Lower Levels of Education Are Associated with Cognitive Impairment in the Old Order Amish

Abstract

Background:

Lower education has been reported to be associated with dementia. However, many studies have been done in settings where 12 years of formal education is the standard. Formal schooling in the Old Order Amish communities (OOA) ends at 8^th grade which, along with their genetic homogeneity, makes it an interesting population to study the effect of education on cognitive impairment.

Objective:

The objective of this study was to examine the association of education with cognitive function in individuals from the OOA. We hypothesized that small differences in educational attainment at lower levels of formal education were associated with risk for cognitive impairment.

Methods:

Data of 2,426 individuals from the OOA aged 54–99 were analyzed. The Modified Mini-Mental State Examination (3MS-R) was used to classify participants as CI or normal. Individuals were classified into three education categories: <8, 8, and >8 years of education. To measure the association of education with cognitive status, a logistic regression model was performed adding age and sex as covariates.

Results:

Our results showed that individuals who attained lowest levels of education (<8 and 8) had a higher probability of becoming cognitvely impaired compared with people attending >8 years (OR = 2.96 and 1.85).

Conclusion:

Even within a setting of low levels of formal education, small differences in educational attainment can still be associated with the risk of cognitive impairment. Given the homogeneity of the OOA, these results are less likely to be biased by differences in socioeconomic backgrounds.

Keywords

Amish cognitive function cognitive impairment dementia education logistic regression Modified Mini-Mental State Examination (3MS)

INTRODUCTION

Dementia is characterized by impairments in the ability to perform daily life activities due to the deterioration of cognitive function. Cognitive functioning is a good indicator of independence and survival in older adults and can be described as the group of mental abilities such as thinking, learning, reasoning, decision making, attention, remembering, and problem solving [1 –3]. Alzheimer’s disease (AD) is the most common form of dementia accounting for 60%to 70%of all cases. Other forms of dementia include vascular, Lewy body, and frontotemporal dementia [4]. The projected number of people with AD in the United States will be 5.8 million in 2020 and this number is expected to increase to 13.8 million by 2050 [5, 6]. Clinical manifestations of AD include cognitive decline, loss of functional independence and a range of behavioral and psychiatric symptoms [7]. Besides age, genetic and environmental factors such as smoking, alcohol use, high blood pressure, obesity, and low physical activity are also determinants of dementia [8].

Mortimer [9] proposed that psychosocial factors such as years of formal education may be associated with dementia, suggesting that it may be a protective factor against diagnosable dementia. Zhang et al. [10] published one of the first large epidemiological studies investigating the association between education and dementia in a group of 5,000 individuals from Shanghai, China. They reported that in the older group of the sample with lack of formal education or lower levels (less than 6 years) the prevalence of dementia was higher. Subsequently, Stern and colleagues found that among older adults in a sample population from North Manhattan NY, the relative risk of developing dementia in those with less than 8 years of education compared to those with more than 8 years of education was 2.02 (95%confidence interval 1.33 to 3.06) [11]. Since then, though there have been some studies with different results, the majority of studies have confirmed this relationship [12]. Significantly, the results of two meta-analyses have established that a low level of education is a risk factor for dementia while a high education level is a protective factor [12 –14]. Furthermore, a recent study used Mendelian randomization analysis with 1,271 SNPs previously associated with educational attainment to confirm the causal relationship of education with AD [15].

The concept of cognitive reserve, broadly defined as the way in which the brain is able to maintain cognitive function despite age or disease related pat-hology, has been cited to explain the association between education and dementia [16 –19]. Bennett and colleagues also reported that higher educational attainment was associated with a reduced presence of neuritic plaques (amyloid-β (Aβ) peptide neighboring dystrophic neurites) and diffuse plaques (agg-regated Aβ peptide) among participants from the Religious Orders Study [20], leading to speculation that perhaps cognitive reserve protects against neurodegeneration.

