Neuropsychologia 48 (2010) 978–988
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/neuropsychologia
The cognitive and neural expression of semantic memory impairment in mild
cognitive impairment and early Alzheimer’s disease
Sven Jouberta,b,∗, Simona M. Brambatib, Jennyfer Ansadob, Emmanuel J. Barbeauc, Olivier Feliciand,e,
Mira Didicd, Jacinthe Lacombea,b, Rachel Goldsteina, Céline Chayerf,g,h, Marie-Jeanne Kergoatb,g,h
aDépartement de psychologie et CERNEC, Université de Montréal, Montreal, Québec, Canada
bCentre de recherche de l’Institut universitaire de gériatrie de Montréal, Montreal, Québec, Canada
cCentre de recherche Cerveau et Cognition, université Paul Sabatier – CNRS, Toulouse, France
dService de Neurologie et de Neuropsychologie, AP-HM Timone, Marseille, France
eLaboratoire Epilepsie et Cognition, INSERM U 751, Université de la Méditerranée, Marseille, France
fClinique de mémoire de l’Hôpital Maisonneuve-Rosemont, Montreal, Québec, Canada
gClinique de cognition de l’Institut universitaire de gériatrie de Montréal, Montreal, Québec, Canada
hDépartement de Médecine, Université de Montréal, Montreal, Québec, Canada
a r t i c l ei n f o
Received 2 February 2009
Received in revised form
10 November 2009
Accepted 24 November 2009
Available online 30 November 2009
Mild cognitive impairment
a b s t r a c t
Semantic deficits in Alzheimer’s disease have been widely documented, but little is known about the
integrity of semantic memory in the prodromal stage of the illness. The aims of the present study
were to: (i) investigate naming abilities and semantic memory in amnestic mild cognitive impairment
(aMCI), early Alzheimer’s disease (AD) compared to healthy older subjects; (ii) investigate the associa-
tion between naming and semantic knowledge in aMCI and AD; (iii) examine if the semantic impairment
was present in different modalities; and (iv) study the relationship between semantic performance and
grey matter volume using voxel-based morphometry. Results indicate that both naming and semantic
knowledge of objects and famous people were impaired in aMCI and early AD groups, when compared
to healthy age- and education-matched controls. Item-by-item analyses showed that anomia in aMCI
and early AD was significantly associated with underlying semantic knowledge of famous people but not
with semantic knowledge of objects. Moreover, semantic knowledge of the same concepts was impaired
in both the visual and the verbal modalities. Finally, voxel-based morphometry analyses revealed that
semantic impairment in aMCI and AD was associated with cortical atrophy in the anterior temporal lobe
(ATL) region as well as in the inferior prefrontal cortex (IPC), some of the key regions of the semantic
down of semantic knowledge of famous people and objects, combined with difficulties in the selection,
manipulation and retrieval of this knowledge.
© 2009 Elsevier Ltd. All rights reserved.
over a lifetime and shared by a same cultural group. It includes
for instance knowledge about objects, people, places, concepts,
facts and language. Semantic deficits have been documented in
Alzheimer’s disease (AD) using a variety of standard clinical neu-
ropsychological tests, such as confrontation naming (Huff, Corkin,
Abbreviations: aMCI, amnestic mild cognitive impairment; MMSE, Mini-Mental
State Examination; AD, Alzheimer’s disease; VBM, voxel-based morphometry.
∗Corresponding author at: Centre de recherche de l’IUGM, 4565, chemin Queen
Mary, Montreal, Québec, H3W 1W5, Canada. Tel.: +1 514 340 3540x3551;
fax: +1 514 340 3548.
E-mail address: firstname.lastname@example.org (S. Joubert).
& Growdon, 1986), visual–verbal semantic matching (Hodges &
Patterson, 1995) and category fluency (Adlam, Bozeat, Arnold,
Watson, & Hodges, 2006; Rosser & Hodges, 1994). Studies which
have investigated semantic breakdown in AD using more sensitive
experimental measures have also shown that subordinate knowl-
edge is more impaired than superordinate knowledge (Chertkow &
Bub, 1990; Hodges, Patterson, Oxbury, & Funnell, 1992), that bio-
et al., 2001; Gonnerman, Andersen, Devlin, Kempler, & Seidenberg,
ple is disproportionately impaired relative to other categories of
conceptual knowledge and to autobiographical memory (Greene &
Hodges, 1996; Thompson, Graham, Patterson, Sahakian, & Hodges,
Clinical research has shown that AD is often preceded by
a clinical phase commonly referred to as amnestic mild cogni-
0028-3932/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.
S. Joubert et al. / Neuropsychologia 48 (2010) 978–988
tive impairment (aMCI) (Gauthier et al., 2006; Petersen, 2003).
Although, as in AD, episodic memory impairment is considered to
be the main clinical feature of aMCI, recent studies have demon-
strated the presence of semantic memory impairment as well
(Adlam, Bozeat, Arnold, Watson, & Hodges, 2006; Ahmed, Arnold,
Thompson, Graham, & Hodges, 2008; Duong, Whitehead, Hanratty,
& Chertkow, 2006; Estevez-Gonzalez et al., 2004; Joubert et al.,
2008; Seidenberg et al., 2009; Vogel, Gade, Stokholm, & Waldemar,
2005). For instance, several studies have shown that aMCI patients
suffered from early breakdown of semantic knowledge of famous
people. Predementia patients who later on converted to AD were
initially more impaired at naming famous faces relative to the non-
converters (Estevez-Gonzalez et al., 2004; Thompson et al., 2002;
Vogel et al., 2005). Recently, Ahmed et al. (2008) demonstrated
that naming photographs of famous people and famous buildings
was significantly more impaired than naming of line drawings of
common objects in aMCI individuals. The authors suggested that
semantic knowledge about unique entities such as famous peo-
ple and buildings may be more vulnerable to damage than more
general knowledge about common objects. We also found such
evidence in aMCI patients and demonstrated that the fine-grained
semantic processing of entities which have unique idiosyncratic
attributes may be affected earlier and to a greater extent than
the semantic processing of generic concepts (Joubert et al., 2008).
Similarly, using experimental measures of object naming, a study
showed in preclinical AD patients that impaired naming of famous
faces was more predictive of future conversion to AD than object
naming (Thompson et al., 2002).
Overall, these studies have demonstrated that naming pictures
of unique entities including famous people, famous buildings and
famous public events is significantly more affected in aMCI indi-
viduals than picture naming of common objects. Although anomia
for famous people has been documented in recent years, lit-
tle is still known about the integrity of semantic knowledge in
aMCI. Most studies in aMCI patients have been descriptive, i.e.,
they have reported that aMCI patients were impaired on nam-
ing tasks, without assessing what mechanisms were associated
with this impairment. The first objective of this study is to eval-
uate the naming deficits and the semantic deficits in aMCI patients
compared to early AD and normal controls. The possibility that
aMCI patients present with genuine semantic memory impair-
ment remains unclear at this point in time, although there is now
some indication that this may be so several years before conver-
sion to dementia (Amieva et al., 2008). The second objective is to
investigate if there is an association between anomia and impaired
semantic knowledge in aMCI. The third objective is to examine if
the same concepts are impaired in different modalities in aMCI,
i.e., visual and verbal, which may provide evidence of a central
breakdown of semantic knowledge. Finally, we also carried out
an imaging study to verify if the semantic deficits observed in
these patients correlated with atrophy in brain regions known to
be important for semantic memory.
