggests a higher proportion of individuals without cognitive impairment in the high‐ education group compared to the low‐education group.

TABLE 2.

Prevalence of dementia and MCI in monolingual and bilingual participants with low and high levels of education.

	Low‐education group			High‐education group			Overall
Diagnosis	Monolinguals (n = 372)	Bilinguals (n = 175)	P	Monolinguals (n = 59)	Bilinguals (n = 628)	P	Monolinguals (n = 431)	Bilinguals (n = 803)	P
Dementia	19 (5.1%)	0 (0%)	.01 *	2 (3.4%)	3 (0.4%)	.006 *	21 (4.9%)	3 (0.4%)	.001 *
MCI	36 (9.7%)	12 (6.9%)	.281	1 (1.7%)	31 (4.9%)	.26	37 (8.5%)	43 (5.3%)	.001 *
NCI	317 (85.2%)	163 (93.1%)	.008 *	56 (94.9%)	594 (94.6%)	.092	373 (86.3%)	757 (94.1%)	.001 *

Abbreviations: MCI, mild cognitive impairment; NCI, no cognitive impairment.

^*Significance, P < .05.

TABLE 3.

Comparison of prevalence of dementia and MCI between low‐educated (≤10 years) and high‐educated (>10 years) elderly.

	Low educated (n = 547)	High educated (n = 687)	P
Dementia	19 (3.5%)	5 (0.7%)	<.001 *
MCI	48 (8.8%)	32 (4.7%)	.003 *
NCI	480 (87.8%)	650 (94.6%)	<.001 *

Abbreviations: MCI, mild cognitive impairment; NCI, no cognitive impairment.

^* Significant, P< .05.

Further, we investigated the difference in prevalence of cognitive impairment among speakers of one versus two languages versus three or more languages (Table 4). Individuals who spoke only one language had a higher prevalence of dementia compared to people who spoke two (P = .003) and three or more languages (P < .001). Among persons with MCI, individuals who spoke one language had a trend toward a higher prevalence of MCI compared to people who spoke two (P = .078) and three or more languages (P = .054).

TABLE 4.

Prevalence of dementia and MCI across different subgroups classified according to number of languages used in daily life.

Diagnosis	One language (n = 431, 35%)	Two languages (n = 305, 25%)	≥ Three languages (n = 498, 40%)	P	One vs two languages	One vs ≥ three languages	Two vs ≥ three languages
Dementia	21 (4.9%)	1 (0.3%)	2 (0.4%)	<.001 *	0.003 *	<0.001 *	0.82
MCI	37 (8.6%)	16 (5.2%)	27 (5.4%)		0.078	0.054	0.90
NCI	373 (86.5%)	288 (92.3%)	469 (95.5%)		0.013 *	<0.001 *	<0.001 *

Abbreviations: MCI, mild cognitive impairment; NCI, no cognitive impairment.

^* Significant, P < .05.

3.3. Comparison of general cognitive function between monolingual and bilingual speakers with NCI and MCI

Results revealed higher ACE‐III total score and scores of attention, memory, fluency, language, and visuospatial subdomains among bilinguals compared to monolingual speakers with NCI and MCI (P < .05). Table 5 shows a comparison of general cognitive function measured using ACE‐III between monolinguals and bilinguals with NCI and MCI.

TABLE 5.

Comparison of general cognitive function scores between monolingual and bilingual speakers with NCI and MCI.

	NCI				MCI
	Monolingual (n = 373)	Bilingual (n = 757)	t	P	Monolingual (n = 37)	Bilingual (n = 43)	t	P
ACE‐III total	90.79 (7.087)	94.13 (4.41)	9.711	<.001 *	67.78 (15.99)	80.02 (9.16)	4.272	<.001 *
Attention (18)	17.3 (1.364)	17.59 (0.828)	4.516	<.001 *	14.43 (3.51)	16.09 (1.56)	2.797	.006 *
Memory (26)	23.39 (2.925)	24.58 (1.76)	8.486	<.001 *	15.16 (5.12)	18.58 (5.43)	2.882	.005 *
Fluency (14)	10.89 (1.862)	11.58 (1.748)	6.101	<.001 *	7.68 (3.11)	9.00 (2.15)	2.240	.028 *
Language (26)	25.19 (1.904)	25.54 (1.239)	3.674	<.001 *	21.86 (6.02)	24.40 (1.80)	2.624	.010 *
Visuospatial (16)	14.02 (2.784)	14.85 (1.843)	5.969	<.001 *	8.65 (4.18)	11.95 (3.43)	3.882	<.001 *

Abbreviations: ACE‐III, Addenbrooke's cognitive examination‐III; MCI, mild cognitive impairment; NCI, no cognitive impairment.

^*Significant, P < .05.

