Neurocase (2001) Vol. 7, pp. 407–417

© Oxford University Press 2001

Verbal Encoding Deficits in a Patient with a Left
Retrosplenial Lesion
Carrie R. McDonald, Bruce Crosson, Edward Valenstein1 and Dawn Bowers
Departments of Clinical and Health Psychology and 1Neurology, University of Florida Brain Institute, Gainesville, Florida, USA

Abstract
Over the past decade, memory impairments associated with retrosplenial damage have received increased attention
among neuroscientists, although the exact role of the retrosplenial region in memory has not been clearly defined.
Evidence from lesion studies and functional neuroimaging has implicated the retrosplenial region in verbal episodic
memory, temporal ordering of information, and topographical memory. In addition, recent positron emission
tomography studies have shown increased activation of the retrosplenial cortex during tasks involving both the
encoding and retrieval of episodic information. The objective of this study was to define more clearly the nature of
memory impairments observed in retrosplenial amnesia. A 47-year-old amnesic male with a left retrosplenial
arteriovenous malformation was examined on neurocognitive tasks of automatic and directed encoding, temporal
ordering of information, and remote memory. Despite normal performance on frontal cognitive tasks, intact memory for
remote information, and a superior IQ, this individual exhibited a profound deficit in the encoding of information,
evidenced by poor release from proactive interference, poor category clustering on word list recall, poor semantic
encoding on a levels of processing task, and mild impairments in temporal ordering. These results imply that the
retrosplenial region plays a role in the verbal encoding of information, which contributes to the profound verbal
memory impairment reported in previous case studies of patients with retrosplenial damage.

Introduction
In 1987, Valenstein et al. reported a landmark case of a
patient with profound amnesia following hemorrhage of an
arteriovenous malformation (AVM) in a region involving the
retrosplenial cortex and the cingulate bundle. This patient
evidenced severe anterograde amnesia and circumscribed
retrograde amnesia, with relatively intact intellectual functioning and memory for remote information. Valenstein et al.
(1987) concluded that their patient’s retrosplenial damage
had produced amnesia by disrupting the portion of Papez’
circuit which traverses retrosplenial cortex, providing one
route from the left hippocampus to the left anterior thalamic
nuclei. Several years later, Rudge and Warrington (1991)
reported similar memory impairments in nine patients with
tumors involving the splenium of the corpus callosum. These
authors proposed that the observed amnesia resulted from a
disconnection between the frontal and the temporal lobes,
due to fornix damage rather than to a thalamohippocampal
disconnection, or damage to the retrosplenial cortex itself.
To date, the exact anatomical explanation for retrosplenial
amnesia remains unclear, although both models seem
plausible explanations for the amnesia associated with
retrosplenial damage.

Clinical case studies
Since these initial reports, neuroscientists have focused more
attention on elucidating the exact nature of memory impairments observed in patients with retrosplenial amnesia. Clinical case studies have shown that patients with retrosplenial
damage exhibit verbal and non-verbal anterograde amnesia
(Takayama et al., 1991; Katai et al., 1992; Iwasaki et al.,
1993; Arita et al., 1995; Yasuda et al., 1997; Sato et al.,
1998; Correa et al., 1999), time-limited retrograde amnesia
(Iwasaki et al., 1993), topographical amnesia (Cammalleri
et al., 1996; Sato et al., 1998), and poor acquisition of
temporal information (Bowers et al., 1988; Correa et al.,
1999). In addition, some case reports suggest that these
memory impairments are material-specific, with loss of
verbal episodic memory more profound following left
retrosplenial damage and loss of memory for spatial relationships more apparent in patients with right retrosplenial
damage (Maddock, 1999). For example, Valenstein et al.’s
patient initially presented with both verbal and visual memory
deficits associated with his left retrosplenial AVM. However,
subsequent testing revealed that verbal tests alone remained
consistently impaired, while tests of non-verbal memory,

Correspondence to: C. R. McDonald, Department of Clinical and Health Psychology, University of Florida, PO Box 100165 HSC, Gainesville,
FL 32610-0165, USA. Tel: ⫹1 352 265 0680 (ext. 46887); Fax: ⫹1 352 265 0468; e-mail: carrierm@ufl.edu

408 C. R. McDonald et al.

with the exception of the Rey–Osterrieth Complex Figure
Test (ROCFT), were typically normal. Conversely, Takahashi
et al. (1997) studied three patients with topographical amnesia
resulting from small, focal hemorrhages involving the right
retrosplenial region. These patients had intact discrimination
and identification of familiar landscapes, but could not
remember familiar spatial relationships between two distant
locations. No verbal memory impairments were noted in
these patients with right retrosplenial damage.
Other studies have described the amnesic syndrome associated with retrosplenial damage as bitemporal in nature. For
example, Correa et al. (1999) presented a study of a 41-yearold male with memory deficits secondary to a glioblastoma
of the splenium of the corpus callosum and retrosplenial
region. Their patient showed a profound impairment in verbal
and non-verbal learning and recall, as well as difficulty with
the temporal ordering of information. In addition, their patient
exhibited characteristics of frontal dysfunction, including
perseveration, increased susceptibility to interference, and
decreased verbal fluency. Katai et al. (1992) presented the
case of a 73-year-old patient with impaired recent memory
for both verbal and non-verbal information following a left
cerebral infarction in the retrosplenial region. In addition
to anterograde amnesia, this patient had severe retrograde
amnesia for information up to 2 years prior to the injury.
More recently, Gainotti et al. (1998) reported the case of a
patient with a bilateral retrosplenial tumor whose memory
impairment was characterized primarily by dense retrograde
amnesia for personal events and poor visual, but relatively
intact verbal learning and recall. The authors note, however,
that their case cannot be considered a ‘pure’ case of retrosplenial amnesia, as the tumor also involved the fornix
bilaterally.
In summary, a handful of cases of retrosplenial damage
have been reported in the literature over the past decade.
These case reports, however, are greatly limited by their
anatomical variability (e.g. most involving additional damage
to the splenium and fornix) and by relatively non-specific
descriptions of the resulting amnesic syndrome.