written informed consent before participation. The research protocol was approved
by the Research Ethics committee of Institut universitaire de gériatrie de Montréal
(IUGM). Considering the cultural sensitivity of semantic tests, all participants were
required to have lived at least the last 40 previous years of their lives in Quebec and
their mother tongue was French.
to the Cognition clinic of the IUGM in Montreal. MCI patients were seen at the
included normal activities of daily living as assessed during a clinical interview and
defined by a score of 0 on the IADL (independent activities of daily living) (Lawton
& Brody, 1969), and an MMSE (Mini-Mental State Examination) score ≥25 (Folstein,
Folstein, & McHugh, 1975). Criteria included a memory complaint corroborated by
an informant and confirmed with formal neuropsychological measures of episodic
memory. An objective memory impairment was defined by impaired performance
using a cut-off score of 1.5 standard deviations (S.D.) below the mean of matched
normal elderly subjects on standard measures of episodic memory. Here, we used
the test de rappel libre/rappel indicé à 16 items (RL/RI 16) (Van der Linden et al.,
2004), a test of episodic memory in French similar to the free and cued selective
reminding test (FCSRT) (Grober, Buschke, Crystal, Bang, & Dresner, 1988). Finally,
rological and Communicative Disorders and Stroke and the Alzheimer’s Disease and
Related Disorders Association (NINCDS-ADRDA) (McKhann et al., 1984). All subjects
had a complete neurological examination and did not show any evidence of other
neurological disease or other potential causes of dementia that could account for
their condition. AD subjects were in an “early” stage of the disease and had a mean
MMSE score of 22.1 (range: 17–25). Healthy elderly controls were 16 adults (10
females and 6 males) showing normal cognitive functioning, such as demonstrated
by their performance on a general neuropsychological examination. Normal con-
trols (NC) were matched to patient groups on the basis of age and educational level.
psychiatric illness, traumatic brain injury, history of alcoholism, untreated medical
or metabolic condition, general anesthesia in the last 12 months, or uncorrected
hearing and vision problems.
Demographic data of the three groups of patients indicate that they did not
differ significantly in terms of age (F(2,44)=0.286, p=0.752; aMCI: mean (M)=73.7,
standard deviation (S.D.)=6.3; AD: M=74.2, S.D.=6.5; NC: M=72.4, S.D.=7.1) nor
in terms of their MMSE score (F(2,44)=38.2, p<0.001; aMCI: M=27.4, S.D.=1.6; AD:
M=22.1, S.D.=2.9; NC: M=28.1, S.D.=1.5). Post-hoc analyses reveal a significant
but not between aMCI and controls (p=0.607).
All three groups of participants underwent a comprehensive general neuropsy-
chological evaluation, which included standard neuropsychological measures of
memory, language, executive functions, visuospatial and visuoperceptual abili-
ties. Episodic memory was assessed with the RL/RI 16 (Van der Linden et al.,
2004), a free/cued word recall test widely used as a measure of verbal learning
in French speaking populations. Visual memory was assessed using the immediate
and delayed recall (20min) conditions of the Rey complex figure (Rey, 1960). Lan-
guage was assessed using the 15-item version of the Boston Naming Test (Calero,
Arnedo, Navarro, Ruiz-Pedrosa, & Carnero, 2002), as well as with the Letter (P) and
Category (animals) fluency tests (Cardebat, Doyon, Puel, Goulet, & Joanette, 1990).
the Trail Making Test (Reitan, 1955). Short-term and working memory were evalu-
ated using the forward and backward span subtests of the Weschler Memory Scale
Rey–Osterrieth figure (Rey, 1960). Finally, the incomplete letters task of the Visual
object and space perception battery (VOSP) (Warrington & James, 1991) was used
to assess primary visuoperceptual processing and the Benton line orientation test
was used to evaluate visuospatial abilities (Benton, Varney, & Hamsher, 1978).
2.3. Semantic memory test
This computerized task was devised by the principal investigator (S.J.) and was
composed of two parts. The first part of the test evaluates the integrity of naming
abilities and semantic memory of common generic concepts (objects), while the
second part investigates the integrity of naming abilities and semantic memory of
(i) Thefirstevaluationconsistedof two parts. Inthefirst part, subjects hadtoname
and identify 40 photographs of objects, half of which were man-made objects
individually on a computer screen and the subject was instructed to name the
object. Once an answer had been provided, a series of four questions assess-
ing semantic knowledge about the concept were presented one-by-one below
the image. Each question assessed different concept attributes. For instance, the
first question assessed categorical knowledge. Once an answer had been pro-
vided, a second question evaluating general perceptual knowledge of the object
was presented. A third and fourth question evaluated functional/encyclopedic
knowledge of the concept. Questions were formulated so that the subject had
to provide yes/no responses. This semantic recognition task was intentionally
chosen to reduce the load of controlled strategic retrieval processes required
S. Joubert et al. / Neuropsychologia 48 (2010) 978–988
to access and manipulate semantic knowledge in memory. There were as many
their own personal experiences, therefore relying on semantic-based informa-
tion rather than on episodic-based information. A correct response was given if
the subject was able to provide the specific name of the photograph that was
sented randomly within each session using the E-Prime software (Psychology
Software Tools, Pittsburg, PA).
(ii) The second part of the semantic memory test involved the presentation of 30
photographs of famous people, including local and international actors, politi-
cians, athletes, singers, etc. One third of the celebrities were famous in the last
15 years (recent), another third were famous between 1970 and 1990, and one
third was famous between 1930 and 1970. As in the first part, photographs of
faces of famous people were presented one at a time on a computer screen. The
a series of questions assessing semantic knowledge about the person were pre-
ual, and the last question concerned the period during which the celebrity was
ticipant had to provide yes/no responses. There were as many positive and false
statements and stimuli were presented randomly within each session. Again,
subjects were explicitly asked to answer the question based on the most widely
culturally accepted answer and not on the basis of their own personal expe-
riences. Practise stimuli were presented before the beginning of each session.
Again, a response was considered correct if the subject was able to provide the
full name (first and last name) of the famous face that was presented. If a partial
response was provided (for example, initials such as JFK or only the last name),
the subjects were encouraged to provide the complete name.
During a subsequent evaluation that took place 2 weeks later, the exact same
stimuli (objects and people) were presented from their names on the computer
screen and spelled aloud at the same time by the experimenter. The purpose of
this second evaluation was to examine if subjects accessed semantic knowledge
about different classes of concepts differently in visual (i.e., photographs) vs. verbal
(names) modalities. During this session, names were once again presented one-by-
one, and subjects had to provide yes/no answers to the same questions described
above (no naming was required during this session). This type of experimental
design allows verifying the consistency of performance over different assessments.
A consistent pattern of impairment across modalities is assumed to reflect central
the proportion of correct responses for objects and famous people between the two
administrations of each portion of the semantic memory test. The order of presen-
tation was counterbalanced so that half of the participants underwent the objects
test first and the other half underwent the test of famous people first. Again, stimuli
were randomized within each part of the test. Photographs were always presented
during the 1st evaluation while names were presented during the 2nd evaluation,
so that verbal presentation of the names would not cue subsequent presentation of
as well as examples of the questions that were asked for objects and famous people.
2.4. Statistical analyses
2.4.1. Neuropsychological tests
One-way analyses of variance were carried out on the neuropsychological tests
for the three groups of subjects. A Levene test was used to assess homogeneity of
variances. When homogeneity of variance was met, significant main effects were
analyzed using Tukey’s honestly significant difference (HSD). If the homogeneity of
main effects were analyzed using Mann–Whitney U-tests.
