Is the Bahasa Malaysia version of the Montreal Cognitive Assessment
(MoCA-BM) a better instrument than the Malay version of the Mini
Mental State Examination (M-MMSE) in screening for mild
cognitive impairment (MCI) in the elderly?
Rosdinom Razalia,⁎, Lim Jean-Lib, Aida Jaffarb, Mahadir Ahmadc, Shamsul Azhar Shahd,
Norhayati Ibrahimc, Normah Che Dinc, Nik Ruszyanei Nik Jaafara,
Marhani Midina, Hatta Sidia, Saharudin Ahmadb
aDepartment of Psychiatry, Faculty of Medicine, Universiti Kebangsaan Malaysia Medical Centre, Kuala Lumpur, Malaysia
bDepartment of Family Medicine, Faculty of Medicine, Universiti Kebangsaan Malaysia Medical Centre, Kuala Lumpur, Malaysia
cHealth Psychology Programme, School of Health Care Management, Faculty of Health Sciences, Universiti Kebangsaan Malaysia, Kuala Lumpur, Malaysia
dDepartment of Community Health, Faculty of Medicine, Universiti Kebangsaan Malaysia Medical Centre, Kuala Lumpur, Malaysia
Introduction: Mild Cognitive Impairment (MCI) is a known precursor to Alzheimer disease, yet there is a lack of validated screening
instruments for its detection among the Malaysian elderly.
Objective: To compare the Bahasa Malaysia version of the Montreal Cognitive Assessment (MoCA-BM) with the Malay version of the Mini
Mental State Examination (M-MMSE) in the detection of MCI among the Malaysian elderly.
Methodology: This is a cross-sectional study conducted at the primary care centre of Universiti Kebangsaan Malaysia, Kuala Lumpur from
December 2011 to mid-January 2012. Subjects aged 60 and above were recruited using systematic sampling method. Cut-off scores of 22/23
for MoCA-BM and 25/26 for M-MMSE were adopted. Kappa value and Pearson’s correlation coefficient were used to ascertain the
correlation between MOCA-BM and M-MMSE. Data were analysed using Mann–Whitney and Chi Square tests.
Results: The mean age of the 180 subjects enrolled was 65.3 years (SD = 5.4). They had a median of 6 years (IqR 25–75 = 5–11) total
formal education. The prevalence of MCI using MoCA-BM and M-MMSE was 55.6% and 32.8% respectively. The odds of developing MCI
were 1.153 (95% CI = 1.055, 1.261; p b 0.05) for every 1 year increase in age, 0.813 (95% CI = 0.690, 0.959; p b 0.05) with every extra
year of education. Increasing age and lower education level were significantly associated with MCI. The MoCA-BM showed good internal
consistency with Cronbach’s alpha of 0.80. It had moderate correlation with M-MMSE (Pearson correlation coefficient = 0.770, p b 0.001)
and moderate agreement for detecting MCI with Kappa values of 0.497 (p b 0.001).
Conclusion: The prevalence of MCI was higher using MoCA-BM compared to M-MMSE. Both instruments showed moderate concordance
for screening MCI with correlation of their scores.
© 2014 Elsevier Inc. All rights reserved.
In 2008, the proportion of Malaysians aged 60 years and
above was 7% or a total of 1.9 million elderly persons .
With the increasing number of elderly, the incidence of age-
related illnesses will also increase. One inevitable effect of
aging is mild cognitive impairment (MCI) which is now
recognised to be a precursor of Alzheimer disease (AD) .
The 2006 International Psychogeriatric Association Expert
Conference on MCI defined MCI as “A cognitive decline
Available online at www.sciencedirect.com
Comprehensive Psychiatry 55 (2014) S70–S75
Publication of this supplement was supported by Universiti Kebangsaan
Malaysian Medical Centre, Kuala Lumpur, Malaysia.
Conflict of interest: None.
⁎Corresponding author. Department of Psychiatry, Faculty of Medi-
cine, Universiti Kebangsaan Malaysia Medical Centre, Jalan Yaacob Latif,
Bandar Tun Razak, Cheras, 56000, Kuala Lumpur, Malaysia. Tel.: +60
391456143; fax: +60 391456681.
E-mail address: email@example.com (R. Razali).