Positron emission tomography (PET) studies
In addition to studies of patients with focal brain lesions,
a growing functional neuroimaging literature provides
converging evidence for the role of the retrosplenial area in
memory (Shallice et al., 1994; Tulving et al., 1994; Fletcher
et al., 1995). Shallice et al. (1994) found that the acquisition
of episodic memories was associated with activity in both
the left pre-frontal cortex and the left retrosplenial area,
whereas retrieval was associated with increased activity in
the right pre-frontal cortex and pre-cuneus. Similar results
were reported by Fletcher et al. (1995) in a study of verbal
episodic encoding. Tulving et al. (1994) found activation of
the retrosplenial cortex during encoding of novel pictures,
but not familiar ones, suggesting that the retrosplenial region
may be one component of a visual/spatial novelty encoding

network in the brain. Similarly, Dolan and Fletcher (1997)
have implicated posterior cortex, specifically the left
hippocampal formation, in the encoding of novel verbal
information.
Based on evidence from case studies and functional
imaging, it seems plausible that the left retrosplenial cortex
and left hippocampal formation, which are richly and
reciprocally connected, may be components of a complex,
distributed encoding network for verbal information. It
remains unclear from these studies, however, whether the
retrosplenial cortex has a specific memory-related function
in and of itself, or whether the amnesia results from a
disconnection of mesial temporal, pre-frontal, and/or diencephalic structures important in this memory encoding system
(Valenstein et al., 1987; Maddock, 1999).

Current case study
The present case study represents an attempt to clarify the
nature of the amnesic disturbance due to retrosplenial damage
by using well-validated clinical and experimental paradigms
in a patient with a discrete left retrosplenial lesion, which
spared the fornix and most of the splenium. This is important,
as few, if any, studies of retrosplenial damage have been
reported in which the fornix and the splenium are intact.
We were primarily interested in examining the extent to
which our patient was impaired in the encoding and retrieval
of verbal information. These parameters were of interest
because recent PET research has implicated the left
retrospenial region in the encoding of verbal episodic information. In addition, numerous case studies of patients with
left retrosplenial damage have shown that these patients
demonstrate poor acquisition and temporal ordering of verbal
information (Valenstein et al., 1987; Bowers et al., 1988;
Correa et al., 1999), which may reflect a more general verbal
encoding deficit. Finally, retrieval of verbal information was
of interest because we wished to rule out the possibility that
poor memory performances were related to impaired retrieval
of information.
To examine these parameters, we evaluated the degree to
which our patient was impaired on tasks of automatic and
directed semantic encoding, temporal ordering of verbal
information, and retrieval of autobiographical/remote
information. Tasks of automatic semantic encoding require
that the subject spontaneously uses semantic cues to facilitate
recall performance, without the instruction of the examiner.
On the other hand, in directed semantic encoding the examiner
instructs the subject to process information at varying levels
of analysis (i.e. perceptual, phonemic, or semantic). Although
the encoding of information was once conceptualized as a
single process, recent views suggest that adequate encoding
of new information involves at least two distinct processes
(Verfaellie and Cermak, 1991). First, the individual must
correctly process the to-be-remembered information. This
stage involves the initial semantic analysis of information.
Second, the individual must manipulate the results of semantic

Retrosplenial amnesia 409

analysis in order to permit adequate storage of the information
(Verfaellie and Cermak, 1991). Failures during this second
stage of encoding are observed in amnesic patients who
correctly analyse semantic information, but continue to show
memory impairments at recall. These patients do not benefit
from their semantic analysis because they are unable to
utilize this information in a strategic manner to improve
memory performance. In addition to evaluating semantic
encoding, temporal ordering of information was of interest
because it is believed to reflect another aspect of encoding—
the ‘time-tagging’ of new information in memory—and has
been found to be impaired in other patients with retrosplenial
damage (Bowers et al., 1988; Correa et al., 1999).
To summarize, we were interested in exploring whether
our patient exhibited a deficit in the semantic encoding
of information, as well as whether such an impairment
represented a breakdown in the initial semantic analysis or
subsequent cognitive manipulation of information. Second,
we were interested in whether other aspects of encoding,
such as temporal ordering of information, were impaired. We
hypothesized that our patient would be impaired on tasks
which involved the semantic encoding of information, as
well as on tasks of the temporal ordering of information in
memory. These hypotheses were based on previous PET
research, as well as findings from previous case studies of
retrosplenial damage. Third, we hypothesized that our patient
would not be impaired in the retrieval of remote memories.
This hypothesis was based on the assumption that our
patient’s memory problems were related to poor encoding of
information and not to a retrieval deficit.

Case report
LC, a 47-year-old, left-handed college-educated engineer
presented to our neuropsychology clinic with complaints of
chronic memory impairment, difficulty sustaining attention,
and poor concentration. Years earlier he had experienced an
acute onset of severe headache pain which lasted for 3 days.
LC was admitted to Shands Teaching Hospital on 8 March
1991, where he underwent a cerebral arteriogram and
magnetic resonance imaging (MRI), which revealed an AVM
medial to the trigone of the left lateral ventricle, filling from
the left anterior cerebral artery (ACA), middle cerebral artery
(MCA), and posterior cerebral artery (PCA) branches. The
arteriogram also revealed minor filling from the MCA on the
right. A computed tomography (CT) scan with and without
contrast revealed a 5 ⫻ 5 ⫻ 3 cm mass with associated
subacute blood. There was no evidence of hydrocephalus.
His past medical history and family history were noncontributory.
On admission, LC’s vital signs were stable, his temperature
was 39°C, and his lungs were clear to auscultation. His
medical record states that he was awake, alert, and oriented
times three with a normal mental status examination. Speech
was fluent with normal repetition. Cranial nerves II–XII were
intact and a motor examination was 5/5 for all groups.