The naming data were analyzed using separate two-way repeated measures
analyses of variance on the percentage of correct responses. First, analyses of vari-
as the between-factor variable (objects, famous people) and stimulus category was
the within factor (NC, aMCI, AD), which allows to consider the variance between
ance with stimulus category as the within-factor, and group as the between-factor
were carried out to confirm results from the item analyses of variance. Decomposi-
tion of significant interactions was carried out using the least significant differences
(LSD) method. We used an alpha level of 0.05 for all statistical tests.
2.4.3. Task difficulty (naming)
A number of studies have addressed the question of task difficulty in the con-
text of assessing semantic knowledge and naming of people and objects (Kay &
Hanley, 2002; Lyons, Kay, Hanley, & Haslam, 2006; Miceli et al., 2000). Lyons et
al. (2006) emphasized the importance of controlling for task difficulty when carry-
ing out studies on category-specific differences: a test may be less sensitive than
another because of possible differences in difficulty between the items, which may
be due to high variability on one test and low variability on another. They suggested
that one way to control item difficulty is to provide two tests of the same format in
which means and standard deviations of control performance is carefully matched
(Lyons et al., 2006). If we still observe differences between tests, then this would
reinforce the finding of differences between two tests. We thus carried out a post-
hoc analysis on the naming data. We selected 10 objects and 10 famous people that
were matched for naming difficulty in our control subjects (80.6% correct naming).
We then compared the naming performance for the same 10 objects and 10 famous
faces in the MCI and in the AD groups. A two-way repeated measures ANOVA of
naming performance was carried out. Post-hoc contrasts using LSD method were
performed when necessary.
2.4.4. Semantic questions
The analyses of performance on tests of semantic knowledge of objects and
famous people were identical to those of naming performance. Overall percentage
of correct answers on semantic questions was determined using a composite score
which was calculated by averaging performance across all semantic probes for each
2.4.5. Consistency between naming and semantic knowledge
Since the same items were used for object naming and semantic knowledge,
one important question we wished to address was to examine if there was a consis-
tency at the item specific level between naming performance and performance on
semantic probes, in order to examine whether anomia in the aMCI and AD groups
is associated with underlying semantic impairment. In order to do this, we exam-
ined naming and semantic performance between each subject at the item-by-item
level. We examined the performance for each item across each subject, and deter-
mined the overall number of occurrences at the single item level when: naming and
semantics was correct; naming and semantics was incorrect; naming was correct
chi-square analyses on the data to determine if a significant association between
naming and semantics was found.
2.4.6. Modality effects
In order to determine if the groups differed in terms of their ability to answer
semantic probes in different modalities of presentation (visual and verbal), sepa-
rate two-way repeated measures ANOVA with one within-factor, group (NC, aMCI,
AD) and one between-factor, modality (visual and verbal) were carried out on the
percentage of semantic questions answered correctly. An ANOVA was carried out
one for the category of famous people, and another for the category of objects. Post-
hoc contrasts were carried out using the LSD method. Finally, analyses of variance
with stimulus category as the within-factor, and group as the between-factor were
carried out to confirm these results.
2.4.7. Consistency between the visual and verbal modality
We examined if there was a consistency at the item-by-item level between
semantic performance in the visual modality and in the verbal modality. In order to
perform this, we examined semantic performance, in both modalities, at the item
specific level and for each subject. We examined the performance for each single
item across each subject, and determined the overall number of instances at the
single item level when: visual and verbal semantics was correct; visual and verbal
semantics was incorrect; visual semantics was correct but verbal semantics incor-
rect; and visual semantics was incorrect but verbal semantics correct. We then ran
chi-square analyses on this data to determine if there was a significant association
between semantic performance in the visual modality and in the verbal modality.
2.4.8. Post-hoc voxel-based morphometry analysis
The behavioral results revealed semantic deficits in both aMCI and AD patients.
In order to map the regions responsible for this semantic deficit in our patient sam-
ple, we correlated the gray matter (GM) volume at a voxel level with the severity of
the semantic deficit in our patient population using voxel-based morphometry. The
semantic performance was represented by the mean Z score obtained in the four
tests assessing semantics, i.e., semantic test of objects and people in both picture
and verbal modality.
It has been recently proposed that semantic cognition is supported by a three-
left temporo-parietal regions (Jefferies & Lambon Ralph, 2006). The anterior tem-
poral lobe (ATL) region would represent an amodal semantic representation store,
tic control’, i.e., the executive-control mechanisms that interact with the semantic
representation to produce an appropriate activation of key knowledge for a specific
task. Based on this model, we expect to observe a significant correlation between
GM integrity in ATL and semantic test performance if the semantic impairment is
parietal regions if the semantic impairment is determined by a deficit of semantic
S. Joubert et al. / Neuropsychologia 48 (2010) 978–988
Neuropsychological data of controls, aMCI and AD patients.
p value for group effect
Immediate free recall of a word list (16)
Immediate total recall of a word list (16)
Delayed free recall of a word list (16)
Delayed total recall of a word list (16)
Rey–Osterrieth immediate recall (36)
Rey–Osterrieth delayed recall (36)
Executive function/working memory
Part C (interference)
Trail Making Test
Digit span forward
Digit span backward
Naming (Boston naming) (15)
Verbal fluency “P” in 2 min
Category fluency “animals” in 2 min
Visual object and space perception battery
Incomplete letters (20)
Rey–Osterrieth figure – copy (36)
Benton line orientation (30)
Results are presented in Mean and S.D. in brackets.
ap<0.05 between control group and patient group.
bp<0.01 between control group and patient group.
cp<0.05 between MCI and AD groups.
dp<0.01 between MCI and AD groups.
126.96.36.199. Subjects. All aMCI and AD subjects that participated in the behavioral study
who were able to undergo a high-quality MRI scan were included in the imaging
study. The scan was acquired within six months from neuropsychological testing.
The group was composed of 25 patients. In the voxel-based morphometry correla-
tion analysis all subjects were entered in the statistical analysis as a single group as
previously described (Brambati et al., 2006; Rosen et al., 2005), and they were not
sub-grouped on the basis of their diagnosis. However, for sample characterization
purposes, the subject group included fourteen aMCI and eleven AD.
188.8.131.52. Image acquisition. MRI scans were obtained on a 3T Siemens Trio MRI
(Siemens, Erlangen, Germany). High resolution anatomical images were acquired
using an optimized MPRAGE protocol (TR=2.3s, TE=2.94ms, TI=900ms, flip
angle=9◦, FOV=256×240, voxel 1mm×1mm×1.2mm) using an 8-channel coil.
Images were acquired in the horizontal plane.
184.108.40.206. Voxel-based morphometry analysis. VBM analysis included image spatial
pre-processing and statistical analysis. Both steps were implemented in the
SPM2 software package (Wellcome Department of Imaging Neuroscience, London;
http://www.fil.ion.ucl.ac.uk/spm) running on Matlab 6.5.1 (MathWorks, Natick,
220.127.116.11. Imagepre-processing. AnatomicalMRIimageswerespatiallypre-processed
using standard procedures (Good et al., 2001). All T1 structural images were seg-
mented, bias corrected and spatially normalized to Montreal Neurological Institute
(MNI) space using the unified segmentation procedure (Ashburner & Friston, 2005).
The VBM analysis was based on modulated gray matter images, whereby the gray
matter value in each voxel is multiplied by the Jacobian determinant derived from
the spatial normalization in order to preserve the total amount of gray matter from
the original images. These modulated gray matter images were smoothed with a
Gaussian kernel (12mm FWHM).
