This article was downloaded by: [Lekeu, Françoise]

On: 25 May 2009

Access details: Access Details: [subscription number 911472312]
Publisher Psychology Press
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Journal of Clinical and Experimental Neuropsychology

Publication details, including instructions for authors and subscription information:
http://www.informaworld.com/smpp/title~content=t713657736

The California Verbal Learning Test and other standard clinical

neuropsychological tests to predict conversion from mild memory impairment to
dementia
Françoise Lekeu abc; Delphine Magis ab; Patricia Marique ab; Xavier Delbeuck d; Sophie Bechet a; Bénédicte
Guillaume a; Stéphane Adam a; Jean Petermans c; Gustave Moonen a; Eric Salmon abc
a
Memory Centre, Department of Neurology, University Hospital, Liège, Belgium b Cyclotron Research Centre,
University of Liège, Liège, Belgium c Geriatric Day Hospital, University Hospital, Liège, Belgium d Memory
Clinic, University Hospital, Lille, France

First Published on: 20 May 2009

To cite this Article Lekeu, Françoise, Magis, Delphine, Marique, Patricia, Delbeuck, Xavier, Bechet, Sophie, Guillaume, Bénédicte,
Adam, Stéphane, Petermans, Jean, Moonen, Gustave and Salmon, Eric(2009)'The California Verbal Learning Test and other
standard clinical neuropsychological tests to predict conversion from mild memory impairment to dementia',Journal of Clinical and
Experimental Neuropsychology,
To link to this Article: DOI: 10.1080/13803390902889606
URL: http://dx.doi.org/10.1080/13803390902889606

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial or
systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or
distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses
should be independently verified with primary sources. The publisher shall not be liable for any loss,
actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly
or indirectly in connection with or arising out of the use of this material.
 JOURNAL OF CLINICAL AND EXPERIMENTAL NEUROPSYCHOLOGY
2009, iFirst, 1–12

The California Verbal Learning Test and other standard

NCEN

clinical neuropsychological tests to predict conversion

from mild memory impairment to dementia

Françoise Lekeu,1,2,3 Delphine Magis,1,2 Patricia Marique,1,2 Xavier Delbeuck,4

POWER OF CVLT IN DEMENTIA PREDICTION

Sophie Bechet,1 Bénédicte Guillaume,1 Stéphane Adam,1 Jean Petermans,3

Gustave Moonen,1 and Eric Salmon1,2,3
1
Memory Centre, Department of Neurology, University Hospital, Liège, Belgium
2
Cyclotron Research Centre, University of Liège, Liège, Belgium
3
Geriatric Day Hospital, University Hospital, Liège, Belgium
4
Memory Clinic, University Hospital, Lille, France
Downloaded By: [Lekeu, Françoise] At: 09:48 25 May 2009

This study describes the neuropsychological assessment of 34 patients with questionable Alzheimer’s disease
(QAD) followed up for 3 years. Several measures were selected from the California Verbal Learning Test (CVLT)
and compared to other cognitive tasks to assess the best neuropsychological indices for (a) detecting early memory
impairment in QAD and (b) predicting conversion to AD. Concerning detection, the results indicated that a recall
measure depending on semantic categorization (short-delay cued recall) signaled a memory deficit in stable QAD
patients, suggesting that episodic and semantic memory problems are involved in the early cognitive impairments
of stable QAD patients. However, the conversion to AD was best predicted by the initial performance at the
recency index (score reflecting high reliance on working memory), corroborating the idea that AD patients (even
at the questionable stage) essentially rely on preserved phonological loop functioning in memory tasks. Finally, an
additional impairment in visuospatial memory (Rey’s figure) provided a good discriminant value to distinguish
converters from stable QAD patients, showing that various cognitive disabilities deteriorate in AD.

Keywords: Questionable Alzheimer’s disease; Mild cognitive impairment; California Verbal Learning Test;
Conversion to dementia; Prediction.

INTRODUCTION patients with isolated memory impairment were

Differentiating people with mild memory loss who
will remain stable from those destined to develop
Alzheimer’s disease (AD) represents a real chal-
lenge, in particular from the perspective of early
cognitive and/or drug support.
To achieve this goal, the number of longitudinal
studies of patients with mild cognitive impairment
(MCI) or questionable Alzheimer’s disease (QAD)
has increased over the last 10 years. In these studies,
tested with a large battery of classical neuropsy-
chological tests to isolate, at the end of a given
follow-up period, which of the initial cognitive
measures were the most sensitive in distinguish-
ing converters—that is, patients who developed
AD—from those patients who remained stable
(Fowler, Saling, Conway, Semple, & Louis, 1997;
Petersen et al., 1999).
A number of these studies have shown that mea-
sures of episodic memory performance may be the

This study was conducted on behalf of the Network for Efficiency and Standardization of Dementia Diagnosis (NEST-DD),
supported by the European Commission (5th framework), and it was finalized under the auspices of the EC-FP6-project DiMI, LSHB–
CT–2005–512146. The work was also supported by grants from the Fonds National de la Recherche Scientifique (FNRS) in Belgium,
the IUAP P6/29, the University Hospital of Liège, and the University of Liège.
Address correspondence to Françoise Lekeu, Memory Centre and Geriatric Day Hospital–University Hospital of Liège, Quai
Godefroid Kurth 45, B-4020 Liège, Belgium (E-mail: francoise.lekeu@hotmail.com).