Univariate GLM analysis for the whole group showed that bilingualism was independently associated with the ACE‐III total score [F (1,1210) = 7.812, P = .005, η_p ² = .006] after adjusting for age, sex, years of education, occupation, diagnosis, and socioeconomic status. Additionally, age was negatively associated with cognition [F (1,1210) = 22.025, P < .001, η_p ² = .018], and years of education was positively associated with cognition [F (1,1210) = 10.244, P = .001, η_p ² = .008]. As expected, the strongest association was found between the diagnosis groups on cognition [F (2,1210) = 450.261, P < .001, η_p ² = .427]. Among individuals with NCI, we not only found an independent effect of bilingualism [F (1,1111) = 4.603, P = .032, η_p ² = .004] but also demonstrated that there were no interaction effects between bilingualism and other potential confounding factors. More specifically, there was no interaction effect of bilingualism with age [F (1,1110) = .032, P = .858, η_p ² < .001], sex [F (1,1110) = .030, P = .864, η_p ² < .001], years of education [F (1,1110) = 1.052, P = .305, η_p ² = .001], socioeconomic status [F (3,1108) = .575, P = .632,  η_p ² = .002], and occupational status [F (9,1102) = 1.102, P = .358, η_p ² = .009]. Among individuals with MCI, there was no significant difference between bilinguals and monolinguals in the GLM analysis [F (1,64) = 2.844, P = .097, η_p ² = .043]. Furthermore, in MCI, there was no interaction effect of bilingualism with age [F (1,63) = .038, P = .847, η_p ² = .001], sex [F (1,63) = .325, P = .571, η_p ² = .005], socioeconomic status [F (3,61) = 1.508, P = .221, η_p ² = .069], or occupational status [F (6,58) = .952, P = .466, η_p ² = .090]. 

To explore the influence of the number of languages spoken on cognitive performance, we compared ACE‐III in individuals who spoke one versus two versus three or more languages. In the whole cohort, there was a main effect of the number of languages [F(2,1227) = 63.486, P < .001], indicating that persons who spoke only one language (mean ACE‐III score 87.19, SD = 13.42) performed poorly compared to individuals who spoke two languages (mean ACE‐III score 92.77, SD = 5.87, P < .001, Cohen's d = .608) as well as three or more languages (mean ACE‐III score 93.58, SD = 5.69, P < .001, Cohen's d = .708). There was no difference in general cognition between individuals who spoke two languages versus three or more languages (P = .516).

4. DISCUSSION

In this study, we found that the prevalence of dementia and MCI was higher in monolinguals compared with bilingual speakers. In our community‐based cohort, older bilinguals with NCI had better cognitive functioning compared to monolinguals independent of age, sex, education, socioeconomic status, and occupational status. Overall findings indicate that bilingualism promotes healthy aging and protects against cognitive decline and dementia, which might be due to additional cognitive benefits noted in bilingual speakers. ⁵⁵ This is attributed to increased neuronal plasticity and better cognitive reserve resulting from practicing more than one language across the lifespan. ¹³ , ⁵⁶

This is the first study investigating the impact of bilingualism on the prevalence of both dementia and MCI in a large population. We addressed the key methodological problems in earlier clinic and community studies to a greater extent in the current study. The door‐to‐door survey offered high‐quality data from a single urban community, eliminating the methodological concerns of clinic‐based studies, such as selection bias and complementary class issues, that is, failure to represent individuals who do not seek help despite memory complaints and those who report to a clinic without any cognitive impairment. ⁵⁷ The participants underwent a comprehensive cognitive examination through culturally and linguistically adapted and validated tests. ⁴⁸ The LUQ was used as a reliable measure of bilingualism to capture the proficiency of each spoken language with respect to understanding, naming, reading, speaking, and writing. ³⁷ , ⁴⁶ The target language for cognitive testing was decided based on the composite score obtained for each language on the LUQ. The clinician (examiner) was also proficient in the target language used for testing. ⁴⁷ Other factors that impact cognitive status in older age, such as education, sex, occupation, and socioeconomic status, were accounted for, as potential confounding factors. Although second language knowledge is associated with immigration status in the West, the non‐immigrant population in India provides an ideal opportunity to explore the association between bilingualism and dementia.

The findings of our study indicated a favorable effect of bilingualism on late‐life cognition in an urban community, demonstrating a significantly lower prevalence of dementia in bilinguals (0.4%) than monolinguals (4.9%). Similarly, the prevalence of MCI was also significantly lower in bilinguals (5.3%) than monolinguals (8.5%). This positive impact of bilingualism on the prevalence of dementia was retained regardless of high or low education status. We also found no difference in the prevalence of MCI and dementia between those who spoke two and three or more languages. Similarly, speaking two or more languages was beneficial to cognition compared to being monolingual status, but there was no added benefit to knowing three or more languages. The findings of the current study are consistent with previous literature showing bilingual effects on the prevalence of dementia and cognitive impairment in community settings. ⁴ , ²² , ²³ , ⁵⁸ The prevalence of preclinical Alzheimer's disease was much lower in middle‐age early bilinguals (Spanish‐Euskera) than monolinguals (Spanish only) residing in Basque Country, Spain. ⁵⁹ Furthermore, active use of multiple languages was beneficial in the Luxembourg cohort of non‐immigrants, demonstrating a lower likelihood of developing cognitive impairment. ⁵⁸