18.104.22.168. Statistical analysis. The mean Z score in semantic tests was entered as
covariate of interest in the ‘covariate model’. Person and object semantic test scores
all subjects, irrespective of their diagnosis, were entered as a single group in the sta-
tistical model. Age and gender were entered as nuisance covariates. Global nuisance
effect was accounted by scaling all images to the same global volume. The corre-
lation was tested using a  t-contrast, assuming that decreased semantic abilities
would be associated with decreased gray matter volumes.
The significance of each effect of interest was determined using the theory of
of p<0.05 corrected for multiple comparisons (SPM family-wise error – FWE) was
accepted. A less conservative threshold of p<0.005 uncorrected was adopted in our
regions of interest (ROIs) including bilateral temporal lobes, the prefrontal cortex
and the temporo-parietal regions (Jefferies & Lambon Ralph, 2006).
3.1. Neuropsychological tests
of subjects and significant main effects are presented in Table 1.
Results indicate that aMCI patients differ significantly from con-
the isolated nature of the memory impairment in this group. Inter-
estingly, the only other neuropsychological test for which aMCI
patients were significantly more impaired than controls was the
cator of semantic decline in aMCI (Adlam et al., 2006; Hodges,
Erzinclioglu, & Patterson, 2006; Murphy, Rich, & Troyer, 2006).
Results of the naming performance are presented in Fig. 1.
A two-way repeated measures ANOVA of naming performance
S. Joubert et al. / Neuropsychologia 48 (2010) 978–988
Fig. 1. Performance of control, aMCI and AD groups at naming and answering
semantic questions from photographs.
performed on percentage correct responses revealed a significant
stimulus category×group interaction (F(2,136)=30.92, p<0.001).
Decomposition of this interaction showed a significant group
effect for the objects test (F(2,136)=8.13, p<0.001) and for the
famous people’s test (F(2,136)=90.29, p<0.001). Post-hoc con-
trasts revealed that the aMCI and the AD groups were significantly
more impaired than the control group at naming objects (p=0.023
and p<0.001, respectively) and famous faces (p<0.001), but the
aMCI and AD groups did not differ between each other. In addi-
tion, decomposition of the stimulus category×group interaction
revealed a significant stimulus category effect. Post-hoc compar-
isons indicate that naming performance was significantly better
for objects than for famous people in all three groups (p<0.001).
Finally, an ANOVA with stimulus category as the within-factor and
group as the between-factor confirmed results of the item analy-
sis. Taken together, results demonstrate that naming objects and
famous people is impaired in both aMCI and AD groups when com-
pared to controls, corroborating similar findings in the literature
(Adlam et al., 2006; Joubert et al., 2008; Thompson et al., 2002).
3.2.1. Task difficulty
for naming difficulty in our control subjects (80.6% correct nam-
famous people in the MCI group, and 70.6% for objects and 46.3%
for famous people in the AD group. The aMCI group and the AD
group were significantly impaired at naming famous faces when
compared to the control group (p<0.001), but the aMCI and AD
Number of observed values at the item specific level for naming and semantic
performance in the aMCI and AD groups. Chi-square tests indicate the significant
CorrectIncorrectTotal Correct IncorrectTotal
a?2(1)=73.1, p<0.001; ?2(1)=1.29, p=0.26.
b?2(1)=21.27, p=0.001; ?2(1)=1.49, p=0.22.
groups did not differ between each other. The aMCI and AD groups
were not significantly impaired at naming objects relative to con-
trols. Moreover, a significant stimulus category effect was found:
famous face naming was significantly more impaired than naming
objects in the aMCI group (F(1,18)=25.06, p<0.001) and in the AD
group (F(1,18)=6.52, p<0.05). Subject analyses overall confirmed
these results. In conclusion, even when object and famous face
naming performance is strictly controlled for difficulty in healthy
subjects, naming famous faces remains significantly more affected
than naming objects in aMCI and early AD, suggesting that differ-
ences between these categories of knowledge are not due to task
difficulty. With respect to semantic performance in control sub-
jects, performance was identical for objects and people (objects:
86.4% and people: 86.5%, see Fig. 1), indicating that semantic ques-
tions were equivalent in terms of difficulty.
3.3. Semantic questions
Analysis of semantic performance expressed in percentage cor-
rect response revealed a significant stimulus category×group
interaction (F(2,136)=6.56, p=0.002). Decomposition of this inter-
action showed a significant group effect for the objects test
(F(2,136)=18.65, p<0.001) and for the famous persons test
(F(2,136)=51.40, p<0.001). Post-hoc contrasts (LSD) showed that
for object semantics, the control group performed better than the
aMCI group (p<0.001) and the AD group (p<0.001), but that the
semantics, the aMCI and the AD groups both differed significantly
from the control group (p<0.001) but again the aMCI did not differ
significantly from the AD group. An ANOVA with stimulus category
as the within-factor and group as the between-factor replicated
these results. Moreover, a significant stimulus category effect was
found. Post-hoc comparisons showed that semantic knowledge of
famous people was significantly more impaired than knowledge of
objects in both the aMCI and AD groups (p=0.028 and p=0.044,
respectively), but not in the control group. In sum, these analyses
show that semantic knowledge for objects and famous people is
impaired in both aMCI and AD groups when compared to controls,
but that famous people knowledge is significantly more affected
than object knowledge (see Fig. 1).
3.4. Consistency between naming and semantic knowledge
Results are presented in detail in Table 2A. A very significant
association between naming and semantic knowledge for famous
people was found in the aMCI group (?2(1)=73.1, p<0.001) and
in the AD group (?2(1)=21.27, p<0.001). However, no significant
association between naming and semantic knowledge for objects
S. Joubert et al. / Neuropsychologia 48 (2010) 978–988
ities of the semantic test.
was found in the aMCI group (?2(1)=1.29, p=0.26) and in the AD
group (?2(1)=1.49, p=0.22).
3.5. Modality effects
Results are illustrated in Fig. 2. An ANOVA for famous peo-
ple knowledge showed a significant modality×group interaction
a significant group effect both for the visual (F(2,58)=40.06,
p<0.001) and the verbal modalities (F(2,58)=23.53, p<0.001).
Post-hoc analyses demonstrated that the aMCI and the AD groups
were more impaired than the control group (p<0.001) in both
the visual and in the verbal modality. The aMCI and AD groups
did not differ between each other in either modality. Decompo-
sition of the group by modality interaction also showed a modality
effect for the aMCI group (F(1,59)=15.80, p<0.001) and for the
AD group (F(1,59)=8.97, p=0.006) but not for the control group
(F(1,59)=0.93, p=0.342), indicating that semantic knowledge of
famous people was more impaired in the visual than in the ver-
bal modality in aMCI and AD patients. Results of the ANOVA for
objects demonstrated no significant modality by group interac-
tion (F(2,78)=1.67, p=0.19). There was an absence of modality
effect (F(2,78)=0.37, p=0.55), although a group effect was found
(F(2,78)=32.19, p<0.001). Post-hoc contrasts showed that the
aMCI and the AD groups were more impaired than the control
group in terms of object knowledge (p<0.001). Analyses of vari-
ance using individual subjects as cases analyses replicated these
results. In sum, semantic knowledge of objects and famous people
Number of observed values at the item specific level for semantic performance in
the visual and in the verbal modalities in the aMCI and AD groups.
Correct IncorrectTotal CorrectIncorrect Total
a?2(1)=95.21, p<0.001; ?2(1)=109.86, p<0.001.
b?2(1)=90.02, p<0.001; ?2(1)=125.14, p<0.001.
is impaired in both the visual and verbal modality in aMCI and AD.