© 2009 Psychology Press, an imprint of the Taylor & Francis Group, an Informa business
http://www.psypress.com/jcen DOI: 10.1080/13803390902889606
 2 LEKEU ET AL.
best indicators for differentiating people who will
develop AD from those who will not (Albert,
Moss, Tanzi, & Jones, 2001; Artero, Tierney,
Touchon, & Ritchie, 2003; Bäckman, Jones,
Berger, Laukka, & Small, 2004; Blackwell et al.,
2004; Chen et al., 2000). However, to increase the
sensitivity and specificity of the predictive mea-
sures, memory performance should be considered
in combination with other “nonmemory” measures
(Albert et al., 2001; Artero et al., 2003; Chen et al.,
2000; Tabert et al., 2006). This is not surprising
when one considers that pure, isolated memory
impairment is rare (Alladi, Arnold, Mitchell,
Nestor, & Hodges, 2006) and that cognitive
impairment in the predementia stage of AD can be
observed in many domains, including episodic
memory, semantic memory, executive functions,
attention, verbal abilities, visuospatial skills, and
perceptual speed (Bäckman et al., 2004; Bäckman,
Downloaded By: [Lekeu, Françoise] At: 09:48 25 May 2009
Jones, Berger, Laukka, & Small, 2005). In the same
vein, some authors recently proposed revising the
diagnostic criteria for MCI to incorporate a change
in activity level and impairment in nonmnesic
cognitive functions (Artero, Petersen, Touchon, &
Ritchie, 2006).
According to Petersen et al. (1999), verbal epi-
sodic memory is the first cognitive domain
impaired in the AD process. Supporting this sug-
gestion, Chen et al. (2000) found that, among a
battery of 16 neuropsychological tests, the delayed
recall of a word list combined with the Trail
Making Test B measure offered the best predictive
accuracy for dementia conversion. In the same
vein, Artero et al. (2003) entered the data from two
prospective studies in a regression analysis to pre-
dict participants at risk for AD conversion. They
confirmed that delayed verbal recall is an import-
ant measure to take into account, but it must be
combined with data from construction and cate-
gorical fluency tests. Nevertheless, some studies
also highlighted the high predictive power of visual
memory performance in differentiating converters
from nonconverters (Albert et al., 2001; Blackwell
et al., 2004; Fowler et al., 1997; Masur, Sliwinski,
Lipton, Blau, & Crystal, 1994).
The differences between these studies concerning
the best predictor of conversion to dementia might
be due to various factors, such as the definition of
MCI and the population studied (selection of per-
sons with memory problems or individuals selected
to represent the general population). Moreover,
the tests used in these kinds of studies are quite
different, and some of them are only used in experi-
mental settings. However, it is important to challenge
the clinical value of well-known standard neuro-
psychological tasks in making an accurate diagnosis.
The California Verbal Learning Test (CVLT) is a
widely used neuropsychological test that assesses
episodic memory (Delis, Freeland, Kramer, &
Kaplan, 1988; Delis, Kramer, Kaplan, & Ober,
1987). The test begins with a 16-item word learning
test (List A) presented over five trials. Immediately
after the last learning trial, a second list of 16
words (List B) is presented as an interference trial,
followed by short (a few minutes) and long (20
minutes) delayed free and cued recall of the List A
and a recognition memory trial. Consequently,
analysis of results on this test can provide informa-
tion about the functioning of several cognitive
domains, since the CVLT requires the contribution
of episodic memory but also, among other things,
of semantic memory and executive abilities of
categorization.
To the best of our knowledge, no follow-up
study has described the initial performance of
patients with MCI or QAD on the CVLT. The
three-year follow-up study done by Albert et al.
(2001) only used the CVLT basic measure of total
recall and found that this measure had little sensi-
tivity and specificity for an early diagnosis of mem-
ory impairment. Moreover, the measure did not
have sufficient discriminant power to differentiate
converters from stable patients. More recently,
Greenaway et al. (2006) used a large number of
memory measures from the CVLT to analyze the
pattern of memory dysfunction across groups of
normal controls (NC), MCI patients, and AD
patients. The results showed that, compared to NC,
MCI patients displayed reduced learning, rapid
forgetting, increased recency recall, more intrusion
errors, and poor recognition discriminability. A dis-
criminant function analysis demonstrated that
delayed recall and total learning were the best mea-
sures for distinguishing among the three groups.
Unfortunately, that study did not present follow-
up data for the MCI patients.
Based upon these observations, our study could
be viewed as an extension of the two studies
described above. Patients with “isolated” memory
impairment and no significant decrease in daily
activities were selected in memory clinics where dif-
ferent neuropsychological tests (but not the CVLT)
were used to assess memory performance. Patients
were referred to the Cyclotron Research Centre to
enter a European multicentre study (Network for
Efficiency and Standardization of Dementia
Diagnosis, NEST-DD), for which the CVLT was
selected as a commonly available test to evaluate
verbal episodic memory. Different CVLT scores
and other baseline neuropsychological tests were
analyzed, searching for the cognitive indices that
would prove most useful for (a) the objective

