Hong Kong J Psychiatry 2009;19:97-102 Original Article

Optimal Cut-off Score on the Chinese Version

of Executive Interview (C-EXIT25) in a Hong
Kong Chinese Population
中文版的執行功能測試（C-EXIT25）在香港華籍人口中的最
佳切點分數
SSM Chan, CHM Li, SLK Pang, CSM Wong, HFK Chiu, LCW Lam

陳秀雯、李曉薇、彭麗君、黃秀雯、趙鳳琴、林翠華

Abstract
Objectives: To determine the optimal cut-off score on the Chinese version of the Executive Interview to
discriminate all-cause dementia patients from non-dementia subjects.
Participants and Methods: A total of 141 community-dwelling elders were assessed with the Chinese
version of the Executive Interview, the Cantonese version of the Mini-Mental State Examination, and Nelson’s
Modified Card Sorting Test. Severity of dementia was determined using the Clinical Dementia Rating Scale.
Results: Higher total scores on the Chinese version of Executive Interview (greater impairment) yielded
a statistically significant negative correlation with Nelson’s Modified Card Sorting Test’s ‘Number of
categories’, but positive correlations with the test’s ‘Errors’, ‘Perseverative errors’, ‘Non-perseverative
errors’, and ‘Percentage of perseverative errors’. The sensitivity and specificity at different cut-off values
on the Chinese version of the Executive Interview used to plot the receiver operating characteristic curve
gave an area under the curve of 0.97 (95% confidence interval, 0.94-0.99; p ≤ 0.01). The cut-off value
of 15 best distinguished Clinical Dementia Rating 0 and 0.5 from Clinical Dementia Rating 1 and 2
(sensitivity = 90.7%; specificity = 87.2%).
Conclusions: The results of the current study and the previous pilot study support that the Chinese version of the
Executive Interview as a potentially useful bedside tool for executive functional assessment in Chinese elders,
by virtue of good internal consistency, inter-rater reliability, concurrent validity and discriminatory power.

Key words: Aged; Dementia; Geriatric assessment; Neuropsychological tests; Psychiatric status rating scales

摘要

目的：探討香港華籍人口中，使用中文版的執行功能測試分辨痴呆症患者及非痴呆症患者的最
佳切點分數。
參與者與方法：共141位居住在社區的老人接受中文版的執行功能測試、廣東話版的簡短智能測
驗，及Nelson的修改式卡片分類測驗。並使用臨床痴呆評估量表量度痴呆症的嚴重程度。
結果：中文版的執行功能測試得分高，與Nelson的修改式卡片分類測驗中的「分類數」呈顯著
負相關，但與「錯誤」、「持續性錯誤」、「非持續性錯誤」，以及「持續性錯誤的百分比」
呈顯著正相關。中文版的執行功能測試在不同切點的敏感性及特異性在ROC曲線下之區域為
0.97（95%置信區間：0.94-0.99；p ≤ 0.01）。切點為15最能分辨臨床痴呆評估量表0-0.5及1-
2（敏感性 = 90.7%；特異性 = 87.2%）。
結論：根據內部一致性、評分者信度、同時效度、以及辨別力，本研究及過往一項先導研究的
結果均顯示，中文版的執行功能測試是用於檢測華籍老年人執行功能一種有用的臨床工具。

關鍵詞：年老、痴呆症、老人檢測、醫科學生、神經心理測試、精神狀態量表
Prof Sandra Sau-man Chan, MRCPsych, FHKAM (Psychiatry), FHKCPsych, China.
Department of Psychiatry, The Chinese University of Hong Kong, Hong Kong, Prof Linda Chiu-wa Lam, FRCPsych, FHKAM (Psychiatry), FHKCPsych, MD,
China. Department of Psychiatry, The Chinese University of Hong Kong, Hong Kong,
Ms Catherine Hiu-mei Li, BSocSc (Hons), MSW, Department of Psychiatry, The China.
Chinese University of Hong Kong, Hong Kong, China.
Ms Shirley Lai-kwan Pang, Hons Dip (Coun & Psy), BSc (Hons), Department of Address for correspondence: Prof Sandra SM Chan, Department of Psychiatry,
Psychiatry, The Chinese University of Hong Kong, Hong Kong, China. G/F, Multicentre, Tai Po Hospital, 9 Chuen On Road, Tai Po, Hong Kong, China.
Ms Corine Sau-man Wong, BCogSc (Hons), Department of Psychiatry, The Tel: (852) 2607 6025; Fax: (852) 2667 1255;
Chinese University of Hong Kong, Hong Kong, China. E-mail: schan@cuhk.edu.hk
Prof Helen Fung-kum Chiu, FRCPsych, FHKAM (Psychiatry), FHKCPsych,
Department of Psychiatry, The Chinese University of Hong Kong, Hong Kong, Submitted: 14 November 2008; Accepted: 5 January 2009