Knowledge of famous people is more impaired in the visual than in
the verbal modality in these patients, while knowledge of objects
is similarly affected in both modalities.
3.6. Consistency between visual and verbal modalities
Results are presented in detail in Table 2B. With respect to
the aMCI group, results indicate that there was a very significant
association between semantic knowledge in the visual and in the
verbal modalities for famous people (?2(1)=95,21, p<0.001) and
for objects (?2(1)=109.86, p<0.001). A very significant associa-
tion between semantic knowledge in the visual and in the verbal
modalities for famous people was also found in the AD group
(?2(1)=92.02, p<0.001) and for objects (?2(1)=125.14, p<0.001).
3.7. Neuroimaging results
No voxels showed significant correlation between the mean Z
score and the GM volume when a threshold of p<0.05 corrected
for multiple comparisons (SPM family-wise error – FWE) was used.
However, when the threshold was lowered to the more permissive
threshold of p<0.005 uncorrected in our ROIs based on previous
ysis revealed a significant positive correlation between semantic
test performance and GM volume in the left anterior temporal
lobe (x=−49, y=14, z=−26, Z score=2.8, p<0.005 uncorrected,
cluster size=91 voxels) and in the left inferior prefrontal gyrus,
pars opercularis (BA 44) (x=−45, y=12, z=9, Z score=2.9, p<0.002
revealed in the temporo-parietal regions even when the threshold
was lowered to p<0.05 uncorrected (see Fig. 3). The removal of the
AD patients from the analysis did not eliminate the positive corre-
lation between semantic test performance and GM volume in the
left anterior temporal lobe and in the left inferior prefrontal gyrus,
pars opercularis (BA 44); in fact the positive correlation persisted
in the MCI group, albeit with a predictably slightly lower level of
significance (left ATL: Z score=2.7, p<0.005, left IFG: Z score=2.5,
Our first objective was to evaluate aMCI subjects’ confrontation
naming abilities. Naming objects and naming famous faces were
both significantly impaired in aMCI and AD patients. Consistent
with previous findings (Ahmed et al., 2008; Joubert et al., 2008),
naming famous faces was disproportionately impaired relative to
S. Joubert et al. / Neuropsychologia 48 (2010) 978–988
Fig. 3. (A) Brain areas in which the GM volume positively correlate with performance on the semantic tests. The threshold for display is p<0.005, uncorrected. Maps of
significant correlation are superimposed on axial and coronal sections of the canonical template of SPM2. (B) Relationships between gray matter volumes in arbitrary units
(y-axis) and semantic deficit severity at the peak voxel.
naming objects in both groups (see Fig. 1). These findings suggest
that confrontation naming tests in which pictures of objects are
required to be named may be less sensitive diagnostic measures
of anomia in aMCI than tests that require naming famous faces.
Moreover, it is particularly important to investigate face naming
abilities since they are a sensitive predictor of future conversion to
AD (Estevez-Gonzalez et al., 2004; Thompson et al., 2002; Vogel et
more difficult nature of the former task. In fact, the difference was
still present even when task difficulty was controlled for, following
the method proposed by Lyons and colleagues (Lyons et al., 2006).
Altogether, these results highlight the importance of including a
famous persons naming task in the screening neuropsychological
battery administered to aMCI patients.
4.2. Semantic memory
In addition to the naming impairment, aMCI and early AD
patients showed impaired semantic knowledge of objects and
famous people. With respect to previous studies which have
focused on naming abilities, the present finding represents a sig-
nificant step forward in understanding the nature of the memory
impairment in very early AD. In addition, previous studies have
not show whether naming difficulties in aMCI are associated with
a breakdown of underlying semantic knowledge. In one seminal
study by Chertkow and Bub (1990), the authors demonstrated that
there was a strict correspondence in AD between naming impair-
ment and impaired knowledge of the items that had to be named,
suggesting that word finding difficulties were at least in part due
to underlying semantic disruption. Here we tested whether the
naming deficits observed in aMCI were associated with deficits
in the semantic knowledge of the concept that had to be named.
Our hypothesis that naming deficits in aMCI were associated with
deficits in semantic memory was partly confirmed. We found that
naming deficits were associated with underlying semantic deficits
in aMCI and AD patients only for the category of famous people.
In contrast, the association between naming and semantic knowl-
edge of objects was not significant. A closer examination of naming
and semantic knowledge of objects presented in Fig. 2A indicates
that there were more occurrences of correct–incorrect naming-
semantic associations, relative to the number of occurrences of
incorrect–incorrect and correct–correct naming-semantic associa-
tions. This imbalance in the proportion of naming-semantic errors
led to the absence of significant association between naming and
knowledge of objects. One possible interpretation of these findings
may not be sufficient to compromise access to the lexical represen-
more generic, as opposed to the lexical representations of names
which are unique and denote single individuals, lexical represen-
tations of objects may be more robust and may require a higher
threshold of semantic impairment before they become affected.
Results concerning semantic knowledge of objects vs. famous
people also offer interesting insights into the nature of the seman-
tic impairment in aMCI and early AD. Contrary to a previous study
in which we used free recall tasks to tap semantic knowledge
about objects and people in aMCI patients (Joubert et al., 2008),
the current study intentionally used semantic questions that pro-
vide maximal contextual information in order to reduce the load
of effortful retrieval processes. It was found using this type of
material that although object and famous person knowledge were
both significantly impaired in aMCI and AD relative to controls,
famous person knowledge was disproportionately affected rela-
tive to object knowledge in both aMCI and AD. In relation to our
previous study, these results indicate that famous person knowl-
edge is more impaired than object knowledge not only in naming
and free recall tests, but also in low effortful semantic recogni-
S. Joubert et al. / Neuropsychologia 48 (2010) 978–988
tion tasks. Knowledge of famous people, which relies to a greater
affected than knowledge of objects in aMCI and AD, objects sharing
more common attributes with other exemplars of their category
and relying to a greater extent upon their sensory and functional
properties (i.e., greater reliance upon sensory experience).
An argument in favor of the central nature of the semantic
impairment in aMCI is the multimodal nature of the deficits that
is observed. Subjects were asked identical semantic questions in
different modalities two weeks apart, in the visual modality during
the first session and verbal during the second. Object and person
semantics were significantly impaired in both modalities in aMCI
a highly significant correspondence between items in the visual
and verbal modalities. Results of the present study corroborate in
aMCI findings of previous studies which have evidenced impaired
access to semantic knowledge from different modalities of input
in different patient populations such as AD (Greene & Hodges,
1996) and SD (Snowden, Thompson, & Neary, 2004). An interest-
ing finding is that person semantics was notably more affected in
the visual than in the verbal modality in the aMCI and AD patients
(see Fig. 2). One possible interpretation is that in addition to the
semantic impairment, aMCI and early AD patients may also suffer
erbate the semantic deficit in the visual modality. The perceptual
analysis of complex visual stimuli such as faces may be impaired
although this deficit may not be serious enough to affect visual
this pattern of performance. There is also an alternative interpreta-
tion to these results, proposed by Haslam and colleagues (Haslam,
Kay, Hanley, & Lyons, 2004; Lyons et al., 2006). They found that
young and elderly healthy subjects were more familiar with and
produced more accurate information in response to the names of
famous people than in response to their faces. The authors sug-
gested that “faces and other visual information about people may
simply be less stable than names” (Haslam et al., 2004, p. 462). In
other words, changes in the appearance of faces may arise with age
(such as changes in hairstyle, facial appearance, etc.), while names
to faces than names.
