Memory loss due to transient hypoperfusion
.in. the medial temporal lobes including
hippocampus
Tanabe H, Hashikawa K, Nakagawa Y, Ikeda M, Yamamoto H, Harada
K, Tsumoto T, Nishimura T, Shiraishi J, Kimura K. Memory loss due to
transient hypoperfusion in the medial temporal lobes including
hippocampus
Acta Neurol Scand 1991: 84: 22-27.
A typical case of transient global amnesia (TGA) was investigated with
single photon emission computed tomography (SPECT) and magnetic
resonance imaging (MRI) with 1.5-tesler scans. During the amnesic eposide,
a marked decrease of cerebral blood flow was observed in the areas
confined to the territory of the bilateral posterior cerebral arteries including
the hippocampus. After the episode, cerebral blood flow returned to
normal and a circumscribed lesion was detected in the middle portion of
CA 1 field of the left hippocampus. The SPECT findings prove direct evidence
that the medial temporal structures are involved in the establishment of
new memories, as well as in process of recalling only recently acquired
memories, but not in retrieval of memories acquired long ago. The MRI
findings indicate that a unilateral partial damage to CA 1 sector of the
hippocampus does not develop a definite memory impairment and that
high-resolution MRI study on the hippocampus is necessary in TGA patients.

Memory is suspected to be processed in certain localized areas of the brain. Particularly, the medial
temporal lobe including the hippocampus is considered to be a critical structure for formation of memory (1). Transient global amnesia (2) is a wellrecognized clinical entity characterized by a severe,
but transient, inability to form new memories without other higher nervous dysfunctions (3). After the
amnesic episode, patients recover completely so as
to tell examiners what happened to them, this being
different from the reported cases (4), who have permanent memory impairment due to irreversible brain
damage. In addition, “control tests” are possible in
TGA patients, since we can give them the same
memory tests during and after the episode. Thus,
visualization of impaired or inactive sites during an
episode of TGA may give a unique opportunity to
localize the sites responsible for the memory-related
processes in the human brain. However, the attack
takes place unexpectedly and fades quickly so that
it is a rare chance to carry out computer-assisted
imaging of the brain using radioactive substrates
during the episode. In fact, there have been no full
reports which performed such imaging techniques
during the episode, while there exist some reports
which performed comprehensive neuropsychological evaluation during the episode (5-10). Here we
22

H. TanabelS3,K. Hashikawa’,
Y. Nakagawa3, M. Ikeda’, H. Yamamoto’,
K. Harada4, T. Tsumoto’,
T. Nishimura’, J. Shiraishi3, K. Kimura’

’

Departments of Neuropsychiatry, Radiology,
* Nuclear Medicine, Neurophysiology,
Biochemical Research Center, Osaka University
Medical School, Faculty of Health and
Sport Sciences, Osaka University, Japan

Key words: transient global amnesia; medial
temporal lobe; cerebral blood flow; single photon
emission computed tomography; hippocampus;
magnetic resonance imaging; memory
Hirotaka Tanabe, Department of
Neuropsychiatry, Osaka University Medical
School, Fukushima, Osaka 553, Japan
Accepted for publication December 6, 1990

report an observation of a TGA patient with single
photon emission computed tomography (SPECT)
scans during and after an amnesic episode.
Case report

The patient is a 63-year-old, right-handed woman
without vascular risk factors expect for labile hypertension. Her past history has been otherwise unremarkable. For a few days before the attack of TGA,
she had been fatigued due to overwork resulting in
a dull headache. Her episode of TGA is reported as
follows: On the morning of February 8 around 7
AM, 1989, she became suddenly perplexed and repeatedly asked her son the same questions over and
over despite repeated explanations. For example,
she asked him where her husband was, evidently not
remembering that he had been hospitalized with a
fracture of his foot one month before. Her speech
and behaviour were otherwise normal. At that point,
her son, who is a medical doctor, took her immediately to our hospital. On the way to our hospital by
car, she repeatedly demanded where they were going,
despite the fact that she was always given the same
reply. At 10 AM, clinical examination was performed. General and neurological examinations were
normal except for her blood pressure of 170/