Cerebellar examination was within normal limits except for
some bilateral end gaze nystagmus noted. Deep tendon
reflexes were 2⫹ and symmetric.
LC received three embolizations of the AVM, followed by
his first stereotactic radiosurgery in 1991. He first received
embolization of PCA feeders which resulted in a reduction
in the volume of the AVM by one third, with a residual
opacified nidus of approximately 4 ⫻ 2 cm. Pre- and postcompensation MRI studies demonstrated a residual flow
within the AVM nidus. The first radiosurgery involved 1000
rads of radiation to the margin of his AVM through 30 and
16 mm collimators.
CT and MRI performed 6 months post-radiosurgery
revealed no definite evidence of further decrease in nidus
size or in overall flow through the AVM. A CT scan performed
3 years later (1 January 1994) still revealed a zone of atrophy
mixed with areas of dystrophic calcification within the deep
portion of the vascular malformation. There was evidence of
increased vascular structures even on a non-enhanced scan.
There was normal ventricular size and a reasonably normal
sulcal pattern, other than areas of encephalomalacia in the
region of the nidus of the AVM. An enlarged posterior
cerebral artery was evident on the left side, presumably
representing the afferent blood supply to the AVM. No acute
hemorrhage was identified. LC received a second stereotactic
surgery in 1995, to remove the remainder of the AVM.
Following the second surgery, MRI revealed a thrombosed
AVM with some low flow state adjacent to the trigone of the
left lateral ventricle. There appeared to be leukomalacic
changes in and around the radiation field associated with
the surgeries. In addition, there was minimal damage to the
splenium of the corpus callosum at the most lateral extent
on the left, although the fornix remained intact bilaterally.
Post-surgical neurological examination revealed intact cranial
nerves, motor reflexes, and sensory findings. LC showed
normal verbal fluency, and an intact orientation, but he was
able to recall only one of three objects after a 5-min delay,
suggesting an impairment in verbal memory. Figure 1 presents
MRIs obtained in 1998 showing the residual of an AVM
affecting retrosplenial cortex and subadjacent white matter.
Figure 2 shows axial slices depicting the full extent of
the retrosplenial lesion mapped according to Damasio and
Damasio’s (1989) anatomic atlas.

Neuropsychological evaluations
LC reportedly began to experience significant memory
problems and progressive difficulties with his job following
the second radiosurgery in 1995, leading to his resignation
as vice-president of engineering in 1997. As a result of the
perceived cognitive changes, LC was referred to our clinic
where he was evaluated on three separate occasions. Detailed
neuropsychological examination results are shown in Table 1.
Formal neuropsychological testing was performed in January
1997 and 2 years later in January 1999. During both
evaluations, LC was alert, attentive, and socially appropriate.

410 C. R. McDonald et al.

Fig. 1. Magnetic resonance images (T1- and T2-weighted images) taken in June 1998 of residual of an arteriovenous malformation in the left hemisphere with
multiple abnormal vessels involving the retrosplenial cortex and subadjacent white matter. (A) T1 sagittal section 5 mm left of midline. (B) T2 horizontal plane
extending through the mid-portion of the third ventricle inferior to the splenium of the corpus callosum.

He showed no signs of a thought disorder and conversational
speech was fluent and grammatically correct. With the
exception of changes in his memory functioning, performances on other tasks of neurocognitive abilities were similar
for both evaluations.
Performances during the first evaluation, however, were
confounded by the fact that LC reported heavy alcohol
intake at the time. Therefore, the degree to which his
neuropsychological deficits on the first evaluation were a
result of his vascular malformation versus alcohol abuse
cannot be determined. According to the patient and spouse
report, LC was not drinking alcohol at the time of the second
evaluation.

First neuropsychological evaluation
The results of both neuropsychological evaluations are
depicted in Table 1. On the Wechsler Adult Intelligence
Scale-Revised (WAIS-R; Wechsler, 1981), LC’s full-scale
IQ was 127 (superior), with a verbal IQ of 134 (very
superior) and a performance IQ of 111 (high average).
Confrontation naming was intact, and no signs of aphasia
were present. In addition, he showed no evidence of spatial
neglect or impaired facial recognition. Performances on
neurocognitive tasks of executive functioning (Trail Making,
Parts A and B, Stroop Interference Test and Auditory
Consonant Trigrams) were also intact. The one exception

Retrosplenial amnesia 411

Fig. 2. Axial slices showing the full extent of the retrosplenial lesion mapped according to Damasio and Damasio’s (1989) anatomic atlas. The lesion is depicted
by the darkened areas.

was a borderline performance on the Halstead–Reitan
Booklet Category Test (Halstead, 1947), which measures
concept formation and flexibility during novel problemsolving (Spreen and Strauss, 1998).

complex figure adequately. However, his approach to the
copy was disorganized, and may have led to poor encoding
of the design. This is consistent with the fact that his
immediate and delayed recall of the same figure were
significantly impaired and lacking in detail.

Examination of memory functions
Verbal memory. LC’s immediate recall of prose passages
was average (Wechsler Memory Scale-Revised; WMS-R),
while his performance after a 30-min delay fell in the low
average range. Performances were well below average on a
task of verbal list learning [California Verbal Learning Test
(CVLT); Delis et al., 1987]. LC’s immediate list recall was
impaired and he did not appear to benefit from repeated
exposure to the list. Delayed free recall was also significantly
impaired, as was category cued recall and item recognition.
A significant number of intrusions and false positives were
also noted across trials.
Non-verbal memory. LC’s immediate recall of geometric
designs was high average (WMS-R), while his recall of the
same figures after a 30-min delay was in the low average
range. Non-verbal memory was also significantly impaired
on the ROCFT (Osterrieth, 1944). LC was able to copy a

Second neuropsychological evaluation
The results of the second neuropsychological evaluation
indicated that LC’s non-verbal memory had significantly
improved, while his verbal memory remained impaired. On
the WAIS-R, LC obtained a full-scale IQ of 134 (very
superior)—close to the score obtained in the initial evaluation.
His verbal IQ was 148 (very superior) and his performance
IQ was 112 (high average). No evidence of spatial neglect
was present (Judgment of Line Orientation), and all
performances on neurocognitive tasks of executive functioning were intact (Stroop Interference Test, Auditory Consonant Trigrams, Booklet Category Test).