© 2009 Hong Kong College of Psychiatrists 97

SSM Chan, CHM Li, SLK Pang, et al

Introduction with formal ECF measurement and other validated global

cognitive tests for elders.13 That study showed that C-
‘Executive control functions’ (ECFs) broadly encompass EXIT25 has high internal consistency (Cronbach’s α =
a range of cognitive skills that are responsible for the 0.79) and inter-rater reliability (r = 0.91, p < 0.01). The
planning, initiating, sequencing, and monitoring of complex correlational properties of the C-EXIT25 to provide indices
goal-directed behaviour. Well-known executive tests like of the Modified Wisconsin Card Sorting Test (MCST)8,14,15
the Wisconsin Card Sorting Test (WCST), the Stroop, the were superior to the Cantonese version of the Mini-Mental
Category Test, and Trails B have been specifically associated State Examination (C-MMSE)16 and the Chinese version
with frontal lobe structural or metabolic changes.1-5 of the Mattis Dementia Rating Scale,17 even after adjusting
Classical ECF measures are often multidimensional and no for age, gender, and educational level. It also discriminates
single measure comprehensively assesses all ECF domains. different stages in Clinical Dementia Rating (CDR).18,19
The ECF has not been a routine domain for clinical Given the small sample size and the inadequate number of
assessment, by virtue of its perceived complex construct by normal controls in the pilot study, we lacked the statistical
most clinicians. The relative paucity of validated bedside power to set up a ROC analysis to examine how well the
instruments is another reason. Although these ECF measures C-EXIT25 discriminates normal from dementia subjects
have a reasonably high level of validity,6 they are not user- at different cut-off points. The objective of this study was
friendly as bedside clinical tools. Consequently, they are to determine the optimal cut-off score on the C-EXIT25 that
often restricted to research endeavours, and their use for discriminates all-cause dementia from non-dementia subjects.
daily clinical assessment of ECF is limited.7-9
It is generally agreed that an ideal bedside instrument Methods
should be user-friendly for most frontline carers (including
non-professionals), feasible (not time-consuming or The details of translation, back-translation, pilot-application
demanding for auxiliary tools), and reliable across and adaptation were reported in another paper.13 In brief,
situations and between raters. For the purpose of screening, the original English version of EXIT2510 was translated
it should also have reasonable validity, which includes to Chinese and back-translated to affirm its face validity.
good sensitivity and specificity. One of the examples is the Items on ‘Number-Letter Task’, ‘Word Fluency Task’,
Folstein’s Mini-Mental State Examination available since ‘Anomalous Sentence Repetition’, ‘Memory/Distraction
1970s, which is a short bedside tool that has become a Task’, ‘Interference Task’, and ‘Serial Order Reversal Task’
popular test for screening out patients with dementia and required major adaptation to fit into our local cultural /