detection of early memory impairment in QAD
versus control participants and (b) the early diagnosis
of patients converting to AD during a three-year
follow-up period.
METHOD
Participants
The participants were 34 patients referred by neu-
rologists working in memory clinics to participate
in the NEST-DD European multicentre study.
They had memory problems confirmed by a rela-
tive, but without any significant impairment in
daily living activities (assessed by Lawton’s scale;
Lawton & Brody, 1969) or global cognitive deteri-
oration. Indeed, all participants demonstrated
independent functioning in activities of daily living
Downloaded By: [Lekeu, Françoise] At: 09:48 25 May 2009
with minor interference of cognitive impairments,
since they all scored 1 at each section of Lawton’s
scale. The patients had undergone neurological,
neuropsychological, and structural neuroimaging
evaluations and were classified as QAD since they
had a Clinical Dementia Rating (CDR) score of
0.5 (Morris, 1993), and they did not meet criteria
for Alzheimer-type dementia (McKhann et al.,
1984). They met the criteria for either amnestic
MCI (with at least one memory test performance
1.5 standard deviations below the mean for
age-matched controls) or multiple-domain MCI
(when cognitive performance was also decreased in
another, nonmemory, domain; Petersen et al.,
2001; Winblad et al., 2004). Exclusion criteria were
dementia, mental retardation, fewer than four
years of education, brain trauma, epilepsy, cancer,
depression, any major systemic disease, or any sub-
stance abuse. At the time of inclusion, all patients
were free of medication that could noticeably
affect brain function.
All QAD patients had Mini-Mental State Exami-
nation (MMSE) scores of 21 or over at baseline
evaluation (MMSE; Folstein, Folstein, & McHugh,
1975). Note that the exclusion criteria included a
CDR score of greater than 0.5, attributed by the
referring neurologist, while the MMSE score at entry
into the study was taken as an experimental variable.
All patients were clinically followed up for 36
months, if they did not convert to AD before the end
of this period. At the final follow-up session, diag-
noses were reviewed on the basis of a multidisciplinary
discussion taking into account neuropsychological,
neurological, functional, and behavioral factors.
The diagnosis of AD was given according to gener-
ally accepted clinical criteria (McKhann et al.,
1984). At 36 months, 17 patients (50%) remained
POWER OF CVLT IN DEMENTIA PREDICTION 3
stable, and 17 patients (50%) had converted to proba-
ble Alzheimer’s disease (PrAD). In this latter group,
2 patients converted within 6 months, 4 patients
converted within 12 months, 6 patients converted
within 18 months, 4 converted within 24 months,
and 1 converted in 36 months. The mean follow-up
duration for the study was 26.8 months.
Normal control participants (NC) were 14 cogni-
tively and neurologically intact elderly adults who
lived in the community and were recruited by word
of mouth. The exclusion criteria were the same as
those used for the patients, and the NC had a CDR
score of zero. All control participants completed the
experimental neuropsychological testing.
All patients and controls gave their informed
consent to participate in the study, which was
approved by the Ethics Committee of the University
Hospital Centre in Liège.
Neuropsychological tests
At the baseline session, all QAD patients and NC
performed a neuropsychological battery that exam-
ined verbal and visual episodic memory as well as
visuoperceptive processing, working memory,
language, and executive functions.
For verbal episodic memory, the French adapta-
tion of the California Verbal Learning Test was
used (CVLT; Delis et al., 1988; Delis et al., 1987).
Performance was carefully examined by using
different index and memory scores.
Short-delay recall
In the CVLT, there are two lists of 16 words (List
A and List B) belonging to four semantic categories.
The measures for List A were: the total number of
correct words produced during Trials 1 to 5; the
short-delay free recall, corresponding to the number
of correct words freely recalled immediately after
the recall of List B; and the short-delay cued recall,
which corresponds to the number of correct words
recalled (over and above the free recall) in response
to semantic cues. Three indices were also calculated:
a constancy learning index, which corresponds to
[(the number of times an item correctly recalled dur-
ing Trials 1 to 4 is recalled during Trial 5)/(total free
recall for Trials 1 to 4)] × 100; a primacy index,
which corresponds to [(cumulative recall of the first
four items of List A during Trials 1 to 5)/(total free
recall for Trials 1 to 5)] × 100; and a recency index,
which corresponds to [(cumulative recall of the last
four items of List A during Trials 1 to 5)/(total free
recall for Trials 1 to 5)] × 100.
The only measure for List B is the total number
of correct words freely recalled.
 4 LEKEU ET AL.
Long-delay recall
After a 20-minute delay, correct words (List A)
were measured in both free- and cued-recall
conditions.
Errors
The total number of intrusions produced during
the recall trials was measured.
Recognition
Recognition was tested after the long-delay
recall. The following measures were obtained: total
number of false recognitions; false recognitions
from List B; a discrimination index, which assesses
recognition memory performance more accurately
by taking into account hits, false recognitions, and
omissions; and finally a response bias, which char-
acterizes the tendency for a participant to respond
Downloaded By: [Lekeu, Françoise] At: 09:48 25 May 2009
“yes” or “no” in recognition memory tasks (a neu-
tral response bias = 0).
Organization in free recall
Concerning organization in free recall, the CVLT
also provides two specific measures: semantic clus-
tering and serial order recall. Semantic clustering
demonstrates how participants organize the word
list during learning (encoding). This measure corre-
sponds to the number of consecutively recalled
words from the same semantic category, taking
into account the total number of words recalled.
Serial order recall is the number of words recalled
in the same order as presented. A proportional
measure (serial cluster ratio) is obtained by dividing
the number of serial order clusters by the maximum
number of order clusters, which in turns depends on
the total number of words recalled. Details about
the formula used for the index calculation have
been published by Delis et al. (1988).
Visual episodic memory was assessed by using the
30-minute delayed recall of the Rey complex figure
(Rey, Osterrieth, & Taylor, 1991).
The other cognitive tests used were: copying of
the Rey complex figure for visuoperceptive pro-
cessing; the forward digit span and block tapping
tests for working-memory evaluation; the cate-
gory (animals: 120 s) and phonemic fluency tests
(P, R, V: 120 s each) for language assessment; and
the Stroop test for evaluation of executive func-
tions (and specifically inhibition; Lezak, 1983).
The MMSE was used to assess global cognitive
impairment (Folstein et al., 1975), and the 21-item
Hamilton scale evaluated depressive symptoms
(Hamilton, 1967).
Data analysis
A one-way analysis of variance (ANOVA), with
each cognitive measure as the dependent variable
and a factor representing the three groups (NC,
stable QAD, converting QAD), was first per-
formed. Planned comparisons were then carried
out between groups (NC vs. stable QAD, stable
QAD vs. converting QAD, NC vs. converting
QAD, whole QAD group vs. NC). Only cognitive
measures that were significantly different (using
Bonferroni corrections) among groups were
entered in forward stepwise regression analyses,
which were followed by associated discriminant
analyses and the receiver operating characteristic
(ROC) curve method. We a priori decided to
introduce only neuropsychological variables in the
subsequent analyses.
Our different regression and discriminant
analyses aimed to answer four questions concern-
ing patients referred to memory clinics: (a) The
first analysis, contrasting normal control and sta-
ble QAD groups, investigated the sensitivity of
the CVLT indices for the detection of early
memory impairment; (b) the analysis contrasting
the stable and converting QAD groups was
designed to isolate which neuropsychological
measure(s) would be the best predictors of demen-
tia in QAD; (c) the analysis contrasting the nor-
mal controls and the converters was designed to
identify the most accurate neuropsychological
measures for early dementia diagnosis (by com-
paring the two previous analyses and this third
analysis); (d) as is done in the literature, we also
contrasted the normal controls and the whole
QAD group, in order to highlight the neuropsy-
chological measures that best detect the QAD
“syndrome.”
RESULTS
Distributional and demographic analyses
Demographic and clinical features for the NC,
stable QAD, and converting groups are presented
in Table 1. There were no statistically significant dif-
ferences between groups with respect to educational
level, F(2, 45) = 0.65, ns, or Hamilton Rating Scale
for Depression (HAM-D) score of depression,
F(2, 45) = 0.25, ns. Sex distribution was statistically
independent from final diagnosis, c2(1) = 2.95, p <
.08, although there were more women in the con-
verting group (71%) than in the stable QAD group
(35%). The mean MMSE score of the whole QAD
group was 24.9 (SD 2.2). As expected, the normal