Examination of memory functions
Verbal memory. Consistent with his performance on the
first evaluation, LC’s immediate recall of prose passages was

412 C. R. McDonald et al.
Table 1. Summary of the neuropsychological data obtained during LC’s first and second clinical evaluations
29 January 1997

WAIS-R
Verbal IQ
Performance IQ
Full-scale IQ
Language
Boston Naming Test

Frontal/executive
Trail Making A
Trail Making B
Stroop Interference Test
Booklet Category Test
Auditory Consonant Trigrams
Memory
WMS-R
Logical Memory I
Logical Memory II
Visual Reproduction I
Visual Reproduction II

CVLT
List A total 1–5
List A Trial 1
List A Trial 5
List B
List A SDFR
List A SDCR
List A LDFR
List A LDCR
Semantic cluster
Serial cluster
Recognition hits
False positives
Total intrusions

ROCFT
Copy
Immediate recall
Delayed recall

27 January 1999

Raw

Scaled score

Raw

Scaled score

89
50
139

134
111
127

97
51
148

148
112
134

59

18

Raw

Percentile

Raw

Percentile

32⬘
70⬘
111
73 ers
3⬘
9⬘
18⬘

16th
31st
69th
16th
67th
93rd
100th

111
52 ers
3⬘
9⬘
18⬘

69th
70th
86th
80th
80th

25
14
36
20

57th
24th
88th
21st

21
10
39
34

34th
14th
98th
87th

Raw

z-score

Raw

z-score

38
6
7
5
2
9
5
9
2.2
3.0
10
13
32

T ⫽ 30
–1
–3
–1
–4
–1
–2
–1
0
1
–3
5

39
4
9
5
3
5
7
8
0.3
1.2
16
6
23

T ⫽ 31
–2
–2
–1
–3
–3
–2
–2
–2
–1
1
2

Raw

Scaled score

Raw

Scaled score

64
11
12

9
3
3

63
26
23

8
6
6

WAIS-R, Wechsler Adult Intelligence Scale-Revised; WMS-R, Wechsler Memory Scale-Revised; CVLT, California Verbal Learning Test; ROCFT,
Rey–Osterrieth Complex Figure Test.
SDFR, short delay free-recall; SDCR, short delay cued-recall; LDFR, long delay free-recall; LDCR, long delay cued-recall.

average, while his recall after a 30-min delay was low
average. LC’s performance on the CVLT was also consistent
with his performance on the initial evaluation. LC’s immediate
list recall was significantly impaired and he did not appear
to benefit from repeated exposure to the list. Delayed free
recall was also significantly impaired, as was category cued
recall. Item recognition, as calculated by recognition hits,
appeared normal. LC also continued to exhibit a significant
number of false positives in the recognition trial, suggesting

a strong positive response bias as opposed to true recognition
accuracy. A significant number of intrusions was also noted
across trials. In addition, use of semantic clustering was
impaired. On both evaluations, LC’s best trial performance
after repeated exposure to the word list was only 62% recall.
Non-verbal memory. LC’s immediate recall of geometric
designs on the WMS-R was very superior and his recall
after a 30-min delay was superior, suggesting a significant

Retrosplenial amnesia 413

improvement in non-verbal memory. LC’s performances
remained impaired on the ROCFT, but were significantly
improved relative to the initial evaluation. LC’s copy of a
complex figure was low average. However, his immediate
and delayed recall of the same figure were borderline. Similar
to the initial evaluation, his copy and reproductions of the
figure were disorganized and lacking in detail.

Experimental procedures
LC returned to our clinic 4 months after his second neuropsychological evaluation at our request to complete additional
memory testing. The purpose of this evaluation was to define
more clearly the nature of LC’s anterograde amnesia using
well-validated experimental paradigms of encoding and
retrieval. We examined LC’s performance on neurocognitive
tasks of automatic and directed encoding [e.g. levels of
processing (LOP), release from proactive interference] and
temporal ordering of information (e.g. temporal ordering for
sentences). In addition, we administered tests of remote
memory for events (Kortenkamp Remote Memory Test;
KRMT) and for personal information (Crovitz Remote
Memory Test; CRMT) to determine if LC showed any
retrieval deficit or evidence of retrograde amnesia.

Encoding tasks
Release from proactive interference. Release from proactive interference was assessed using Wickens’ (1970)
category shift manipulation. This paradigm was administered
in order to assess LC’s automatic use of semantic cues to
facilitate memory performance at recall. In this procedure,
subjects are presented with eight blocks consisting of five
trials of words. Each trial consists of the presentation of
three words read aloud by the examiner, followed by a
distraction task which requires the subjects to count backwards in threes, beginning with a specified number in order
to prevent rehearsal. After the distraction task, subjects are
requested to recall the previously heard words. For four of
the eight blocks, words in all trials belong to a single
taxonomic category. In the other four blocks, the taxonomic
category is shifted in the fifth trial. Taxonomic categories
include occupations, furniture, body parts, transportation,
fruit, clothing, musical instruments, and geographical landmarks. Wickens proposed that if there is a decline in
performance across early trials followed by a significant
improvement in the shift trial, subjects are engaging in
category encoding and ‘releasing’ from proactive interference. Failure to release from proactive interference, on the
other hand, is represented by no improvement in the shift
trial and suggests poor semantic analysis of the information.
LC’s performance in the shift trials (66% correct) did not
differ significantly from his performance in the non-shift
trials (58% correct). As can be seen in Fig. 3, LC and agematched controls showed no difference in their performances
in the non-shift condition (58 and 60% correct, respectively).

Fig. 3. The number of items recalled by LC and age-matched controls across
shift (———) and non-shift (— — —) trials of a release from proactive
interference paradigm.