monitoring disease progress. In terms of ECF assessment, linguistic context. This was because direct translation was
a validated bedside measure of executive dyscontrol known not appropriate for the following reasons: (1) The basic
as the Executive Interview (EXIT25) contains 25 items linguistic unit in the English language is the alphabet, while
and requires 15 minutes to administer.10 It is a structured ideographic characters are used in Chinese, there being no
interview and the comprehensive range of symptoms of grammatically equivalent expressions in Chinese. (2) The
executive failure elicited by it include: perseveration, translated items were not common colloquial expressions
imitation behaviour, echopraxia, echolalia, intrusions, frontal and demanded a sophisticated level of comprehension. (3)
release signs, lack of spontaneity, prompting, disinhibited The task becomes too easy even for highly impaired subjects,
and utilisation behaviours. It was reported to have high after direct translation. For example, in the “Number-Letter
internal consistency (α = 0.85) and inter-rater reliability Task” (Item 1), the alphabets A to E are changed to the non-
(r = 0.91). Scores of the EXIT25 also correlate well with numerical chronological words of “甲，乙，丙，丁，戊”
WCST categories (r = 0.54), Trail Making Part B (r = 0.64), since none of the study participants understand English, and
the Test of Sustained Attention (Time, r = 0.82; Errors, r = these chronological words are well-known to most local
0.83), and Lezak’s Tinker Toy Test (r = 0.57). Scores range Chinese without reliance on a sophisticated education. In
from 0 to 50 with high scores indicating impairment. A score the “Memory/Distraction Task” (Item 6), the 3 items used
of 15/50 best discriminates normal elderly from all-cause are “Apple”, “Train”, “Newspaper” in Chinese language, as
dementia (sensitivity = 0.93, specificity = 0.83, area under in the 3-item registration and retrieval test in C-MMSE.16
the receiver operating characteristic [ROC] curve [AUC] = These 3 items have the same number of syllables in
0.93, n = 200). Chinese and are common daily commodities from different
There is dearth of data concerning ECF profiling in categories. The original EXIT25 uses forward and backward
Chinese elders. Since performance in standardised ECF tests spelling of the word “Cat” after the initial registration, so
are often subject to educational level and cultural biases,11,12 as to distract subjects before retrieval of the 3 registered
validation of a user-friendly bedside instrument in the items. Again, due to different language structure, we added
Chinese language may provide an alternative convenient questions about the appearance of a cat such as “how many
tool for such profiling in our locality, which is suitable for legs does a cat have?” or “how many ears?” or “how many
daily clinical practice. We have previously reported the tails” to serve as a distraction, and we trust that the level of
results of a pilot study on the correlational properties of difficulty relative to our local elders was compatible with
the Chinese version of Executive Interview (C-EXIT25) that in the original version targeted at English-speaking