Given that genetics is a strong risk factor for dementia, the relative genetic homogeneity of the Amish community [21] makes it easier to investigate non-genetic factors that may be associated with cognitive impairment such as education. Although the Amish possess relatively low educational attainments, they are not burdened by illiteracy. By choice, formal schooling in most old-order Amish (OOA) communities ends at the 8^th grade and is geared toward learning basic math, reading, and writing [22]. However, although most of them leave the school after completing 8^th grade, we found differences in educational attainments. Some Amish leave school before completing 8^th grade once they reach their 17^th birthday and some of them attend 9^th grade if they are still not 17 years old. As a group, the OOA are literate and productive within a variety of occupations.

The protective effects of education are robust [18 , 24], but have typically been studied in settings where compulsory education is 12 years. Less is known about the impact of education on cognitive abilities in settings with limited formal educational opportunities [25 –28]. The objective of this study is to examine the association of education with cognitive function in individuals from the genetically homogenous OOA communities using a cohort of 2,426 subjects. This sample size is almost 5 times larger than previous studies of education and dementia in the Amish. Our primary hypothesis is that differences in educational attainment, even within the narrow range observed in the OOA, will be associated with variations in risk for cognitive impairment based on the Revised Modified Mini-Mental State Examination (3MS-R) score [29].

METHODS

Study population

The Collaborative Amish Aging and Memory Pro-ject (CAAMP) is a population-based study of aging and age-related traits (including AD) conducted by investigators at University of Miami and Case Western Reserve University in the OOA communities of Indiana and Ohio initiated in 1996. All studies have been undertaken after Institutional Review Board (IRB) review and approval. Participants were residents from the Amish communities located in Holmes County, Ohio and Adams, Elkhart, and LaGrange counties in Indiana. All participants were recruited and ascertained as previously described by Pericak-Vance et al. [30], Hahs et al. [21], and Edwards et al. [31, 32] using information from Amish communities, published public directories, and referrals from previously enrolled participants.

The initial screening visit obtained basic sociodemographic information, family history, medical history, and tested memory and orientation using the revised version of the 3MS-R [29]. All individuals were enrolled by study personnel who went door to door and conducted the various cognitive, functional, and behavioral examinations. Based on the results of the screening visit, interviewers revisited participants to obtain a blood sample for DNA and RNA extraction and to conduct follow-up assessments [32, 33]. For individuals age 65 to 79, only those with 3MS-R scores < 87 (indicative of potential cognitive impairment) were revisited. As a result of this two-visit strategy, not all individuals with information on education and results of the 3MS-R testing have blood samples available for genetic testing. Informed consent was obtained from each participant or a proxy which was usually a close family member.

Procedures

We analyzed data from 2,426 individuals aged 54–99 whose sociodemographic information and raw scores from the new version of the 3MS-R were available. Cut-off scores to classify participants as cognitively impaired or normal were determined by stratifying participants into six age groups (54–69, 70–74, 75–79, 80–84, 85–89, and 90–99) to take into account the effects of age on the 3MS-R score as established in Tschanz et al. [29]. Then, using the mean and standard deviation of the normative sample for each age group and the 9^th percentile as cut-off point (sensitivity ∼76%; specificity ∼86%) [29], 337 individuals were classified as possibly cognitively impaired (cases) and 2,089 as cognitively normal (controls). We also used educational attainment to stratify participants into three groups: less than 8 years of education (low); 8 years of education (medium), and greater than 8 years of education (high). The cutoff of 8 years was used because, as mentioned before, the majority of the Amish complete 8 years of schooling (more details in the Introduction section) before they start working, particularly in agricultural, manufacturing, and skilled labor jobs. While approximately 75%of individuals in the sample completed 8^th grade, some of them reached age 17 and left school before completing the 8^th grade. In addition, some individuals repeated grades prior to reaching age 17 and left school before completing the 8^th grade; we did not count repeating a grade as a year of school. Only participants that went on to achieve a ninth-grade level of education or higher were categorized in the high educational attainment group (more than 8 years).