For controls, however, the change in semantic class in trial
5 of the category shift condition was associated with a
significant increase in recall as compared with the no-shift
condition (25% increase). LC did not show a significant
increase with the shift condition (8.3% increase). This
performance suggests a failure to release from proactive
interference, which is likely to reflect poor semantic encoding
of the taxonomic information.
LOP. The LOP paradigm (Craik and Lockhart, 1972) was
administered in order to determine whether LC would benefit
from depth of encoding for verbal information. In this task,
the examinee is presented with 36 words on flashcards and
is directed to encode each word based on either the word’s
perceptual (e.g. Is this word printed in upper-case letters?),
phonemic (e.g. Does this word rhyme with daisy?), or
semantic (e.g. Is this a type of flower?) features. Free recall
for the words is assessed after a 20-min delay, followed by
a multiple-choice recognition trial. Depth of encoding in this
task ranges from very shallow (perceptual) to deep (semantic)
analysis of the information. Deeper levels of analysis should

414 C. R. McDonald et al.

Fig. 4. Performances of LC and age-matched controls on item recognition in
a levels of processing task. Error bars represent the standard deviations for
the control group.

theoretically produce more elaboration, and stronger memory
for information than shallow encoding of information at
recall, especially over the long term.
LC’s performance on the LOP was significantly impaired
relative to a group of eight age-matched controls. LC was
able to recall freely only one of the 36 words presented on
the LOP after a 20-min delay. His recognition of these
words was 50% for the words introduced with a perceptual
orienting question, 42% for words introduced with a phonological orienting question, and only 50% for words introduced
with a semantic orienting question. In contrast, the control
group showed the expected pattern, correctly identifying 66%
of the perceptually, 79% of the phonologically, and 93% of
the semantically cued words. Recognition performances on
the LOP for LC and the control group are shown in Fig. 4.
Temporal memory. To assess LC’s ability to encode
information in a temporal context, we gave a sentence
discrimination task similar to that used by previous investigators (Bowers et al., 1988). A set of eight target sentences
is shown (set 1) on flashcards, followed 30 min later by the
presentation of a second set of eight different target sentences
(set 2). Two minutes later a recognition test is given. For the
recognition test, the 16 sentences from the two target sets
are randomly intermixed with 16 distractor sentences, and
subjects are instructed to state whether the sentences had
previously been seen (recognition judgment). If so, they then
indicate whether the sentence had occurred in the first or
second set of sentences (i.e. temporal order judgment). Each
stimulus is shown for a 2-s duration. Two separate scores
are obtained for the task—recognition accuracy and accuracy
of temporal order judgments. Recognition accuracy is based
on the number of targets correctly identified as having been
previously presented (true positives) in conjunction with the
number of distractors correctly identified as not having been
presented (true negatives). Percentage recognition accuracy
is then derived from the formula: ‘true positives ⫹ true
negatives/32,’ where 32 equals the total number of targets
and distractors shown on the recognition test. Accuracy on
temporal order judgments is based only on those targets that
are correctly identified as having been presented (i.e. true

Fig. 5. Percentage of target words recognized by LC and age-matched controls
relative to their temporal ordering judgments of the same words. Error bars
represent the standard deviations for the control group.

positives). This score is computed from the formula: ‘number
of correct temporal order judgments/true positives’.
The results of LC’s performance on the temporal
ordering task are depicted in Fig. 5, beside the performance
of the control group. As shown, LC’s recognition accuracy
was 81%, which is close to that of a control group
(mean ⫽ 85%, SD ⫽ 12.9). In contrast, his temporal
order judgments were somewhat lower (64%), falling 1.5
SD below the mean of the control group (mean ⫽ 80%,
SD ⫽ 10.7).
Retrograde memory. The KRMT (Kortenkamp, personal
communication) was used to assess LC’s memory for information acquired before his illness. The KRMT is a 15-item
multiple-choice format that includes questions about worldwide events that received significant media coverage during
the 1970s (n ⫽ 5), 1980s (n ⫽ 5), and 1990s (n ⫽ 5). These
questions cover events in sport and politics, as well as
national disasters that are generally familiar to the American
public. LC’s performance on the KRMT was within normal
limits (95% correct), suggesting intact remote memory for
public events.
The CRMT (Crovitz and Schiffmann, 1974) was used to
assess LC’s memory for personal events occurring throughout
his life. The CRMT is an open-ended interview that requires
subjects to recall detailed information about specific events
that occurred during childhood (age 1–18 years), early
adulthood (age 19–30 years), and middle adulthood (age 40–
60 years) (i.e. Recall a specific event from your childhood
about a birthday party....). Responses on the CRMT are
scored qualitatively, according to the degree of detail provided
by the subject for each personal event. Responses are also
compared across decades to determine if subjects are able to
provide equally detailed information about events occurring
in each of the age ranges identified. LC’s performance on
the CRMT suggests intact memory for autobiographical
information throughout his life. He was able to recall and
elaborate on specific events occurring during childhood and
adolescence, young adulthood, and middle adulthood with
appropriate detail and affective response. Therefore, LC did
not appear to exhibit a generalized or time-limited retrograde
amnesia.

Retrosplenial amnesia 415

Discussion
The results of this case study are largely consistent with our
hypotheses and with other cases of retrosplenial amnesia
reported in the literature (Valenstein et al., 1987). On the
initial clinical evaluation, our patient exhibited severe anterograde amnesia, marked by poor learning and recall of both
verbal and visual information. At follow-up, however, LC’s
memory impairment was primarily for verbal information,
which is consistent with his left-sided lesion. The exception
was that LC continued to perform poorly on the ROCFT at
follow-up. Although this finding suggests residual non-verbal
memory impairment, research has shown the ROCFT to
discriminate poorly between patients with right versus left
hemisphere dysfunction (Barr et al., 1997). This is probably
related to the fact that the scoring of the ROCFT is largely
influenced by the presence or absence of details within the
figure, which is more dependent on intact left hemisphere
functioning. Consistent with this interpretation, LC maintained the gestalt of the figure in his reproductions, but
omitted many of the details. It is also possible that LC
maintained a more subtle visual memory impairment that
was detected only by the ROCFT—the more difficult of the
two non-verbal tasks with higher organizational demands.
The results of the experimental testing revealed three major
findings. First, our patient exhibited a profound deficit in the
semantic encoding of information, evidenced by poor release
from proactive interference on Wickens’ paradigm and poor
semantic encoding on a LOP task. These results were
complimented by poor semantic clustering on the CVLT
evidenced in the clinical evaluations. While LC was able to
retain an intact semantic analysis of verbal information (i.e.
he was able to answer correctly semantic orienting questions
related to verbal stimuli in the LOP paradigm), he appeared
unable to utilize the results of this semantic analysis to
improve effectively his memory performance. Therefore, our
patient exhibited an impairment in the cognitive manipulation
of semantic information, suggesting a deficit in the second
stage of encoding. Inability to utilize cognitive strategies
spontaneously to facilitate encoding of information has been
described in patients with frontal lobe damage (Squire, 1982;
Shimamura, 1995). On the other hand, patients with frontal
pathology are generally able to employ semantic encoding
strategies when instructed by the examiner (Butters and
Cermak, 1980), as this method facilitates retrieval of the
to-be-remembered information. Our patient, however, was
impaired on tasks of both automatic and directed encoding,
perhaps suggesting involvement of both frontal and
hippocampal systems.
Our second major finding from the experimental testing
was that LC showed a mild impairment in the temporal
ordering of verbal information. While his ability to order
verbal information in memory was less impaired than his use
of semantic encoding strategies, he performed correctly in
only 64% of the trials, compared with the 80% accuracy
obtained by age-matched control subjects. Previous research-