98 Hong Kong J Psychiatry 2009, Vol 19, No.3

 Optimal Cut-off Score on the C-EXIT25

elders. Such semantic contexts also introduce the same testing procedures.
interference with the subsequent retrieval of registered Details of the assessment entailed the following:
items, so that some impaired participants add words related (a) Demographics (gender, age, educational level, marital
to “Cats” during subsequent recall. As for the “Interference status, living arrangements, and physical health
Task” (Item 7), the original version uses “Brown” printed screen), which were all charted on a standard data
in black while we use “Red” (in Chinese) printed in black. entry form.
It is adapted this way, since the Chinese expression for the (b) The C-EXIT25 described above.
English word “Brown” has 2 Chinese characters, and to (c) The C-MMSE,16 which is an 11-item brief bedside
local elders of elementary educational level, such characters cognitive test that comprises questions on time / place
are not as well-known as the word “Red” expressed in a orientation, memory (immediate and delayed recall),
single Chinese character. Lastly, the “Serial Order Reversal ideomotor praxia, language and visuo-spatial tasks. It
Task” (Item 22) was totally revamped since the original has been validated in Chinese and has been widely
English version asked subjects to recite months of the years used in daily clinical practice and research.
backward, while most of the calendar nomenclatures in the (d) The MCST,8,14,15 which utilises 2 identical subsets of
Chinese language are simple numerical values, making the 24 response cards out of the 128 original cards used
task too simple for most impaired elders. Although there in the WCST, given out in a particular sequence, such
are some forms of calendar nomenclature that use distinct that no consecutive cards share the same attribute
non-numerical chronological phrases or words, they are (form, colour, and number). The test also allows
too difficult for even highly educated local elders and subjects to decide the initial sequence of sorting
hence were considered not applicable. Taking into account principles in the first 3 categories. The number of
all these linguistic and cultural limitations as well as the consecutive correct sorts is reduced from 10 to 6 for
tentative construct of this task (testing serial order reversal a category to be achieved. The test is halted when
on well-learnt sequences that demand moderately sustained the 3 categories are completed twice or when all 48
attention), we asked subjects to do serial-2 subtractions response cards have been exhausted. The MCST has
starting from 20. The serial-2 subtraction (a common proved popular with researchers investigating ECF
bedside test) taxes executive function since it requires and is able to discriminate between patients with
reorganising a well-learnt set with good working memory frontal lobe lesions and normal controls, but is rather
and sustained attention, which cannot be completed solely equivocal for differentiating subjects with the former
by relying on semantic retrieval of memory. The length of and those with non-frontal lobe lesions.20 Despite its
this task is similar to the “Calender” version of the original shortcomings compared to the full WCST (lack of
EXIT25, and being reasonably free of educational bias — it detailed psychometric profiling, and normative), it
does not demand sophisticated scholastic skills. was nevertheless chosen for this study, because in our
The initial translated Chinese version was administered prior experience the full WCST was too cumbersome
to 15 consenting elders conveniently recruited from a for our local elderly with a limited level of education.
psychogeriatric outpatient clinic for a pilot run. Minor Moreover, the MCST has been experimented with
final modifications were made to the questionnaire before local community-dwelling Chinese people, whose
administering it formally to our recruits. performance profile correlated well with relevant
demographic factors.14 It is thus a feasible alternative
Subjects and Assessments to the full version of the WCST and has local reference
Elders, aged 65 years or above, and capable of giving written data.
informed consent (or with available next-of-kin to give This study protocol was approved by the Joint CUHK-
consent in case of mental incapacity), were conveniently New Territories East Cluster Clinical Research Ethics
selected from several psychogeriatric clinics. The latter Committee (CRE-2003.260; CRE-2004.103).
included clinics of The Chinese University of Hong Kong
(CUHK) and several local community elderly facilities. Data Analyses
Recruitment was in 2 periods (2003-2004 and 2005-2007). The ROC curve was plotted using measures of sensitivity
The subjects were selected irrespective of their diagnoses or and specificity at various C-EXIT25 cut-off values.
co-morbid medical conditions. Dementia staging, defined The ROC curve is used to demonstrate the overall
by CDR18,19 was assigned by an independent psychiatrist discriminatory power of assessment tools over a range of
blinded to the C-EXIT25 scores. The assessment battery cut-off values. The optimal cut-off value is indicated by the
(detailed below) was administered by 2 research assistants part of the curve closest to the upper left corner. The AUC
who received relevant training in the pilot phase of the is a measure of the diagnostic power of the test. A perfect
study, during which they achieved satisfactory inter-rater test will have an AUC of 1.0, and AUC of 0.5 means the
reliability (Pearson correlation, 0.91; p < 0.01). Exclusion test performs no better than chance. All statistical analyses
criteria were: stage 3 (severe) dementia on CDR, inability were performed with the Statistical Package for the Social
to follow simple commands, inability to speak Cantonese, Sciences (Windows version 16.0; SPSS Inc, Chicago [IL],
or sensory / physical deficits that precluded performing the US).