Moreover, tasks requiring effortful, strategic access to semantic
memory have been found to be altered in elders with aMCI (Duong
et al., 2006). The questions assessing the semantic properties of
concepts in our experimental procedure were intentionally pre-
sented in a recognition task, in an effort to reduce the executive
load or the processes involved in the retrieval of semantic infor-
mation required in free recall tasks. Subjects had to answer yes
or no to questions that offered semantic contextual information,
which did not require much effort in terms of retrieval, selec-
tion among competing semantic alternatives, and manipulation of
semantic information. Therefore, we believe the semantic deficits
observed in our aMCI and AD patients cannot be attributed solely
to semantically oriented executive deficits. Rather, our results sug-
gest that a deficit at the level of the representation of the concept
is responsible for the deficits observed in semantic testing in aMCI
and early AD, and that this central deficit may be combined with
additional difficulties in the selection, manipulation and retrieval
4.3. Neural correlates of semantic impairment
Semantic cognition most likely relies upon a network of cor-
tical regions including the ATLs, the left inferior prefrontal cortex,
aim of our VBM study was to investigate if the semantic impair-
ment observed in our aMCI and early AD patients correlated with
atrophy in this semantic network of regions underlying semantic
cognition. Results first show that the degree of gray matter volume
in the left ATL and in the left inferior frontal gyrus correlates with
the performance in the semantic tests in aMCI and AD. Regarding
the role of the ATL, several independent lines of evidence suggest
that the ATL plays a critical role in the representation of an amodal
semantic store (Patterson, Nestor, & Rogers, 2007). Evidence of the
role of the ATL region in semantic memory has come in large part
from the study of SD patients (Patterson et al., 2007). SD patients
present with a progressive cross-modal semantic memory impair-
ment along with atrophy predominating in the left inferior and
anterior lateral temporal lobe (Hodges, Salmon, & Butters, 1992).
There is also evidence to suggest that the ATL is more involved in
processing concepts at a subordinate level rather than at a more
general superordinate level (Rogers et al., 2006). Moreover, the
integrity of the ATL region correlates with the integrity of semantic
abilities (Brambati et al., 2006). Thus, the ATL has been described as
key region in the semantic processing of concepts at a subordinate,
abstract and amodal level (Patterson et al., 2007). Furthermore,
converging evidence of the role of the ATL in conceptual knowl-
edge has come from functional neuroimaging studies. In fact, an
area of the left ATL almost overlapping with the one found in our
study (x=−50, y=8, z=−22) was found to be a common area of
activation for the fine-grained semantic processing of objects and
famous people (Gorno-Tempini & Price, 2001; Rogers et al., 2006).
In summary, our results thus show that the performance in seman-
tic tests in our patient sample is associated with the GM volume
in left ATL structures. These neuroanatomical results also provide
not explain the semantic deficit observed in aMCI and early AD and
that the observed semantic impairment is at least in part due to a
breakdown of stored conceptual knowledge.
The semantic test performance in aMCI and AD also correlated
with the GM volume in the left prefrontal cortex (IPC) at the level
of the inferior frontal gyrus. This area has been previously associ-
ated with the control mechanisms that interact with the semantic
representation to produce an appropriate activation of key knowl-
edge in order to successfully achieve a semantic task (Jefferies
& Lambon Ralph, 2006). Many fMRI studies have in fact demon-
strated the involvement of this area during tasks requiring control
of semantic information (Garavan, Ross, Li, & Stein, 2000; Peers et
al., 2005). Other neuroimaging studies have identified the anterior
and inferior prefrontal region as playing a key role in retrieving,
maintaining, monitoring and manipulating conceptual representa-
tions distributed in other regions of the brain (Gabrieli, Poldrack,
& Desmond, 1998; Martin & Chao, 2001; Poldrack et al., 1999).
Even though in the present study we tried to reduce the involve-
ment of executive mechanisms in the semantic task by presenting
yes–no type questions in a recognition mode, our task may still
have required to some degree the active generation and retrieval
of semantic knowledge, which is compatible with atrophy found
in the IPC in our patient sample. Finally, no significant correlation
was found between semantic test performance and the GM vol-
ume in the temporo-parietal region, a region which has also been
assumed to be associated with the semantic cognition network,
even when the threshold was lowered to a very permissive level
(p<0.05 uncorrected). In conclusion, this study is the first to pro-
vide evidence that the integrity of the ATL and the IPC correlates
with scores of semantic abilities of objects and people in aMCI and
4.4. Semantic memory in normal vs. pathological aging
Semantic memory does not have the same trajectory of decline
in normal aging as other memory systems such as episodic mem-
ory and working memory. In normal aging, the most robust and
S. Joubert et al. / Neuropsychologia 48 (2010) 978–988
well-documented age-related memory decline concerns episodic
memory (Verhaeghen & Cerella, 2002; Verhaeghen & Marcoen,
1993; Wechsler, 1997), while accumulated semantic knowledge
remains stable or even improves throughout the lifespan (Allen,
Sliwinski, & Bowie, 2002; Allen, Sliwinski, Bowie, & Madden, 2002;
Hoyer & Verhaeghen, 2006; Smith, 1996). In contrast, our results
and those of previous studies indicate that semantic deficits may
present as a very early manifestation of cognitive decline in incip-
ient AD. Amieva et al. (2008) showed that the first measurable
decline in cognitive performance began as early as 12 years before
the development of dementia and was observed on a category
fluency task, which is strongly associated with semantic knowl-
edge. Predementia patients who later on developed AD were also
at the outset more impaired at naming famous faces relative to
the non-converters (Estevez-Gonzalez et al., 2004; Thompson et
al., 2002; Vogel et al., 2005). Semantic tests may thus help discrim-
go unnoticed and this may be due in part to their mild nature and
to the limited availability of clinically oriented semantic tests in
a clinical setting. Future longitudinal studies are needed to deter-
mine if aMCI individuals presenting with semantic deficits are at
greater future risk of converting toward AD and if semantic tests
may contribute to improving early clinical diagnosis of AD.
This study demonstrated that aMCI patients present with nam-
ing deficits, as well as a genuine semantic impairment concerning
objects and famous people. The pattern of semantic decline in
MCI is similar to that found in early AD but contrasts with that
found in normal aging. Semantic impairment was associated with
distinct patterns of cortical atrophy in the ATL region previously
documented to be affected in SD and activated in semantic related
ment in aMCI and early AD may result from degraded semantic
representations combined with selective deficits in the selection,
manipulation and retrieval of semantic information.
This work was supported by a grant from the Alzheimer Society
of Canada (ASC) to SJ and by a Chercheur-Boursier Junior 1 award
to SJ. We are very thankful to statistician Francine Giroux for her
valuable help with analyses and for very useful suggestions, and
to Annie Dugas for her help in recruiting patient populations. We
would also like to thank all the patients and control subjects who
participated in this study.
List of the stimuli that were used in the semantic memory test
Aspirateur (vacuum cleaner)
Mètre de mesure (tape measure)
Ours Polaire (polar bear)
Rasoir électrique (electric razor)
Sèche-cheveux (hair dryer)
Tondeuse à Gazon (lawnmower)
Tronc ¸onneuse (chainsaw)
John F. Kennedy
Louis de Funès
Pierre Elliott Trudeau
Reine Elizabeth II
Examples of questions that were used in the semantic memory
test (verbal modality)
Est-ce que c’est un animal qui vit dans l’Arctique? (Is this an arctic
Est-ce que c ¸a a un bec? (Does this animal have a beak?)
Est-ce que c ¸a peut vivre dans l’eau? (Can this animal live in the water?)
