Alcoholism: Clinical and Experimental Research Vol. 32, No.

7
July 2008

Episodic and Working Memory Deficits in Alcoholic

Korsakoff Patients: The Continuity Theory Revisited
Anne Lise Pitel, Hélène Beaunieux, Thomas Witkowski, François Vabret, Vincent de la
Sayette, Fausto Viader, Béatrice Desgranges, and Francis Eustache

Background: The exact nature of episodic and working memory impairments in alcoholic Kor-
sakoff patients (KS) remains unclear, as does the specificity of these neuropsychological deficits
compared with those of non-Korsakoff alcoholics (AL). The goals of the present study were
therefore to (1) specify the nature of episodic and working memory impairments in KS, (2) deter-
mine the specificity of the KS neuropsychological profile compared with the AL profile, and (3)
observe the distribution of individual performances within the 2 patient groups.
Methods: We investigated episodic memory (encoding and retrieval abilities, contextual mem-
ory and state of consciousness associated with memories), the slave systems of working memory
(phonological loop, visuospatial sketchpad and episodic buffer) and executive functions (inhibi-
tion, flexibility, updating and integration abilities) in 14 strictly selected KS, 40 AL and 55 control
subjects (CS).
Results: Compared with CS, KS displayed impairments of episodic memory encoding and
retrieval, contextual memory, recollection, the slave systems of working memory and executive
functions. Although episodic memory was more severely impaired in KS than in AL, the single
specificity of the KS profile was a disproportionately large encoding deficit. Apart from organiza-
tional and updating abilities, the slave systems of working memory and inhibition, flexibility and
integration abilities were impaired to the same extent in both alcoholic groups. However, some
KS were unable to complete the most difficult executive tasks. There was only a partial overlap
of individual performances by KS and AL for episodic memory and a total mixture of the 2
groups for working memory.
Conclusions: Korsakoff’s syndrome encompasses impairments of the different episodic and
working memory components. AL and KS displayed similar profiles of episodic and working
memory deficits, in accordance with neuroimaging investigations showing similar patterns of brain
damage in both alcoholic groups.
Key Words: Korsakoff’s Syndrome, Chronic Alcoholism, Episodic Memory, Working Mem-
ory, Executive Functions.