0010-440X/$ – see front matter © 2014 Elsevier Inc. All rights reserved.
level but that does not interfere notably with activities of daily
life”. To clinically diagnose MCI, there has to be memory
complaints, preferably confirmed by someone else, objective
memory impairment using any of the neurocognitive tools,
normal general cognitive function and intact activities of daily
living. The individual should have no history of dementia .
groups are amnestic MCI–single domain, amnestic MCI–
multiple domains, non-amnestic MCI–single domain and non-
amnestic MCI has the highest risk of progression to dementia,
to Alzheimer’s Disease (AD) . Therefore, there is a critical
criterion . For better diagnostic accuracy, a combination of
clinical features, neuropsychological testing, biomarkers and
available cognitive impairment screening tools, the MMSE
is the most widely used and has been translated and validated
into many languages including the Malay language .
The Montreal Cognitive Assessment (MoCA) was devel-
oped by Nasreddine in 2005 as a screening test for MCI in
subjects who have normal MMSE scores. It has been translated
into various languages and still shows excellent sensitivity and
specificity in detecting MCI despite some cultural and country
specific modifications. Using a cut-off point of 25/26, the
Japanese MoCA (MoCA-J) demonstrated a sensitivity of
93.0% and specificity of 87.0% . In Korea, a cutoff score
is easy to use, thus making it a suitable and sensitive tool for
MCI detection by frontline clinicians.
Sensitivity and specificity of MoCA have been shown to
be higher than those of the MMSE in the detection of MCI
and mild AD . Using a cutoff score of 26, the sensitivity
for MMSE and MoCA to detect MCI was 18% and 90%
respectively. In the mild AD group, the MMSE had a
sensitivity of 78%, while the MoCA had 100%.
Such encouraging findings have created interests among
local authors to translate and test MoCA for use in the local
population.Thus,this study aimstodetermine thecorrelation
and agreement between MoCA-BM and M-MMSE in
detecting MCI among the elderly attending an urban primary
care centre. Bahasa Malaysia was chosen as it is the official
and national language of Malaysia, widely spoken among the
Malays and other ethnic groups in the country.
A cross-sectional study was conducted among elderly
attendees at the ‘Pusat Perubatan Primer UKM’ (PPPUKM),
the primary care clinic of Universiti Kebangsaan Malaysia,
located at Bandar Tasik Selatan in Kuala Lumpur over a 6-
week period from early December 2011 to mid-January
2012. A systematic sampling method was adopted whereby
every 5th patient registered at the clinic was approached for
this study. Inclusion criteria included those aged 60 and
above, with subjective complaints of mild memory impair-
ment, with no impairment in daily functioning, literate in
Bahasa Malaysia and not having severe visual or hearing
impairment. Those with history of neurological diseases
affecting cognitive functions (such as stroke and dementia)
were excluded. Information sheet and verbal explanation
were given before individual written consent was taken.
Two sets of measuring instruments were used in this
study: the Montreal Cognitive Assessment—the Bahasa
Malaysia version (MoCA-BM) and the Malay version of the
Mini Mental State Examination (M-MMSE).
MoCA is a one-page 30-point test that can be completed
in 10 min. It tests for 8 key cognitive domains: attention and
concentration, executive function, memory, language, visuo-
constructional skills, conceptual thinking, calculations and
orientation. A cut-off score of 26 or more was used to
differentiate between normal and those with MCI .
For this study, the English version was translated into
Bahasa Malaysia and changes were made to make it more
relevant and culturally acceptable to the study population. A
cut-off point of 22/23 was adopted, as in other studies
involving Asian populations in Korea and Hong Kong
[9,11], considering the lower level of educational achieve-
ments among the elderly in Malaysia. Furthermore, 1 point
was added to the total MoCA-BM score if the patients had 12
or less years of formal education, as suggested in
Nasreddin’s original study .
2.2.2. The Mini Mental State Examination (MMSE)
The original version was developed by Folstein in 1975
and is still being used widely as a screening tool for
dementia. The original instrument consisted of 11 items with
a total score of 30. It tests for orientation, registration,
attention, calculation, recall, naming, repetition, 3 stage
command, reacting, writing and copying. A score of 21 or
less is suggestive of dementia when corrected for gender and
education. As the MMSE was originally invented as a tool to
detect and monitor the progress of dementia, a cut-off score
for MCI was never suggested. A meta-analysis of all the
studies available using MMSE to detect MCI found no
standardized cut-off score but one which ranged from 23 to
29 . There are 3 Malay language versions of MMSE (M-
MMSE-7, M-MMSE-3 and MMSE-S) with cut-off scores of
20/21, 17/18 and 16/17 respectively which are significantly
influenced by gender and educational level differences .
