18616ournal of NeurologAy, Neurosurgery, and Psychiatry 1996;61:186-193

Memory without context: amnesia with

confabulations after infarction of the right
capsular genu
Armin Schnider, Klemens Gutbrod, Christian W Hess, Gerhard Schroth

Department of
Neurology
A Schnider
K Gutbrod
C W Hess
Division of
Neuropsychological
Rehabilitation
A Schnider
K Gutbrod
Division of
Neuroradiology,
University Hospital,
Bern, Switzerland
G Schroth
Correspondence to:
Neurologische
Universitatsklinik,
Inselspital, CH-3010 Bern,
Dr med Armin Schnider,

Switzerland.
Received 17 January 1996
and in revised form
29 April 1996
Accepted 3 May 1996

ment in the conscious, deliberate recall of
information".' However, various behavioural
patterns indicate that the "amnestic syndrome" is not a single entity. Some patients
are much more impaired in freely recalling
than in recognising previously presented information, deny any failure of their memory, and
sometimes confabulate spontaneously and in
memory tests. This type of amnesia has been
documented in patients with orbitofrontal and
basal forebrain damage after rupture of an
aneurysm of the anterior communicating
artery,2-7 in some patients with paramedian
thalamic infarction,89 and in the early course
of Wernicke-Korsakoff syndrome.'0 By contrast, some patients also fail to recognise and
apparently feel entirely unfamiliar with previously presented information (indicating deficient information storage), are very concerned
about the failure of their memory, and do not
spontaneously confabulate. Distinct examples
of this type of memory failure were seen in
patients with severe amnesia after focal medial
temporal lobe damage. "-13
Patients with orbitofrontal damage,'4 1'
Wernicke-Korsakoff syndrome,'617 and thalamic lesions'8 have also been shown to fail to
store the temporal order and the context of
information acquisition, a disorder which-in
subtler form-was also shown in patients with
dorsal prefrontal lesions sparing the orbitofrontal area.'920 Unfortunately, most reports
did not provide any clinical observations as to
the relevance of the disorder (behaviour in
everyday life, insight, confabulations, etc).
Although a loss of the temporal tag in memory
was suggested by Van der Horst2' a long time
ago as a possible mechanism of spontaneous
confabulations, no data supporting this link
have been presented so far.
We describe a patient who had severe
amnesia after an infarct involving the right
inferior capsular genu and anterior thalamic
nucleus. Amnesia after capsular genu lesions
has been described but the reports either failed
to further characterise the type of memory failure or to describe its impact on the patients'
lives.2223 Our patient failed most recognition
tasks because she felt falsely familiar with most
distracter items. She had no insight into the
failure of her memory and confabulated spon(J Neurol Neurosurg Psychiatry 1996;61:186-193)
taneously and in response to questions, but
confabulations seemed to be composed of eleKeywords: amnesia; confabulation; capsular genu; ments of true events. In a series of experiments
recognition; context memory
we tested the hypothesis that this memory disorder resulted from an inability to store the
Anterograde amnesia can be defined as a "pro- temporal and spatial context of information
found defect in new learning [with] impair- acquisition and was independent of the capac-

Abstract
Objective-To explore the mechanism of
an amnesia marked by confabulations
and lack of insight in a patient with an
infarct of the right inferior capsular genu.
The confabulations could mostly be
traced back to earlier events, indicating
that the memory disorder ensued from an
inability to store the temporal and spatial
context of information acquisition rather
than a failure to store new information.
Methods-To test the patient's ability to
store the context of information acquisition, two experiments were composed in
which she was asked to decide when or
where she had learned the words from
two word lists presented at different
points in time or in different rooms. To
test her ability to store new information,
two continuous recognition tests with
novel non-words and nonsense designs
were used. Recognition of these stimuli
was assumed to be independent of the
context of acquisition because the patient
could not have an a priori sense of familiarity with them.
Results-The patient performed at
chance in the experiments probing knowledge of the context of information acquisition, although she recognised the
presented words almost as well as the
controls. By contrast, her performance
was normal in the recognition tests with
non-words and nonsense designs.
Conclusion-These findings indicate that
the patient's amnesia was based on an
inability to store the context of information acquisition rather than the information itself. Based on an analysis of her
lesion, which disconnected the thalamus
from the orbitofrontal cortex and the
amygdala, and considering the similarities between her disorder, WernickeKorsakoff syndrome, and the amnesia
after orbitofrontal lesions, it is proposed
that contextual amnesia results from
interruption of the loop connecting the
amygdala, the dorsomedial nucleus, and
the orbitofrontal cortex.

Memory without context: amnesia with confabulations after infarction of the right capsular genu

ity to learn new information. The patient's
lesion interrupted the connections between
the thalamus and the prefrontal (including
orbitofrontal) cortex. We propose an anatomical model for temporal order amnesia and
confabulations, which integrates the evidence
from our patient and reports of patients with
Wernicke-Korsakoff syndrome and orbitofrontal damage.