T HE COMBINATION OF heavy alcohol consumption

and thiamine deficiency is the most common etiology of
Korsakoff’s syndrome (Korsakoff, 1889). It is classically
described as a disproportionate impairment of episodic mem-
ory, compared with other aspects of cognitive function (Kop-
elman, 1995, 2002). Episodic memory is currently described
as the memory system in charge of the encoding, storage and
retrieval of personally experienced events, associated with a
precise spatial and temporal context of encoding. Episodic
memory includes a specific state of consciousness (autonoetic
From the Inserm-EPHE-Universite´ de Caen ⁄ Basse-Normandie,
Unite´ U923, GIP Cyceron, CHU Côte de Nacre (ALP, HB, TW,
FVa, VdlS, FVi, BD, FE), Service d’Alcoologie, CHR Cle´menceau,
(FVa), and De´partement de Neurologie, CHU Côte de Nacre (VdlS,
FVi) Caen, France.
Received for publication November 9, 2007; accepted March 10, 2008.
Reprint requests: Prof. Francis Eustache, Inserm-EPHE-Universite´
de Caen ⁄ Basse-Normandie, Unite´ U923, Laboratoire de Neuropsy-
chologie, CHU Côte de Nacre, 14033 Caen Cedex, France; Fax:
+33 (0)2 31 06 51 98; E-mail: neuropsycho@chu-caen.fr
Copyright  2008 by the Research Society on Alcoholism.
DOI: 10.1111/j.1530-0277.2008.00677.x
Alcohol Clin Exp Res, Vol 32, No 7, 2008: pp 1–13
consciousness) giving rise to the impression of re-experiencing
or reliving the past and mentally traveling back in subjective
time (Tulving, 2001). The specific features of the episodic
memory disorder in Korsakoff’s syndrome remain unclear,
but 2 main hypotheses have been put forward.
It was initially suggested that the amnesia might stem from
a deficit in retrieval processes (Irle et al., 1990; McDowall,
1981; Warrington and Weiskrantz, 1978; d’Ydewalle and Van
Damme, 2006). Another hypothesis centered on a deficit in
encoding processes (van Asselen et al., 2005; Cermak and
Butters, 1972; Kopelman, 1985) which was suspected of dis-
proportionately affecting contextual memory (Kopelman,
1989; Mayes et al., 1985, 1992; McKone and French, 2001).
Indeed, compared with healthy controls, patients with
Korsakoff amnesia have been shown to encode temporal
(Kopelman, 1985; Kopelman et al., 1997) and spatial con-
textual information (Kopelman et al., 1997; Mayes et al.,
1991; Shoqeirat and Mayes, 1991) less efficiently. Further-
more, alcoholic Korsakoff patients may have difficulty
integrating these 2 contextual attributes with the target
item (Postma et al., 2006), suggesting a deficit in the
1
2 PITEL ET AL.
incidental binding together of item and contextual informa-
tion to form complex memories (Chalfonte et al., 1996).
This inability to encode the spatiotemporal context and ⁄ or
retrieve episodic information has led to a third hypothesis
characterizing the amnesia in this pathology as a deficit in
conscious recollection abilities (McKone and French, 2001).
d’Ydewalle and Van Damme (2006) assessed the state of con-
sciousness associated with memories using an R ⁄ K ⁄ G para-
digm (Gardiner et al., 2002) which makes it possible to
differentiate between autonoetic and noetic consciousness.
This experimental paradigm consists of a recognition task in
which subjects have to indicate whether they remember the
specific episode (R answer - autonoetic consciousness), know
that this episode happened to them (K answer - noetic con-
sciousness) or guess that they may have experienced this epi-
sode (G answer). Korsakoff patients were found to provide
fewer R answers than controls, indicating a deterioration in
autonoetic consciousness in this pathology.
At present, none of these hypotheses (i.e., deficits in retrie-
val, encoding or recollection processes) has definitively been
adopted to explain the amnesia in Korsakoff’s syndrome.
Thus, further studies are required to gain a better understand-
ing of the impaired episodic memory components of this
amnesic syndrome.
Working memory also needs to be investigated afresh in
Korsakoff patients, with reference to the theoretical frame-
work put forward recently by Baddeley (Baddeley, 2000,
2003). According to this author, working memory is com-
posed of 3 slave systems plus a central executive. The first 2
slave systems of working memory to be described were the
phonological loop and the visuospatial sketchpad (Baddeley,
1986), which are thought to be in charge of the short-term
storage of phonological and visuospatial information. The
newly identified slave system of working memory is supposed
to be responsible for maintaining multimodal information
and is known as the episodic buffer. The central executive is
considered by the author to play a similar role to that of exec-
utive functions (Baddeley, 1996).
Slave systems of working memory have seldom been
assessed in Korsakoff’s syndrome. The phonological loop
and the visuospatial sketchpad are reportedly preserved in
alcoholic Korsakoff patients (Joyce and Robbins, 1991; Kop-
elman, 1991; Pollux et al., 1995). However, a more recent
study revealed impaired performances on both digit and
block spans (Brand et al., 2003). As for the episodic buffer, it
has yet to be investigated in this disease. Executive functions,
such as organization, inhibition, flexibility, categorization and
updating abilities, have been more frequently assessed and are
generally reported to be impaired in KS (Brand et al., 2005;
Brokate et al., 2003; Jacobson et al., 1990; Kopelman, 1995;
Oscar-Berman et al., 2004).
Further investigations are required to examine the different
components of episodic memory and working memory in the
light of their various current neuropsychological conceptions
(Baddeley, 2000, 2003; Eustache and Desgranges, 2008; Tulv-
ing, 2001, 2002; Wheeler et al., 1997). Such studies would not
only help us to describe episodic and working memory disor-
ders in Korsakoff’s syndrome but would also allow us to
determine the specificity of these memory deficits compared
with those that are already present in non-alcoholic Korsak-
off patients.
In effect, heavy and prolonged alcohol consumption leads
to brain damage (Chanraud et al., 2007; Rosenbloom et al.,
2003) and cognitive deficits (Parsons and Nixon, 1993; Sulli-
van et al., 2000) in alcoholics without Korsakoff’s syndrome.
Together with several other neuropsychological disorders,
impairments of factual information learning (Glenn and Par-
sons, 1992; Tivis et al., 1995), the slave systems of working
memory (Pitel et al., 2007b) and executive functions
(Ambrose et al., 2001; Ihara et al., 2000; Noel et al., 2001) are
classically described in non-Korsakoff alcoholic patients.
In a recent study (Pitel et al., 2007a), we identified the com-
ponents of episodic memory and executive functions that are
impaired in non-Korsakoff alcoholics. Results confirmed defi-
cits in learning abilities and indicated impairments in encod-
ing and retrieval processes, contextual memory and
autonoetic consciousness (Tulving, 2001). Non-Korsakoff
alcoholics also presented executive dysfunctions, such as
impairments of organization, inhibition, flexibility, updating
and integration abilities. Our findings revealed that the epi-
sodic memory impairments in non-Korsakoff alcoholics
should not be regarded as the consequences of these executive
dysfunctions, but rather as ‘‘genuine’’ episodic memory defi-
cits. As some of these neuropsychological impairments had
previously been regarded as characteristic of Korsakoff’s syn-
drome, our results called into question the specificity of the
neuropsychological profile of alcoholic Korsakoff patients
compared with non-Korsakoff alcoholics.
Concerning episodic memory performance, a few studies
have made direct comparisons of performances between these
2 groups, showing that alcoholic Korsakoff patients are more
severely impaired than non-Korsakoff alcoholics (Emsley
et al., 1996; Krabbendam et al., 2000). However, no study has
qualitatively compared the profiles of impaired episodic mem-
ory components in these 2 groups, taking into account one of
the recent definitions of this memory system (Tulving, 2001;
Wheeler et al., 1997). As far as the slave systems of working
memory are concerned, a single study (Joyce and Robbins,
1991) has compared scores on a span task assessing the visuo-
spatial sketchpad. The authors found that alcoholic Korsak-
off patients and non-Korsakoff alcoholics did not
significantly differ from controls on a spatial span task. On
the other hand, comparisons of the executive functions of
these 2 groups have revealed that organizational abilities may
be more severely impaired in Korsakoff patients than in non-
Korsakoff alcoholics (Hildebrandt et al., 2004; Jacobson
et al., 1990; Joyce and Robbins, 1991; Oscar-Berman et al.,
2004). Alcoholic Korsakoff patients have also been shown to
perform more poorly than non-Korsakoff alcoholics on flexi-
bility (Brokate et al., 2003; Jacobson et al., 1990) and updat-
ing tasks (Brokate et al., 2003; Hildebrandt et al., 2004). By
contrast, an earlier investigation reported similar levels of
EPISODIC AND WORKING MEMORY IN KORSAKOFF PATIENTS
inhibition abilities in both these groups (Krabbendam et al.,
2000) and failed to find any impairment of inhibition in either
the non-Korsakoff alcoholics or the alcoholic Korsakoff
patients. These comparative studies have provided some use-
ful preliminary information about episodic memory and
working memory impairments in alcoholic Korsakoff patients
compared with non-Korsakoff alcoholics, but further investi-
gations are still required.
Findings suggesting that some of the episodic memory
impairments and executive dysfunctions in alcoholic Korsak-
off patients may already be present in non-Korsakoff alcohol-
ics support the continuity hypothesis (Ryback, 1971). This
theory postulates that cognitive disturbances in non-Korsak-
off alcoholics and alcoholic Korsakoff patients lie along a
continuum of mild to moderate deficits (Bowden, 1990; But-
ters and Brandt, 1985; Parsons, 1998). According to this
hypothesis, alcohol has a neurotoxic effect on brain function
and there is therefore a correlation between cognitive perfor-
mance and drinking history (Oscar-Berman et al., 2004).
However, researchers have so far failed to demonstrate any
such correlation between neuropsychological deficits and
drinking history (Ihara et al., 2000; Noel et al., 2001; Schaef-
fer and Parsons, 1987), notably because it is difficult to gain
an accurate picture of the drinking history of alcoholic
patients, let alone alcoholic Korsakoff patients (Emsley et al.,
1996). Even though their drinking history may not explain
the heterogeneity of alcoholics’ cognitive abilities, the theory
of a continuum between non-Korsakoff alcoholics and alco-
holic Korsakoff patients may at least have a heuristic value
(Butters and Brandt, 1985) and could be explored separately
for each cognitive function with, for example, episodic mem-
ory on the one hand and working memory ⁄ executive func-
tions on the other. For some authors, this continuum may
even extend all the way from social drinking (Parsons, 1998)
to alcohol dementia (Bowden, 1990; Butters and Brandt,
1985; Kopelman, 1995).
The present study therefore had 3 main goals. The first
objective was to identify the characteristics of episodic mem-
ory impairments and working memory deficits in Korsakoff’s
syndrome, with reference to current definitions of these cogni-
tive functions. The second objective was to determine the
specificity of the episodic and working memory impairments
in alcoholic Korsakoff patients compared with those of non-
Korsakoff alcoholics. We took the duration of abstinence
and the intelligence levels of the Korsakoff patients into
account, as these factors may lead to the existence of sub-
groups of alcoholic Korsakoff patients and therefore influ-
ence comparison results. Moreover, we conducted not only
between-group analyses [alcoholic Korsakoff patients vs.
non-Korsakoff alcoholics vs. control subjects (CS)], in line
with previous comparative studies, but also, more originally,
within-group analyses (e.g. encoding vs. retrieval) to compare
the memory profiles of the 2 patient groups. The third and
final objective was to observe the distribution of individual
performances in terms of episodic memory and working
memory in the 2 patient groups. In effect, statistical mean-
3
based analyses hide the variability of patients’ results. Observ-
ing the distribution of individual performances would make it
possible to compare disability profiles and identify any over-
lap in the distribution of scores across the 2 groups. Just such
an overlap in the distribution of cognitive test scores has
already been noted in non-Korsakoff alcoholics and healthy
controls (Parsons, 1998), and further investigations are now
required, in order to ascertain whether this overlap extends to
alcoholic Korsakoff patients.
MATERIALS AND METHODS
Subjects
The sample consisted of 109 subjects: 14 alcoholic Korsakoff
patients (KS), 40 non-Korsakoff alcoholics (AL) and 55 CS. The
3 groups were matched for age and number of years of schooling.
However, KS differed from AL on the Mini Mental State Exami-
nation (Folstein et al., 1975) and the 3 groups differed on the
sum of the 3 free recalls in the French version of the free and
cued selective reminding test (FCSRT) (Grober and Buschke,
1987; Grober et al., 1988). Two subtests of the WAIS III (Wesch-
ler, 2001) were administered to measure intelligence. On the
vocabulary subtest, reflecting verbal IQ, KS performed signifi-
cantly lower than AL who, in turn, obtained poorer scores than
CS (Brand et al., 2003; Oscar-Berman et al., 2004). On the matrix
subtest, reflecting performance IQ, AL and KS were similarly
impaired, compared with CS (Brand et al., 2003; Joyce and Rob-
bins, 1991). Demographic information and other characteristics of
the participants are provided in Table 1. No participants were
aged over 70, were taking psychotropic medication or presented
psychiatric problems or had any additional medical history (head
injury, coma, epilepsy, depression, etc.) which might have affected
cognitive function. In agreement with the local ethics committee,
all of them gave their informed verbal consent, collected prior to
their inclusion in the study by the medical doctors who also car-
ried out the investigations in this study. They ascertained first
(notably for KS) that participants were able to understand the
goals and modalities of the experiment and second that they
agreed to participate. For KS, wherever possible, informed con-
sent was collected from guardians or caregivers as well as from
the patients themselves.
Alcoholic Korsakoff Patients. Korsakoff patients were diag-
nosed with reference to the clinical DSM IV (American Psychiatric
Association, 1994) criteria of ‘‘amnesia due to substance abuse’’. All
KS patients had a history of heavy drinking (longer than 20 years),
presented disproportionately severe memory disorders (as showed by
the FCSRT in Table 2) compared with other cognitive functions and
their memory impairments had social repercussions. The conse-
quences of their cognitive impairments were such that none of the
KS were able to go back to their previous jobs and all of them lived
in sheltered accommodation or were inpatients waiting for a place in
an institution.
The background information for the KS came from family mem-
bers and medical records. For each patient, the selection was made
according to a codified procedure in French officially registered cen-
ters by 2 senior neurologists (V.D.L.S. and F.V.). Each patient’s case
was examined by a multidisciplinary team made up of specialists in
cognitive neuropsychology and behavioral neurology. Clinical and
neuroimaging investigations ruled out other possible causes of mem-
ory impairments (particularly focal brain damage). Four KS had pre-
sented full-blown Wernicke’s encephalopathy with the classic triad of
symptoms (oculomotor disorders, confusion and cerebellar ataxia).
In 7 other KS, Wernicke’s encephalopathy may have preceded Kor-
sakoff’s syndrome, but this episode was characterized by an incom-
4 PITEL ET AL.