TABLE 1
Demographic variables of study participants
NC Stable QAD Converting QAD
Characteristics (N = 14) (N = 17) (N = 17)
No. women (%) 8 (57) 6 (35) 12 (71)
Age (in years) 66.0 (6.6) 66.6 (6.9) 72.0 (5.9)a
Education 12.0 (3.1) 12.3 (5.2) 10.8 (2.5)
(in years)
Ham-D 3.2 (3.8) 2.5 (2.8) 2.6 (2.7)
MMSE 29.1 (0.9) 25.7 (2.0)b 24.1 (2.1)a
Note. NC = normal controls. QAD = questionable Alzheimer’s
disease. Ham-D = Hamilton Rating Scale for Depression;
MMSE = Mini Mental State Examination. Standard deviations
in parentheses.
a
Significantly different from controls and stable QAD group.
b
Significantly different from controls.
controls had higher MMSE scores than the
Downloaded By: [Lekeu, Françoise] At: 09:48 25 May 2009
patients, F(1, 45) = 53.4, p < .0001. Moreover, the
converters had lower MMSE scores at entry than
the stable QAD participants, F(1, 45) = 7.43, p <
.01. The converting patients were also older at
entry than the two other groups, as is well estab-
lished in the literature. MMSE score was entered
as a potential predictive variable in all regression
analyses. Since age differed between groups, and
more importantly between converters and stable
QAD patients, we checked that introducing age in
all relevant regression and discriminant analyses
did not change our results. Age was never retained
as a discriminant variable.
Neuropsychological data analysis
In the stable QAD group, visual working memory
(Block Tapping Test), several results on the CVLT
(total free recall, short- and long-delay free recall,
short- and long-delay cued recall, free recall of List B,
total and List B false recognitions, discrimination
and recency indices, constancy learning, and
semantic clustering), and fluency scores were
impaired compared to those for controls (Table 2).
Similar deficits were observed in the converters
compared to the controls. Moreover, visual epi-
sodic memory and some CVLT indices (total free
recall, short- and long-delay free recall, short- and
long-delay cued recall, total and List B false recog-
nitions, discrimination and recency indices) were
lower in the converters than in the stable group. By
using a Bonferroni correction (p < .002), only those
variables that differed strongly between groups
were entered into the regression equation. Note
that the effect sizes were positive for all neuropsy-
chological variables (Table 2).
POWER OF CVLT IN DEMENTIA PREDICTION 5
Detection of memory deficit (versus normal
forgetting) in elderly participants
When comparing NC versus stable QAD
patients, a total of four variables (MMSE, free
recall of List B, short-delay free recall and short-
delay cued recall of List A) were entered into the
regression analysis, and only MMSE and short-
delay cued recall were retained as potentially signi-
ficant explanatory variables, F(2, 30) = 13.08,
adjusted R2 = .43, p < .0001. These two variables
were subsequently entered into a discriminant ana-
lysis. The results showed that the MMSE score
correctly classified 87% of participants, with a
sensitivity of 93% and a specificity of 81%. The
short-delay cued recall score correctly classified
80% of participants, with a sensitivity of 86% and a
specificity of 75%. The combination of these two
variables allowed us to correctly classify 87% of
participants (sensitivity 93% and specificity 81%).
By using ROC curves, we plotted the sensitivity
against 1-specificity for these two cognitive scores
(Figure 1). The figure confirmed that the MMSE
had better discriminant validity for the distinction
between the NC and stable QAD group (the area
under the curve was 0.94; SD = 0.04) than the
short-delay cued recall score (the area under the
curve was 0.83; SD = 0.08).
Discrimination of QAD converters versus
nonconverters
When comparing stable QAD versus converting
QAD patients, a total of four variables (delayed
recall of Rey complex figure, total free recall 1 to 5,
long-delay free recall, and recency index) were
entered in the regression analysis. The status of
converters compared with stable QAD partici-
pants was best predicted by the visual memory
measure of recall of Rey’s figure and the recency
index of the CVLT, F(2, 30) = 13.08, adjusted
R2 = .43, p < .0001. These two predictive variables
allowed us to correctly classify 75% of patients,
with a sensitivity of 72% and a specificity of 80%.
Taken alone, the recall of Rey’s figure correctly
classified 76% of patients, with a sensitivity and
specificity of 76%, while the recency index cor-
rectly classified 72% of patients, with a sensitivity
of 73% and a specificity of 72%. By using ROC
curves, we plotted the sensitivity against 1-specificity
for these two cognitive scores (Figure 2). The
figures showed that the score for the recall of
Rey’s figure had only a slightly better discriminant
validity for the distinction between the stable and
converter QAD groups (area under the curve was
0.84; SD = 0.07) than the recency index (area
under the curve was 0.83; SD = 0.07).
 6 LEKEU ET AL.

TABLE 2
Mean performance of the control, stable QAD, and converting QAD groups on the different neuropsychological tests

Stable Converting
Control group QAD QAD p value Post hoc Effect
Neuropsychological testing (N = 14) (N = 17) (N = 17) (2-tailed) (p < .05) size

Working memory Digit span 6.1 (1.3) 5.6 (2.1) 5.5 (1.8) .62 — .021
Block Tapping Test 5.9 (1.8) 4.7 (1.3) 4.5 (1.7) .04 STCV < NC .131
Visual episodic Delayed recall (/36) 18.3 (6.2) 15.9 (6.6) 7.9 (4.5) .0001 NCST > CV .383
memory (Rey’s figure)
Verbal episodic (A) Total free recall 54.6 (10.6) 43.3 (12.7) 3.3 (7.4) .0001 NC > ST > CV .479
memory (CVLT) Trials 1–5
(A) Short-delay free recall 11.1 (3.0) 7.1 (3.8) 3.6 (2.3) .0001 NC > ST > CV .499
(A) Short-delay cued recall 12.3 (2.1) 8.3 (3.7) 6.0 (2.1) .0001 NC > ST > CV .477
(A) Long-delay free recall 11.5 (2.7) 7.9 (4.1) 3.8 (2.6) .0001 NC > ST > CV .496
(A) Long-delay cued recall 12.2 (2.3) 9.0 (3.8) 6.0 (2.5) .0001 NC > ST > CV .419
(B) Free recall 7.7 (3.0) 4.7 (2.0) 3.6 (2.2) .0001 STCV < NC .337
Total intrusion errors 7.6 (6.8) 8.2 (6.6) 13.9 (12.6) .11 — .091
Total false recognitions 0.9 (0.9) 3.2 (3.0) 5.5 (4.0) .0006 NC > ST > CV .283
False recognitions, List B 0.4 (0.5) 1.7 (1.4) 2.6 (1.4) .00007 NC > ST > CV .351
Discrimination index 95.9 (2.8) 89.3 (8.0) 83.3 (9.3) .0002 NC > ST > CV .329
Downloaded By: [Lekeu, Françoise] At: 09:48 25 May 2009

Response bias 0.3 (0.5) 0.4 (0.5) 0.4 (0.4) .84 — .007
Primacy index 29.2 (7.6) 27.5 (6.5) 26.3 (12.2) .67 — .018
Recency index 24.7 (7.3) 28.4 (8.5) 41.3 (1.3) .00001 NCST > CV .407
Constancy learning 85.4 (8.0) 74.9 (13.0) 73.7 (13.3) .05 STCV < NC .164
Semantic clustering 2.6 (0.9) 1.7 (1.0) 1.5 (.9) .05 STCV < NC .198
Serial order clustering 1.8 (1.0) 1.4 (0.8) 1.2 (1.3) .31 — .054
Language functions Semantic fluencya 35.4 (8.6) 28.2 (8.6) 24.5 (5.6) .0005 STCV < NC .29
Phonemic fluencya 18.6 (4.7) 13.9 (5.0) 15.5 (4.2) .05 NC > ST .139
Executive functions Stroop test (interference 20.8 (14.4) 25.6 (13.6) 29.7 (5.9) .13 — .088
index)
Visuoperceptive Copy (/36) 31.7 (5.7) 32.8 (2.9) 31.5 (3.5) .56 — .027
function (Rey’s figure)

Note. QAD = questionable Alzheimer’s disease. CVLT = California Verbal Learning Test. NCST > CV = performance in normal con-
trol and stable groups is better than that in converting group; NC > ST > CV = performance in normal controls is better than that in
stable and converting groups, and performance in stable group is better than that in converting group; STCV < NC = performance in
stable and converting groups is worse than that in normal control group; NC > ST = performance in normal control group is better
than that in stable group; all p < .05. Standard deviations in parentheses.
a
Z scores are reported for phonemic and semantic fluency.