Transient global amnesia
100 mmHg. The patient was disoriented in time but
not space and person. She showed complete inability to form any new memories, but no confabulation
was noted. For instance, she could readily name
three objects, but could recall none of the objects
after distraction or a delay of a couple of minutes.
Moreover, she could not remember even the fact that
she had seen them. Thus, the patient could not recall nor recognize the events which she had just
experienced. In contrast to such severely impaired
recent memory, she gave a good account of events
from the remote past at least several years before the
attack. Also, her immediate memory was preserved:
digit spans were seven forward and five backward.
She could execute serial subtractions of the digit “7”
rapidly and flawlessly, and could abstract proverbs
satisfactorily. Social interaction and affective behaviour were also normal. Except for memory, all other
clinically evaluated higher cortical functions were
normal.
The acute phase lasted until 11:30 AM, when she
felt as if she had awoken and she began to partially
remember events as they were occurring. At this
point, she was concerned about what had happened
during the past four and a half hours and could
successfully perform the three object recalling test.
However, the tasks such as story recall and paired
associate leaning still disclosed a definite impairment of recent memory; these tasks were normally
performed when she was re-tested one week later
(Index scores ofthe Wechsler Memory Scale-Revised
(1 l), performed several months afterwards, were
normal: Attention-Concentration, 117; Verbal Memory; 110; Visual Memory, 116; General Memory,
112; and Delayed Memory, 108). Thus the severe
anterograde amnesia began to regress at 11:30 AM
and her memory function appeared totally recovered
at 3 PM about eight hours after the onset of TGA.
Her memory for the events which occurred from
11:30 AM to 3 PM was patchy and vague. She recovered completely from the retrograde amnesia,
which had covered at least 10 months with fragments of memory during the acute phase; she could
recall all of the events which had happened previous
to retiring for the night before. At 4 PM, she went
to her husband’s office and finished her work as
usual.
A computed tomographic (CT) scan, performed
at 11:15 AM during the episode, revealed no abnormalities. Electroencephalogram (EEG) and external
cerebrovascular Doppler ultrasound examinations,
carried out one week later, were normal. SPECT
using
N-isopropyl( ‘231)-P-iodoamphetamine(IMP-SPECT) scans (12) were performed both during and after the TGA episode (Fig. 1 and Table 1).
During the episode, a marked decrease of cerebral
blood flow was observed in the areas confined to the

Table 1. 1-123 IMP distribution in the brain during & after TGA attack
After

1 week

During
_
Whole brain

54

53

5 month

_

_

~

rt

It

rt

It

_
rt

56

_
It

Slice 1

Cerebellum

84

85

82

79

73

80

Slice 2

Temporal
Medial
Lateral

47
61

45
64

59
62

61
60

67
75

71
73

88
65
68
71

85
69
63
70

74
70
74
67

80
73
71
66

81
16
77
75

79
74
73
74

Slice 3

Thalamus
Frontal
Occipital
Temporal
(lateral)

(average counts/voxel/administration dose (mCi))
For quantitative analysis, 12 mm thick 3 slices were made by summation of original
4 mm thick consecutive 3 slices, respectively. Slice 1 corresponds to the level of the
cerebellum, slice 2 that of the midbrain, and slice 3 that of the basal ganglia. Fourteen
symmetrical regions of interest (16 mmX 16 mm) were analyzed, 2 in the slice 1 for the
cerebellar hemisphere, 4 in the slice 2 for the lateral and medial portions of the temporal
lobe, and 8 in the slice 3 for the frontal lobe, lateral portion of the temporal lobe, the
occipital lobe and the thalamus. The average counts for 4 mm cubic voxel were calculated for each region of interest. For the correction of the small difference of 1-123 IMP
administration doses, each value was divided by administration dose (mCil.