Previous research revealed superior executive function and experience‐dependent plasticity in older bilinguals, resulting in a greater cognitive reserve. ⁵⁶ In our cohort, healthy older bilinguals with NCI had better general cognitive abilities compared to monolinguals, independent of the social determinants of health, such as age, sex, education, socioeconomic status, and occupation. Bilingual speakers scored better on attention, memory, fluency, language, and visuospatial subdomains of ACE‐III compared to monolinguals. Similarly, other studies also reported higher scores on general cognitive function measured using Mini‐Mental State Examination in healthy bilingual older adults in comparison to matched monolinguals. ⁶⁰ , ⁶¹ The superior general cognitive function in healthy older bilinguals than monolinguals implies functional advantage, that is, advantage of lifelong exposure and practicing more than one language. The constant switching between languages and inhibition of a certain language activates executive function and attention control mechanisms in healthy older bilinguals. The observed effect of bilingualism on general cognitive function in individuals with MCI also indicates a potential advantage, similar to the significant difference in ACE‐III score reported in another clinical study conducted in India. ¹⁸

Several sociocultural factors, such as immigration status, ethnicity, culture, and interactional contexts, affect the relationship between bilingualism and the risk of dementia, resulting in variable results across societies. ²⁶ The bilingual population in India is characterized by a normative practice of bilingualism and extensive code switching in interactional contexts. According to recent models of bilingualism, such as the Systems Framework of Bilingualism ³³ and the Adaptive Control Hypothesis, ⁶² throughout their lifespan bilingual speakers use contextual cues and switch to an appropriate language while interacting with diverse language groups, which results in an improvement in cognitive control mechanism. In India, there exists a long‐term social interaction in a rich bilingual environment that shapes the anatomic and functional substrates of cognition, possibly helping bilinguals to maintain late‐life cognitive performance and protect against cognitive decline. ¹³ , ³² Therefore, a reduced prevalence of dementia and MCI with a better baseline cognitive performance in bilinguals is a credible observation indicating that bilingualism is a protective factor for dementia in India.

A number of factors, including socioeconomic status, primary lifetime occupational skills, and education, have been identified as confounding variables that interfere with understanding the actual effect of bilingualism. Education is an independent protective factor for dementia and is a major confounder when exploring the protective effects of bilingualism. ²⁶ Since lower education was associated with higher prevalence of MCI and dementia in the current study, participants were categorized into low (≤10 years of education) and high (>10 years of education) education groups to assess the independent benefits of bilingualism on dementia prevalence. A significant observation was that dementia prevalence was higher in monolinguals of both educational levels compared to bilinguals. Studies also suggest levels of occupational complexity are associated with improved late‐life cognitive abilities promoting healthy aging and cognitive reserve. ⁶³ Nevertheless, the present study found no moderating effects of occupation on the association between bilingualism and general cognitive function among healthy older individuals. The participants in this study included older individuals with engagement in occupations requiring different levels of complexity; however, all were similarly exposed to substantial bilingual interactional contexts on a daily basis in social settings. These findings are in line with previous studies reporting the effect of bilingualism on dementia onset independent of education and occupation. ¹¹ , ¹⁸ Moreover, our results indicate that bilingualism and socioeconomic status do not interact with respect to cognition in old age in our study context. Despite the fact that bilingualism in our study cohort independently influenced cognition in old age and promoted healthy aging, this is likely to vary depending on the context.

There are some limitations to this study. It has been reported that self‐rated proficiency scales either overestimate or underestimate one's proficiency. ⁶⁴ As this is a large‐scale community study, objective measures of language use, frequency of language switching, and age of acquisition were not documented. However, an extensively validated LUQ was used to measure bilingualism. Previous studies have found moderate to high correlations between self‐rated proficiency scales and objective measures. ⁶⁵ Bilingualism was treated as a dichotomous variable in the current study, since the aim was to determine the prevalence of MCI and dementia between the two groups. Finally, since the current study is cross‐sectional, the association between bilingualism and the risk of cognitive impairment and dementia needs to be investigated in the future through larger longitudinal studies.

To conclude, the current study is the first of its kind that investigated the prevalence of cognitive impairment and dementia in a linguistically diverse urban community using culturally and linguistically adapted screening tools. The findings of the study reveal that actively speaking more than one language provides protection against cognitive impairment in linguistically diverse communities with enriched language interactions. Our study has implications for public health interventions, highlighting the importance of promoting bilingualism and multilingualism as potential cognitive reserve factors that may delay or slow down cognitive decline.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest. Author disclosures are available in the Supporting information.

CONSENT STATEMENT

This study was approved by the Institutional Research Ethics Committee of the National Institute of Mental Health and Neurosciences, Bengaluru. Formal written informed consent for participation was obtained from all participants and their caregivers or legally accepted representatives.

Supporting information

S1. Language combinations of different language subgroups

Supporting Information

Venugopal A, Paplikar A, Varghese FA, et al. Protective effect of bilingualism on aging, MCI, and dementia: A community‐based study. Alzheimer's Dement. 2024;20:2620–2631. 10.1002/alz.13702