Hong Kong J Psychiatry 2009, Vol 19, No.3 99

SSM Chan, CHM Li, SLK Pang, et al

Results errors’, ‘Failure to maintain set’, ‘Unique errors’) across

CDR stages 0, 0.5, 1, and 2 are shown in Table 1.
A total of 141 elders consented to participate in this study and
completed all the assessments. There was a predominance Correlations between Parameters of the Modified
of female subjects across all groups of dementia staging Wisconsin Card Sorting Test and the Chinese
(female, 103; male, 38). The mean age of all the subjects Version of Executive Interview
was 76.8 (standard deviation, 7.0) years. The mean age Table 2 shows that higher total scores on the C-EXIT25
and years of education of subgroups classified by the CDR (greater impairment) correlates negatively with ‘Number
staging are shown in Table 1. The CDR of 0, 0.5, 1, and 2 of categories’ but positively with ‘Errors’, ‘Perseverative
denotes ‘normal cognitive function’, ‘very mild dementia’, errors’, ‘Non-perseverative errors’, and ‘Percentage of
‘mild dementia’, and ‘moderate dementia’, respectively. perseverative errors’ at statistically significant levels. The
observed negative correlation between total score on the C-
Descriptive Profiles of the Cognitive Measures EXIT25 and ‘Failure to maintain set’ did not reach statistical
(Chinese Version of Executive Interview, Cantonese significance.
Version of Mini-Mental State Examination,
Modified Wisconsin Card Sorting Test) Optimal Cut-off Score on the Chinese Version of
The mean total scores of the C-EXIT25 and C-MMSE as Executive Interview that Distinguishes Subjects
well as performance indices of the MCST (‘Number of with Clinical Dementia Rating of 1 or Above from
categories’, ‘Total number of errors’, ‘Non-perseverative Others
errors’, ‘Perseverative errors’, ‘Percentage of perseverative Table 3 shows the sensitivity and specificity at different

Table 1. Age, years of education, and score attained for the C-MMSE and performance indices of the MCST in CDR 0,
CDR 0.5, CDR 1, and CDR 2.*
CDR staging† All (n = 141)
0 (n = 45) 0.5 (n = 37) 1 (n = 34) 2 (n = 25)
Age (years) 74.0 (6.5) 76.6 (6.8) 79.3 (6.4) 78.9 (7.6) 76.8 (7.0)
Years of education 4.5 (4.9) 1.7 (3.1) 2.2 (3.3) 2.6 (4.7) 2.9 (4.2)
C-MMSE 27.0 (2.5) 22.5 (3.7) 18.9 (3.9) 14.0 (4.1) 21.7 (5.8)
C-EXIT25 8.4 (3.4) 12.4 (4.3) 20.1 (4.9) 24.7 (4.6) 15.1 (7.5)
MCST
Number of categories 3.9 (1.9) 2.4 (1.3) 1.4 (1.0) 1.2 (0.9) 2.4 (1.8)
Number of errors 16.2 (10.8) 24.8 (9.2) 30.6 (6.2) 32.2 (6.1) 24.9 (10.7)
Non-perseverative errors 8.4 (5.6) 12.6 (4.9) 15.9 (7.7) 14.2 (9.0) 12.4 (7.3)
Perseverative errors 7.7 (8.6) 12.1 (9.2) 14.5 (8.7) 18.2 (11.4) 12.5 (10.0)
FMS 0.8 (0.9) 1.1 (1.4) 0.9 (1.2) 0.6 (0.9) 0.8 (1.1)
Unique errors 1.6 (2.4) 1.8 (2.3) 2.7 (3.8) 2.9 (3.6) 2.1 (3.0)
% of perseverative error 36.7 (34.9) 43.9 (20.4) 46.6 (22.3) 53.9 (28.4) 44.3 (24.2)
Abbreviations: C-MMSE = Cantonese version of Mini-Mental State Examination; CDR = Clinical Dementia Rating; C-EXIT25 =
Chinese version of Executive Interview; FMS = Failure to maintain set; MCST = Modified Wisconsin Card Sorting Test.
*
Data are shown as mean (standard deviation).
†
CDR 0, 0.5, 1, and 2 denotes ‘normal cognitive function’, ‘very mild dementia’, ‘mild dementia’, and ‘moderate dementia’,
respectively.

Table 2. Correlations between C-EXIT25 scores and MCST performance indices.

No. of Errors Non-perseverative Perseverative % of perseverative Unique errors FMS
categories errors errors errors
Pearson -0.53 *
0.56 *
0.37 *
0.33 †
0.19† 0.19† -0.14
correlation
Abbreviations: C-EXIT25 = Chinese version of Executive Interview; FMS = Failure to maintain set; MCST = Modified Wisconsin
Card Sorting Test.
*
p < 0.01.
†
p < 0.05.