187

remote personal and public events back to
childhood years but confabulated on events
occurring after the stroke. She seemed to
recognise the personnel on the ward but often
confused their names or confabulated on the
circumstances of their meeting. For example,
when asked whether she recognised one of the
examiners, she explained: "You are Dr S (correct). We did physiotherapy this morning (she
had had physiotherapy, but not with Dr S),
then we worked on the computer (she had regular computer training, but not that morning
and never with Dr S) where you showed me a
camel and other animals composed of dots
(referring to a perceptual priming test taken
two weeks previously)". Thus her confabulations appeared to consist of freely arranged
elements of previous actual events.
She was discharged to her own home three
months after the stroke. She needed regular
assistance from her daughter because she was
unable to organise her own household. She
would forget to prepare the meals and would
confuse appointments and events. For example, she retained a sample of urine every morn-

Case report
A 62 year old housewife with an 11 year
school education was admitted to the
Department of Neurology for confusion and
memory disturbances insidiously evolving over
five days. She had a four year history of mild
hypertension and diabetes mellitus, which had
been well controlled with atenolol, diltiazem,
and glibenclamid. At admission, she did not
know why she was at the hospital and denied
any memory problem. Physical examination
was normal except for high blood pressure
(190/95). Blood glucose was moderately raised
(9-8 mM) and lipid electrophoresis showed a
decreased high density lipoprotein fraction. ing for several days stating that she had an
Cardiac examination was normal. Carotid appointment with her family doctor (she had
Doppler ultrasound showed discrete plaques actually had an appointment some days previat the carotid bifurcation. Brain CT showed ously). Convinced that she had a good meman isolated hypodense lesion involving the ory, she never used a shopping list and bought
genu of the internal capsule and possibly the the same items several times, thus amassing
anterior thalamus on the right side. A diagnosis for example, a considerable stock of toilet
of right capsular genu infarction, presumably paper which she thought she had run out of.
due to diabetes and hypertension, was made, Her daughter thought that the patient's memand the patient was put on aspirin.
ory disturbance extended two to three years
She remained severely amnestic. She never backwards. The patient was repeatedly astonacknowledged any memory disturbance and ished at the presence of "new" buildings in her
maintained that she was in the hospital for her town that had been constructed during the
diabetes. She gave precise accounts of past previous two to three years or at the decease of
friends who had actually died in the same

period.
Table I Memory assessment
NEUROPSYCHOLOGICAL EVALUATIONS

Time after onset
3 months

7 months

Task/Reference

Score

Percentile

Score

Percentile

Digit span forward65
Corsi block tapping66
Wechsler memory scale65:
Logical memory
Immediate
Delay
Word associating learning:
Immediate
Delayed
California verbal learning test26:

5
6

20
80

5
6

20
80

3
1

< 1
< 1

7
2

<1
<1

10
2

< 1
< 1

11
3

<1

2
4
2
5
12
4
41

< 1
< 1
< 1
< 1
< 1

45
< 1

3
9
4
4
12
0
65

84
<1

3-5

< 1

9

20

1
5
2
12
6

< 1
< 1
< 1
< 1

2
2
2
7
4

< 1
< 1
< 1
< 1
< 1

21
19
14

-

21
21
17

9
9
4

-

9
9
3

Trial 1
Trial 5
Long delay free recall
Long delay cued recall
Recognition
Perseverative errors
Intrusions
Rey-Osterrieth complex figure test:

Delayed Recall67:
Rey visual design learning test68:
Trial I
Trial 5
Delayed Recall

Recognition
Recognition, false positives (out of 15)
Autobiographical memory interview30:
Personal semantic memory
Childhood (maximum 21)
Early adult life (maximum 21)
Recent facts (maximum 21)
Autobiographical
Childhood (maximum 9)
Early adult life (maximum 9)
Recent facts (maximum 9)

8

5
<1
<1
< 1
< 1
<1

-

Neuropsychological evaluations were performed three and seven months after the
stroke with essentially similar results. The
patient had low average intelligence (WAISR24 VIQ 92; PIQ 90; FIQ 90). Language and
associated functions as well as visuoconstructive and visuospatial capacities, phasic alertness, cognitive speed, and interference control
were normal. Whereas verbal fluency25 was
normal, design fluency26 was slightly
decreased. Concept identification27 was
impaired, primarily due to many perseverative
errors.
MEMORY TESTING

Formal tests disclosed a severe verbal and
non-verbal anterograde amnesia (table 1).
Unlike "typical" amnestic patients, however,
she repeatedly surprised the examiners with
details from previous tests. What appeared to
be confabulations were-with virtually no
exception-intrusions of material presented in
previous tests or semantically related material.
For example, she produced even more "intrusion errors" in the California verbal learning
test (CVLT)28 at the follow up examination
after seven months, when form B was used,