From a clinical viewpoint, this distinction is
clearly the most important. The cutting score was
11.6 for Rey’s figure recall, and it was 34.6 for the
recency index.
Detection of very early AD versus elderly
controls
For the comparison between NC and converters,
a total of 12 variables were entered in the analysis
(MMSE, delayed recall of Rey complex figure,
total free recall 1 to 5, free recall List B, short-delay
free and cued recall, long-delay free and cued
recall, false recognitions, recency index, discrimi-
nation index, semantic fluency). The status of
prodromal AD was best predicted by a model com-
posed of the MMSE, the recency index, and the
short-delay cued recall measure, F(3, 23) = 49.22,
adjusted R2 = .84, p < .00001. Taken alone, the
MMSE score correctly classified 100% of patients,
with a sensitivity of 100% but a specificity of
87.5%. Taken alone, the short-delay cued recall
measure correctly classified 93% of patients, with a
sensitivity of 94% and a specificity of 92%. The
recency index alone correctly classified 86% of
patients, with a sensitivity of 93% and a specificity
of 81%. The combination of the three variables
allowed us to correctly classify 100% of partici-
pants, with a sensitivity and a specificity of 100%.
By using ROC curves, we plotted the sensitivity
against 1-specificity for these three cognitive scores
(Figure 3). These figures showed that the short-
delay cued recall score (area under the curve was
0.98; SD = 0.02) had a slightly better discriminant
validity for the distinction between the NC and
converting QAD groups than the MMSE score
 POWER OF CVLT IN DEMENTIA PREDICTION 7

ROC Curve ROC Curve

1,00 a 1,00

,75 ,75
Sensitivity

Sensitivity
,50 ,50

Reference Line

,25 SDCR ,25

MMSE

0,00 0,00
0,00 ,25 ,50 ,75 1,00 0,00 ,25 ,50 ,75 1,00
1 - Specificity 1 - Specificity
Diagonal segments are produced by ties. Diagonal segments are produced by ties.
Downloaded By: [Lekeu, Françoise] At: 09:48 25 May 2009

Figure 1. Receiver-operating characteristic (ROC) curve for

Mini-Mental State Examination (MMSE) and short-delay cued ROC Curve
recall (SDCR) scores for discrimination between normal controls
b 1,00
(NC) and stable questionable Alzheimer’s disease (QAD) groups.

(area under the curve was 0.97; SD = 0.03) or the ,75

recency index score (area under the curve was 0.91;
SD = 0.05).
Sensitivity

When converters where compared to normal

elderly controls, the cutting score was 27 for the
MMSE, 32.4 for the recency index, and 9.1 for the
short-delay cued recall.
Initial detection of the entire QAD group
versus elderly controls
To comply with the literature, we finally
performed a comparison between the whole QAD
group and the controls, and a total of nine varia-
bles were entered in the analysis (MMSE, semantic
fluency, total free recall 1 to 5, free recall List B,
short-delay free and cued recall, long-delay free
and cued recall, discrimination index). The status
of QAD compared to NC was best predicted by a
model composed of the MMSE and the short-
delay cued recall, F(2, 44) = 31.38, adjusted R2 =
.56, p < .00001. Taken alone, the MMSE score cor-
rectly classified 89% of patients with a sensitivity of
97% but a specificity of 77%. The short-delay cued
recall measure alone correctly classified 83% of
patients, with a sensitivity of 88% but a specificity
of 71%. The combination of both variables allowed
us to correctly classify 91% of participants, with a
sensitivity of 97% and a specificity of 81%. By
using ROC curves, we plotted the sensitivity
against 1-specificity for these two cognitive scores
(Figure 4). This figure showed that the MMSE had
,50
,25
0,00
0,00 ,25 ,50 ,75 1,00
1 - Specificity
Diagonal segments are produced by ties.
Figure 2. Receiver-operating characteristic (ROC) curves for
delayed recall of Rey’s figure (a) and recency index (b) for
discrimination between stable and converter questionable
Alzheimer’s disease (QAD) groups.
a better discriminant validity for the distinction
between the whole QAD group and NC (area
under the curve was 0.96; SD = 0.02) than did the
short-delay cued recall score (area under the curve
was 0.90; SD = 0.04).
Additionally, we performed the same analyses
on QAD patients with a MMSE score of 23 and
more, allowing us to equate MMSE score in stable
QAD participants (26.1 ± 1.6) and converters
(25.1 ± 1.7), and the results were essentially similar
(data not shown).
 8 LEKEU ET AL.

ROC Curve ROC Curve

a 1,00 1,00

,75 ,75
Sensitivity

Sensitivity
,50 ,50

Reference Line
Reference Line
,25 SDCR
,25 SDCR
MMSE
MMSE

0,00
0,00 ,25 ,50 ,75 1,00 0,00
0,00 ,25 ,50 ,75 1,00
1 - Specificity
1 - Specificity
Diagonal segments are produced by ties.
Diagonal segments are produced by ties.
Downloaded By: [Lekeu, Françoise] At: 09:48 25 May 2009

ROC Curve
b 1,00 Figure 4. Receiver-operating characteristic (ROC) curve for
Mini-Mental State Examination (MMSE) and short-delay cued
recall (SDCR) scores for discrimination between normal con-
trols (NC) and the whole questionable Alzheimer’s disease
(QAD) group.
,75

controls using the same tests and the recency index

Sensitivity

,50 of the CVLT, with an impressive correct classifica-

,25
0,00
0,00 ,25 ,50 ,75
1 - Specificity
Diagonal segments are produced by ties.
Figure 3. Receiver-operating characteristic (ROC) curves for
Mini-Mental State Examination (MMSE) and short-delay cued
recall (SDCR) scores (a) and recency index (b) for discrimina-
tion between normal controls (NC) and converter questionable
Alzheimer’s disease (QAD) groups.
DISCUSSION
This study showed that patients with a stable syn-
drome of questionable AD can be distinguished
from controls by their MMSE scores and their
performance on the short-delay cued recall com-
ponent of the CVLT. These two scores provided a
correct classification rate of 87%, demonstrating
their importance for detecting mild cognitive
impairments in our population of elderly partici-
pants. Converters could be differentiated from
tion rate of 100% (corresponding to both sensitiv-
ity and specificity of 100%). Finally, converters
could be distinguished from stable QAD partici-
pants by their delayed recall performance for
Rey’s complex figure and the recency index of the
CVLT, but the correct classification rate remained
modest (75%).
1,00 The data were obtained in relatively small
samples of patients, but they are quite significant.
Although replications studies are required before
generalizing our observations, the data are con-
gruent with the literature showing that different
memory tests are useful for detecting mild cogni-
tive impairments in elderly populations and that
direct comparisons are mandatory to know the
relative discriminant values of the tests (Ivanoiu
et al., 2005). Although combining scores was neces-
sary to achieve optimum discrimination, ROC
curves demonstrated that each cognitive test had
good discriminant value, and they are discussed
individually.
Age is one important risk factor for AD, and
this was confirmed in our study (as it has been in
many previous reports on MCI), since converters
were older than the other two groups (Kryscio,
Schmitt, Salazar, Mendiondo, & Markesbery,
2006). However, age was not retained as a predic-
tive variable.