100 Hong Kong J Psychiatry 2009, Vol 19, No.3

 Optimal Cut-off Score on the C-EXIT25

Table 3. Sensitivity and specificity at different cut-off concurrent validity for the C-EXIT25 has been discussed in a
values on the C-EXIT25.* previous pilot study that compared the scores it yielded with
those from the MCST (an alternative measure of ECF).13 In
C-EXIT25 cut- Sensitivity (%) Specificity (%)
our pilot study,13 it was shown that the total C-EXIT25 score
off score
correlates moderately well at statistically significant levels
3 100 7.3
with ‘Number of categories’ (Pearson correlation = –0.53, p
4 100 9.7 < 0.01) and other performance indices of the MCST. Such
5 100 14 observations are consistent with the original validation
6 100 20.2 study of the EXIT25 in a Caucasian population, in which
7 100 23.2 the Pearson correlation coefficient between the total score
8 100 26.8 and the WCST category was -0.54.10
9 100 32.9 Our results need to be interpreted in the context
10 100 46.3 of methodological limitations, which mainly arise from
11 100 58 sampling. For instance, subjects were conveniently sampled
12 99 65.3 from a community setting. Hence, we could not be sure of
their representativeness with respect to the clinical population
13 97.4 70.7
and community. However, the correlational properties of
14 94 79.3 the C-EXIT25 to MCST performance indices in samples
15 90.7 87.2 collected at 2 different time-points (the current study vs the
16 87.2 92 pilot validation study13) retained a similar pattern, i.e. higher
17 83 95.7 scores on the C-EXIT25 correlated negatively with the
18 78 97.6 MCST category to a moderate degree (Pearson correlation
19 70.4 98.8 coefficient approximated to -0.5, p < 0.01). This suggests the
20 61.9 99.4 concurrent validity of the C-EXIT25 in elders was consistent
21 57.5 100 in different samples that were not randomly selected from the
22 52.6 100 reference population. The non-restrictive subject inclusion
criteria adopted in this study (i.e. inclusion irrespective
23 39 100
of diagnoses, co-morbid medical conditions or dementia
24 33.9 100
subtype) allowed our results to be generalised to real-
25 29.6 100 life clinical settings. Yet they also introduced unmeasured
26 23.7 100 confounders that might have influenced subject performance
27 18.6 100 in the respective cognitive tests. Future studies should be
28 15.3 100 conducted in larger samples to examine the discriminant
29 8.5 100 property of the C-EXIT25 in subtypes of dementia or
30 6.8 100 subgroups classified by their different medical / psychiatric
31 5.1 100 co-morbidities.
32 1.7 100
34 0 100
Conclusion
Abbreviation: C-EXIT25 = Chinese version of Executive
Interview.
The results of the current study and the previous pilot study
*
Area under the receiver operating characteristic curve = support that the C-EXIT25 is a potentially useful bedside
0.97 (95% confidence interval, 0.94 - 0.99). tool for ECF assessment in Chinese elders, by virtue of
good internal consistency, inter-rater reliability, concurrent
validity, and discriminatory power.

cut-off values on the C-EXIT25. Its AUC was 0.97 (95% Acknowledgement
confidence interval, 0.94-0.99; p ≤ 0.01). The cut-off value
of 15 best distinguishes CDR 0 and 0.5 from CDR 1 and 2 We would like to thank Prof DR Royall for his kind
(sensitivity = 90.7%; specificity = 87.2%). permission to adapt the original EXIT-25 to Chinese and all
his generous advice and support.
Discussion
Declaration of interest
The results of the present study support the discriminant
validity of the C-EXIT25 in Chinese elders, given that the This study is supported by the US National Institute of
AUC was close to 1.0 in the ROC analysis. A score of 15/50 Health’s Global Health Research Initiative Program For New
on the C-EXIT25 best discriminates CDR 0 / 0.5 from CDR Foreign Investigators (Grant Number: 5-R01TW0007256-
1 / 2 cases with optimal specificity plus sensitivity. The 03).