Statistical analysis

Initially we evaluated age, sex, and education level to inform subsequent analyses. Age differences between cases and controls were evaluated using an independent-samples t-test. We tested whether sex and education level were related to cognitive status using Chi-Square tests. Subsequently, to measure the association of education with cognitive status (impaired versus non-impaired) a logistic regression model was performed adding age and sex as covariates to control for differences in these parameters.

Given the previously reported association of higher education with higher 3MS-R scores among cognitively normal people [29, 34], a secondary goal of this study was to see if this pattern still holds at low levels of formal education in our Amish population. To accomplish this, we analyzed the group of cognitively normal individuals (2,089), sorting them into three groups based on the three education levels and tested for differences in 3MS-R scores and age using a t-test as follows: level 1 versus level 3, level 1 versus level 2, and level 2 versus level 3. We also checked whether 3MS-R score was related to sex using a Chi-Square test. Finally, using linear regression, we measured the association of 3MS-R score (dependent variable) with level of education (independent variable), adding age and sex as covariates. Statistical tests were conducted using the Statistical Analysis System 9.3 (SAS 9.3) with significance threshold set to 0.05.

RESULTS

A detailed description of the sample by education, age, and sex is presented in Table 1. Of 337 cognitively impaired individuals there were 121 men (36%) and 216 women (64%). The number of cognitively normal individuals was 2,089, of which 848 (41%) were men and 1,241 (59%) were women. Chi square test results showed that there was no relationship between the proportion of men/women and cognitive status (p = 0.10). Mean age among the 337 cognitively impaired individuals (mean, 78.82; SD, 7.16) was significantly greater than that for 2,089 cognitively normal (mean, 73.76; SD, 6.99); (t [2424] = –12.29, p < 0.0001). The levels of education for the cognitively normal cohort were higher than the cognitively impaired cohort (see Fig. 1). Chi-square test of differences in education was statistically significant at the 0.05 alpha level; p < 0.0001. Noticeably, the proportion of individuals with more than 8 years of education in the cognitively normal group (18.09%) was more than double the proportion of individuals in the cognitively impaired group (8.01%).

Table 1

Demographics of final data set used for the analysis

	Cognitively Impaired	%	Cognitively Normal	%	Total	%
Number of participants	337	13.9%	2089	86.1%	2426
Number of Males	121	36%	848	41%	969	40%
Number of Females	216	64%	1241	59%	1457	60%
Mean Age	78.82		73.76		74.47
Age SD	7.16		6.99		7.23
Education
<8 y	52	15%	142	7%	194	8%
8 y	258	77%	1569	75%	1827	75%
>8 y	27	8%	378	18%	405	17%

Significantly, the age among cognitively impaired (mean, 78.82; SD, 7.16) exceeded that for cognitively normal (mean, 73.76; SD, 6.99); t [2424] = –12.29, p < 0.0001, and therefore we controlled for age and sex in the logistics regression model.

Fig. 1

Levels of education by cognitive status. Levels of education were higher in cognitively normal individuals. Chi-square test of differences in education was statistically significant at the 0.05 alpha level; p < 0.0001. Noticeably, the proportion of individuals with more than 8 years of education in the cognitively normal group (18.09%) was more than double the proportion of individuals in the cognitively impaired group (8.01%).

After controlling for age and sex, the logistic regression model (Table 2) showed that participants with less than 8 years of education were 2.96 times as likely to be cognitively impaired than those with more than 8 years education (p < 0.0001, 95%confidence interval for the odds ratio: 1.75–5.00). Participants with 8 years education were 1.85 times as likely to be cognitively impaired than those with more than 8 years education (p = 0.0042, 95%CI: 1.22–2.82). These results suggest a dose-response effect, where increasing levels of education reduce the risk of cognitive impairment, even in this cohort of individuals with lower levels of formal educational opportunity. Age remained in the final model as a risk factor (p < 0.0001 and positive sign coefficient), while sex was excluded for not adding a significant contribution to the model. As an example, Table 3 shows the estimated probabilities of becoming cognitively impaired for a 70-year-old individual with different levels of education, calculated using the model equation (Equation 1). Noticeably, the probability of becoming cognitively impaired is 2.6 times for a person with less than 8 years of education versus more than 8 years (see Table 3).