188

Schnider, Gutbrod, Hess, Schroth

than after three months; almost all of these task was composed with the items of two temadditional errors were intrusions from form A, poral order and spatial context memory tasks
which she had learned four months previously. (described below). Each version of the recogMotor learning in a mirror drawing task29 was nition tasks was also given to two age matched
normal. Retrograde memory was intact except controls.
for the recent past: on the questionnaire of
Kopelman et a130 she achieved perfect scores TEMPORAL ORDER JUDGEMENT (EXPERIMENT 2)
for the first two periods (childhood, adoles- To test the ability to time label newly acquired
cence, table 1). Only for the latest period information, two semantically matched lists
(recent old memory), were her scores deficient with 10 common concrete words were comand she confabulated on several items. In a posed. Both lists were learned in the same
famous faces recognition test she was shown manner one hour apart. All 10 words from a
pictures of the Swiss Bundesrate (members of list were orally presented and the patient was
the government) of the past five decades and requested to recall all words aloud. To ensure
asked to name them. Whereas her perfor- learning of all words, the order of presentation
mance was similar to age matched controls for was rotated by two words from trial to trial.
previous decades, she failed to name the Learning trials were repeated until all words
Bundesrate presently in office (most of them were recalled in at least three trials. One hour
have been in office for several years); her per- after the last learning trial of list 2, free recall
formance was more than 1 SD below the of all words from both lists was tested. Then, a
mean of five age matched controls. However, list of printed word pairs with one word from
she correctly recognised their names from list 1 and one word from list 2 was presented.
All words from list 1 were paired with all
three choices.
words from list 2, once with the word from list
1 on the left, once with the word from list 2 on
the left (2 x 10 x 10 words = 200 word
Methods
Because the patient's confabulations and pairs). She was requested to indicate for each
intrusions in memory tests could mostly be word pair which word had appeared in the first
traced back to earlier events or tests, we word list. One hour later, recognition of all
hypothesised that her amnesia was based on words was tested by presenting each word with
an inability to store the temporal order and three semantic foils. This experiment was perspatial context of information acquisition formed three months after the stroke. Because
rather than the information itself. Thus she this task seemed very easy and was expected to
would store new information but later fail to yield a ceiling effect, only two age matched
recall or recognise it because she would be control subjects were tested.
ignorant as to what stored information pertained to a particular situation. We composed SPATIAL AND PERSONAL CONTEXT
experiments to explore whether she did indeed (EXPERIMENT 3)
store new information and whether she had Even if the patient failed to time label stimuli,
knowledge about the context of information recognition would still be expected to be better if she recalled the context (place, person) of
acquisition.
stimulus presentation. For example, recognition of priming task items should succeed
FREE VERSUS CONSTRAINED RECOGNITION
based on knowledge about the stimulus con(EXPERIMENT 1)
The patient achieved bad scores in recognition text (the priming task), independent of the
tasks due to many false positives. This experi- time of presentation. Two aspects of context
ment explored whether her spontaneous memory were tested: attribution of a stimulus
behaviour in recognition tests indeed reflected to the place where it had been presented (spavery poor knowledge about target items or tial context) and to the person who had prewhether she had better (implicit) knowledge sented it (personal context). The same word
about the target items than suggested by her lists as in experiment 2 (performed four
spontaneous behaviour. We composed a months previously) were used. The word lists
recognition test with 16 line drawings from were presented by two examiners in their disSnodgrass and Vanderwart3' which had been tinctly different offices at the same time: after
used in a perceptual priming task (test not every second learning trial of a list, the patient
detailed here; her performance was normal). was brought to the other examiner's office for
Each target item was presented with three two learning trials of the other word list until
semantic foils (original drawings vertically criterion was reached for both word lists. After
arranged and printed on 21 x 30 cm white one hour, recall was tested. Then 200 word
sheets of paper). In the first run, she was asked pairs composed in the same way as in experito indicate the items that had been presented ment 2 were presented by a neutral examiner
in the "test involving the drawings with dashed in one of the examiners' offices and the patient
lines" (alluding to the masking of the drawings was requested to indicate for each word pair
in the priming task) with no mention that only which word had been presented in this room
one item per series was correct (free recogni- ("Which of these two words have you learned
tion). Then a second run was made, in which in this room today?"). Thirty minutes later, she
she was allowed to select only one item per was again presented with the 200 word pairs
page (constrained recognition). This test was by one of the examiners in a neutral room and
applied two days, one week, and one month requested to indicate "which of these two
after the priming test. A similar recognition words have you learned from me today?". This

Memory without context: amnesia with confabulations after infarction of the right capsular genu

experiment was performed seven months after
the stroke. As in experiment 1, because this
task seemed very easy and was expected to
yield a ceiling effect, we tested only two age
matched controls.
NEW LEARNING (EXPERIMENT 4)
If the patient's amnesia was based solely on

failed contextual labelling of information, she
should be able to learn new information and to
retrieve it if the time and context of presentation could not be confused and would be irrelevant for recognition. This hypothesis could
not be tested with meaningful stimuli because
their recollection might be particularly dependent on contextual knowledge. We used the
non-verbal learning test of Sturmn and Willmes
(NVLT)32 with nonsense designs and a verbal
learning test (VLT) with non-words. Both
tests consisted of 160 cards, each containing
one stimulus. All cards were shown for three
seconds, and the patient was asked to indicate
for each stimulus whether she had seen it
before. Both tests actually comprised eight
series of 20 randomly ordered cards, each
series containing eight recurring and 12 novel
stimuli. The total number of recurrences was
56. The average interval between subsequent
presentations of a stimulus was 60 seconds.
The score was calculated as total number of
correct recognitions minus total number of
false positives. The normal values provided by
Sturm and Willmes32 are based on a healthy
control population of n = 400 for the NVLT
and n = 314 for the VLT. Normal means
(SD) were: NVLT 32-9 (10 1); VLT 37-7
(10-7).
Results
EXPERIMENT 1: FREE VERSUS CONSTRAINED
RECOGNITION