The California Verbal Learning Test and the
influence of semantic abilities and working
memory
Although four neuropsychological scores emerged
as being the best at discriminating between groups,
other variables also distinguished our populations.
Our group of patients with stable cognitive impair-
ment (stable QAD) performed worse than controls
on several neuropsychological tests (visual span,
CVLT learning score, all recall performances on the
CVLT, false recognitions and the related discrimi-
nation index, lower constancy in CVLT learning,
decreased semantic clustering, and decreased
verbal fluency). However, only the MMSE score
and three CVLT memory scores were retained
when multiple comparisons were done.
The CVLT contains many memory retrieval
scores, and our study found the short-delay cued
Downloaded By: [Lekeu, Françoise] At: 09:48 25 May 2009
recall to be the best at discriminating both stable
and converting QAD participants from controls.
To the best of our knowledge, no previous study
has characterized performance on the different
parts of the CVLT in order to detect isolated mem-
ory impairment and predict conversion to dementia.
Cued recall may particularly well capture the dis-
tinction between patients and controls, because the
control participants make better use of semantic
information to guide cued retrieval, as suggested
by their better semantic clustering abilities on the
CVLT and better verbal fluency (see Table 2). These
results highlight the importance of a semantic recall
measure in memory tests to detect early memory
impairments (Adam et al., 2007; Bäckman et al.,
2005; Blackwell et al., 2004). According to the
recent literature, the MCI population is character-
ized by deficits in cognitive domains beyond epi-
sodic memory, and especially in semantic memory
(Adlam, Bozeat, Arnold, Watson, & Hodges, 2006;
Alladi et al., 2006; Dudas, Clague, Thompson,
Graham, & Hodges, 2005; Duong, Whitehead,
Hanratty, & Chertkow, 2006; Hodges, Erzinclioglu, &
Patterson, 2006; Ribeiro, de Mendonça, & Guerreiro,
2006; Thompson, Graham, Patterson, Sahakian, &
Hodges, 2002). The decrease we observed in verbal
fluency, and essentially in semantic fluency, in
stable QAD (before correction for multiple com-
parisons) suggests that executive dysfunction (and
impaired access to semantic memory) might occur
in some of those “stable” patients. Semantic fluency
was impaired in the converting QAD group, but it
was not retained in the discriminant analyses, sug-
gesting that the fluency deficit was less significant
than the memory impairment in our population of
converters. In summary, our results suggest that
the short-delay cued recall measure of the CVLT is
POWER OF CVLT IN DEMENTIA PREDICTION 9
good at detecting memory impairment, but would
not be the best measure for predicting dementia.
Numerous researchers have used the CVLT in
the context of AD, and they have consistently
found a decline in memory performance on many
scores compared to normal control participants
(Greenaway et al., 2006; Kohler, 1994). In a previ-
ous longitudinal study, the Dementia Rating Scale
performance (Mattis, 1973), a score of global cog-
nitive performance, and the long-delay free recall
score showed the best classification power between
AD patients and controls, but the total learning
score for the five free recall trials was also a good
indicator of AD (Salmon et al., 2002). In addition
to impaired learning and retrieval, heightened
recency effects, ineffective use of semantic clustering,
and increases in intrusion errors also typified the
CVLT performance of patients with mild AD
(Greenaway et al., 2006; Simon, Leach, Winocur, &
Moscovitch, 1994). In our analysis, the recency
index was the only CVLT measure retained to dis-
tinguish converters from stable patients with
QAD. The tendency to preferentially retrieve the
last items of List A during the five learning trials
characterizes a strategy based on working memory
that is favored by patients with prodromal AD.
The MMSE score
The discussion of multiple cognitive impairments
in stable QAD is relevant to our results showing
that MMSE performance is significantly impaired
in our QAD patients. MMSE score was an experi-
mental variable in our study. Although the test was
given to patients in memory clinics before their
entry into the study, “priming” or “test–retest”
effects were somewhat reduced because the envir-
onment (the Cyclotron Research Centre) was com-
pletely new for all participants. The composite
score was lower in the converting QAD group than
in the two other groups, but it was also lower in the
stable QAD group than in controls. This is not
unexpected, since the MMSE comprises different
orientation/memory/attentional questions, and
MMSE impairment was described in a previous
longitudinal study of 82 MCI patients, 54 of whom
converted to dementia during a 42-month follow-up
period (Huang et al., 2003). The MMSE is a widely
used screening test; it has a ceiling effect and thus
lacks sensitivity for patients with mild dementia
(Tierney, Szalai, Dunn, Geslani, & McDowell,
2000). Three large epidemiological studies have
reported good sensitivity, specificity, and likeli-
hood ratio for the positive test of the MMSE in
dementia diagnosis (Grut, Fratiglioni, Viitanen, &
 10 LEKEU ET AL.
Winblad, 1993; Monsch et al., 1995; Ritchie &
Fuhrer, 1992). But our results are closer to those
obtained by Benson, Slavin, Tran, Petrella, and
Doraiswamy (2005). In this study, the MMSE dis-
tinguished MCI patients from control participants
(cutoff score of 27), with good specificity (72%) but
poor sensitivity (57%), and mild AD patients (cut-
off score of 24) from normal controls with high
specificity (100%) and modest sensitivity (75%).
Our stable QAD patients had a mean MMSE score
of 25.7, and thus the MMSE allowed us to distin-
guish them from normal controls with both good
specificity and good sensitivity (81% and 93%,
respectively). As in the Benson et al. study, AD
patients (converters in our study) were distin-
guished from normal controls with high specificity
(87.5%) and sensitivity (100%).
Downloaded By: [Lekeu, Françoise] At: 09:48 25 May 2009
Delayed recall of Rey’s complex figure
In our study, delayed recall of Rey’s complex figure
(along with the recency index) showed the best dis-
criminant value between converting and stable
QAD groups, providing a 75% correct classifica-
tion. The performance of stable QAD patients on
the visual episodic memory test was not signifi-
cantly different from that of controls, while their
performance on the CVLT was decreased for many
scores. Visual memory has been reported to be
involved in the preclinical phase of Alzheimer’s
disease (Kawas et al., 2003). A similar distinction
between performance on visual and verbal memory
tests was observed in a previous study comparing
converting versus stable QAD patients (Albert
et al., 2001; Blackwell et al., 2004; Fowler et al.,
1997; Masur et al., 1994) and another visual mem-
ory test, the CANTAB Paired Associate Learning
Test, was shown to differentiate between converter
and nonconverter QAD groups (Blackwell et al.,
2004; Fowler, Saling, Conway, Semple, & Louis,
2002). In another study, visual recognition mem-
ory performance was altered in MCI participants
compared to normal controls but MCI partici-
pants’ performance was better than that of AD
patients (Barbeau et al., 2004). Moreover, MCI
patients who failed on this task had lower scores
for free recall and received less benefit from cueing
on the free and cued Selective Reminding Test.
Although the structure of the tests is not the same,
those studies suggest that a visual episodic memory
impairment might help to distinguish prodromal
AD from stable “mild cognitive impairment.”
Finally, our results are in agreement with the
results of two prospective studies (Borroni et al.,
2006; Tierney et al., 1996). In the Tierney’s study,
120 memory-impaired patients without dementia
were followed for two years The results showed
that the neuropsychological measures that best
predicted AD conversion at the two-year follow-
up included the delayed recall of Rey’s figure.
Similarly, a two-year follow-up study of MCI
patients found that results for the delayed recall of
Rey’s figure were initially inferior in converter
than in nonconverter MCI participants (Borroni
et al., 2006).
Few CVLT variables did not differ between
groups. The number of intrusions was more varia-
ble in very early AD (SD = 12) than in stable QAD
and in controls (SD = 6), and this probably
explains the absence of difference. The response
bias for yes and no answers during recognition was
similar between groups, as was the recall of the
first four words and the trend to repeat the order of
the presentation. It is interesting to see that seman-
tic clustering did not differ between stable QAD
and converters, consistent with our discussion on
the absence of discriminative value of cued recall.
In conclusion, the value of the CVLT test is that
it provides several scores that depend on specific
strategies and different cognitive domains. Numer-
ous scores were found to differ between QAD
patients and normal controls, including those that
have classically been used in the literature, such as
the learning and delayed total recall measures.
However, the interest of our study was to highlight
the importance of a score depending on prior and
spontaneous semantic categorization (cued recall)
to objectively demonstrate memory deficits in
stable QAD patients. In addition, an increased
reliance on working memory (expressed by the
recency index) correctly distinguished converter
from nonconverter patients. A high recency index
should be considered as a risk factor for subsequent
development of dementia during the neuropsycho-
logical investigation of participants with mild
cognitive complaints.
Original manuscript received 11 January 2008
Revised manuscript accepted 12 March 2009
First published online day month year
REFERENCES
Adam, S., Van der Linden, M., Ivanoiu, A., Juillerat, A.-C.,
Bechet, S., & Salmon, E. (2007). Optimization of
encoding specificity for the diagnosis of early AD:
The RI-48 task. Journal of Clinical and Experimental
Neuropsychology, 29, 477–487.
Adlam, A. L., Bozeat, S., Arnold, R., Watson, P., &
Hodges, J. R. (2006). Semantic knowledge in mild
cognitive impairment and mild Alzheimer’s disease.
Cortex, 42, 675–684.