Table 1. Main Features of Participants

Alcoholic Korsakoff Non-Korsakoff Control Group

patients (N = 14) alcoholics (N = 40) subjects (N = 55) Statistics p-value comparisons
Age 54.35 ± 6.82 49.41 ± 6.82 50.46 ± 10.95 F(2,106) = 1.49 0.22 ⁄
Years of schooling 11.71 ± 3.96 10.50 ± 2.63 11.58 ± 3.65 F(2,106) = 1.38 0.25 ⁄
Mini Mental State Examination 22.23 ± 3.62 27.10 ± 2.46 ⁄ t(52) = 5.65 <0.001* AL > KS
FCSRTa 7.21 ± 3.79 28.13 ± 7.44 33.49 ± 4.99 F(2,106) = 111.98 <0.001* CS > AL > KS
Vocabulary subtestb (WAIS III) 5.21 ± 2.75 7.40 ± 3.26 11.20 ± 2.26 F(2,106) = 38.34 <0.001* CS > AL > KS
Matrix subtestb (WAIS III) 7.64 ± 2.67 8.38 ± 2.37 11.02 ± 2.07 F(2,106) = 21.86 <0.001* CS > (AL = KS)

Mean ± standard deviation. KS: alcoholic Korsakoff patients; AL: non-Korsakoff alcoholics; CS: control subjects.
a
Free and cued selective reminding test (sum of the 3 free recalls).
b
Standard T scores.
*Significant effect of group (t-test for the Mini Mental State Examination and univariate ANOVA for the free and cued selective reminding test
and the WAIS III subtests).

Table 2. Clinical Description of the Alcoholic Korsakoff Patients

Alcoholic Length of FCSRTa:

Korsakoff Schooling Gayet-Wernicke sobriety raw FCSRTa:
patients Gender Age (years) encephalopathy (months) scores z-scores
1 M 54 8 Complete 1 3 )6.24
2 M 56 12 Complete 24 9 )5.01
3 M 58 7 Partial 1 4 )6.03
4 M 48 15 Absent 2 11 )4.60
5 M 50 12 Partial 6 5 )5.83
6 M 69 20 Absent 1 7 )5.42
7 M 63 16 Absent 52 13 )4.19
8 F 52 6 Partial 1 6 )5.63
9 F 61 14 Partial 100 10 )4.81
10 M 47 9 Complete 80 12 )4.40
11 M 44 9 Partial 2 2 )6.44
12 F 49 15 Partial 1 11 )4.60
13 M 55 12 Complete 1 6 )5.63
14 M 55 9 Partial 4 2 )6.44

a
FCSRT: Free and cued selective reminding test (sum of the 3 free recalls).

plete clinical picture (e.g. nystagmus). In 3 other KS, there was no

history of Wernicke’s encephalopathy at all. The amnesia in these last
KS had developed insidiously but over a period of <8 weeks (Cut-
ting, 1978). Details of the KS sample are provided in Table 2.
Non-Korsakoff Alcoholics and Control Participants. The AL
patients and CS were the same as in our previous study (Pitel et al.,
2007a). There were 20 men in the control group and 33 in the alco-
holic group. AL patients were recruited by clinicians, in accordance
with the DSM IV criteria of alcohol dependence (American Psychi-
atric Association, 1994) while they were receiving alcohol with-
drawal treatment as inpatients at Caen University Hospital. All
patients were still at an early stage of abstinence (11.18 ±
11.75 days of sobriety prior to inclusion,), as it is now clear that epi-
sodic memory and working memory recover with abstinence and
may even return to normal (Fein et al., 2006; Rourke and Grant,
1999). We only selected patients who had already been weaned off
alcohol, as verified by means of the Cushman score (Cushman
et al., 1985), to decrease the likelihood of acute alcohol withdrawal
effects. AL were interviewed to specify the age at which they had
had their first alcoholic drink (16.67 ± 4.13), their age at the onset
of alcoholism (28.63 ± 11.04), the length of time they had drunk to
excess (20.32 ± 8.08 years) and their usual daily alcohol consump-
tion (18.91 ± 10.16 ‘‘standard drinks’’ per day, i.e., any drink that
contains approximately 10 g of pure alcohol). CS were interviewed
to check that their alcohol consumption did not exceed the recom-
mendations of the World Health Organization (no more than 21
weekly standard drinks for men and 14 for women, and no more
than 4 at any 1 time).
Procedures
Episodic Memory Assessment. The different components of epi-
sodic memory were assessed by means of 2 complementary memory
tests (Pitel et al., 2007a).
Encoding and retrieval processes: In order to preferentially tax
either encoding or retrieval processes, we used a specifically
designed memory test, inspired by the Double Memory test (Bus-
chke et al., 1995). In this task, entitled the ‘‘Spondee’’ (for SPON-
taneous-DEEp) test, the subjects had to learn 2 different lists of
16 words belonging to 16 different categories and encoded in 2
different ways. The first list was encoded spontaneously according
to the strategies that the subjects were able to implement on their
own: the words were displayed 4 · 4 on a sheet and subjects had
to point to words read aloud by the experimenter. A recognition
phase was then carried out, where patients had to recognize the
16 target words hidden among distractors. Target and distractor
words were presented visually 1 at a time and subjects had to say
if they recognized them. Unlike the first word list, the second one
was deeply encoded. The words were still displayed 4 · 4 on a
sheet, but subjects had to point to them in response to their
semantic category. They were then asked to recall as many words
as possible, in any order and without any time limit.
This is a method which had already been used in a previous study
by our research team (Chételat et al., 2003) to assess encoding and
retrieval processes. Recognition following the spontaneous encoding
subtest, where the target was presented in order to compensate for
any potential retrieval deficits, but where encoding was not sup-
ported, was assumed to be mainly dependent on encoding ability. In
EPISODIC AND WORKING MEMORY IN KORSAKOFF PATIENTS
contrast, free recall in the wake of the deep encoding subtest, which
represented the opposite situation, given that encoding was rein-
forced whereas retrieval was self-initiated, was assumed to reflect
mainly retrieval ability. For simplicity’s sake, these tests will be desig-
nated hereafter by the process they preferentially taxed (i.e., ‘‘encod-
ing’’ and ‘‘retrieval’’ subtests).
Factual, temporal and spatial memory: The factual, temporal and
spatial components of episodic memory were measured by means of
an original Ecological Contextual Memory test (ECM test). The
ECM test consisted in learning 6 pairs of unrelated words (Danger-
Beauty, Student-Tongue, etc.) presented at different times (3 pairs in
the morning and 3 in the afternoon) and places (different for each
subject and each pair of words) in the space of a single day. Each
time, subjects had to read the word pair, construct a sentence con-
taining each word and recall both words to ensure that they had been
encoded semantically (i.e., deeply and deliberately). Subjects were
told that they had to learn all 6 word pairs but not that they had to
learn the context. At the end of the day, the subjects performed a rec-
ognition task of the factual, temporal, and spatial information. The
target word pair was displayed at the same time as 3 distractors (1
word pair that was semantically linked with the target, one that was
phonologically linked and one that had no link at all). For each word
pair, subjects also had to recognize the temporal and spatial context
of encoding, in the presence of a distractor which corresponded to
the encoding context of a different word pair. The ‘‘total recognition
score’’ corresponded to the recognition of the whole episode: which
words were presented and when and where this episode took place.
State of consciousness associated with memories: The subjective
reporting of state of consciousness was assessed using the ‘‘Remem-
ber ⁄ Know ⁄ Guess’’ paradigm (Gardiner et al., 2002). In the recogni-
tion task of the ECM test described above, subjects had to indicate
whether they (1) remembered the specific episode with the impression
of reliving some of the details (R answer), (2) just knew that this epi-
sode had happened to them but did not remember any specific event
or detail (K answer), or (3) guessed that they might have experienced
this episode, but neither remembered nor knew it (G answer). These
responses were given independently for each type of content (factual,
temporal and spatial). The scores corresponded to the combined per-
centages of R, K, and G answers for each correctly recognized con-
tent. Only R answers reflected the autonoetic state of consciousness
specific to episodic memory.
Working Memory Assessment. The different components of
working memory were evaluated by means of specific neuropsycho-
logical tasks (Pitel et al., 2007b).
Slave systems: The slave systems of working memory were assessed
by means of 3 computerized passive storage tasks. The phonological
loop and the visuospatial sketchpad were evaluated by verbal span
and spatial span tasks respectively. The episodic buffer was assessed
by means of a multimodal span task (Quinette et al., 2006): patients
were asked to memorize increasingly long strings of letters (verbal
span), locations (spatial span) and letters placed in an array (multi-
modal span) and had to recall them immediately afterwards. The
final score corresponded to the number of correctly reported
sequences.
Central executive: The central executive of working memory was
assessed using 5 different types of task involving executive functions.
Organization and the ability to generate strategies were assessed by 2
verbal fluency tasks (Cardebat et al., 1990). The fluency score corre-
sponded to the number of correct words supplied in the letter (p) and
categorical (animal) fluency tasks. Inhibition ability was assessed by
means of the Stroop test (Stroop, 1935). The number of colors
named in the interference condition was recorded to gauge inhibition
ability. The scores were age-corrected. Flexibility ability was assessed
by the alternate response subtest of the attentional assessment test
(Zimmermann and Fimm, 1993), updating abilities by the n-back
(N-2) paradigm (Quinette et al., 2003) and integration ability by a
5
computerized test (Quinette et al., 2006) which required subjects to
integrate multimodal information before memorizing it. For the last
3 executive tasks, the percentage of correct answers was recorded.
Statistical Analysis
Since there was no significant difference between the cognitive per-
formances of the men and women in each groups, we pooled the data
without taking account of the sex ratio.
The first step in the statistical analysis consisted in standardizing
all the subjects’ neuropsychological scores in the form of individual
z-scores. These z-scores, computed for each subject and each vari-
able, were based on the mean performance of the CS group which,
by definition, had mean scale scores of zero and standard deviations
of one. Thus, episodic memory and working memory variables were
all on the same scale, allowing us to carry out between- and within-
group analyses.
We then carried out repeated-measures analyses of variance
(ANOVA) on the episodic memory and working memory variables.
In the case of episodic memory, we examined encoding versus retrie-
val processes, the different components of contextual memory and
the state of consciousness associated with memories. The slave sys-
tems of working memory and the central executive were investigated
separately. Post hoc tests (LSD Fisher) were subsequently performed,
in order to conduct between-group (KS vs. AL vs. CS) and within-
group comparisons (e.g., encoding vs. retrieval). The observation of
interaction effects and comparisons of post hoc analyses provided
information about the existence of similar or contrasting profiles of
neuropsychological impairments in the 2 patient groups.
To determine whether the results of the comparisons between
AL and KS occurred as a result of the heterogeneity of the KS
sample, we formed subgroups of KS on the basis of the duration
of sobriety and degree of intelligence. The short-term abstinent
KS (N = 8) had a period of sobriety less than or equivalent to
2 months, whereas the long-term abstinent KS (N = 6) had a
period of sobriety of more than 2 months. Concerning intelli-
gence, we chose a cut-off score corresponding to the mean
(Weschler, 2001) minus a 1.65 standard deviation (10–
1.65 · 3 = 5.05) for both the vocabulary and matrix subtests. We
thus identified a subgroup of KS with impaired intelligence
(N = 7) and another one with preserved intelligence (N = 7).
These subgroups were compared by means of Mann–Whitney
tests.
Lastly, we examined the distribution of the individual perfor-
mances of the 2 patient groups on the episodic and working memory
tasks. To this end, we calculated 2 mean z-scores for each patient: a
mean z-score for episodic memory (computed from impaired episodic
memory scores) and a mean z-score for working memory (computed
from impaired working memory scores). We subjected the results of
the episodic memory test used at inclusion to the same analysis, in
order to emphasize the validity of our patient selection.
A probability level of 0.05 was adopted for all analyses.
RESULTS
Repeated-Measures Analyses of Variance
Episodic Memory. The repeated-measures analysis of var-
iance carried out on encoding and retrieval scores showed a
significant effect of group [F(2,106) = 72.71; p < 0.001] and
test [F(1,106) = 6.573; p = 0.01], as well as an interaction
effect [F(2,106) = 6.57; p = 0.002]. Post hoc tests concerning
between-group comparisons revealed that, for both scores,
KS were more severely impaired than AL and CS, with AL
performing more poorly than CS as well. Within-group
6 PITEL ET AL.