Table 2

Odds Ratio Estimates (Effect of Education on CI)

Effect	Point Estimate	95%Wald Confidence Limits
Age	1.09	1.07	1.11
<8 y versus >8 y	2.96	1.75	5.00
8 y versus >8 y	1.85	1.22	2.82

Odds Ratio Estimates of logistics regression model (2.96 and 1.85) and 95%confidence intervals show that 8 years and >8 years of education exert a protective effect against cognitive impairment compared to <8 years of education. Results are adjusted for age and sex.

Table 3

Model prediction of cognitive impairment for a 70-year-old individual

Level of Education	Log-odds	Odds	Probability
<8 y	–1.86	0.16≈1 : 6.43	0.13
8 y	–2.33	0.10≈1 : 10	0.09
>8 y	–2.95	0.05≈1 : 20	0.05

Model prediction of cognitive impairment for a 70-year-old individual shows that the probability of CI is 2.6 times (0.13 / 0.05) for a person with < 8 years versus > 8 years of education.

$\begin{matrix} Log - odds of CI = - 9.035 + 0.087 * Age + 1.084 * EDU 1 + 0.616 * EDU 2 \end{matrix}$ (1) (dummy variables notation: EDU1 = 1 if the person achieved less than 8 years of education; = 0 otherwise. EDU2 = 1 if the person achieved 8 years of education; = 0 otherwise.)

For the evaluation of the selected regression model, we calculated the goodness of fit (GOF) using different methods. Results displayed in Supplementary Table 2 indicate that this is a good model with no lack of fit.

To evaluate the association of higher education with higher 3MS-R scores (our secondary goal), we proceeded to analyze the group of subjects classified as cognitively normal, sorted into the education level sub-groups. The 3MS-R scores box plots for each sub-group are shown in Fig. 2. The mean 3MS-R scores were 87.22 (<8 years of education), 89.61 (8 years), and 91.14 (>8 years), and the differences between them were statistically significant (p < 0.0001). As observed in Fig. 2, higher 3MS-R scores were related to higher education levels. Age was also found to be significantly different (p < 0.0001), but unlike 3MS-R score, this relationship was negative with higher levels of age associated with lower levels of education (Fig. 2); for example, the mean age of individuals with >8 years of education was 71.24 compared to 78.16 for participants with <8 years of education. In regards to sex, the proportions of men/women were 35/65%, 41/59%, and 39/61%for <8, 8, and >8 years of education respectively, and they were not related to the level of education (Chi-square test p = 0.2197). Supplementary Table 1 depicts 3MS-R score, sex, and age parameters sorted by education level within the group of cognitively normal individuals.

Fig. 2

Distribution of 3MS-R score and age by education level (EDU) in the cognitively normal group. Higher 3MS-R scores were associated to higher education levels while higher levels of age were associated with lower levels of education.

To quantify the relationship found between level of education and 3MS-R score in the cognitively normal group, we calculated the linear regression model adding age and sex as covariates (equation 2). The sex coefficient was not statistically significant; therefore, it was not included in the final model. Notably, individuals of the same age had an average 3MS-R score that was 1.96 lower in the group with less than 8 years of education compared to the group with more than 8 years. Negative coefficients for the dummy variables EDU1 and EDU2 (Equation 2) indicate that the predicted 3MS-R score is lower for people with less than 8 years and 8 years compared to more than 8 years of education. Finally, as seen in Supplementary Figure 1, an analysis of covariance, shows the effect of education in the 3MS-R score while controlling for age.