Albert, M. S., Moss, M. B., Tanzi, R., & Jones, K.
(2001). Preclinical prediction of AD using neuro-
psychological tests. Journal of the International Neu-
ropsychological Society, 7, 631–639.
Alladi, S., Arnold, R., Mitchell, J., Nestor, P. J., &
Hodges, J. R. (2006). Mild cognitive impairment:
Applicability of research criteria in a memory clinic
and characterization of cognitive profile. Psychologi-
cal Medicine, 36, 507–515.
Artero, S., Petersen, R., Touchon, J., & Ritchie, K.
(2006). Revised criteria for mild cognitive impair-
ment: Validation within a longitudinal population
study. Dementia and Geriatric Cognitive Disorders,
22, 465–470.
Artero, S., Tierney, M. C., Touchon, J., & Ritchie, K.
(2003). Prediction of transition from cognitive
impairment to senile dementia: A prospective, longi-
tudinal study. Acta Psychiatrica Scandinavica, 107,
390–393.
Bäckman, L., Jones, S., Berger, A. K., Laukka, E. J., &
Small, B. J. (2004). Multiple cognitive deficits during
the transition to Alzheimer’s disease. Journal of
Internal Medicine, 256, 195–204.
Downloaded By: [Lekeu, Françoise] At: 09:48 25 May 2009
Bäckman, L., Jones, S., Berger, A.-K., Laukka, E. J., &
Small, B. J. (2005). Cognitive impairment in preclini-
cal Alzheimer’s disease: A meta-analysis. Neuropsy-
chology, 19, 520–531.
Barbeau, E., Didic, M., Tramoni, E., Felician, O.,
Joubert, S., Sontheimer, A., et al. (2004). Evaluation
of visual recognition memory in MCI patients. Neu-
rology, 62, 1317–1322.
Benson, A. D., Slavin, M. J., Tran, T. T., Petrella, J. R., &
Doraiswamy, P. M. (2005). Screening for early
Alzheimer’s disease: Is there still a role for the Mini-
Mental State Examination? Journal of Clinical
Psychiatry, 7, 62–67.
Blackwell, A. D., Sahakian, B. J., Vesey, R., Semple, J. M.,
Robbins, T. W., & Hodges, J. R. (2004). Detecting
dementia: Novel neuropsychological markers of pre-
clinical Alzheimer’s disease. Dementia and Geriatric
Cognitive Disorders, 17, 42–48.
Borroni, B., Anchisi, D., Paghera, B., Vicini, B.,
Kerrouche, N., Garibotto, V., et al. (2006). Com-
bined 99mTc-ECD SPECT and neuropsychological
studies in MCI for the assessment of conversion to
AD. Neurobiology of Aging, 27, 24–31.
Chen, P., Ratcliff, G., Belle, S. H., Cauley, J. A., DeKosky,
S. T., & Ganguli, M. (2000). Cognitive tests that best
discriminate between presymptomatic AD and those
who remain nondemented. Neurology, 55, 1847–1853.
Delis, D. C., Freeland, J., Kramer, J. H., & Kaplan, E.
(1988). Integrating clinical assessment with cognitive
neuroscience: Construct validation of the California
Verbal Learning Test. Journal of Consulting and
Clinical Psychology, 56, 123–130.
Delis, D. C., Kramer, J. H., Kaplan, E., & Ober, B. A.
(1987). California Verbal Learning Test: Adult version
manual: Research edition. San Antonio, TX: The
Psychological Corporation.
Dudas, R. B., Clague, F., Thompson, S. A., Graham, K. S.,
& Hodges, J. R. (2005). Episodic and semantic mem-
ory in mild cognitive impairment. Neuropsychologia,
43, 1266–1276.
Duong, A., Whitehead, V., Hanratty, K., & Chertkow, H.
(2006). The nature of lexico-semantic processing defi-
cits in mild cognitive impairment. Neuropsychologia,
44, 1928–1935.
POWER OF CVLT IN DEMENTIA PREDICTION 11
Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975).
Mini-Mental State: A practical method for grading
the state of patients for the clinician. Journal of
Psychiatry Research, 12, 189–198.
Fowler, K. S., Saling, M. M., Conway, E. L., Semple,
J. M., & Louis, W. J. (1997). Computerised neu-
ropsychological tests in the early detection of demen-
tia: Prospective findings. Journal of the International
Neuropsychological Society, 3, 139–146.
Fowler, K. S., Saling, M. M., Conway, E. L., Semple,
J. M., & Louis, W. J. (2002). Paired associate
performance in the early detection of DAT. Journal
of the International Neuropsychological Society, 8,
58–71.
Greenaway, M. C., Lacritz, L. H., Binegar, D., Weiner,
M. F., Lipton, A., & Munro Cullum, C. (2006). Patterns
of verbal memory performance in mild cognitive impair-
ment, Alzheimer disease, and normal aging. Cognitive
and Behavioral Neurology, 19, 79–84.
Grut, M., Fratiglioni, L., Viitanen, M., & Winblad, B.
(1993). Accuracy of the Mini-Mental Status Exami-
nation as a screening test for dementia in a Swedish
elderly population. Acta Neurologica Belgica, 87,
312–317.
Hamilton, M. (1967). Development of a rating scale for
primary depressive illness. British Journal of Social
Clinical Psychology, 6, 278–296.
Hodges, J. R., Erzinclioglu, S., & Patterson, K. (2006).
Evolution of cognitive deficits and conversion to
dementia in patients with mild cognitive impairment:
A very-long-term follow-up study. Dementia and
Geriatric Cognitive Disorders, 21, 380–391.
Huang, C., Wahlund, L.-O., Almkvist, O., Elehu, D.,
Svensson, L., Jonsson,T., et al. (2003). Voxel- and
VOI-based analysis of SPECT CBF in relation to
clinical and psychological heterogeneity of mild cog-