By contrast with the controls, the patient's free
recognition was very unreliable: she felt familiar
with all items that had actually been presented
(perceptual priming items: 16 of 16 possible
true positives; temporal order judgement and
context memory tests: 20 of 20), but also with
many or most distracter items (perceptual
priming items: six of 48 possible false positives
after two days, 36 after one week, 39 after one
month; temporal order judgement and context
memory tests: 23 and 25 of 60 possible false

Table 2 Context memory
Patient

Temporal order judgment (experiment 2):
Free recall (maximum 20)
8
Correct temporal order judgments (maximum 200) 103
Constrained recognition after one hour
(maximum 20)
16
Spatial and personal context memory (experiment 3):
Free recall (maximum 20)
12
Correct spatial context judgments (maximum 200) 109
Correct personal context judgments (maximum 200) 116
Constrained recognition after one hour
(maximum 20)
18

Control 1

Control 2

12
180

11
181

19

20

14
197
200

14
200
200

20

20

Free recall was tested one hour after learning of list 2. Constrained recognition was tested as
explained in experiment 1.

189

positives after one hour). The controls
achieved perfect scores in these tests and had
no false positives. The patient's recognition
was much more precise in the constrained
condition (perceptual priming items: 15, 13,
and 14 of 16 possible true positives after two
days, one week and one month; temporal
order and context memory tests: 16 and 18 of
20 possible true positives after one hour).
These results indicated that she had stored the
test items considerably better than expected
from free recognition.
EXPERIMENT 2: TEMPORAL ORDER JUDGEMENT

Table 2 summarises the results. Recall after
one hour, immediately before making the temporal order judgement, was not much worse
than in the controls (however, she produced
many intrusions). By contrast, her temporal
order judgements were random (52% correct;
chance level, 50%), whereas the controls'
judgements were 90% accurate (comparison
of the patient and the two controls, x2 (2) =
114, P = 0-0001). In constrained recognition
one hour after the temporal order judgement,
she performed only slightly below the controls,
indicating that her poor temporal order judgement did not simply reflect poor item storage.
EXPERIMENT 3: SPATIAL AND PERSONAL
CONTEXT

Table 2 summarises the results. The patient's
free recall before testing judgement of context
was comparable with the controls. Her judgement of context was random (spatial context,
108/200, 55%) or near random (personal context, 116/200, 58%; random range, 100 (SD
13)), whereas the controls judged spatial and
personal context of the words 99 to 100%
correctly. Constrained recognition after one
hour was comparable with the controls.
EXPERIMENT 4: NEW LEARNING

The patient's performance was normal in both
tests: she achieved a score of 25 in the NVLT
(29 correct; four false positives; percentile, 37)
and 37 in the VLT (43 correct; six false positives; percentile, 56).
LESION ANALYSIS

Lesion analysis was based on an MRI (1.5
Tesla; STIR: TR = 3200 ms, TEf = 19 ms,
TI = 165 ms) performed seven months after
the stroke (figs 1 and 2). Axial images were
parallel to the AC-PC line, and coronal
sections were perpendicular to it. The
damaged structures were determined based on
comparison of the MRI with the atlases of
Nieuwenhuys et al 33 and Van Buren and
Borke.4 The lesion involved the inferior part
of the genu of the right internal capsule and
the right anterior thalamic nucleus (figs 1 and
2, horizontal sections). It extended all along
the anterior surface of the thalamus and
ventrally below the level of the anterior
commissure, which was spared, and included
the anterior and inferior thalamic peduncle.
The inferior part of the column of the fornix
was possibly involved (fig 1, coronal sections),
whereas the mamillothalamic tract was spared.

Schnider, Gutbrod, Hess, Schroth

190

I

I
Figure 1 MRI seven months after stroke demonstrating the lesion of the right capsular genu and anterior thalamus (bold arrowhead). Axial slices are
parallel, coronal slices perpendicular to the AC-PC line. Note the atrophy of the anterior limb of the right internal capsule (thin arrows), which contains
the projections of the dorsomedial thalamic nucleus with the prefrontal and orbitofrontal cortex.

There was severe atrophy of the anterior limb
of the right internal capsule. The dorsomedial
and dorsolateral nucleus were spared. The
lesion most likely represents infarction in the
distal territory of the posterior choroidal
branches.35 There were some additional tiny
lesions of the white matter not exceeding typical age related changes.

AFI1
Figure 2 Details from figure 1 (axial slices). The lesion (+) involves the capsular genu
and the anterior thalamic nucleus (the small circle (0) indicates the position of the left
anterior nucleus). Note the atrophy of the anterior limb of the right internal capsule (thin
arrows) including the anterior thalamic peduncle (left picture).