territory of the bilateral posterior cerebral arteries
including the hippocampal formation (arrows in the
left column of Fig. 1). After the episode, this abnormal finding disappeared and cerebral blood flow
(CBF) returned to normal (An IMP-SPECT scan,
performed 5 months later, showed almost the same
flow distribution images as just after TGA). Thus,
we demonstrated that a transient reduction of flow
in the both medial temporal lobes including the hippocampal formation took place during an episode of
TGA.
A magnetic resonance imaging (MRI) with 1.5 T
scan, carried out one month after the TGA attack,
disclosed a circumscribed lesion in the left hippocampus which could not be detected by CT scan
(Fig. 2). The brain were cut in 5 mm-thick horizontal
and coronal sections. The lesion was observed on
only one horizontal and coronal slice. By reference
to detailed atlases of the hippocampus (13-15), the
lesion was found to be situated in the approximately
middle portion of the CA 1 field.
Discussion

TGA occurs in subjects of both sexes, mainly among
those aged 50years or more, and the amnesic episode is characterized as follows (16, 17). Patients
suffer sudden anterograde and retrograde memory
loss without other neurological signs or symptoms.

23

_

Tanabe et al.

r ebellum

Occipital

,Frontal

’0cci pit a 1
.*

Fig. 1. SPECT using N-i~opropyl(”~I)-p-iodoamphetamine
(IMP-SPECT) during and after the episode of TGA. The images taken
during the episode are shown on the left, those taken one week after the attack on the middle and schematic diagrams of three afterTGA images on the right. Data acquisitions were carried out with a newly developped 4-head rotating gamma camera SPECT with a
low energy high resolution collimator (12). The spatial resolution was 9.8 mm in full width at half at center. Data acquisition were started
10 min. after the intravenous injection of about 6 mCi 1-123 I M P and continued for 35 min. The raw data were reconstructed using a
standard filtered-back projection algorithm with a pre-treatment by Wiener filter. The attenuation correction were performed by correction method. Using the set of 2 light projectors for position setting, we could obtain almost the same anatomical slices for two studies.
The images shown here were selected from 42 sequential slices at 4 mm pitch. All images were displayed with same colour scale. In
the images, pink corresponds to the highest level of cerebral blood flow and blue to the lower level. Compared with the afer-TGA images, the images taken at 1 PM during the episode demonstrate a marked hypoperfusion confined to the territory of the posterior cerebral arteries with preserved blood flow in other areas including the frontal cortex and thalamus. White arrows indicate areas with
reduced blood flow. The images of the second raw approximately correspond to the level of the horizontal slice of MRI (see Fig. 2).
Note the location of the medial temporal lobes including the hippocampus.

24

Transient global amnesia

A

C

Fig. 2. MRI with 1.5 T scans, carried out one month after the TGA attack, showing a small circumscribed lesion in the left hippocampus.
(A) Top: TI-weighted coronal image (TR/TE: 420/50). Bottom: T2-weighted coronal image (TR/TE: 2000/90). Arrows indicate lesions.
The rectangle outlines the region shown at magnification in C. (B) T1-weighted horizontal image (TR/TE: 420/50). The level of this image
approximately corresponds to the SPECT images of the second raw of Fig. 1. Black arrow indicates the lesion and white arrow the
hippocampal fissure (HF). (C) Top: close view of the left hippocampus. Bottom: the corresponding diagram. Big black arrows indicate
the lesion in the CA 1 field and asterisk represents it. CS, collateral sulcus; EC, entorhinal cortex; T, temporal horn of the lateral ventricle;
S, subiculum; D, dentate gyrus; CAI, Field CA 1 of the hippocampus; CA 3, Field CA 3 of the hippocampus.