Hong Kong J Psychiatry 2009, Vol 19, No.3 101

SSM Chan, CHM Li, SLK Pang, et al

References cognitive impairment: the executive interview. J Am Geriatr Soc

1992;40:1221-6.
11. Anthony JC, LeResche L, Niaz U, von Korff MR, Folstein MF. Limits
1. Milner B. Effects of different brain lesions on card sorting. Arch
of the ‘Mini-Mental State’ as a screening test for dementia and delirium
Neurol 1963;9:90-100.
among hospital patients. Psychol Med 1982;12:397-408.
2. Drewe EA. The effect of type and area of brain lesion on Wisconsin
12. Nelson A, Fogel BS, Faust D. Bedside cognitive screening instruments.
card sorting test performance. Cortex 1974;10:159-70.
A critical assessment. J Nerv Ment Dis 1986;174:73-83.
3. Malloy P, Rasmussen S, Braden W, Haier RJ. Topographic evoked
potential mapping in obsessive-compulsive disorder: evidence of 13. Chan SM, Chiu FK, Lam CW. Correlational study of the Chinese
frontal lobe dysfunction. Psychiatry Res 1989;28:63-71. version of the executive interview (C-EXIT25) to other cognitive
4. Marenco S, Coppola R, Daniel DG, Zigun JR, Weinberger DR. measures in a psychogeriatric population in Hong Kong Chinese. Int J
Regional cerebral blood flow during Wisconsin Card Sorting Test in Geriatr Psychiatry 2006;21:535-41.
normal subjects studied by xenon-133 dynamic SPECT: comparison of 14. Chan CW, Lam LC, Wong TC, Chiu HF. Modified Card Sorting Test
absolute values, percent distribution values, and covariance analysis. Performance among community dwelling elderly Chinese people.
Psychiatry Res 1993;50:177-92. Hong Kong J Psychiatry 2003;13:2-7.
5. Bench CJ, Frith CD, Grasby PM, Friston KJ, Paulesu E, Frackowiak 15. de Zubicaray GI, Smith GA, Chalk JB, Semple J. The Modified Card
RS, et al. Investigations of the functional anatomy of attention using Sorting Test: test-retest stability and relationships with demographic
the Stroop test. Neuropsychologia 1993;31:907-22. variables in a healthy older adult sample. Br J Clin Psychol
6. Royall DR, Lauterbach EC, Cummings JL, Reeve A, Rummans TA, 1998;37:457-66.
Kaufer DI, et al. Executive control function: a review of its promise 16. Chiu HF, Lee H, Chung WS, Kwong PK. Reliability and validity of
and challenges for clinical research. A report from the Committee the Cantonese version of MMSE — a preliminary study. Hong Kong J
on Research of the American Neuropsychiatric Association. J Psychiatry 1994;4:25-8.
Neuropsychiatry Clin Neurosci 2002;14:377-405. 17. Chan AS, Choi A, Chiu H, Lam L. Clinical validity of the Chinese
7. Haaland KY, Vranes LF, Goodwin JS, Garry PJ. Wisconsin Card version of Mattis Dementia Rating Scale in differentiating dementia of
Sort Test performance in a healthy elderly population. J Gerontol Alzheimer’s type in Hong Kong. J Int Neuropsychol Soc 2003;9:45-55.
1987;42:345-6. 18. Hughes CP, Berg L, Danziger WL, Coben LA, Martin RL. A new clinical
8. Nelson HE. A modified card sorting test sensitive to frontal lobe scale for the staging of dementia. Br J Psychiatry 1982;140:566-72.
defects. Cortex 1976;12:313-24. 19. Morris JC, Storandt M, Miller JP, McKeel DW, Price JL, Rubin EH,
9. Robinson LJ, Kester DB, Saykin AJ, Kaplan EF, Gur RC. Comparison et al. Mild cognitive impairment represents early-stage Alzheimer
of two short forms of the Wisconsin Card Sorting Test. Arch Clin disease. Arch Neurol 2001;58:397-405.
Neuropsychol 1991;6:27-33. 20. de Zubicaray GI, Ashton R. Nelson’s (1976) Modified Card Sorting
10. Royall DR, Mahurin RK, Gray KF. Bedside assessment of executive Test: a review. Clin Neuropsychol 1996;10:245-54.

102 Hong Kong J Psychiatry 2009, Vol 19, No.3