ers have found similar temporal ordering impairments in
patients with retrosplenial damage (Bowers et al., 1988;
Correa et al., 1999) and have hypothesized that this defect
occurs in the early stages of acquisition, such that incoming
stimuli are not initially encoded or ‘time-tagged’ in terms of
their temporal properties. It is interesting that our patient
also exhibited a significant number of intrusions on the
CVLT, a finding which may also reflect an inability to ‘timetag’ information according to discrete episodes (i.e. list A
versus list B words). This finding is consistent with the
hypothesis that our patient exhibited a rather generalized
deficit in the encoding of new verbal information captured
by tests of semantic and temporal encoding.
Our third major finding was that LC was unimpaired in
the retrieval of remote memories. This suggests that LC’s
retrosplenial damage resulted in a severe anterograde amnesia,
with no apparent retrograde amnesia or retrieval deficit.
Evidence against a retrieval deficit hypothesis of LC’s impairments is also reflected by the fact that category cueing and
item recognition on the CVLT did not facilitate LC’s recall
of verbal episodic information.
In summary, the results of this case study suggest that the
retrosplenial region plays a role in memory functioning,
specific to the encoding of new verbal information. Our
patient was impaired in initial ‘time-tagging’ of information,
as well as semantic encoding at the level of cognitive
manipulation, and not at the level of the initial semantic
analysis. Therefore, our case is consistent with previous cases
of amnesia following retrosplenial damage, in addition to
recent functional imaging research implicating the retrosplenial cortex in verbal episodic encoding. It is noteworthy
that PET studies have cited activated voxels during verbal
episodic encoding in Brodmann’s (BA) 29, as well as in
areas 23 and 31, which the authors refer to as the retrosplenial
portion of the posterior cingulate gyrus (Shallice et al., 1994;
Fletcher et al., 1995). Our patient’s most extensive damage
was to BAs 29 and 30, although the damage extended more
anteriorly and superiorly into BAs 23 and 31. Brodmann
describes areas 29 and 30 as human retrosplenial cortex
[Brodmann (1994) cited in Morris et al. (2000)], which
is most consistent with the location of our lesion. Other
researchers, however, also include areas 23 and 31 as part
of retrosplenial cortex (Fletcher et al., 1995; Maddock, 1999).
Therefore, our patient’s damage did overlap with activations
from PET research, although expert opinions are divided as
to whether activations in areas 23 and 31 should be defined
as in the retrosplenial cortex (Vogt et al., 2000).
Valenstein et al. (1987) suggested that lesions of the
cingulum and retrosplenial cortex may cause amnesia by
disrupting input from the anterior nucleus of the thalamus
to the hippocampus. Von Cramon and Schuri (1992) further
proposed that disruption of a septal–hippocampal route
including the posterior cingulum, the supracommissural
hippocampus and the proximal fornix may contribute to cases
involving more persistent forms of amnesia. Additionally,
Rudge and Warrington (1991) have argued that amnesia may

416 C. R. McDonald et al.

result from a disconnection between the frontal and temporal
lobes, due to fornix damage, which prevents the cognitive
mediational memory system (within the frontal lobes) from
providing meaningful input to the medial temporal memory
system.
The results of our study suggest that retrosplenial damage
causes severe and persistent anterograde amnesia, characterized by poor temporal ordering and semantic encoding
in the absence of identifiable frontal lobe pathology. While
a frontotemporal disconnection may have contributed to our
patient’s verbal amnesia, he performed normally on tests of
frontal executive ability which are believed to rely on
dorsolateral pre-frontal cortex (Fuster, 1997). Second, our
patient demonstrated poor recognition memory more
characteristic of patients with medial temporal dysfunction
(Scoville and Milner, 1957) than of patients with frontal lobe
impairment. Third, LC did not benefit from category cueing
as would be expected following frontal lobe pathology or
possibly from a frontotemporal disconnection. Therefore, we
feel that LC’s amnesia more likely reflects a disconnection
of diencephalic from medial temporal structures, similar to
the mechanisms described by Valenstein et al. (1987) in their
original paper.
Other factors may have also contributed to LC’s severe
anterograde amnesia. It is possible that LC’s alcohol use
contributed largely to the memory impairments observed in
the first evaluation. This explanation, however, is unlikely as
his profile closely resembles the profiles from other case
reports of patients with retrosplenial damage (Valenstein
et al., 1987; Correa et al., 1999). Furthermore, his verbal
memory deficits were markedly impaired at the time of the
second and third evaluations, despite reported long-term
abstinence from alcohol. As mentioned, frontal lobe involvement cannot be ruled out as dorsolateral pre-frontal cortex
is richly connected with medial temporal structures and
retrosplenial cortex and may have contributed to features of
our patient’s amnesic syndrome.
Finally, we cannot completely rule out the possibility that
slight damage to the splenium did not contribute to LC’s
memory impairment. Although the fornix was intact bilaterally, there was mild damage to the most lateral extent of
the splenium on the left. Few studies of memory impairment
following splenial damage, however, have been reported in
the literature. Rudge and Warrington (1991) studied nine
patients with splenial tumors and reported severe memory
deficits, as well as impairments in visual perception, in all
patients. This study was limited, however, by significant
anatomic variability in the placement of the tumors and the
authors concluded that it was actually damage to the fornix
subjacent to the splenium that accounted for the persisting
amnesia. To our knowledge, no studies have been reported
in the literature of amnesia following damage restricted to
the splenium of the corpus callosum. Additionally, damage
to the splenium in our patient was very minimal, making this
hypothesis even more unlikely.
It is important to highlight that the fornix was intact