For this study, MMSE scores of 26 or below would indicate
S71 R. Razali et al. / Comprehensive Psychiatry 55 (2014) S70–S75
possible mild cognitive impairment as suggested by Tasha
et al. .
2.3. Research procedure
2.3.1. Translation of MoCA into Bahasa Malaysia
Prior to the initiation of the study, permission to translate
MoCA was requested from the original author of MoCA, Dr
Ziad Nasreddine. The original English version of MoCA was
first translated into Bahasa Malaysia by 2 appointed
independent language experts before it was back-translated
into English by another 2 appointed independent language
experts. The back-translated versions were then compared
with the original MoCA to ensure accuracy before a
harmonised version was produced.
A panel ofexpertsconsistingofa public healthphysician, a
geriatric psychiatrist, a group of psychologists and family
together. It was assessed sentence by sentence to ensure
accuracy of translation, comprehensibility of the instruction
and cultural relevance for its use in a local setting. Several
i). The original English version used 5 different nouns
— “Face”, “Velvet”, “Church”, “Daisy” and “Red”
to assess memory and delayed recall. As “baldu”
(translation of “velvet”) was unfamiliar to many
Bahasa Malaysia speakers, it was replaced with
“kapas” (translation of “cotton”). “Church” was
replaced with “school” to avoid religious sensitivity
as Malaysia is a multicultural country which
practises various religions. “Daisy” was changed
to “rose” as the latter is more common and sounds
similar in Bahasa Malaysia. “Red” was changed
to “blue” (“biru”) to prevent association with
“Rose”. Only “face” remained unchanged from
the original version.
ii). For language assessment, the name “John” was
changed to “Johan” as it is a common local name.
iii). For fluency test, letter “S” was used instead in the
translated version as there were more words
beginning with the letter “S” than “F” in Bahasa
iv). All other domains in the Bahasa Malaysia version
were translated verbatim. Following these changes,
the pre-final version of MoCA-BM underwent
2.3.2. Pilot testing
MoCA-BM was tested on 20 selected attendees at the
PPPUKM using the inclusion and exclusion criteria.
Preliminary results showed good internal consistency with
a Cronbach’s alpha of 0.80 based on the standardised items.
This tool took approximately 10–15 min to complete. No
difficulties were encountered during the testing and subjects
gave favourable feed-back when questioned regarding
comprehension and ease of the questionnaire. It was then
decided that no further amendments were needed and the
MoCA-BM was ready for use.
2.3.3. Data collection
A self-administered questionnaire on sociodemographic
data was given to subjects prior to the administration of
MoCA-BM and M-MMSE.
2.4. Statistical analysis
All analysis was done using the Statistical Package for
Social Sciences (SPSS) software version 20. P values less
than 0.05 were taken as representing significant difference
for all analyses. Descriptive analysis was used to depict the
sociodemographic data of the subjects to determine the
prevalence of MCI using both the MoCA-BM and M-
MMSE. Pearson correlation coefficient was used to correlate
the findings of both MoCA-BM and M-MMSE. The
agreement between M-MMSE and MoCA-BM to detect
MCI was obtained from the Kappa value.
A total of 180 patients were included in this study. They
have all completed the questionnaires fully as required.
3.1. Prevalence of MCI
The prevalence of MCI using cut-off points of 22/21 for
MoCA-BM and 26/25 for M-MMSE was 55.6% and 32.8%
respectively (Table 1).
3.2. MoCA-BM and M-MMSE scores
(SD = 5.2) whereas M-MMSE scores ranged from 9 to 30
with a mean of 26.8 (SD = 3.4). As shown in Fig. 1, subjects
who scored low in the MoCA-BM also had similar scoring
trends for the M-MMSE. There was moderate correlation
between MoCA-BM and M-MMSE using Pearson correla-
tion with correlation coefficient of 0.770 and this finding was
statistically significant with p b 0.001 (Fig. 2).