comparisons showed that the encoding score of KS was lower
than the retrieval score, unlike AL, who presented similar
encoding and retrieval impairments (Fig. 1A).
Concerning contextual memory, results also indicated a sig-
nificant effect of group [F(2,105) = 53.56; p < 0.001] and test
[F(3,315) = 32.29; p < 0.001], as well as interaction effects
[F(6,315) = 11.27; p < 0.001]. Results of the post hoc
between-group comparisons indicated that, on factual recog-
nition, KS had lower results than AL and CS, whose perfor-
mances were broadly similar. Concerning temporal, spatial
and total recognition, AL differed from CS, with KS present-
ing even poorer results than AL. Results of the post hoc
within-group comparisons were the same for both patient
groups: temporal recognition and recognition of the whole
episode were similarly and significantly more impaired than
spatial recognition which, in turn, was more severely impaired
than factual recognition (Fig. 1A).
With regard to the state of consciousness associated with
memories, there was also a significant effect of group
[F(2,105) = 23.55; p < 0.001] and test [F(2,210) = 40.15;
p < 0.001], together with a significant interaction effect
[F(4,210) = 18.16; p < 0.001]. Post hoc between-group
comparisons revealed that KS performed more poorly than
AL, who also obtained lower performances than CS on the
number of R and G answers, although there was no differ-
ence in the number of K answers. Post hoc within-group com-
parisons showed that in both patient groups, the number of
R and G answers was more badly affected than the number
of K answers (Fig. 1A).
Raw episodic memory scores are reported in Table 3.

Fig. 1. Impaired z scores of episodic memory (A) and working memory (B) in alcoholic Korsakoff patients and non-Korsakoff alcoholics compared with
control subjects. Control subjects had mean scores of zero and standard deviations of 1. Mean z-scores are shown for each patient group and each
impaired variable. (A) Even though episodic memory was more severely impaired in alcoholic Korsakoff patients than in non-Korsakoff alcoholics, the 2
groups presented similar profiles of impaired episodic memory components, except for encoding, which was specifically disproportionately impaired in alco-
holic Korsakoff patients. (B) The tasks assessing the slave systems of working memory and inhibition, flexibility and integration abilities were impaired to the
same extent in both patient groups. However, organization and up-dating abilities were more severely impaired in the alcoholic Korsakoff patients.
EPISODIC AND WORKING MEMORY IN KORSAKOFF PATIENTS 7

Table 3. Episodic Memory Performances of Alcoholic Korsakoff Patients, Non-Korsakoff Alcoholics, and Control Subjects (Raw Data)

Cognitive Alcoholic Korsakoff Non-Korsakoff Post hoc

functions Variables patients alcoholics Control subjects comparisons
Encoding Recognition after spontaneous encoding 37.95 ± 13.39 76.25 ± 15.77 84.31 ± 13.39 CS >
AL > KS
Retrieval Free recall after deep encoding 16.07 ± 7.23 40.47 ± 17.73 53.64 ± 17.50 CS >
AL > KS
Contextual memory Factual recognitiona 56.41 ± 23.11 97.50 ± 8.05 99.09 ± 3.82 (CS =
AL) > KS
Temporal recognitiona 51.28 ± 17.30 77.91 ± 19.39 92.12 ± 11.94 CS >
AL > KS
Spatial recognitiona 55.13 ± 19.70 80.00 ± 20.04 88.48 ± 16.32 CS >
AL > KS
Total recognitiona 19.23 ± 17.80 62.08 ± 26.14 83.03 ± 20.41 CS >
AL > KS
State of R answersa 0.54 ± 0.80 1.93 ± 0.82 2.25 ± 0.63 CS >
AL > KS
consciousness K answersa 0.74 ± 0.81 0.47 ± 0.50 0.46 ± 0.40 ns
associated with G answersa 1.69 ± 1.16 0.58 ± 0.61 0.29 ± 0.40 CS > AL > KS
memories

Mean ± standard deviation. KS: alcoholic Korsakoff patients; AL: non-Korsakoff alcoholics; CS: control subjects. R = ‘‘remember’’ answers in
the R ⁄ K ⁄ G paradigm; K = ‘‘know’’ answers in the R ⁄ K ⁄ G paradigm; G = ‘‘guess’’ answers in the R ⁄ K ⁄ G paradigm. ns: no significant difference.
a
Data from 1 KS were missing.