Discussion
This patient's amnesia was disabling in everyday life and severe by traditional measures of
memory functions (table 1). She did not seem
to realise her failures; she was unconcerned
and denied any memory problem. Some
behavioural features suggested that she stored
information considerably better than implied
by her test scores: although she often confused
previously met people and earlier events, she
seemed to be familiar with them; her spontaneous confabulations could mostly be traced
back to actual events, and the confabulations
produced in memory tests were mostly intrusions from earlier tests. These behavioural
findings motivated a series of experiments to
test whether her memory disorder was based
on an inability to store the context (temporal
order, space, etc) of information acquisition
rather than the information itself.
The patient's bad performance in recognition tasks resulted from false familiarity with
most distracter items. This impairment is reminiscent of the patient of Delbecq-Derouesne
et al36 who showed particularly bad recognition with relatively preserved recall (however,
with abundant confabulations!) after extensive, primarily right sided frontal and anterior
temporal damage after rupture of an aneurysm
of the anterior communicating artery.
Delbecq-Derouesne et al explained the dis-

Memory without context: amnesia with confabulations after infarction of the right capsular genu

crepancy between recall and recognition with
different strategies: recall requires an active
searching strategy whereas recognition is
based on a familiarity judgement.'6 This interpretation is supported by our finding that the
patient's recognition significantly improved
when she was forced to use a searching strategy for recognition, too (forced choice of one
item only; experiment 1). Delbecq-Derouesne
et al concluded that their patient's bad
recognition "does not result from any lack of
knowledge but rather from his inability to
distinguish his experience of remembering
from another kind of misleading subjective
experience triggered by the distracters" but
did not experimentally substantiate this idea.36
We hypothesised that our patient's false
familiarity with distracter items in recognition
tasks was based on an inability to coencode
the time and the spatial context of the
occasion in which information was acquired.
Any type of information stored without these
attributes cannot be related to a particular
event and will easily be confused with elements of different events. Experiments 2 and 3
were consistent with our hypothesis: she
recalled the temporal order and the context
(place and person) in which word lists were
presented at or near chance. This interpretation predicted that recognition would be normal if the context of information acquisition
was irrelevant. We tested this prediction with a
recognition task involving novel non-words
and nonsense designs32 because the patient
could not have an a priori sense of familiarity
with these stimuli (experiment 4). She performed in the normal range for both types of
stimuli, confirming that her memory disorder
was not based on an inability to store new
information.
Our findings offer an interpretation of
confabulations, the mechanism for which has
been a matter of debate. Amnesia, albeit
typically present, does not correlate with the
severity of confabulations.'7-'9 Our patient
indicates that an inability to store item specific
information is not even necessary for confabulations (experiment 4). A combination of
amnesia with prefrontal damage has been
postulated.5 40 41 Although confabulations are
often associated with deficient performance on
frontal lobe tasks, it is unclear what "frontal
dysfunction" would explain confabulations.
Impaired "self monitoring" has been
alleged,'942 but the physiological basis of "self
monitoring" is unclear. Van der Horst speculated a long time ago that spontaneous confabulations resulted from an inability to maintain
temporal order in memory.2' Our study lends
experimental support to this idea. We propose
that confabulations are based on an inability to
temporally and contextually label information,
resulting in confusion of elements of memory
that do neither temporally nor contextually
belong together. Confabulations may thus not
reflect a subject's desire "to fill gaps in memory"43 but rather mirror the subject's spontaneous recollection of information stored with
insufficient contextual clues.
How does the patient's lesion account for

191

her memory disorder? Her lesion involved the
connections of the thalamus with the hippocampus (anterior nucleus lesion) and with
the amygdala and orbitofrontal cortex (capsular
genu lesion). This lesion thus interrupted
both limbic loops considered important for
memory functions (hippocampus-fornix-anterior thalamic nucleus-cingulum-hippocampus;
amygdala-dorsomedial thalamic nucleusorbitofrontal cortex-amygdala).44 Whether
damage to the anterior thalamic nucleus is sufficient to cause amnesia in humans has not
been established due to scarcity of patients
with isolated lesions but was suggested by animal experiments.45 As the anterior nucleus has
strong connections with the hippocampus
(through the fomix and cingulum) but no significant connections with the amygdala and
orbitofrontal cortex,448 it may be assumed
that the behavioural syndrome emanating
from damage to the anterior nucleus would be
qualitatively similar to hippocampal rather
than orbitofrontal damage. We suggest that
the patient's anterior nucleus lesion was
responsible for her relatively inferior, albeit
normal, recognition of nonsense designs as
compared with non-words (experiment 4).
However, it is unlikely that the anterior
nucleus lesion accounted for the main features
of her amnesia-that is, the confabulations,
false familiarity in recognition tasks, and lack
of insight-because these are not characteristics of the amnesia after hippocampal damage.'2"3 We suggest that these features were
due to damage of the loop connecting the
amygdala, the orbitofrontal cortex, and the
dorsomedial nucleus. Whether amygdala damage alone may produce amnesia is unknown
because no single case with a previously
healthy brain and isolated damage to the
amygdala has ever been reported. Patients
with inherited, slowly progressive damage to
the amygdala (thus allowing continuous compensation of the damage) have been reported
to show relatively mild, albeit significant,
memory deficits.4950 Animals' performance in
recognition tasks was markedly worse after
combined lesions of the hippocampus and
amygdala, indicating a significant role of the
amygdala in memory processes.5' But this
finding also seems to depend on the specific
type of recognition task used and has been
strongly contested.52 However, animal experiments may be of limited value in discussing
our patient's amnesia because they depend on
recognition tasks (our patient's recognition
performance was normal under specific circumstances; experiment 4) and ignore some
behavioural features that are so essential to
human amnesia, such as free recall, confabulations, and insight.
Paramedian thalamic infarction may
produce an amnesia which is initially accompanied by confabulations. Chronic confabulations have also been described.9 Chronic
amnesia after paramedian thalamic infusion
requires a relatively anterior thalamic lesion
involving the area which is normally supplied
by the polar artery.53-55 Such lesions regularly
involve not only the mamillothalamic tract but