Est-ce que c ¸a saute? (Can this animal jump?)
Est-ce que c’est un animal sauvage? (Is it a wild animal?)
Est-ce que c ¸a a des griffes? (Does this animal have claws?)
Est-ce que c ¸a produit du fromage? (Can you produce cheese from this
Est-ce que c ¸a se mange? (Can you eat the meat from this animal?)
Est-ce que c’est un instrument de jardinage? (Is this and instrument that is
used for gardening?)
Est-ce qu’il y en a qui sont rondes? (Is it round-shaped?)
Est-ce qu’il y en a qui font de la musique? (Can it be used to play music?)
S. Joubert et al. / Neuropsychologia 48 (2010) 978–988
Est-ce qu’il y en a qui s’utilisent à deux personnes? (Can it be used by two
Est-ce que cette personne est un homme politique? (Was this person a
Est-ce que cette personne est américaine? (Is this person American?)
Est-ce que cet homme a été Premier Ministre du Canada? (Was this person
Prime Minister of Canada?)
Est-ce qu’il était apprécié par les Montréalais? (Was this person
appreciated by Montrealers?)
Est-ce que cette personne a été au pouvoir dans les années 40? (Was this
person in power in the 1940s?)
Est-ce que cette personne est une chanteuse? (Is this person a singer?)
Est-ce que cette personne est originaire de France? (Is this person native
Est-ce que son mari était son imprésario? (Was her husband her
Est-ce qu’elle est enfant unique? (Was she a single child?)
Est-ce que cette personne est devenue connue dans les années 60? (Did
this person become famous in the 1960s?)
Est-ce que cette personne est une danseuse? (Is this person a dancer?)
Est-ce que cette personne est américaine? (Is this person American?)
Est-ce que cette femme décédée dans un accident de voiture? (Did this
woman die in a car accident?)
Est-ce qu’elle a eu une histoire d’amour avec le Président des États-Unis?
(Did she have an affair with the President of the United States?)
Est-ce que cette personne était très connue dans les années 80? (Was this
person very famous in the 1980s?)
edge in mild cognitive impairment and mild Alzheimer’s disease. Cortex, 42(5),
Ahmed, S., Arnold, R., Thompson, S. A., Graham, K. S., & Hodges, J. R. (2008). Nam-
ing of objects, faces and buildings in mild cognitive impairment. Cortex, 44(6),
Allen, P. A., Sliwinski, M., & Bowie, T. (2002). Differential age effects in semantic
and episodic memory. Part II: Slope and intercept analyses. Experimental Aging
Research, 28(2), 111–142.
Amieva, H., Le Goff, M., Millet, X., Orgogozo, J. M., Peres, K., Barberger-Gateau, P., et
al. (2008). Prodromal Alzheimer’s disease: Successive emergence of the clinical
symptoms. Annals of Neurology, 64(5), 492–498.
Ashburner, J., & Friston, K. J. (2005). Unified segmentation. Neuroimage, 26(3),
Benton, A. L., Varney, N. R., & Hamsher, K. D. S. (1978). Visuospatial judgement: A
clinical test. Archives of Neurology, 35, 364–367.
Brambati, S. M., Myers, D., Wilson, A., Rankin, K. P., Allison, S. C., Rosen, H. J., et al.
(2006). The anatomy of category-specific object naming in neurodegenerative
diseases. Journal of Cognitive Neuroscience, 18(10), 1644–1653.
Calero, M. D., Arnedo, M. L., Navarro, E., Ruiz-Pedrosa, M., & Carnero, C. (2002). Use-
fulness of a 15-item version of the Boston Naming Test in neuropsychological
Series B, Psychological Sciences and Social Sciences, 57(2), P187–P191.
Cardebat, D., Doyon, B., Puel, M., Goulet, P., & Joanette, Y. (1990). Evocation lexicale
formelle et sémantique chez des sujets normaux. Performances et dynamiques
de la production en fonction du sexe, de l’âge et du niveau d’étude. Acta Neuro-
logica Belgia, 90, 207–217.
Chertkow, H., & Bub, D. (1990). Semantic memory loss in dementia of
Alzheimer’s type. What do various measures measure? Brain, 113(Pt 2), 397–
Duong, A., Whitehead, V., Hanratty, K., & Chertkow, H. (2006). The nature of lexico-
semantic processing deficits in mild cognitive impairment. Neuropsychologia,
Estevez-Gonzalez, A., Garcia-Sanchez, C., Boltes, A., Otermin, P., Pascual-Sedano, B.,
Gironell, A., et al. (2004). Semantic knowledge of famous people in mild cogni-
tive impairment and progression to Alzheimer’s disease. Dementia and Geriatric
Cognitive Disorders, 17(3), 188–195.
Folstein, M., Folstein, S., & McHugh, P. (1975). “Mini-mental state”. A practical
method for grading the cognitive state of patients for the clinician. Journal of
Psychiatric Research, 12(3), 189–198.
Friston, K. J., Holmes, A., Worsley, K. J., Poline, J.-B., Frith, C. D., & Frackowiak, R.
S. (1994). Statistical parametric maps in functional imaging: A general linear
approach. Human Brain Mapping, 2(4), 189–210.
Fung, T. D., Chertkow, H., Murtha, S., Whatmough, C., Peloquin, L., Whitehead, V., et
al. (2001). The spectrum of category effects in object and action knowledge in
dementia of the Alzheimer’s type. Neuropsychology, 15(3), 371–379.
in language and memory. Proceedings of the National Academy of Sciences of the
United States of America, 95(3), 906–913.
Garavan, H., Ross, T. J., Li, S. J., & Stein, E. A. (2000). A parametric manipulation of
central executive functioning. Cerebral Cortex, 10(6), 585–592.
Mild cognitive impairment. Lancet, 367(9518), 1262–1270.
Gonnerman, L. M., Andersen, E. S., Devlin, J. T., Kempler, D., & Seidenberg, M. S.
(1997). Double dissociation of semantic categories in Alzheimer’s disease. Brain
and Language, 57(2), 254–279.
Good, C. D., Johnsrude, I. S., Ashburner, J., Henson, R. N., Friston, K. J., & Frackowiak,
R. S. (2001). A voxel-based morphometric study of ageing in 465 normal adult
human brains. Neuroimage, 14(1 Pt 1), 21–36.
Gorno-Tempini, M. L., & Price, C. J. (2001). Identification of famous faces and build-
ings: A functional neuroimaging study of semantically unique items. Brain,
124(Pt 10), 2087–2097.
Greene, J. D., & Hodges, J. R. (1996). The fractionation of remote memory. Evi-
dence from a longitudinal study of dementia of Alzheimer type. Brain, 119(Pt
Grober, E., Buschke, H., Crystal, H., Bang, S., & Dresner, R. (1988). Screening for
dementia by memory testing. Neurology, 38(6), 900–903.
specific or modality-neutral? Cortex, 40(3), 451–466.
Hodges, J. R., Erzinclioglu, S., & Patterson, K. (2006). Evolution of cognitive deficits
long-term follow-up study. Dementia and Geriatric Cognitive Disorders, 21(5–6),
in the course of Alzheimer’s disease? Neuroanatomical and diagnosic implica-
tions. Neuropsychologia, 33, 441–459.
Hodges, J. R., Patterson, K., Oxbury, S., & Funnell, E. (1992). Semantic dementia.
Progressive fluent aphasia with temporal lobe atrophy. Brain, 115(Pt 6), 1783–
Hodges, J. R., Salmon, D. P., & Butters, N. (1992). Semantic memory impairment in
Alzheimer’s disease: Failure of access or degraded knowledge? Neuropsycholo-
gia, 30(4), 301–314.