Working Memory. The repeated-measures ANOVA car-

ried out on performances on the 3 span tasks found a signifi-
cant group effect [F(2,106) = 35.14; p < 0.001] but no effect
of test [F(2,212) = 0.40; p = 0.67] and no interaction effect
[F(44,212) = 0.39; p = 0.82]. Post hoc between-group compar-
isons revealed that KS and AL had lower results than CS in
the 3 span tasks, but there was no difference between the 2
patient groups. The within-group comparisons revealed that
all 3 span tasks were impaired to the same extent.
Not all the KS carried out all the executive tests, as some of
the tasks were too difficult for those with the most severe
dysexecutive syndrome. As a consequence, the following anal-
ysis was carried out on the available data for each test
(Table 4). The repeated-measures ANOVA showed a signifi-
cant group effect [F(2,102) = 26.30; p < 0.001], but no effect
of test [F(4,408) = 2.18; p = 0.07] and no interaction effect
[F(8,408) = 1.05; p = 0.40]. KS and AL performed more
poorly than CS on each executive task. The 2 patient groups
differed from each other on the fluency task and the n-back
paradigm, with KS being more severely impaired than AL,
but not on the inhibition, flexibility and integration tests. The
executive tasks were homogeneously impaired in the AL
group, while in the KS group, the inhibition test was less
badly affected than the fluency and n-Back tasks.
Working memory results are summarized in Table 4 (raw
scores) and Fig. 1B.
Subgroups Analyses
Duration of Abstinence. Short-term abstinent KS did not
differ from long-term abstinent KS on any of the episodic or
working memory scores.
Intelligence. KS with impaired intelligence did not differ
from KS with preserved intelligence on any of the episodic or
working memory scores.
Individual Descriptive Analyses
Mean episodic memory z-scores were calculated from the 8
impaired scores (encoding and retrieval, factual, temporal,
spatial and total recognition, R and G answers), as were mean
working memory z-scores (verbal, visuospatial and multi-
modal span tasks, fluency, Stroop, alternate response, n-Back
and integration tests).
Individual analyses of KS and AL episodic memory perfor-
mances showed only a partial overlap between AL with the
poorest performances and KS with the best ones (Fig. 2A).
Conversely, the analyses of working memory performances
painted a very mixed picture (Fig. 2B), with some KS and AL
presenting preserved working memory, others mild to moder-
ate impairments. Two KS performed noticeably worse than
the other KS and AL on the working memory tasks.

Table 4. Working Memory Performances of Alcoholic Korsakoff Patients, Non-Korsakoff Alcoholics, and Control Subjects (Raw Data)

Alcoholic Korsakoff Non-Korsakoff Post hoc

Cognitive functions patients alcoholics Control subjects comparisons
Slave systems of Phonological loop 4.43 ± 1.45 4.40 ± 1.10 5.81 ± 0.98 CS > (AL = KS)
working memory Visuospatial sketchpad 3.64 ± 1.08 3.98 ± 1.05 5.05 ± 0.78 CS > (AL = KS)
Episodic buffer 3.57 ± 1.09 3.65 ± 0.98 4.71 ± 0.76 CS > (AL = KS)
Executive functions Organization 29.14 ± 10.32 44.08 ± 12.99 53.91 ± 14.57 CS > AL > KS
Inhibitiona 31.75 ± 14.46 32.80 ± 11.12 42.81 ± 11.42 CS > (AL = KS)
Flexibilityb 71.00 ± 17.47 82.65 ± 14.84 90.80 ± 10.29 CS > (AL = KS)
Updatingb 63.39 ± 14.15 78.49 ± 16.67 90.95 ± 7.40 CS > AL > KS
Integrationc 53.93 ± 18.42 60.25 ± 16.60 71.27 ± 12.41 CS > (AL = KS)

Mean ± standard deviation. KS: alcoholic Korsakoff patients; AL: non-Korsakoff alcoholics; CS: control subjects.
a
Data from 1 KS were missing.
b
Data from 3 KS were missing.
c
Data from 2 KS were missing.
8 PITEL ET AL.

Fig. 2. Distribution of individual performances for episodic memory (A), working memory (B), and the memory task used at inclusion (C) in alcoholic Kor-
sakoff patients and non-Korsakoff alcoholics. Non-Korsakoff alcoholics are shown in gray and alcoholic Korsakoff patients in black. (A) Mean episodic mem-
ory z-scores were computed for each subject from the 8 impaired episodic memory scores (encoding and retrieval scores, factual, temporal, spatial and total
recognition, R and G answers). There was only a partial overlap between the worst-performing non-Korsakoff alcoholics and the best-performing alcoholic
Korsakoff patients. (B) Mean working memory z-scores were computed for each subject from the 8 impaired working memory scores (verbal, visuospatial
and multimodal span tasks, fluency, Stroop, alternate response, n-Back, and integration tasks). There was a total mixture of the Korsakoff and non-Korsakoff
groups, but 2 Korsakoff patients seemed to be particularly severely impaired. (C) Sum of the free recalls on the Free and Cued Selective Reminding Test
used at inclusion. There is a perfect dichotomy between the non-Korsakoff and Korsakoff groups, even though the results of the non-Korsakoff alcoholics
with the poorest performances were similar to those of the alcoholic Korsakoff patients with the best ones.
EPISODIC AND WORKING MEMORY IN KORSAKOFF PATIENTS
Lastly, the same analysis conducted on individual perfor-
mances on the inclusion test (FCSRT) showed a perfect
dichotomy between the AL and KS groups, even though the
results of the AL with the poorest performances were similar
to those of the KS with the best ones (Fig. 2C).
DISCUSSION
The objectives of the present study were firstly to gain a
better understanding of episodic and working memory
impairments in KS, secondly to compare the episodic and
working memory profiles of KS and AL, and lastly to analyze
the distribution of individual performances in the 2 patient
groups. We examined these cognitive functions with reference
to the theoretical frameworks for episodic memory and work-
ing memory proposed by Tulving (2001) and Baddeley (2000,
2003) respectively. Our results therefore reflect these concep-
tual choices.
The subgroup analyses revealed that there were no signifi-
cant differences in episodic and working memory perfor-
mances either between short-term and long-term abstinent
KS or between KS with impaired and preserved intelligence.
These results therefore suggest that the variability in the intel-
ligence and the duration of abstinence of the KS sample did
not create subgroups of amnesic patients with heterogeneous
profiles of episodic and working memory impairments.
Episodic Memory
KS performed poorly on factual recognition after sponta-
neous encoding in the Spondee test, suggesting that their
encoding abilities may be impaired (van Asselen et al., 2005;
Cermak and Butters, 1972). Without incentives to undertake
deep processing of the information, KS may have used ineffi-
cient, superficial encoding strategies (Butters and Cermak,
1980). In addition to this deficit in encoding processes, KS
may also have impaired retrieval abilities (Warrington and
Weiskrantz, 1978; d’Ydewalle and Van Damme, 2006) but
this observation has to be interpreted with caution. In effect,
it can be difficult to assess retrieval abilities when there is an
encoding impairment, as the retrieval deficit can be inter-
preted as a consequence of amnesia rather than as its cause
(Mayes et al., 1981; Meudell and Mayes, 1984). That said,
our retrieval score corresponded to free recall after deep
encoding, which is assumed to reflect retrieval ability when
encoding has been reinforced. Thus, both encoding and retrie-
val abilities were impaired in KS, the former being more
severely impaired than the latter.
In the ECM test, KS presented impaired scores on the rec-
ognition task for factual, temporal and spatial information.
KS were more severely impaired in temporal and spatial rec-
ognition than in factual recognition, thereby lending weight
to the hypothesis of a disproportionate impairment of inci-
dental contextual memory encoding in Korsakoff’s syndrome
(Kopelman et al., 1997; Mayes et al., 1992). Furthermore,
contrary to previous observations (Postma et al., 2006),
9
recognition was not impaired to the same extent for temporal
and spatial information, with temporal recognition being the
most severely impaired (Downes et al., 2002). Results of the
ECM test also confirm the impairment of autonoetic con-
sciousness in Korsakoff’s syndrome (d’Ydewalle and Van
Damme, 2006), suggesting a deficit in recollection abilities
(McKone and French, 2001) which may be ascribed to both
the encoding and retrieval impairments.
The comparison of episodic memory performances in the 2
patient groups indicated that, even if the extent of the deficits
was altogether different in KS and AL, given that KS system-
atically obtained lower scores than AL (Emsley et al., 1996),
there were similarities in their profiles of episodic memory def-
icits. In addition to the severity of the impairments, the main
distinguishing feature of episodic memory deficits in KS may
be the disproportionate impairment of encoding abilities com-
pared with retrieval ones (van Asselen et al., 2005; Cermak
and Butters, 1972).
Concerning the individual results, there was only a partial
overlap of the 2 patient groups when using the mean episodic
memory z-score, involving the AL with the worst perfor-
mances and the KS with the best ones. These findings do not,
however, undermine the validity of our patient selection, as
the same analysis, performed on the results of the traditional
FCSRT used at inclusion, showed that the 2 groups were per-
fectly separated despite the proximity of the AL with the
worst performances and those with the best ones. The com-
parison of these 2 individual analyses indicates that the
FCSRT is an appropriate task for conducting a differential
diagnostic between alcoholic Korsakoff patients and non-
Korsakoff alcoholics. The relative discrepancies between these
2 individual analyses may be due to the fact that the 2 scores
involve different memory tasks. For the FCSRT, the score
exclusively involves free recalls, whereas the mean episodic
memory z-score includes recognition tasks (regarding contex-
tual memory), which can be successfully performed using
familiarity processes that are relatively well-preserved in
amnesic patients (Huppert and Piercy, 1978; Verfaellie and
Treadwell, 1993; Yonelinas and Jacoby, 1995).
These results confirm the between-group comparisons
showing that there are similarities in the profiles of the epi-
sodic memory disorders in AL and KS. They also show that
some individual AL may resemble KS. These findings empha-
size the need for all AL subjects to undergo a neuropsycho-
logical assessment, focusing on episodic memory. Those with
equivocal episodic memory performances, similar to those of
KS, may be regarded as ‘‘borderline’’ alcoholics at risk of
developing Korsakoff’s syndrome (Fig. 3A) and should
receive particular attention and preventive action before their
memory disorders have harmful repercussions on their daily
lives.
This observation of individual episodic memory perfor-
mances suggests that we should maybe revisit the theory of
continuity between AL and KS. Episodic memory perfor-
mances may actually lie along a continuum, all the way from
preserved to mild disabilities in social drinkers (Parsons,
10 PITEL ET AL.