They retain personal identity and are able to carry
on normal activities. However, since the patients are
unable to acquire new information, they ask repeated
questions about their immediate environment and
situation. The retrograde memory loss, extending
back for days, months or in some cases even years
before the attack, gradually or quickly shrinks as the
patients recover. No or only a brief period of permanent retrograde amnesia persists, but memory for
events during the acute phase is permanently impaired. The amnesic episodes generally last several
hours. Thus, our patient exhibited the typical features of TGA.
TGA is not a rare illness, but the etiology is not
known yet (18). The two theories proposed most
often are cardiovascular and epileptic causes. The

latter is, however, supported by few authors (19) and
two recent studies prove strong evidence against the
epileptic etiology. In the first, EEG being recorded
at onset of TGA showed no epileptic discharge (20).
In the second, EEG from scalp and sphenoidal electrodes during TGA revealed no electrophysiologic
dysfunction (21). On the other hand, there have been
no reports as yet which proved the former, cerebrovascular etiology, although most authors think that
the amnesic episode is caused by transient ischemia
affecting the both medial temporal lobes, in particular, the bilateral hippocampal formations (16, 18,
22-25).
In the present study, we performed IMP-SPECT
scannings in a typical case of TGA during and after
an episode. During the episode, a marked hypoper25

Tanabe et al.

fusion confined to the territory of the posterior cerebral arteries was observed with preserved blood
flow in other areas including the thalamus and frontal cortex. After the episode, this abnormal finding
disappeared and CBF returned to normal. Thus, we
demonstrated transient decrease of blood flow in the
both medial temporal lobes including the hippocampal formation during an episode of TGA, although
it is unclear whether the hypoperfusion developed
primarily or secondary coupled with hypometabolism caused by mechanisms not as yet elucidated,
such as acute arterial dyscontrol (17) and Leao’s
spreading depression (26).
MRI with 1.5 T scannings, performed one month
after the TGA attack, revealed a focal lesion in the
left hippocampus which could not be shown by CT
scan. The anatomical study demonstrated that the
lesion is situated in the CA 1 sector of the hippocampus. Unfortunately, it is unknown whether the lesion
had already existed before the attack or developed
due to the attack. Considering selective vulnerability or the phenomenon “delayed neuronal death”
(27) of the CA 1 field to mild ischemia in both animals (28-30) and humans (31, 32), the latter possibility seems to be more probable. It is, however,
unclear why the lesion developed only on the left
side. In any case, the MRI finding of our patient
indicates the necessity of high-resolution MRI study
on the hippocampus of TGA patients.
How could our patient have an MRI-verified lesion and no CBF-change after TGA? Usually, a
lesion has abnormal low flow. We think that it is a
matter of high resolution MRI as compared to lower
resolution SPECT (partial volume or compton scatter effect?).
In humans, there is only one case of amnesia following a lesion limited to the hippocampus (31).
This amnesic patient R.B. had an ischemic injury
that resulted in a bilateral lesion limited to the entire
field CA 1 of the hippocampus. The anterograde
amnesia of R.B. was permanent, whilst that of our
patient was transient. Hence, it appears that a unilateral partial damage to the CA1 sector does not
produce permanent disability to lay down new information. From a different viewpoint, R.B.’s recent
memory impairment is not so severe as those of our
patient and the cerebrated patient H.M., in both of
whom the bilateral temporal lobes are affected, although transiently in our patient. Consequently, as
pointed out by Kritchevsky et al. (9) and demonstrated by the present SPECT study, the area of
dysfunction in TGA must involve at least more than
just the bilateral CA 1 region of the hippocampus.
The present results indicate that the medial temporal structures including the hippocampus play an
essential role in the formation of new memories,
because the patient had no memory traces of events
26