A total of 56/180 (31.1%) subjects fell into the MCI range
for both MoCA-BM and M-MMSE whereas 44/180 (24.4%)
of those who tested normal in the M-MMSE actually tested
positive for MCI using MoCA-BM. Of the total sample
studied, 77/180 (42.8%) had normal scores for both MoCA-
BM and M-MMSE. The Kappa value indicating agreement
for diagnosis of MCI using MoCA-BM versus M-MMSE
was 0.497 with p b 0.001(Table 2).
Prevalence of MCI.
MCINo MCI Total
S72R. Razali et al. / Comprehensive Psychiatry 55 (2014) S70–S75
3.3. Goodness of fit of MoCA
Goodness of fit was performed using the Hosmer–
Lemenshow Test and area under the receiver operator curve
(ROC). The X2value using the Hosmer-Lemenshow Testwas
8.25(p = 0.41),indicatingthatthemodelfitsandtheobserved
event rates matched the expected event rates in the different
subgroups of the model population. Area under the ROC was
0.82 (95% CI = 0.76, 0.88) with significant difference of
p b 0.001 (Fig. 3). This showed that the MoCA-BM was a
good test to differentiate MCI from non-MCI cases.
3.4. Sociodemographic profile and MCI
From the total of 180 subjects analysed, there were
significant associations between MCI and age, educational
achievement and gender based on the BM-MOCA. The
mean age was 65.3 years with a standard deviation of 5.4. As
their ages were not normally distributed, the median age was
63 years (IqR 25–75 = 61–68). The median age of those
with MCI was significantly higher than those without any
cognitive impairment (p b 0.001), whereby those with MCI
had a median age of 65 years (IqR 25–75 = 62–70) whereas
those without cognitive impairment were aged 62.5 years
(IqR 25–75 = 60–64)
Table 3 shows the association of MCI with formal
education (p b 0.001) and gender (p b 0.023). There was a
widerangeinthetotalyearsof education (0to18years),with
a median of 6 years (IqR 25–75 = 5–11). The highest levels
of education attained were primary school (44.4%) and
secondary school (37.2%). Subjects with MCI had a median
of 6 years (IqR 25–75 = 3–7) whereas those without MCI
had 10 years (IqR 25–75 = 6–11) of education.
The ethnic distribution of subjects reflected Malaysian
multi-ethnicity. Malays made up 55% of the study
population followed by the Chinese and Indians at 21.1%
and 22.2% respectively. There were more females with MCI
(65.8%) compared to males (48.6%).
In this study, the prevalence of MCI using the MoCA-BM
was high with more than half (55.6%) of subjects screening
positive for MCI. Similarly, high prevalence of MCI (64.7%)
was also detected among elderly attendees of medical clinics
in an urban hospital using MMSE cut-off score of 21/30 to
28/30, which was significantly associated with low educa-
tion , as also shown in this study. Lower level of
education is a known risk factor for MCI .
Another possible factor contributing to the high preva-
lence of MCI in our population was the tool itself. Most
studies had used the original English version MoCA as a
screening tool. Most of our Bahasa Malaysia-speaking
subjects were more fluent speaking in its colloquial form,
with a poorer command of the language might have scored
lower than those who were more linguistically capable, thus
contributing to the higher rate of MCI false positives.
The cut-off score of 22/23 might not be suitable for our
population. Further validation of the MoCA-BM is needed to
ascertain a cut-off score that is specific for our population.
Unlike other studies which employed clinical as well as
neuropsychological testings to diagnose MCI, this study
Fig. 1. Comparison of BM-MoCA and M-MMSE scores.
Fig. 2. Correlation between BM-MoCA and M-MMSE.
S73 R. Razali et al. / Comprehensive Psychiatry 55 (2014) S70–S75
only used the MoCA-BM and M-MMSE to screen for MCI,
which might have led to its higher prevalence.
This study found a moderate strength correlation between
MoCA-BM and M-MMSE scores with a Pearson correlation
coefficient of 0.770 (p b 0.005). This positive linear
relationship between MOCA-BM and M-MMSE scores
was comparable to those reported by the original author of
MoCA, which obtained high correlation between both tools
with correlation coefficient of 0.87 (p b 0.001) . The
lower correlation coefficient in this study could be due to the
types of items in the MoCA-BM which required more
scholastic skills such as recognition of alphabets, arithmetic
and drawing to complete when compared with the M-
MMSE. The MoCA-BM and M-MMSE had a Kappa value
of 0.495, indicating moderate agreement . The agree-
ment difference could be due to the difference in the
functions of the instruments themselves, in which MoCA
was designed specifically for screening MCI  whereas
MMSE was initially designed to detect dementia.