nitive impairment. NeuroImage, 19, 1137–1144.
Ivanoiu, A., Adam, S., Van der Linden, M., Salmon, E.,
Juillerat, A.-C., Mulligan, R., et al. (2005). Memory
evaluation with a new cued recall test in patients with
mild cognitive impairment and Alzheimer’s disease.
Journal of Neurology, 252, 47–55.
Kawas, C. H., Corrada, M. M., Brookmeyer, R.,
Morrison, A., Resnick, S. M., Zonderman, A. B.,
et al. (2003). Visual memory predicts Alzheimer’s dis-
ease more than a decade before diagnosis. Neurology,
60, 1089–1093.
Kohler, S. (1994). Quantitative characterization of
verbal learning deficits in patients with Alzheimer’s
disease. Journal of Clinical and Experimental Neu-
ropsychology, 16, 749–753.
Kryscio, R. J., Schmitt, F. A., Salazar, J. C., Mendiondo,
M. S., & Markesbery, M. D. (2006). Risk factors for
transitions from normal to mild cognitive impairment
and dementia. Neurology, 66, 828–832.
Lawton, M. P., & Brody, E. M. (1969). Assessment of
older people: Self-maintaining and instrumental
activities of daily living. Gerontologist, 9, 179–186.
Lezak, M. D. (1983). Neuropsychological assessment
(2nd ed.). New York: Oxford University Press.
Masur, D. M., Sliwinski, M., Lipton, R. B., Blau, A. D., &
Crystal, H. A. (1994). Neuropsychological prediction
of dementia and the absence of dementia in healthy
elderly persons. Neurology, 44, 1427–1432.
Mattis, S. (1973). Dementia Rating Scale: Professional
manual. Odessa, FL: Psychological Assessment
Resources.
 12 LEKEU ET AL.
McKhann, G., Drachman, D., Folstein, M., Katzman, R.,
Price, D., & Stadlan, E. M. (1984). Clinical diagnosis of
Alzheimer’s disease: Report of NINCDS/ADRDA
Work Group under the auspices of Department of
Health and Human Services Task Force on Alzheimer’s
disease. Neurology, 34, 939–944.
Monsch, A. U., Foldi, N. S., Ermini-Funfschilling, D. E.,
Berres, M., Taylor, K. I., Stahelin, H. B., et al.
(1995). Improving the diagnostic accuracy of the
Mini-Mental State Examination. Acta Neurologica
Belgica, 92, 145–150.
Morris, J. C. (1993). The Clinical Dementia Rating
(CDR): Current version and scoring rules. Neurology,
43, 2412–2414.
Petersen, R. C., Doody, R., Kurz, A., Mohs, R. C.,
Morris, J. C., Rabins, P. V., et al. (2001). Current
concepts in mild cognitive impairment. Archives of
Neurology, 58, 1985–1992.
Petersen, R. C., Smith, G. E., Waring, S. C., Ivnik, R. J.,
Tangalos, E. G., & Kokmen, E. (1999). Mild cogni-
tive impairment: Clinical characterization and out-
come. Archives of Neurology, 56, 303–308.
Rey, A., Osterrieth, P. A., & Taylor, L. B. (1991). Rey
Downloaded By: [Lekeu, Françoise] At: 09:48 25 May 2009
Osterrieth complex figure test and Taylor version. In
O. Spreen & E. Strauss (Eds.), A compendium of
neuropsychological tests: Administration, norms,
and commentary (pp. 157–167). New York: Oxford
University Press.
Ribeiro, F., de Mendonça, A., & Guerreiro, M. (2006).
Mild cognitive impairment: Deficits in cognitive
domains other than memory. Dementia and Geriatric
Cognitive Disorders, 21, 284–290.
Ritchie, K., & Fuhrer, R. (1992). A comparative study of
the performance of screening tests for senile dementia
using receiver operating characteristics analysis.
Journal of Clinical Epidemiology, 45, 627–637.
Salmon, D. P., Thomas, R. G., Pay, M. M., Booth, A.,
Hofstetter, B. A., Thal, L. J., et al. (2002). Alzheimer’s
disease can be accurately diagnosed in very mildly
impaired individuals. Neurology, 59, 1022–1028.
Simon, E., Leach, L., Winocur, G., & Moscovitch, M.
(1994). Intact primary memory in mild to moderate
Alzheimer disease: Indices from the California Verbal
Learning Test. Journal of Clinical and Experimental
Neuropsychology, 16, 414–422.
Tabert, M. H., Manly, J. J., Liu, X., Pelton, G. H.,
Rosenblum, S., Jacobs, M., et al. (2006).
Neuropsychological prediction of conversion to
Alzheimer disease in patients with mild cognitive
impairment. Archives of General Psychiatry, 63,
916–924.
Thompson, S. A., Graham, K. S., Patterson, K., Sahakian,
B. J., & Hodges, J. R. (2002). Is knowledge of famous
people disproportionately impaired in patients with
early and questionable Alzheimer’s disease? Neu-
ropsychology, 16, 344–358.
Tierney, M. C., Szalai, J.-P., Dunn, E., Geslani, D., &
McDowell, I. (2000). Prediction of probable Alzheimer
disease in patients with symptoms suggestive of memory
impairment. Value of the Mini-Mental State Examina-
tion. Archives of Familial Medicine, 9, 527–532.
Tierney, M. C., Szalai, J. P., Snow, W. G., Fisher, R. H.,
Nores, A., Nadon, G., et al. (1996). Prediction of
probable Alzheimer’s disease in memory-impaired
patients: Prospective longitudinal study. Neurology,
46, 661–665.
Winblad, B., Palmer, K., Kivipelto, M., Jelic, V.,
Fratiglioni, L., Wahlund, L.-O., et al. (2004). Mild
cognitive impairment—beyond controversies, towards
a consensus: Report of the International Working
Group on Mild Cognitive Impairment. Journal of
Internal Medicine, 256, 240–246.