Fig. 3. The harmful effects of alcoholism on episodic memory (A) and working memory (B). (A) Episodic memory performances may range from pre-
served to mild disabilities in social drinkers (Parsons, 1998) to mild to moderate impairments in non-Korsakoff alcoholics and severe deficits in alcoholic Kor-
sakoff patients. SD: Social drinkers; AL: Non-Korsakoff alcoholics; BAL: ‘‘Borderline’’ non-Korsakoff alcoholics; KS: Alcoholic Korsakoff patients. (B) There
was a total mixture of Korsakoff and non-Korsakoff alcoholics for working memory. Some alcoholic Korsakoff patients with particularly poor performances
could be regarded as meeting the diagnostic criteria for alcohol dementia (Kopelman, 1995). AL: Non-Korsakoff alcoholics. KS: Alcoholic Korsakoff patients.
AAD: Alcoholics with alcohol dementia.

1998) to mild to moderate impairments in non-Korsakoff
alcoholics and severe deficits in alcoholic Korsakoff patients
(Fig. 3A). However, the main contributing factor may not be
drinking history (Ihara et al., 2000; Noel et al., 2001; Schaef-
fer and Parsons, 1987), as originally suggested more than
thirty years ago (Ryback, 1971). Rather, the heterogeneity of
the effects of episodic memory performances in KS and AL
may be due to other factors, such as associated thiamine defi-
ciency (Thomson and Marshall, 2006) or genetically and ⁄ or
experiential vulnerability to the neurotoxic effects of alcohol
(Parsons, 1998).
In short, KS had episodic memory impairments affecting
the different components of this memory system: encoding,
retrieval, contextual memory, and recollection. Whereas epi-
sodic memory was obviously quantitatively more severely
impaired in KS than in AL, the 2 groups had qualitatively
similar profiles of deficits, the single specificity in the KS being
the disproportionate impairment of encoding abilities com-
pared with retrieval ones.
Working Memory
Regarding the slave systems of working memory, our data
confirm a deficit in the phonological loop and the visuospatial
sketchpad in KS (Brand et al., 2003). The present study also
provides arguments in favor of an impairment of the episodic
EPISODIC AND WORKING MEMORY IN KORSAKOFF PATIENTS
buffer in KS. This deficit of the slave systems of working
memory was uniform in the KS and the AL groups, as neither
the phonological loop, the visuospatial sketchpad nor the epi-
sodic buffer seemed to be disproportionately affected. More-
over, the slave systems were impaired to the same extent in
the AL and KS, suggesting that these deficits may not be spe-
cific to KS.
It is also worth noting that some of the executive tasks
were too difficult for KS with the most severe dysexecutive
syndrome and that 2 KS exhibited individual performances
which were noticeably inferior to those of AL and other
KS. Korsakoff’s syndrome did not occur insidiously in
these 2 patients and they were neither the oldest nor those
with the shortest period of abstinence. The refined clinical
criteria for alcohol-related dementia (Oslin et al., 1998)
stress that its diagnostic pattern may encompass Korsak-
off’s syndrome, which can be regarded as a ‘‘specific dis-
ease manifesting itself in alcohol-related dementia.’’ Other
authors have suggested that there are neuropsychological
and neuroimaging arguments in favor of a continuum
between Korsakoff’s syndrome and alcohol dementia (But-
ters and Brandt, 1985; Emsley et al., 1996) rather than
sharp differences, confirming that a differential diagnosis
may be difficult to establish. Thus, KS with the most
severe working memory impairments may be regarded as
presenting alcohol dementia (Butters and Brandt, 1985;
Fig. 3B). The goal of the present study was not to provide
a definitive answer but rather to prompt further research
into the nosological distinction between Korsakoff’s
syndrome and alcohol dementia (Bowden, 1990).
Statistical analyses of the available executive data have
confirmed impairments of organization, inhibition, flexibil-
ity and updating abilities in KS (Brokate et al., 2003; Hil-
debrandt et al., 2004; Jacobson et al., 1990; Oscar-Berman
et al., 2004) and suggest an additional deficit in integration
abilities. Unlike AL, KS did not exhibit homogenous exec-
utive dysfunctions, as organization and updating abilities
were more severely impaired than inhibition ones. Organi-
zation (Hildebrandt et al., 2004; Jacobson et al., 1990;
Joyce and Robbins, 1991; Oscar-Berman et al., 2004) and
updating abilities (Brokate et al., 2003; Hildebrandt et al.,
2004) were more severely impaired in KS than in AL but
the 2 groups did not differ in terms of inhibition, flexibil-
ity, and integration abilities. Our findings therefore suggest
that some of the executive dysfunctions may be shared by
AL and KS. The individual results concerning working
memory confirmed that there was a total mixture of the 2
patient groups, with patients’ performances ranging from
normal to severely impaired.
In short, KS had an overall working memory disorder
affecting both the slave systems and the central executive.
Working memory deficits may not constitute relevant diag-
nostic criteria for Korsakoff’s syndrome, as only the sever-
ity of the organization and updating deficits distinguished
KS from AL. These results do not confirm the assumption
that executive dysfunction is ‘‘a particular property of the
11
Korsakoff state’’ (Joyce and Robbins, 1991) but rather
support the suggestion that at least some of the executive
impairments in KS may be due to another pathological
mechanism, notably the harmful effects of chronic alcohol-
ism on the frontal lobes (Brokate et al., 2003; Kopelman
and Stanhope, 1998).
CONCLUSION
The present study reinvestigated episodic memory and
working memory disorders in KS in the light of recent neuro-
psychological conceptions. Results showed firstly that the dif-
ferent components of these memory systems were impaired in
KS and secondly, that most of these neuropsychological defi-
cits were already present in AL, either in an equivalent form
or a less severe one. Regarding episodic memory, alcoholic
Korsakoff patients may therefore be considered as non-Kor-
sakoff alcoholics, whose existing deficits due to the neurotoxic
effects of ethanol on the Papez circuit have been exacerbated
by the thiamine deficiency, leading to the amnesic syndrome.
Working memory impairments do not seem to be specific to
Korsakoff’s syndrome and may simply reflect the effects of
chronic alcohol consumption on frontocerebellar circuitry
(Sullivan et al., 2003). These findings lend weight to a revised
continuity theory, whereby there is an ‘‘episodic memory con-
tinuity’’ between AL and SK and a ‘‘working memory conti-
nuity’’ among AL, KS, and alcohol dementia patients.
Further investigations are now needed to establish whether
these putative forms of neuropsychological continuity are
mirrored by a neuroanatomical continuity, given that neuroi-
maging and neuropathological investigations have revealed
brain damage in the hippocampi (Sullivan and Marsh, 2003;
Sullivan et al., 1995), thalami (Chanraud et al., 2007; Visser
et al., 1999), mamillary bodies (Harding et al., 2000; Jauhar