$\begin{matrix} 3 MS score = 111.25 - 0.28 * AGE - 1.96 * EDUI - 0.75 * EDU 2 \end{matrix}$ (2)

(dummy variables notation: EDU1 = 1 if the person achieved less than 8 years of education; = 0 otherwise. EDU2 = 1 if the person achieved 8 years of education; = 0 otherwise)

DISCUSSION

Based on a logistic regression model in which we adjusted for age and sex differences between cases and controls, our results show that individuals in the Amish community who attained the lowest level of education (less than eight years) had a higher probability of becoming cognitively impaired (probable dementia) compared with people attending more than 8 years of schooling (OR = 2.96 CI: 1.75–5.00). Moreover, even the intermediate level of education (8 years) showed a higher risk compared to the referenced level of more than eight years (OR = 1.85, CI: 1.22–2.82). In this model, only age and education remained in the final model and sex was not included. This is not surprising given that our univariate analysis did not yield a significant difference in the men/women proportion between cases and controls.

It is interesting to point out that not all suspected dementia cases had undergone clinical evaluation at the time of this study; however, among those evaluated, there were 130 people with a clinical consensus diagnosis of dementia. 106 (82%) of them were under the 3MS 9^th percentile cut off resulting in a sensitivity close to what we expected. Education level proportion was 18%(<8), 75%(8), and 7%(>8), very similar to the distribution of the 337 cognitively impaired individuals classified as such, from the entire dataset, using the 3MS (15%, 77%, and 8%. See Table 1).

Our findings are consistent with Johnson et al. [33], who used a cohort of 516 people and the Mini-Mental Status Exam (MMSE) [35] to identify individuals with possible cognitive impairment in the same or a similar population of OOA. It is worth mentioning that our study had a substantially larger dataset and employed the 3MS-R test, an expanded version of the MMSE and 3MS which has demonstrated sound psychometric qualities including increased sensitivity in screening for dementia [29].

Our findings support the cognitive reserve hypothesis as higher levels of education were associated with better cognitive function, providing a significant protective effect against the risk of probable dementia in the rather homogeneous genetic Amish community. Our results are significant as they further support the importance of education as one of several potentially protective factors that increase resilience against dementia [36]. Many of these protective factors are rooted in lifestyles, several of which may be modifiable [37]. While this effect may be tied to other factors (e.g., literacy, limited burden from disadvantage, etc.), it merits further understanding and study.

One important aspect to consider when evaluating these results is that lower education (<8 years), rather than being a direct influence could be a marker for something else. Factors mentioned by participants as reasons for not completing 8th grade such as illness, family hardship (e.g., need to get a job to help the family or stay at home to help on the farm) and repeating a grade, might be sources of residual confounding that cannot be ruled out.

Our study has some limitations that need to be considered before interpreting the results. Even though, the 3MS-R is a good tool for initial dementia screening, it does not provide a clinical diagnosis of dementia; nevertheless, its sensitivity in detecting dementia is very high, ranging from 100 for the age group 65–69 to 65 for the age group 90 and above. Another limitation is that data was collected over a period of 23 years (since 1996) with the corresponding limitations to calculate prevalence rates to compare with general population data. Finally, the assumption of independence samples for the t-test to compare age differences between cases and controls may be partially questionable given the homogeneity of the Amish; however, this does not have any effect on the final results showing the protective effect of education.

In summary, we found that even with lower overall levels of formal education, we still see the protective effect against dementia. Our results are also less likely to be biased by diverse socioeconomic backgrounds, given the homogeneity of the OOA.

Footnotes

ACKNOWLEDGMENTS

We thank the Amish families for allowing us into their communities and for participating in our study. We also acknowledge the contributions of the Anabaptist Genealogy Database and Swiss Anabaptist Genealogy Association. This study is supported by National Institutes of Health/National Institute of Aging, grant 1RF1AG058066 (to Jonathan L. Haines, Margaret Pericak-Vance, and William K. Scott). Finally, we acknowledge the resources provided by the Department of Population and Quantitative Health Sciences, School of Medicine at Case Western Reserve University and the John P. Hussman Institute for Human Genomics at University of Miami, Miller School of Medicine.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Grasset

, Matthews

, Peres

, Foubert-Samier

, Helmer

, Dartigues

, Brayne

(2018) Evolution of dementia diagnosis over time (1988–2013): Evidence from French and English cohorts. Implication for secular trends analyses. Alzheimers Dement (Amst) 10, 490–497.