Hoyer, W. J., & Verhaeghen, P. (2006). Memory aging. In J. E. Birren, & K. W. Schaie
Huff, F. J., Corkin, S., & Growdon, J. H. (1986). Semantic impairment and anomia
in Alzheimer’s disease: The breakdown of semantic knowledge. Brain and Lan-
guage, 28, 235–249.
Jefferies, E., & Lambon Ralph, M. A. (2006). Semantic impairment in stroke apha-
sia versus semantic dementia: A case-series comparison. Brain, 129(Pt 8),
Joubert, S., Felician, O., Barbeau, E. J., Didic, M., Poncet, M., & Ceccaldi, M. (2008).
Patterns of semantic memory impairment in Mild Cognitive Impairment.
Behavioural Neurology, 19(1–2), 35–40.
Joubert, S., Felician, O., Barbeau, E., Ranjeva, J. P., Christophe, M., Didic, M., et al.
(2006). The right temporal lobe variant of frontotemporal dementia: Cogni-
tive and neuroanatomical profile of three patients. Journal of Neurology, 253(11),
Joubert, S., Felician, O., Barbeau, E., Sontheimer, A., Guedj, E., Ceccaldi, M., et al.
(2004). Progressive prosopagnosia: Clinical and neuroimaging results. Neurol-
ogy, 63(10), 1962–1965.
disorder: Further category-specific semantic dissociation. Cognitive Neuropsy-
chology, 19(2), 113–133.
Lam, L. C., Ho, P., Lui, V. W., & Tam, C. W. (2006). Reduced semantic fluency as an
Geriatric Cognitive Disorders, 22(2), 159–164.
Lawton, M., & Brody, E. (1969). Assessment of older people: Self-maintaining and
instrumental activities of daily living. Gerontologist, 9(3), 179–186.
Lyons, F., Kay, J., Hanley, J. R., & Haslam, C. (2006). Selective preservation of memory
for people in the context of semantic memory disorder: Patterns of association
and dissociation. Neuropsychologia, 44(14), 2887–2898.
Martin, A., & Chao, L. L. (2001). Semantic memory and the brain: Structure and
processes. Current Opinion in Neurobiology, 11(2), 194–201.
McKhann, G., Drachman, D., Folstein, M., Katzman, R., Price, D., & Stad-
lan, E. M. (1984). Clinical diagnosis of Alzheimer’s disease: Report of the
NINCDS-ADRDA Work Group under the auspices of Department of Health
and Human Services Task Force on Alzheimer’s disease. Neurology, 34(7),
Miceli, G., Capasso, R., Daniele, A., Esposito, T., Magarelli, M., & Tomaiuolo, F.
(2000). Selective deficit for people’s names following left temporal damage: An
impairment of domain-specific conceptual knowledge. Cognitive Neuropsychol-
ogy, 17(6), 489–516.
Murphy, K. J., Rich, J. B., & Troyer, A. K. (2006). Verbal fluency patterns in
amnestic mild cognitive impairment are characteristic of Alzheimer’s type
dementia. Journal of the International Neuropsychological Society, 12(4), 570–
The representation of semantic knowledge in the human brain. Nature Reviews.
Neuroscience, 8(12), 976–987.
Peers, P. V., Ludwig, C. J., Rorden, C., Cusack, R., Bonfiglioli, C., Bundesen, C., et
al. (2005). Attentional functions of parietal and frontal cortex. Cerebral Cortex,
988 Download full-text
S. Joubert et al. / Neuropsychologia 48 (2010) 978–988
Petersen, R. C. (2003). Conceptual overview. In R. C. Petersen (Ed.), Mild cognitive
Poldrack, R. A., Wagner, A. D., Prull, M. W., Desmond, J. E., Glover, G. H., & Gabrieli,
J. D. (1999). Functional specialization for semantic and phonological processing
in the left inferior prefrontal cortex. Neuroimage, 10(1), 15–35.
Regard, M. (1981). Cognitive rigidity and flexibility: A neuropsychological study.
Unpublished Ph.D dissertation, University of Victoria, Victoria.
Reitan, R. M. (1955). The relation of the Trail Making Test to organic brain damage.
Journal of Consulting Psychology, 19, 393–394.
Rey, A. (1960). Test de la Figure complexe de Rey. Paris: Les Éditions du Centre de
Rogers, T. T., Hocking, J., Noppeney, U., Mechelli, A., Gorno-Tempini, M. L., Patterson,
K., et al. (2006). Anterior temporal cortex and semantic memory: Reconciling
findings from neuropsychology and functional imaging. Cognitive, Affective &
Behavioral Neuroscience, 6(3), 201–213.
B. L. (2005). Neuroanatomical correlates of behavioural disorders in dementia.
Brain, 128(Pt 11), 2612–2625.
Rosser, A., & Hodges, J. R. (1994). Initial letter and semantic category fluency in
Alzheimer’s disease. Huntington’s disease, and progressive supranuclear palsy.
Journal of Neurology, Neurosurgery, and Psychiatry, 57(11), 1389–1394.
Seidenberg, M., Guidotti, L., Nielson, K. A., Woodard, J. L., Durgerian, S., Zhang, Q., et
Journal of the International Neuropsychological Society, 15(1), 9–18.
Smith, A. D. (1996). Memory. In J. E. Birren, & K. W. Schaie (Eds.), Handbook of the
psychology of aging (pp. 236–250). San Diego: Academic Press.
Snowden, J. S., Thompson, J. C., & Neary, D. (2004). Knowledge of famous faces and
names in semantic dementia. Brain, 127(Pt 4), 860–872.
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? Neuropsychology, 16(3), 344–
Van der Linden, M., Coyette, F., Poitrenaud, J., Kalafat, M., Calicis, F., Wyns, C., et
al. (2004). L’épreuve de rappel libre/rappel indicé à 16 items (RL/RI). In M. Van
der Linden, S. Adam, A. Agniel, C. Baisset-Mouly, F. Bardet, F. Coyette, B. Des-
M. Kalafat, J. Poitrenaud, F. Sellal, & C. Thomas-Antérion (Eds.), L’évaluation des
troubles de la mémoire. Présentation de quatre tests de mémoire épisodique (avec
leur étalonnage) (pp. 25–47). Marseille: Solal.
of meta-analyses. Neuroscience and Biobehavioral Reviews, 26(7), 849–857.
Verhaeghen, P., & Marcoen, A. (1993). Memory aging as a general phenomenon:
Episodic recall of older adults is a function of episodic recall of young adults.
Psychology and Aging, 8(3), 380–388.
Vogel, A., Gade, A., Stokholm, J., & Waldemar, G. (2005). Semantic memory impair-
ment in the earliest phases of Alzheimer’s disease. Dementia and Geriatric
Cognitive Disorders, 19(2–3), 75–81.
Warrington, E. K., & James, M. (1991). The visual object and space perception battery.
Unpublished manuscript, Bury St. Edmunds.
Wechsler, D. (1997). Weschler Memory Scale (WMS III): Administration and scoring
manual. San Antonio, TX: The Psychological Corporation.
Weschler, D. (2001). Echelle clinique de memoire de Weschler MEM III (WMS-III).
Unpublished manuscript, Paris.
Whatmough, C., Chertkow, H., Murtha, S., Templeman, D., Babins, L., & Kelner,
N. (2003). The semantic category effect increases with worsening anomia in
Alzheimer’s type dementia. Brain and Language, 84(1), 134–147.