and Montaldi, 2000) and frontal lobes (Cardenas et al., 2007;
Reed et al., 2003) in alcoholics both with and without major
neurological complications.
ACKNOWLEDGMENTS
This study was funded by ATC Alcool (Inserm). We would
like to thank Elizabeth Portier for revising the English style.
REFERENCES
Ambrose ML, Bowden SC, Whelan G (2001) Working memory impairments
in alcohol-dependent participants without clinical amnesia. Alcohol Clin
Exp Res 25:185–191.
American Psychiatric Association (1994) Diagnostic and Statistical Manual of
Mental Disorders. American Psychiatric Association, Washington.
van Asselen M, Kessels RP, Wester AJ, Postma A (2005) Spatial working
memory and contextual cueing in patients with Korsakoff amnesia. J Clin
Exp Neuropsychol 27:645–655.
Baddeley A (1986) Working Memory. Oxford University Press, Oxford.
Baddeley A. (1996) The fractionation of working memory. Proc Natl Acad Sci
U.S.A. 93:13468–13472.
Baddeley A (2000) The episodic buffer: a new component of working mem-
ory? Trends Cogn Sci 4:417–423.
12
Baddeley A (2003) Working memory: looking back and looking forward. Nat
Rev Neurosci 4:829–839.
Bowden SC. (1990) Separating cognitive impairment in neurologically asymp-
tomatic alcoholism from Wernicke-Korsakoff syndrome: is the neuropsy-
chological distinction justified? Psychol Bull 107:355–366.
Brand M, Fujiwara E, Borsutzky S, Kalbe E, Kessler J, Markowitsch HJ
(2005) Decision-making deficits of Korsakoff patients in a new gambling
task with explicit rules: associations with executive functions. Neuropsychol-
ogy 19:267–277.
Brand M, Fujiwara E, Kalbe E, Steingass HP, Kessler J, Markowitsch HJ
(2003) Cognitive estimation and affective judgments in alcoholic Korsakoff
patients. J Clin Exp Neuropsychol 25:324–334.
Brokate B, Hildebrandt H, Eling P, Fichtner H, Runge K, Timm C (2003)
Frontal lobe dysfunctions in Korsakoff’s syndrome and chronic alcoholism:
continuity or discontinuity? Neuropsychology 17:420–428.
Buschke H, Sliwinski M, Kuslansky G, Lipton RB (1995) Aging, encoding
specificity, and memory change in the Double Memory Test. J Int Neuro-
psychol Soc 1:483–493.
Butters N, Brandt J (1985) The continuity hypothesis: the relationship of long-
term alcoholism to the Wernicke-Korsakoff syndrome. Recent Dev Alcohol
3:207–226.
Butters N, Cermak LS (1980) Alcoholic Korsakoff’s Syndrome: An Informa-
tion-Processing Approach to Amnesia. Academic Press, New York.
Cardebat D, Doyon B, Puel M, Goulet P, Joanette Y (1990) Formal and
semantic lexical evocation in normal subjects. Performance and dynamics of
production as a function of sex, age and educational level. Acta Neurol Belg
90:207–217.
Cardenas VA, Studholme C, Gazdzinski S, Durazzo TC, Meyerhoff DJ
(2007) Deformation-based morphometry of brain changes in alcohol depen-
dence and abstinence. Neuroimage 34:879–887.
Cermak LS, Butters N (1972) The role of interference and encoding in the short-
term memory deficits of Korsakoff patients. Neuropsychologia 10:89–95.
Chalfonte BL, Verfaellie M, Johnson MK, Reiss L (1996) Spatial location
memory in amnesia: binding item and location information under incidental
and intentional encoding conditions. Memory 4:591–614.
Chanraud S, Martelli C, Delain F, Kostogianni N, Douaud G, Aubin HJ,
Reynaud M, Martinot JL (2007) Brain morphometry and cognitive perfor-
mance in detoxified alcohol-dependents with preserved psychosocial func-
tioning. Neuropsychopharmacology 32:429–438.
Chételat G, Desgranges B, de la Sayette V, Viader F, Berkouk K, Landeau B,
Lalevee C, Le Doze F, Dupuy B, Hannequin D, Baron JC, Eustache F
(2003) Dissociating atrophy and hypometabolism impact on episodic mem-
ory in mild cognitive impairment. Brain 126:1955–1967.
Cushman P, Forbes R, Lerner W, Stewart M (1985) Alcohol withdrawal syn-
dromes: clinical management with lofexidine. Alcohol Clin Exp Res 9:103–
108.
Cutting J (1978) The relationship between Korsakoff’s syndrome and ‘alco-
holic dementia’. Br J Psychiatry 132:240–251.
Downes JJ, Mayes AR, MacDonald C, Hunkin NM (2002) Temporal order
memory in patients with Korsakoff’s syndrome and medial temporal amne-
sia. Neuropsychologia 40:853–861.
Emsley R, Smith R, Roberts M, Kapnias S, Pieters H, Maritz S (1996) Mag-
netic resonance imaging in alcoholic Korsakoff’s syndrome: evidence for an
association with alcoholic dementia. Alcohol Alcohol 31:479–486.
Eustache F, Desgranges B (2008) MNESIS: towards the integration of current
multisystem models of memory. Neuropsychol Rev, Feb 29 2008, Epub
ahead of print
Fein G, Torres J, Price LJ, Di Sclafani V (2006) Cognitive performance in
long-term abstinent alcoholic individuals. Alcohol Clin Exp Res 30:1538–
1544.
Folstein MF, Folstein SE, McHugh PR (1975) ‘‘Mini-mental state’’. A practi-
cal method for grading the cognitive state of patients for the clinician. J Psy-
chiatr Res 12:189–198.
Gardiner JM, Ramponi C, Richardson-Klavehn A (2002) Recognition mem-
ory and decision processes: a meta-analysis of remember, know, and guess
responses. Memory 10:83–98.
PITEL ET AL.
Glenn SW, Parsons OA (1992) Neuropsychological efficiency measures in
male and female alcoholics. J Stud Alcohol 53:546–552.
Grober E, Buschke H (1987) Genuine memory deficits in dementia. Dev Neu-
ropsychol 3:13–36.
Grober E, Buschke H, Crystal H, Bang S, Dresner R (1988) Screening for
dementia by memory testing. Neurology 38:900–903.
Harding A, Halliday G, Caine D, Kril J (2000) Degeneration of anterior tha-
lamic nuclei differentiates alcoholics with amnesia. Brain 123:141–154.
Hildebrandt H, Brokate B, Eling P, Lanz M (2004) Response shifting and
inhibition, but not working memory, are impaired after long-term heavy
alcohol consumption. Neuropsychology 18:203–211.
Huppert FA, Piercy M (1978) Dissociation between learning and remembering
in organic amnesia. Nature 275:317–318.
Ihara H, Berrios GE, London M (2000) Group and case study of the dysexec-
utive syndrome in alcoholism without amnesia. J Neurol Neurosurg Psychi-
atry 68:731–737.
Irle E, Kaiser P, Naumann-Stoll G (1990) Differential patterns of memory loss
in patients with Alzheimer’s disease and Korsakoff’s disease. Int J Neurosci
52:67–77.
Jacobson RR, Acker CF, Lishman W. (1990) Patterns of neuropsychological
deficit in alcoholic Korsakoff’s syndrome. Psychol Med 20:321–334.
Jauhar P, Montaldi D (2000) Wernicke-Korsakoff syndrome and the use of
brain imaging. Alcohol Alcohol Suppl 35:21–23.
Joyce EM, Robbins TW (1991) Frontal lobe function in Korsakoff and non-
Korsakoff alcoholics: planning and spatial working memory. Neuropsycho-
logia 29:709–723.
Kopelman MD (1985) Rates of forgetting in Alzheimer-type dementia and
Korsakoff’s syndrome. Neuropsychologia 23:623–638.
Kopelman MD (1989) Remote and autobiographical memory, temporal con-
text memory and frontal atrophy in Korsakoff and Alzheimer patients.
Neuropsychologia 27:437–460.
Kopelman MD (1991) Frontal dysfunction and memory deficits in the alco-
holic Korsakoff syndrome and Alzheimer-type dementia. Brain 114(Pt
1A):117–137.
Kopelman MD (1995) The Korsakoff syndrome. Br J Psychiatry 166:154–
173.
Kopelman MD (2002) Disorders of memory. Brain 125:2152–2190.
Kopelman MD, Stanhope N (1998) Recall and recognition memory in