Fisher

, Chacon

, Chaffee

(2019) Chapter 2 - Theories of cognitive aging and work. In Work Across the Lifespan, Anonymous Academic Press, pp. 17–45.

Smith

, Ryan

(2016) Chapter 16 – PychologicalVitality in the Oldest Old. In Handbook of the Psychology of Aging, K Warner Schaie, Sherry Willis, ed. Academic Press, pp. 303–319.

Sherzai

, Sherzai

(2019) Preventing Alzheimer’s: Our most urgent health care priority. Am J Lifestyle Med 13, 451–461.

Alzheimer’s Association (2020) 2020 Alzheimer’s disease facts and figures. Alzheimers Dement 16, 391–460.

Hebert

, Weuve

, Scherr

, Evans

(2013) Alzheimer disease in the United States (2010–2050) estimated using the 2010 census. Neurology 80, 1778–1783.

Kawaharada

, Sugimoto

, Matsuda

, Tsuboi

, Sakurai

, Ono

(2019) Impact of loss of independence in basic activities of daily living on caregiver burden in patients with Alzheimer’s disease: A retrospective cohort study. Geriatr Gerontol Int 19, 1243–1247.

Panegyres

, Gray

(2010) Dementia risk factors for Australian baby boomers. Neurol Int 2, e13.

Mortimer

(1988) Do psychosocial risk factors con- tribute to Alzheimer’s disease? In Etiology of Dementia of Alzheimer’s type, Henderson AS, Henderson JS, eds. John Wiley & Sons, Chichester, pp. 39–52.

10.

Zhang

, Katzman

, Salmon

, Jin

, Cai

, Wang

, Qu

, Grant

, Yu

, Levy

(1990) The prevalence of dementia and Alzheimer’s disease in Shanghai, China: Impact of age, gender, and education. Ann Neurol 27, 428–437.

11.

Stern

, Gurland

, Tatemichi

, Tang

, Wilder

, Mayeux

(1994) Influence of education and occupation on the incidence of Alzheimer’s disease. JAMA 271, 1004–1010.

12.

Sharp

, Gatz

(2011) Relationship between education and dementia: An updated systematic review. Alzheimer Dis Assoc Disord 25, 289–304.

13.

Caamano-Isorna

, Corral

, Montes-Martinez

, Takkouche

(2006) Education and dementia: A meta-analytic study. Neuroepidemiology 26, 226–232.

14.

Valenzuela

, Sachdev

(2006) Brain reserve and dementia: A systematic review. Psychol Med 36, 441–454.

15.

Raghavan

, Vardarajan

, Mayeux

(2019) Genomic variation in educational attainment modifies Alzheimer disease risk. Neurol Genet 5, e310.

16.

Stern

(2002) What is cognitive reserve? Theory and research application of the reserve concept. J Int Neuropsychol Soc 8, 448–460.

17.

Stern

(2006) Cognitive reserve and Alzheimer disease. Alzheimer Dis Assoc Disord 20, 112–117.

18.

Stern

(2009) Cognitive reserve. Neuropsychologia 47, 2015–2028.

19.

Stern

(2011) Elaborating a hypothetical concept: Comments on the special series on cognitive reserve. J Int Neuropsychol Soc 17, 639–642.

20.

Bennett

, Wilson

, Schneider

, Evans

, Mendes de Leon

, Arnold

, Barnes

, Bienias

(2003) Education modifies the relation of AD pathology to level of cognitive function in older persons. Neurology 60, 1909–1915.

21.

Hahs

, McCauley

, Crunk

, McFarland

, Gaskell

, Jiang

, Slifer

, Vance

, Scott

, Welsh-Bohmer

, Johnson

, Jackson

, Pericak-Vance

, Haines

(2006) A genome-wide linkage analysis of dementia in the Amish. Am J Med Genet B Neuropsychiatr Genet 141B, 160–166.

22.