Age and education were significantly associated with
MCI. This association between MCI and increasing age has
also been shown in Swedish  and Singaporean 
studies. However, there have been suggestions that the
prevalence for MCI should remain stable across all ages as
the diagnosis should be based on age and education specific
norms . Sattler (2012) reported that education level is an
important marker for cognitive reserve as it reflects the
extent of early cognitive stimulation of the brain and
cognitive abilities. It has also been reported that high early
education combined with high late life cognitive activity is
beneficial and protective in preventing MCI .
The study population consisted of elderly outpatients on
treatment for various medical illnesses in an urban primary
care centre. So the findings from this study cannot be
generalized to the elderly community who often prefer to live
in the rural areas. MCI-amnestic type was the main type of
MCI experienced by these subjects, hence the findings might
not be reflective of other types of MCI.
MoCA-BM required a good command of formal Bahasa
Malaysia for it to be fully understood and comprehended.
Despite Bahasa Malaysia being the most widely spoken
language among all races, some of these older generation
Chinese and Indians find difficulty in understanding some of
the instructions in MoCA as it is not their mother-tongue
language spoken at home. Most often they speak the
colloquial form of the language in their daily life. A
validated Tamil and Mandarin versions of the MoCA would
perhaps be able to overcome this difficulty.
This study did not determine the cut-off score for MoCA-
BM but instead had adopted a similar score used in another
Asian MCI study. This could have contributed to the higher
prevalence of MCI in our population as education, culture
and ethnicity can affect the performance of the MoCA. A
future formal validation study of the MoCA-BM is required
to determine an appropriate cut-off score for our population
based on the receiver operator curve (ROC).
6. Conclusion and recommendations
In general, the MoCA-BM was an acceptable and easy to
use tool for MCI screening in the urban elderly. The
prevalence of MCI was higher using MoCA-BM as
compared to M-MMSE. The MoCA-BM and M-MMSE
have shown moderate correlation in their scoring trend.
Association between MCI and level of education.
N = 180 MCI, n (%)X2 df p value
Level of education
Agreement of BM-MoCA and M-MMSE to detect MCI.
N = 180 BM-MoCATotal (%)
MCI (%)No MCI (%)
No MCI (%)
Fig. 3. ROC to assess Goodness of Fit for BM-MoCA.
S74 R. Razali et al. / Comprehensive Psychiatry 55 (2014) S70–S75
There was only moderate but significant concordance Download full-text
between MoCA-BM and M-MMSE in detecting MCI among
the elderly patients attending PPPUKM. This was due to the
fact that the MMSE was never invented to screen for MCI
compared to the MoCA which was invented to do so. Further
changes to some of the MoCA-BM items are needed for the
instrument to be more acceptable to Bahasa Malaysia-
speaking Malaysian population.
its reliability and validity in a population already prediagnosed
as having MCI using more strict diagnostic criteria and
be determined for it to have high sensitivity and specificity as a
screening instrument for MCI. It is also recommended that
MoCA be translated and validated into other local languages as
Malaysia is a melting pot of multiethnic groups of people.
Lastly, increasing age and lower levels of education were
significantly associated with MCI. Therefore, elderly in-
dividuals with these risk factors should be screened
periodically for MCI and early signs of dementia.
The authors would like to thank all clinic staff of
PPPUKM for their cooperation and assistance during the
course of this study. This study was funded by the University
Fundamental Grant and had been approved by the Ethics
Committee of Universiti Kebangsaan Malaysia as well as the
National Medical Research Registry.
 WHO. World health statistics 2010: Malaysia [report on the internet]
[Internet]. 2010 [cited 29 November 2011]. Available from: http://
 Maioli F, Coveri M, Pagni P, Chiandetti C, Marchetti C, Ciarrocchi R,
et al. Conversion of mild cognitive impairment to dementia in elderly
subjects: a preliminary study in a memory and cognitive disorder unit.
Arch Gerontol Geriatr 2007;44(Suppl 1):233-41.
 Gauthier S, Reisberg B, Zaudig M, Petersen RC, Ritchie K, Broich K,
et al. Mild cognitive impairment. Lancet 2006;367(9518):1262-70.