patients with focal frontal, temporal lobe and diencephalic lesions. Neuro-
psychologia 36:785–795.
Kopelman MD, Stanhope N, Kingsley D (1997) Temporal and spatial context
memory in patients with focal frontal, temporal lobe, and diencephalic
lesions. Neuropsychologia 35:1533–1545.
Korsakoff SS (1889) Etude médico-psychologique sur une forme des maladies
de la mémoire. Revue philosophique de la France et de l’étranger 14:401–
530.
Krabbendam L, Visser PJ, Derix MM, Verhey F, Hofman P, Verhoeven W,
Tuinier S, Jolles J (2000) Normal cognitive performance in patients with
chronic alcoholism in contrast to patients with Korsakoff’s syndrome.
J Neuropsychiatry Clin Neurosci 12:44–50.
Mayes AR, MacDonald C, Donlan L, Pears J, Meudell PR (1992) Amnesics
have a disproportionately severe memory deficit for interactive context. Q J
Exp Psychol A 45:265–297.
Mayes AR, Meudell PR, MacDonald C (1991) Disproportionate intentional
spatial-memory impairments in amnesia. Neuropsychologia 29:771–784.
Mayes AR, Meudell PR, Pickering A (1985) Is organic amnesia caused by
a selective deficit in remembering contextual information? Cortex 21:167–
202.
Mayes AR, Meudell PR, Som S (1981) Further similarities between amnesia
and normal attenuated memory: effects with paired-associate learning and
contextual shifts. Neuropsychologia 19:655–664.
McDowall J (1981) Effects of encoding instructions on recall and recognition
in Korsakoff patients. Neuropsychologia 19:43–48.
McKone E, French B (2001) In what sense is implicit memory ‘‘episodic’’?
The effect of reinstating environmental context. Psychon Bull Rev 8:806–
811.
EPISODIC AND WORKING MEMORY IN KORSAKOFF PATIENTS
Meudell PR, Mayes AR (1984) Patterns of confidence loss in the cued recall of
normal people with attenuated recognition memory: their relevance to a
similar amnesic phenomenon. Neuropsychologia 22:41–54.
Noel X, Van der Linden M., Schmidt N, Sferrazza R, Hanak C, Le Bon O,
De Mol J, Kornreich C, Pelc I, Verbanck P (2001) Supervisory attentional
system in nonamnesic alcoholic men. Arch Gen Psychiatry 58:1152–1158.
Oscar-Berman M, Kirkley SM, Gansler DA, Couture A (2004) Comparisons
of Korsakoff and non-Korsakoff alcoholics on neuropsychological tests of
prefrontal brain functioning. Alcohol Clin Exp Res 28:667–675.
Oslin D, Atkinson RM, Smith DM, Hendrie H (1998) Alcohol related demen-
tia: proposed clinical criteria. Int J Geriatr Psychiatry 13:203–212.
Parsons OA (1998) Neurocognitive deficits in alcoholics and social drinkers: a
continuum? Alcohol Clin Exp Res 22:954–961.
Parsons OA, Nixon SJ (1993) Neurobehavioral sequelae of alcoholism. Neurol
Clin 11:205–218.
Pitel AL, Beaunieux H, Witkowski T, Vabret F, Guillery-Girard B, Quinette
P, Desgranges B, Eustache F (2007a) Genuine episodic memory deficits and
executive dysfunctions in alcoholic subjects early in abstinence. Alcohol Clin
Exp Res 31:1169–1178.
Pitel AL, Witkowski T, Vabret F, Guillery-Girard B, Desgranges B, Eustache
F, Beaunieux H (2007b) Effect of episodic and working memory impair-
ments on semantic and cognitive procedural learning at alcohol treatment
entry. Alcohol Clin Exp Res 31:238–248.
Pollux PM, Wester A, De Haan EH (1995) Random generation deficit in alco-
holic Korsakoff patients. Neuropsychologia 33:125–129.
Postma A, Van AM, Keuper O, Wester AJ, Kessels RP (2006) Spatial and
temporal order memory in Korsakoff patients. J Int Neuropsychol Soc
12:327–336.
Quinette P, Guillery B, Desgranges B, de la Sayette V, Viader F, Eustache F
(2003) Working memory and executive functions in transient global amne-
sia. Brain 126:1917–1934.
Quinette P, Guillery-Girard B, Noel A, de la Sayette V, Viader F, Desgranges
B, Eustache F (2006) The relationship between working memory and epi-
sodic memory disorders in transient global amnesia. Neuropsychologia
11:1640–1658.
Reed LJ, Lasserson D, Marsden P, Stanhope N, Stevens T, Bello F, Kingsley
D, Colchester A, Kopelman MD (2003) FDG-PET findings in the Wer-
nicke-Korsakoff syndrome. Cortex 39:1027–1045.
Rosenbloom M, Sullivan EV, Pfefferbaum A (2003) Using magnetic reso-
nance imaging and diffusion tensor imaging to assess brain damage in alco-
holics. Alcohol Res Health 27:146–152.
Rourke SB, Grant I (1999) The interactive effects of age and length of
abstinence on the recovery of neuropsychological functioning in chronic
male alcoholics: a 2-year follow-up study. J Int Neuropsychol Soc
5:234–246.
Ryback RS (1971) The continuum and specificity of the effects of alcohol on
memory. A review. Q J Stud Alcohol 32:995–1016.
13
Schaeffer KW, Parsons OA (1987) Learning impairment in alcoholics using an
ecologically relevant test. J Nerv Ment Dis 175:213–218.
Shoqeirat MA, Mayes AR (1991) Disproportionate incidental spatial-memory
and recall deficits in amnesia. Neuropsychologia 29:749–769.
Stroop J (1935) Studies of interference in serial verbal reactions. J Exp Psychol
18:643–662.
Sullivan EV, Harding AJ, Pentney R, Dlugos C, Martin PR, Parks MH, Des-
mond JE, Chen SH, Pryor MR, De Rosa E, Pfefferbaum A. (2003) Disrup-
tion of frontocerebellar circuitry and function in alcoholism. Alcohol Clin
Exp Res 27:301–309.
Sullivan EV, Marsh L (2003) Hippocampal volume deficits in alcoholic Kor-
sakoff’s syndrome. Neurology 61:1716–1719.
Sullivan EV, Marsh L, Mathalon DH, Lim KO, Pfefferbaum A (1995) Ante-
rior hippocampal volume deficits in nonamnesic, aging chronic alcoholics.
Alcohol Clin Exp Res 19:110–122.
Sullivan EV, Rosenbloom MJ, Pfefferbaum A (2000) Pattern of motor and
cognitive deficits in detoxified alcoholic men. Alcohol Clin Exp Res 24:611–
621.
Thomson AD, Marshall EJ (2006) The natural history and pathophysiology
of Wernicke’s encephalopathy and Korsakoff’s psychosis. Alcohol Alcohol
41:151–158.
Tivis R, Beatty WW, Nixon SJ, Parsons OA (1995) Patterns of cognitive
impairment among alcoholics: are there subtypes? Alcohol Clin Exp Res
19:496–500.
Tulving E (2001) Episodic memory and common sense: how far apart? Philos
Trans R Soc Lond B Biol Sci 356:1505–1515.
Tulving E (2002) Episodic memory: from mind to brain. Annu Rev Psychol
53:1–25.
Verfaellie M, Treadwell JR (1993) Status of recognition memory in amnesia.
Neuropsychology 8:292.
Visser PJ, Krabbendam L, Verhey FR, Hofman PA, Verhoeven WM, Tuinier
S, Wester A, Den Berg YW, Goessens LF, Werf YD, Jolles J (1999) Brain
correlates of memory dysfunction in alcoholic Korsakoff’s syndrome.
J Neurol Neurosurg Psychiatry 67:774–778.
Warrington EK, Weiskrantz L (1978) Further analysis of the prior learning
effect in amnesic patients. Neuropsychologia 16:169–177.
Weschler D (2001) Weschler Adult Intelligence Scale. EAP, Paris.
Wheeler MA, Stuss DT, Tulving E (1997) Toward a theory of episodic mem-
ory: the frontal lobes and autonoetic consciousness. Psychol Bull 121:331–
354.
d’Ydewalle G, Van Damme I (2006) Memory and the Korsakoff syndrome:
not remembering what is remembered. Neuropsychologia 45:905–920.
Yonelinas AP, Jacoby LL (1995) The relation between remembering and
knowing as bases for recognition: effects of size congruency. Journal of
Memory and Language 34:622–643.
Zimmermann P, Fimm B (1993) Testbatterie zur Erfassung von
Aufmerksamkeitsstorungen. Psytest, Freiburg.