Weyer

, Hustey

, Rathbun

, Armstrong

, Anna

, Ronyak

, Savrin

(2003) A look into the Amish culture: What should we learn? J Transcult Nurs 14, 139–145.

23.

Richards

, Sacker

(2003) Lifetime antecedents of cognitive reserve. J Clin Exp Neuropsychol 25, 614–624.

24.

Evans

, Beckett

, Albert

, Hebert

, Scherr

, Funkenstein

, Taylor

(1993) Level of education and change in cognitive function in a community population of older persons. Ann Epidemiol 3, 71–77.

25.

Darwish

, Farran

, Assaad

, Chaaya

(2018) Cognitive reserve factors in a developing country: Education and occupational attainment lower the risk of dementia in a sample of Lebanese older adults. Front Aging Neurosci 10, 277.

26.

Farfel

, Nitrini

, Suemoto

, Grinberg

, Ferretti

, Leite

, Tampellini

, Lima

, Farias

, Neves

, Rodriguez

, Menezes

, Fregni

, Bennett

, Pasqualucci

, Jacob Filho

, Brazilian Aging Brain Study Group (2013) Very low levels of education and cognitive reserve: A clinicopathologic study. Neurology 81, 650–657.

27.

Prince

, Acosta

, Ferri

, Guerra

, Huang

, Llibre Rodriguez

, Salas

, Sosa

, Williams

, Dewey

, Acosta

, Jotheeswaran

, Liu

(2012) Dementia incidence and mortality in middle-income countries, and associations with indicators of cognitive reserve: A 10/66 Dementia Research Group population-based cohort study. Lancet 380, 50–58.

28.

Mejia-Arango

, Gutierrez

(2011) Prevalence and incidence rates of dementia and cognitive impairment no dementia in the Mexican population: Data from the Mexican Health and Aging Study. J Aging Health 23, 1050–1074.

29.

Tschanz

, Welsh-Bohmer

, Plassman

, Norton

, Wyse

, Breitner

, Cache County Study Group (2002) An adaptation of the modified mini-mental state examination: Analysis of demographic influences and normative data: The cache county study. Neuropsychiatry Neuropsychol Behav Neurol 15, 28–38.

30.

Pericak-Vance

, Johnson

, Rimmler

, Saunders

, Robinson

, D’Hondt

, Jackson

, Haines

(1996) Alzheimer’s disease and apolipoprotein E-4 allele in an Amish population. Ann Neurol 39, 700–704.

31.

Edwards

, Gilbert

, Jiang

, Gallins

, Caywood

, Creason

, Fuzzell

, Knebusch

, Jackson

, Pericak-Vance

, Haines

, Scott

(2011) Successful aging shows linkage to chromosomes 6, 7, and 14 in the Amish. Ann Hum Genet 75, 516–528.

32.

Edwards

, Gilbert

, Hicks

, Myers

, Jiang

, Cummings

, Guo

, Gallins

, Konidari

, Caywood

, Reinhart-Mercer

, Fuzzell

, Knebusch

, Laux

, Jackson

, Pericak-Vance

, Haines

, Scott

(2013) Linkage and association of successful aging to the 6q25 region in large Amish kindreds. Age (Dordr) 35, 1467–1477.

33.

Johnson

, Rybicki

, Brown

, D’Hondt

, Herpolsheimer

, Roth

, Jackson

(1997) Cognitive impairment in the Amish: A four county survey. Int J Epidemiol 26, 387–394.

34.

Bravo

, Hebert

(1997) Age- and education-specific reference values for the Mini-Mental and modified Mini-Mental State Examinations derived from a non-demented elderly population. Int J Geriatr Psychiatry 12, 1008–1018.

35.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

36.

Scarmeas

, Stern

(2004) Cognitive reserve: Implications for diagnosis and prevention of Alzheimer’s disease. Curr Neurol Neurosci Rep 4, 374–380.

37.

Perneczky

(2019) Dementia prevention and reserve against neurodegenerative disease. Dialogues Clin Neurosci 21, 53–60.