bilaterally in our patient. This information is paramount as
the fornix, which has often been implicated in memory
functioning itself, is most often damaged in patients with
retrosplenial lesions. Fornix involvement in memory would
seem likely as this structure forms a vital bridge between
medial temporal and medial diencephalic regions implicated
in anterograde amnesia (Aggleton and Brown, 1999).
Research from the animal literature (Squire and Zola-Morgan,
1991), as well as from the human clinical literature
(Garcia-Bengochea and Friedman, 1987), however, has not
provided convincing evidence that fornix damage produces
amnesia. In fact, Aggleton and Brown (1999) provide a
review of the animal and human literature and conclude that
the number of studies of fornix damage with intact memory
functioning far outnumber the cases in which fornix damage
has been associated with amnesia. Other researchers, however,
have found evidence of memory impairments in patients with
fornix damage (Heilman and Sypert, 1977; Hodges and
Carpenter, 1991). The present case provides evidence that
not all of the memory deficits observed following retrosplenial
damage can be attributed to damage of the fornix. Nevertheless, additional case reports of patients with circumscribed
retrosplenial damage, as well as additional functional imaging
research are needed to provide more conclusive evidence
regarding the role of retrosplenial cortex in memory and
cognition.

Acknowledgement
This study was supported by NS93211, MH54632, P50MH32384 and CRC RR00082.

References
Aggleton JP, Brown MW. Episodic memory, amnesia, and the hippocampal–
anterior thalamic axis. Behavioral and Brain Sciences 1999; 22: 425–44.
Arita K, Uozumi T, Ogasawara H, Sugiyama K, Ohba S, Pant B et al. A case
of pineal germinosa presenting with severe amnesia. No-Shinkei-Geke
1995; 23: 271–5.
Barr WB, Chelune GJ, Hermann BP, Loring DW, Perrine K, Strauss E et al.
The use of figural reproduction tests as measures of nonverbal memory in
epilepsy surgery candidates. Journal of the International Neuropsychological
Society 1997; 3: 435–43.
Bowers D, Verfaellie M, Valenstein E, Heilman KM. Impaired acquisition of
temporal information in retrosplenial amnesia. Brain and Cognition 1988;
8: 47–66.
Butters N, Cermak L. Alcoholic Korsakoff’s syndrome: an informationprocessing approach to amnesia. New York: Academic Press, 1980: 69–95.
Cammalleri R, Gangitano M, D’Amelio M, Raieli, V, Raimondo D, Camarda R.
Transient topographical amnesia and cingulate cortex damage: a case report.
Neuropsychologia 1996; 4: 321–6.
Correa DD, Relkin NR, Ruge MI, Schiff N, DeAngeles LM, Hirsch J.
Retrosplenial amnesia: a multimodal investigation. Poster session presented
at the annual meeting of Cognitive Neuroscience, Washington, DC, 1999.
Craik F, Lockhart R. Levels of processing: A framework for memory research.
Journal of Verbal Learning and Verbal Behavior 1972; 11: 671–84.
Crovitz HF, Schiffmann H. Frequency of episodic memories as a function of
their age. Bulletin of Psychological Society 1974; 4: 517–8.
Damasio H, Damasio AR. Lesion analysis in neuropsychology. New York:
Oxford University Press, 1989.
Delis DC, Kramer JH, Kaplan E, Ober BA. California Verbal Learning Test,
research edition. Manual. San Antonio: Psychological Corporation, 1987.
Dolan RJ, Fletcher PC. Dissociating prefrontal and hippocampal function
in episodic memory encoding. Nature 1997; 388: 582–5.