 Petersen RC, Stevens JC, Ganguli M, Tangalos EG, Cummings JL,
DeKosky ST. Practice parameter: early detection of dementia: mild
cognitive impairment (an evidence-based review). Report of the
Quality Standards Subcommittee of the American Academy of
Neurology. Neurology 2001;56(9):1133-42.
 Petersen RC. Mild cognitive impairment as a diagnostic entity. J Intern
 Levey A, Lah J, Goldstein F, Steenland K, Bliwise D. Mild cognitive
impairment: an opportunity to identify patients at high risk for
progression to Alzheimer's disease. Clin Ther 2006;28(7):991-1001.
 Ibrahim NM, Shohaimi S, Chong HT, Rahman AHA, Razali R, Esther
E, et al. Validation study of the Mini-Mental State Examination in a
Malay-speaking elderly population in Malaysia. Dement Geriatr Cogn
 Fujiwara Y, Suzuki H, Yasunaga M, Sugiyama M, Ijuin M, Sakuma N,
et al. Brief screening tool for mild cognitive impairment in older
Japanese: validation of the Japanese version of the Montreal Cognitive
Assessment. Geriatr Geront Int 2010;10(3):225-32.
 Lee JY, Lee Dong Woo, Cho SJ, Na DL, Jeon Hong Jin, Kim SK, et al.
Brief screening for mild cognitive impairment in elderly outpatient
clinic: validation of the Korean version of the Montreal Cognitive
Assessment. J Geriatr Psychiatry Neurol 2008;21(2):104-10.
 Nasreddine ZS, Phillips NA, BÃ©dirian V, Charbonneau S,
Whitehead V, Collin I, et al. The Montreal Cognitive Assessment,
MoCA: a brief screening tool for mild cognitive impairment. J Am
Geriatr Soc 2005;53(4):695-9.
 Wong Adrian, Yun YX, Kwan Pauline WL, Chan Anne YY, Lam
Wynnie WM, Wang Ki, et al. The validity, reliability and clinical
utility of the Hong Kong Montreal Cognitive Assessment (HK-MoCA)
in patients with cerebral small vessel disease. Dement Geriatr Cogn
 Mitchell AJ. A meta-analysis of the accuracy of the mini-mental state
examination in the detection of dementia and mild cognitive
impairment. J Psychiatr Res 2009;43(4):411-31.
 Tasha S, Nadia G, Clive H. The Montreal Cognitive Assessment:
validity and utility in a memory clinic setting. Can J Psychaitry 2007;
 Razali Rosdinom, Baharudin Azlin, Jaafar Nik Ruszyanei Nik, Sidi
Hatta, Rosli Abdul Hadi, Hooi Khoo Boo,et al. Factorsassociated with
mild cognitive impairment among elderly patients attending medical
clinics in Universiti Kebangsaan Malaysia Medical Centre. Sains
 Lopez OL, Jagust WJ, Dulberg C, Becker JT, DeKosky ST, Fitzpatrick
A, et al. Risk factors for mild cognitive impairment in the
Cardiovascular Health Study Cognition Study: part 2. Arch Neurol
 Viera A, Garrett J, et al. Understanding interobserver agreement: the
Kappa statistic. Fam Med 2005;37(5):360-3.
 Hanninen T, Hallikainen M, Tuomainen S, Vanhanen M, Soininen H.
Prevalence of mild cognitive impairment: a population-based study in
elderly subjects. Acta Neurol Scand 2002;106(3):148-54.
 Chan M, Tay L, Chong M. Amnestic mild cognitive impairment and
early Alzheimer's disease in an Asian memory clinic — evidence for a
clinical spectrum. Dement Geriatr Cogn Disord 2011;1(1):113-23.
 Luck T, Riedel-Heller SG, Kaduszkiewicz H, Bickel H, Jessen F,
Pentzek M, et al. Mild cognitive impairment in general practice: age-
specific prevalence and correlate results from the German study on
ageing, cognition and dementia in primary care patients (AgeCoDe).
Dement Geriatr Cogn Disord 2007;24(4):307-16.
 Sattler C, Toro P, Schönknecht P, Schröder J. Cognitive activity,
education and socioeconomic status as preventive factors for MCI and
Alzheimer's disease. Psychiatry Res 2012;196(1):90-5.
S75 R. Razali et al. / Comprehensive Psychiatry 55 (2014) S70–S75