which had happened during the attack when an hypoperfusion was observed in those structures. Also,
the results suggest that these structures in the brain
may not be involved in retrieval of remote past memory and in process of immediate memory, because
the patient had ability to recall the remote past and
immediate memory even during the attack. These
results are generally consistent with the previous
studies on patients with permanent amnesia due to
irreversible focal brain damage (33-37). In these
studies, a profound and permanent loss of recent
memory (severe permanent anterograde amnesia)
with a variable degree of permanent retrograde amnesia has been reported in patients who had bilateral
medial temporal lobe lesions due to various etiologies.
During the TGA, our patient appeared perplexed
and asked series of questions repeatedly as if to try
to orient herself, but she did not definitely realize her
inability to acquire new information, like the abovementioned patients with severe permanent anterograde amnesia. However, our patient spontaneously
described a blank in her memory precisely after having regained the ability to encode new memory, while
these patients cannot realize their memory blanks.
Moreover, in our patient, disappearance of the hypoperfusion in the medial temporal lobes corresponded to resolution of the amnesia. Hence, the
present study directly supports the hypothesis that
the medial temporal structures are involved in the
establishment of new memory as well as in process
of recalling only recently acquired memories, but not
in retrieval of memories acquired long ago.
Acknowledgements
We thank Dr. D. G. Amaral for his helpful advice

References
1. SQUIRELR. Memory and brain. New York; Oxford University Press, 1987.
2. FISHERCM, ADAMSRD. Transient global amnesia. Acta
Neurol Scand 1964: (Suppl 9): 1-83.
3. CAPLANLR. Transient global amnesia: criteria and classification. Neurology 1986: 36: 441.
4. MILNER
B, CORKIN
S, TEUBER
HL. Further analysis of the
hippocampal amnesic syndrome: 14-year follow-up study of
H.M. Neuropsychologia 1968: 6: 215-234.
5. REGARDM, LANDIST. Transient global amnesia: neuropsychological dysfunction during attack and recovery in two
“pure” cases. J Neuro Neurosurg Psychiatry 1984: 47: 668672.
6. GALLASSI
R, LORUSSOS,
STRACCIARI
A. Neuropsychological findings during a transient global amnesia attack and its
following. Ital J Neurol Sci 1986: 7: 45-49.
7. STRACCIARI
A, REBUCCI
GG, GALLASSI
R. Transient global amnesia: neuropsychological study of a “pure” case. J
Neurol 1987: 234: 126-127.
M, SQUIRE
LR, ZOUZOUNIS
JA. Transient
8. KRITCHEVSKY
global amnesia: characterization of anterograde and retrograde amnesia. Neurology 1988: 38: 213-219.

Transient global amnesia
9. KRICHEVSKY
M, SQUIRE
LR. Transient global amnesia: evidence for extensive, temporally graded retrograde amnesia.
Neurology 1989: 39: 213-218.
10. HODGES
JR, WARDCD. Observations during transient global amnesia: a behavioural and neuropsychological study of
five cases. Brain 1989: 112: 595-620.
11. WECHSLER
D. Wechsler Memory Scale-Revised. New York:
Psychological Corporation, 1987.
K, HASHIKAWA
K, ETANIH, UEHARA
A, MORI12. KIMURA
WAKI H, KOZUKA
T, ISAKA
Y, MATSUMOTO
M, KAMADA

H, TABUCHI
H. A new apparatus for brain
T, MORINAGA
imaging: four-head rotating gamma camera single photon emmision computed tomograph. J Nucl Med 1990: 31: 603-609.
R, VOOGDJ, VAN HUIJZENC. The human
13. NIEUWENHUYS
central nervous system: a synopsis and atlas. New York:
Springer-Verlag, 1978.
14. NAIDICHTP, DANIELS
DL, HAUGHTON
VM, WILLIAMS
A, POJUNAS
K, PALACIOS
E. Hippocampal formation and
related structures of the limbic lobe: anatomic-MR correlation. I. Surface features and coronal sections. Radiology 1987:
162: 747-754.
15. AMARALDG. Memory: anatomical organization of candidate brain regions. In: PLUMF ed. Handbook of physiology
(vol. 5). Bethesda MD: American Physiological Society, 1987.
LM. Transient global amnesia: and
16. ROWANAJ, PROTASS
electroencephalographic findings in 10 cases. Neurology 1979:
29: 869-872.
TS, KJAERM, MARQUARDSEN
J, DE
17. HINGEHH, JENSEN
FINEOLIVARIUS
B.The prognosis of transient global amnesia: results of a multicenter study. Arch Neurol 1986: 43:
673-616.
18. CAPLAN
LR. Transient global amnesia. In: FREDERIKS
JAM
ed. Handbook of Clinical Neurology (vol. 45). Amsterdam:
Elsevier, 1985: 205-218.
19. FISHERCM. Transient global amnesia: precipitating activities and other observations. Arch Neurol 1982: 39: 605-608.
20. COLEAJ, GLOORP, %PLAN R. Transient global amnesia:
the electroencephalogram at onset. Ann Neurol 1987: 39 605608.
KJ, LORING
DW, KINGDW, NICHOLSFT. The
21. MEADOR
P3 evoked potential and transient global amnesia. Arch Neurol 1988: 45: 465-467.
22. LOGAN
M, SHERMAN
DG. Transient global amnesia. Stroke
1983: 14: 1005-1007.
23. COLOMBO
A, SCARPAM. Transient global amnesia: pathogenesis and prognosis. Eur Neurol 1988: 28: 111-1 14.