Retrosplenial amnesia 417
Fletcher PC, Frith CD, Grasby PM, Shallice T. Brain systems for encoding
and retrieval of auditory verbal memory: An in vivo study in humans. Brain
1995; 118: 401–16.
Fuster JM. The prefrontal cortex: anatomy, physiology, and neuropsychology
of the frontal lobe. New York: Lippincott-Raven, 1997.
Gainotti G, Almonti S, Di Betta AM, Silveri MC. Retrograde amnesia in a
patient with retrosplenial tumour. Neurocase 1998; 4: 519–26.
Garcia-Bengochea F, Friedman WA. Persistent memory loss following
section of the anterior fornix in humans. Surgical Neurology 1987; 27:
361–4.
Halstead WC. Brain and intelligence. Chicago: University of Chicago Press,
1947.
Heilman KM, Sypert GW. Korsakoff’s syndrome resulting from bilateral
fornix lesions. Neurology 1977; 27: 490–3.
Hodges JR, Carpenter K. Anterograde amnesia with fornix damage following
removal of IIIrd ventricle colloid cyst. Journal of Neurology, Neurosurgery,
and Psychiatry 1991; 54: 633–8.
Iwasaki S, Arihara T, Torii H, Hiraguti M, Kitamoto F, Nakagawa A et al.
A case of splenial astrocytoma with various neuropsychological symptoms
(in Japanese). No-To-Shinkei 1993; 45: 1067–73.
Katai S, Maruyama T, Hashimoto T, Yanagisawa N. A case of cerebral
infarction presented as retrosplenial amnesia. Rinsho-Shinkeigaku 1992;
32: 1281–7.
Maddock RJ. The retrosplenial cortex and emotion: new insights from
functional neuroimaging of the human brain. Trends in Neuroscience 1999;
22: 310–6.
Morris R, Paxinos G, Petrides M. Archetectonic analysis of the human
retrosplenial cortex. Journal of Comparative Neurology 2000; 421: 14–28.
Osterrieth R. Le test de copie d’une figure complexe. Archives de Psychologie
1944; 30: 206–356.
Rudge P, Warrington EK. Selective impairment of memory and visual
perception in splenial tumours. Brain 1991; 114: 349–60.
Sato K, Sakajiri K, Komai K, Takamori M. A patient with amnestic syndrome
with defective route finding due to left posterior cerebral artery territory
infarction. No-To-Shinkei 1998; 50: 69–73.
Scoville WB, Milner B. Loss of recent memory after bilateral hippocampal lesions. Journal of Neurology, Neurosurgery, and Psychiatry 1957;
20: 11–21.
Shallice T, Fletcher P, Frith CD, Grasby P, Frackowiak SJ, Dolan RJ. Brain
regions associated with acquisition and retrieval of verbal episodic memory.
Nature 1994; 368: 633–5.
Shimamura A. Memory and frontal lobe dysfunction. In: Gazzaniga M, editor.
The cognitive neurosciences. Massachusetts: MIT Press, 1995: 803–13.
Spreen O, Strauss E. A compendium of neuropsychological tests, 2nd edn.
New York: Oxford University Press, 1998.
Squire LR. Comparisons between forms of amnesia: Some deficits are unique
to Korsakoff’s syndrome. Journal of Experimental Psychology: Learning,
Memory, and Cognition 1982; 8: 560–71.
Squire LR, Zola-Morgan S. The medial temporal lobe memory system. Science
1991; 253: 1380–6.
Takahashi N, Kawamura M, Shiota J, Kasahata N, Hirayama K. Pure
topographic disorientation due to right retrosplenial lesion. Neurology 1997;
49: 464–9.
Takayama Y, Kamo H, Ohkawa Y, Akiguchi I, Kimura J. A case of retrosplenial
amnesia. Rinsho-Shinkeigaku 1991; 31: 331–3.
Tulving E, Markowitsch HJ, Kapur S, Habib R. Novelty encoding networks
in the human brain: Positron emission tomography data. Neuroreport 1994;
5: 2525–8.
Valenstein E, Bowers D, Verfaellie M, Heilman KM, Day A, Watson RT.
Retrosplenial amnesia. Brain 1987; 110: 1631–46.
Verfaellie M, Cermak LS. Neuropsychological issues in amnesia. In: Martinez
JL, Kesner RP, editors. Learning and memory: a biological view, 2nd edn.
San Diego: Academic Press, 1991.
Vogt BA, Absher JR, Bush G. Human retrosplenial cortex; where is it and is
it involved in emotion. Trends in Neuroscience 2000; 23: 195–7.
Von Cramon DY, Schuri U. The septo-hippocampal pathways and their
relevance to human memory: A case report. Cortex 1992; 28: 411–22.
Wechsler D. Wechsler Adult Intelligence Scale-Revised. Manual. Cleveland:
Psychological Corporation, 1981.
Wickens D. Encoding categories of words: An empirical approach to meaning.
Psychological Review 1970; 77: 1–15.
Yasuda Y, Watanabe T, Tanaka H, Tadashi I, Akiguchi I. Amnesia following
infarction in the right retrosplenial region. Clinical Neurology and
Neurosurgery 1997; 99: 102–5.

Received on 5 September, 2000; resubmitted on 21 May, 2001;
accepted on 8 June, 2001

Verbal encoding deficits in a patient with
a left retrosplenial lesion
C. R. McDonald, B. Crosson, E. Valenstein and
D. Bowers
Abstract
Over the past decade, memory impairments associated with retrosplenial
damage have received increased attention among neuroscientists, although the
exact role of the retrosplenial region in memory has not been clearly defined.
Evidence from lesion studies and functional neuroimaging has implicated
the retrosplenial region in verbal episodic memory, temporal ordering of
information, and topographical memory. In addition, recent positron emission
tomography studies have shown increased activation of the retrosplenial cortex
during tasks involving both the encoding and retrieval of episodic information.
The objective of this study was to define more clearly the nature of memory
impairments observed in retrosplenial amnesia. A 47-year-old amnesic male
with a left retrosplenial arteriovenous malformation was examined on
neurocognitive tasks of automatic and directed encoding, temporal ordering
of information, and remote memory. Despite normal performance on frontal
cognitive tasks, intact memory for remote information, and a superior IQ, this
individual exhibited a profound deficit in the encoding of information,
evidenced by poor release from proactive interference, poor category clustering
on word list recall, poor semantic encoding on a levels of processing task,
and mild impairments in temporal ordering. These results imply that the
retrosplenial region plays a role in the verbal encoding of information, which
contributes to the profound verbal memory impairment reported in previous
case studies of patients with retrosplenial damage.

Journal
Neurocase 2001; 7: 407–17

Neurocase Reference Number:
O232

Primary diagnosis of interest
Verbal anterograde amnesia

Author’s designation of case
LC

Key theoretical issue
d Verbal encoding deficit following damage to the left retrosplenial cortex
d Episodic memory impairments resulting from disruption of Papez circuit

Key words: retrosplenial amnesia; verbal encoding deficit; Papez circuit;
episodic memory

Scan, EEG and related measures
Magnetic resonance imaging (T1- and T2-weighted sequences)

Standardized assessment
Wechsler Adult Intelligence Scale-Revised (WAIS-R), Boston Naming Test,
Trail Making Test (A and B), Stroop Interference Test, Halstead–Reitan
Booklet Category Test, Auditory Consonant Trigrams, Wechsler Memory
Scale-Revised (WMS-R; selected subtests), California Verbal Learning Test
(CVLT), Rey–Osterrieth Complex Figure Test (ROCFT)

Other assessment
Release from proactive interference (Wickens, 1970), levels of processing
(LOP; Craik and Lockhart, 1972), Kortenkamp Remote Memory Test (KRMT),
Crovitz Remote Memory Test (CRMT; Crovitz and Schiffmann, 1974)

Lesion location
d Left retrosplenial area and subadjacent white matter, left splenium

Lesion type
Hemorrhagic arteriovenous malformation (AVM)

Language
English