Schnider, Gutbrod, Hess, Schroth

192

also the internal medullary lamina, which contains the connections of the dorsomedial
nucleus with the amygdala (unidirectional
afferents) and the orbitofrontal cortex (reciprocal connections).535556 The amnesia due to
Wernicke-Korsakoff syndrome, which is often
initially indicated by confabulations, also
seems to best correlate with the extent of damage to the dorsomedial nucleus'0; this correlation might be dependent on the anterior
extension of damage-that is, interruption of
the fibres from the amygdala on the way to the
orbitofrontal cortex. The lesions associated
with amnesia after rupture of an aneurysm of
the anterior communicating artery involve the
basal forebrain (septal nuclei, medial forebrain
bundle, and other subcortical structures), the
medial orbitofrontal cortex, or both.2356
Patients with combined lesions of the basal
forebrain and the orbitofrontal cortex seem to
produce markedly more intrusions (confabulations) in memory tests than patients with subcortical basal forebrain lesions sparing the
orbitofrontal cortex.57 Our patient's lesion,
which caused a type of amnesia similar to
medial thalamic and orbitofrontal damage,
disconnected the dorsomedial nucleus from
both the amygdala and the prefrontal (including orbitofrontal) cortex as it involved the inferior and anterior thalamic peduncles,46 55 S as
shown by the atrophy of these fibres in the
MRI seven months after the stroke (figs 1 and
2). Her lesion thus provides a link between the
lesions causing amnesia after paramedian thalamic infarction and orbitofrontal damage. We
suggest that her amnesia, which-similar to
the amnesia after paramedian thalamic infarction and orbitofrontal damage-was characterised by confabulations and lack of insight,
was based on disconnection of the loop connecting the amygdala, the orbitofrontal cortex,
and the dorsomedial thalamic nucleus.
We can only speculate why a unilateral right
sided lesion was sufficient to cause such a pervasive memory disorder. Initial confabulations
were described after a unilateral right paramedian thalamic infarct.58 Patients with surgical
removal of the right frontal lobe were found to
be more impaired in making recency judgements about newly acquired information than
patients with left frontal ablation.6t
Additionally, talkative behaviour (hyperverbalisation) was demonstrated after right, but not
left, frontal damage.6' It is therefore possible
that patients with a similar lesion on the left
side would not be as amnestic as our patient
and would not confabulate.
Finally, why would the connections
between the amygdala, the prefrontal cortex,
and the dorsomedial nucleus be so important
for memory? Animal studies have shown that
both the amygdala and the prefrontal cortex
process stimulus-reward associations and may
thus be important for the filtering out of
behaviourally relevant information. A temporal integrative function of these structures has
been proposed.6263 Even though the rewards
and punishments used in animal experiments
hardly rival the complexities of human emotions, we propose based on these studies-

that the attribution of an "emotional" or
behavioural significance to a stimulus by the
amygdala and the prefrontal cortex may be the
correlate of the temporal and contextual label
of information. Because the amygdala and the
orbitofrontal cortex have direct connections,
the role of the dorsomedial nucleus is unclear.
Recent studies concerning the visual system
suggest a mechanism by which the dorsomedial-orbitofrontal connection might be critical.
The visual cortex does not passively absorb
information from neurons in the lateral geniculate body but seems to synchronise their activity.64 This feedback probably leads to stimulus
amplification and may be essential for visual
feature detection. A similar function may pertain to the orbitofrontal cortex: it might synchronise relay in the dorsomedial nucleus, a
mechanism which might be important for
"emotional" feature detection.
We thank Professor DY von Cramon for helpful discussion, Dr
J Bogousslavsky for advice on the localisation of the lesion, Dr W
Sturm and Dr K Willmes for allowing us to use their recognition
test, and Dr D Dreier for performing the MRI. The study was
supported by grant No 32-40 432.94 from the Swiss National
Science Foundation.
1 Bauer RM, Tobias B, Valenstein E. Amnesic disorders. In:
Heilman KM, Valenstein E, eds. Clinical neuropsychology.
New York: Oxford University Press, 1993:523-602.
2 Alexander MP, Freedman M. Amnesia after anterior communicating artery aneurysm rupture. Neurology, 1984;
34:752-7.
3 Damasio AR, Graff Radford NR, Eslinger PJ, Damasio H,
Kassel N. Amnesia following basal forebrain lesions. Arch

Neurol 1985;42:263-71.