24. MATIAS-GUIU
J, PICOM, BONAVENTURA
I, MARTINR,
MONASTERIO
J. Platelet aggregation in transient global amnesia. Ital J Neurol Sci 1989: 10: 171-174.
M, ANZALONE
N, MORABITO
A, LANDIG. A
25. GUIDOTTI
case-control study of transient global amnesia. J Neurol Neurosurg Psychiatry 1989: 52: 320-323.
J, JORGENSEN
MB. Leao’s spreading depression in
26. OLESEN
the hippocampus explains transient global amnesia: a hypothesis. Acta Neurol Scand 1986: 73: 219-220.
27. KIRINOT. Delayed neuronal death in the gerbil hippocampus following ischemia. Brain Res 1982: 239: 57-69.
28. WESTERBERG
E, MONAGHAN
DT, KALIMO H, COTMON
CW, WIELOCHTW. Dynamic changes of Excitatory amino
acid receptors in the rat hippocampus following transient
cerebral ischemia. J Neurosci 1989: 9: 798-805.
29. ARAKIT, KATOH, KOGURE
K. Selective neuronal vulnerability following transient cerebral ischemia in the gerbil: distribution and time course. Acta Neurol Scand 1989: 80: 548553.
30. AUERRN, JENSENML, WHISHAW
IQ. Neurobehavioral
deficit due to the ischemic brain damage limited to half of the
CA 1 sector of the hippocampus. J Neurosci 1989: 9: 16411647.
31. ZOLA-MORGAN
S, SQUIREL, AMARAL
DG. Human amnesia and the medial temporal region: enduring memory impairment following a bilateral lesion limited to field CA 1 of
the hippocampus. J Neurosci 1986: 6: 2950-2967.
32. PETITOCK, FELDMANN
E, PULSINELLI
WA, PLUMF.
Delayed hippocampal damage in humans following cardiorespiratory arrest. Neurology 1987: 1281-1286.
33. DEJONGRN, ITABASHI
HH, OLSONJR. Memory loss due
to hippocampal lesions: report of a case. Arch Neurol 1969:
20: 339-348.
34. BENSONBF, MARSDEN
CD, MEADOWJC. The amnesic
syndrome of posterior cerebral artery occlusion. Acta Neurol
Scand 1974: 50: 133-145.
35. CUMMINGS
JL, TOMIYASU
U, READS, BENSON
F. Amnesia with hippocampal lesions after cardiopulmonary arrest.
Neurology 1984: 34: 679-681.
C, DEROUESNE
C, SIGNORET
JL, GRAYF,
36. DUYCKAERTS
ESCOUROLLE
P. Bilateral and limited amygdalohoppocampal lesions causing a pure amnesic syndrome. Ann Neurol 1985: 18: 314-319.
37. SCOVILLE
WB, MILNERB. Loss of recent memory after
bilateral hippocampal lesions. J Neurol Neurosurg Psychiatry 1957: 20: 11-21.

21