4 DeLuca J, Cicerone KD. Confabulation following
aneurysm of the anterior communicating artery. Cortex

1991;27:417-23.
5 DeLuca J. Predicting neurobehavioral patterns following
anterior communicating artery aneurysm. Cortex 1993;
29:639-47.
6 Hanley JR, Davies ADM, Downes JJ, Mayes AR. Impaired
recall of verbal material following rupture and repair of
an anterior communicating artery aneurysm. Cognative

Neuropsychology 1994;11:543-78.
nicating artery: a review of neuroanatomical and neuropsychological sequelae. _7 Clin Exp Neuropsychol 1995;
17:100-21.
8 Guberman A, Stuss D. The syndrome of bilateral paramedian thalamic infarction. Neurology 1983;33:540-6.
9 Hodges JR, McCarthy RA. Autobiographical amnesia
resulting from bilateral paramedian thalamic infarction.
A study in cognitive neurobiology. Brain 1993;116:
921-40.
10 Victor M, Adams RD, Collins GH. The Wernicke-Korsakoff
syndrome. 2nd ed. Philadelphia: FA Davis, 1989.
11 Scoville WB, Milner B. Loss of recent memory after bilateral
hippocampal lesions. .7 Neurol Neurosurg Psychiatry 1957;
20:11-21.
12 Schnider A, Regard M, Landis T. Anterograde and retrograde amnesia following bitemporal infarction. Behav
Neurol 1994;7:87-92.
13 Schnider A, Gutbrod K, Ozdoba C, Bassetti C. Very severe
amnesia with acute onset after isolated hippocampal
damage due to systemic lupus erythematosus. _7 Neurol
Neurosurg Psychiatry 1995;59:644-6.
14 Shoqeirat MA, Mayes AR. Disproportionate incidental
spatial-memory and recall deficits in amnesia.
Neuropsychologia 1991 ;29:749-69.
15 Eslinger P, Damasio AR. Behavioral disturbances associanterior
7 DeLuca J, Diamond BJ. Aneurysm of the anterior commu-

ated with rupture of
Semin Neurol 1984;4:385-9.

communicating artery.

16 Shimamura AP, Squire LR. A neuropsychological study of
fact memory and source amnesia. _7 Exp Psychol Leanm

Mem Cogn 1987;13:464-73.

17 Hunkin NM, Parkin AJ. Recency judgements in Wemicke-

Korsakoff and post-encephalitic amnesia: influences of
proactive interference and retention interval. Cortex

1993;29:485-99.

18 Parkin AJ, Rees J, Hunkin NM, Rose PE. Impairment of
memory following discrete thalamic infarction.

Neuropsychologia 1994;32:39-5 1.

19 Milner B, Petrides M, Smith ML. Frontal lobes and the
temporal organization of memory. Humani Neurobiology
1 985;4: 137-42.
20 Shimamura AP, Janowsky JS, Squire LR. What is the role
of frontal lobe damage in memory disorders? In: Levin
HS, Eisenberg HM, Benton AL, eds. Frontal lobe function
and dysfunction. New York: Oxford University Press,
1991:173-95.
21 Van der Horst L. Uber die Psychologie des Korsakow-

Memory without context: amnesia with confabulations after infarction of the right capsular genu
syndroms. Monatsschrift fur Psychiatrie und Neurologie
1932;83:65-84.
22 Markowitsch HJ, von Cramon DY, Hofmann E, Sick C,
Kinzler P. Verbal memory deterioration after unilateral
infarct of the internal capsule in an adolescent. Cortex
1990;26:597-609.
23 Tatemichi TK, Desmond DW, Prohovnik I, Cross GTI,
Mohr JP, Stern Y. Confusion and memory loss from capsular genu infarction. Neurology 1992;42: 1966-79.
24 Tewes U. HA WIE-R: Hamburger- Wechsler Intelligenztest fur
Erwachsene. Handbuch und Testanweisung. Bern: Huber,
1991.
25 Perret E. Gehirn und Verhalten. Neuropsychologie des
Menschen. Bern: Huber Verlag, 1973.
26 Regard M, Strauss E, Knapp P. Children's production on
verbal and non verbal fluency tasks. Perc Mot Skills
1982;55:839-44.
27 Heaton RK. Wisconsin card sorting test. Odessa: Psychological Assessment Resources, 1981.
28 Delis DC, Kramer JH, Kaplan E, Ober BA. The California
verbal learning test. New York: Psychological Corporation, 1987.
29 Schnider A, Gutbrod K, Hess CW. Motion imagery in
Parkinson's disease. Brain 1995;118:485-93.
30 Kopelman MD, Wilson BA, Baddeley AD. The autobiographical memory interview: a new assessment of autobiographical and personal semantic memory in amnesic
patients. J Clin Exp Neuropsychol 1989;11:724-44.
31 Snodgrass JG, Vanderwart M. A standardized set of 260
pictures: norms for name agreement, image agreement,
familiarity, and visual complexity. J7 Exp Psychol Hum
Learn Mem 1980;6:174-215.
32 Sturm W, Willmes K. NVLT bzw. VLT-Nonverbaler und
Verbaler Lerntest. Modling: Dr G Schuhfried GmbH,
1995.
33 Nieuwenhuys R, Voogd J, van Huijzen C. The human central nervous system. 3rd ed. Berlin: Springer, 1988.
34 Van Buren JM, Borke RC. Variations and connections of the
human thalamus. Vol 2: variations of the human diencephalon. Berlin: Springer, 1972.
35 Plets C, De Reuck J, Van der Eecken H, Van den Bergh R.
The vascularization of the human thalamus. Acta Neurol
Belg 1970;70:687-770.
36 Delbecq-Derouesne J, Beauvois MF, Shallice T. Preserved
recall versus impaired recognition. Brain 1990;113:
1045-74.
37 Mercer B, Wapner W, Gardner H, Benson DF. A study of
confabulation. Arch Neurol 1977;34:429-33.
38 Kapur N, Coughlan AK. Confabulation and frontal lobe
dysfunction. J7 Neurol Neurosurg Psychiatry 1980;43:
461-3.
39 Benson DF. The neurology of thinking. New York: Oxford
University Press, 1994.
40 Stuss DT, Alexander MP, Lieberman A, Levine H. An
extraordinary form of confabulation. Neurology 1978;28:
1166-72.
41 Kopelman MD. Two types of confabulation. J Neurol
Neurosurg Psychiatry 1987;50: 1482-7.
42 Shapiro BE, Alexander MP, Gardner H, Mercer B.
Mechanisms of confabulation. Neurology 1981;31:
1070-6.
43 American Psychiatric Association. DSM-IV. Diagnostic and
statistical manual of mental disorders. 4th ed. Washington,
DC: American Psychiatric Association, 1994.
44 Mishkin M. A memory system in the monkey. Phil Trans R
Soc Lond 1982;298:85-95.
45 Aggleton JP, Sahgal A. The contribution of the anterior
thalamic nuclei to anterograde amnesia. Neuropsychologia
1993;31:1001-19.

193

46 Jones EG. The thalamus. New York: Plenum Press, 1985.
47 Raisman G, Cowan WM, Powell TPS. An experimental
analysis of the efferent projection of the hippocampus.
Brain 1966;89:83-108.
48 Mufson EJ, Pandya DN. Some observations on the course
and composition of the cingulum bundle in the rhesus
monkey. Jf Comp Neurol 1984;225:31-43.
49 Tranel D, Hyman BT. Neuropsychological correlates of
bilateral amygdala damage. Arch Neurol 1990;47:
349-55.
50 Babinsky R, Calabrese P, Durwen HF, et al. The possible
contribution of the amygdala to memory. Behav Neurol
1993;6: 167-70.
51 Mishkin M. Memory in monkeys severely impaired by
combined but not by separate removal of amygdala and
hippocampus. Nature 1978;273:297-8.
52 Zola-Morgan S, Squire LR, Amaral DG. Lesions of the
amygdala that spare adjacent cortical regions do not
impair memory or exacerbate the impairment following
lesions of the hippocampal formation. Jf Neurosci 1989;
9:1922-36.
53 Von Cramon DY, Hebel N, Schuri U. A contribution to
the anatomical basis of thalamic amnesia. Brain 1985;
108:993-1008.
54 Gentilini M, De Renzi E, Crisi G. Bilateral paramedian
thalamic artery infarcts: report of eight cases. Jf Neurol
Neurosurg Psychiatry 1987;50:900-9.
55 Graff-Radford NR, Tranel D, Van HGW, Brandt JP.
Diencephalic amnesia. Brain 1990;113:1-25.
56 Markowitsch HJ, Von Cramon DY, Schuri U. Mnestic performance profile of a bilateral diencephalic infarct patient
with preserved intelligence and severe amnesic disturbances. Jf Clin Exp Neuropsychol 1993;15:627-52.
57 Irle E, Wowra B, Kunert HJ, Hampl J, Kunze S. Memory
disturbances following anterior communicating artery
rupture. Ann Neurol 1992;31:473-80.
58 Amaral DG, Price JL, Pitkiinen A, Carmichael ST.
Anatomical organization of the primate amygdaloid complex. In: Aggleton JP, ed. The amygdala. Neurobiological
aspects of emotion, memory, and mental dysfunction. New
York: Wiley aiss Inc, 1992: 1-66.
59 Foglia P, Perini M, Vanzulli F. Pure amnesia in a case of
right thalamic lesion. ItalJ Neurol Sci 1991;12:211-3.
60 Milner B, Corsi P, Leonard G. Frontal-lobe contribution
to recency judgements. Neuropsychologia 1991;29:
601-18.
61 Kolb B, Taylor L. Affective behavior in patients with localized cortical excisions: role of lesion site and side. Science
1981;214:89-91.
62 Fuster JM. The prefrontal cortex. Anatomy, physiology, and
neuropsychology of the frontal lobes. 2nd ed. New York:
Raven Press, 1989.
63 Kesner RP. Learning and memory in rats with an emphasis
on the role of the amygdala. In: Aggleton JP. The amygdala. Neurobiological aspects of emotion, memory, and mental
dysfunction. New York: Wiley Liss Inc 1992:379-99.
64 Sillito AM, Jones HE, Gerstein GL, West DC. Featurelinked synchronization of thalamic relay cell firing
induced by feedback from the visual cortex. Nature 1994;
369:479-82.
65 Wechsler D. A standardized memory scale for clinical use. Jf
Psychol 1945;19:87-95.
66 Milner B. Interhemispheric differences and psychological
processes. BritMed Bull 1971;27:273-7.
67 Rey A. Test de copie et de reproduction de mimoire de figures
giometriques complexes. Paris: Centre de Psychologie,
1958.
68 Rey A. L'examen clinique en psychologie. Paris: Presses universitaires de Frances, 1964.