T D G Positron Emission Computed
Tomography in a Study of Aphasia
E. Jeffrey Metter, MD,"t Claude G. Wasterlain, MD,*t David E. Kuhl, MD,"
Wayne R. Hanson, P h D , t and Michael E. Phelps, PhD"

Positron emission computed tomography (PECT) using ~HF-2-fluoro-2-deoxy-~-glucose
(FDG) was used to investigate the correlations between clinical status, anatomy (as described by CT), and metabolism in five patients with
stable aphasia resulting from ischemic cerebral infarction. Local cerebral metabolic activity was diminished in an
area larger than t h e area of infarction demonstrated by CT. I n one patient, FDG PECT revealed a metabolic lesion
that probably caused the aphasic syndrome and was not apparent by CT. T h e data suggest that reliance on C T in
delineating the extent of t h e brain lesion in aphasia or other neuropsychological defects can be misleading; FDG
PECT may provide important additional information. T w o patients with similar metabolic lesions had very different clinical syndromes, showing that even when currently available methods are combined, major gaps remain in
clinicoanatomical correlations in aphasia.
Metter EJ, Wasterlain CG, Kuhl DE, Hanson WR, Phelps ME: IXFDGpositron emission computed
tomography in a study of aphasia. Ann Neurol 10:173-183, 1981

Recent developments allow clinical evaluation of the
local cerebral metabolic rate for glucose (LCMRglc)
using 'HF-2-fluoro-2-deoxy-D-glucose (FDG) [8, 17,
30, 32, 331. This method is derived from the autoradiographic technique devised by Sokoloff et a1
[39]. FDG is transported between blood and brain by
the same carrier system as glucose. Once inside the
cell, it is phosphorylated by hexokinase; but since
FDG is a poor substrate for hexose isomerase, it cannot be transformed to fructose-6-phosphate and
further metabolized. The FDG-6-phosphate is trapped
within the cell, where its accumulation is proportional to the rate of glucose utilization. With the use
of an isotopic label (radioactive IRF),the distribution
of F D G can be quantified by positron emission computed tomography (PECT) to give maps of LCMRglc
in humans.
PECT is a noninvasive scanning method using
positron-emitting radionuclides [28, 291 and is similar to the emission computed tomographic technique
developed previously for use with single-photonemitting radionuclides [14, IS]. It produces quantitative cross-sectional images of brain radioactivity in
humans following intravenous injection of a radioactively labeled compound. T h e LCMRglc method
has been extended to humans [17, 32, 331, corrected
for dephosphorylation of FDG-(,-phosphate [8],

validated [30], and shown applicable to the study of
patients with stroke and epilepsy [16, 181.
Since the metabolic rate for glucose has been
directly correlated with neuronal firing rates 1381
and with brain function [7, 18, 30, 31, 371,
F D G PECT might add another dimension to the
many recent studies correlating aphasic syndromes
with anatomical localization as determined by computed tomography (CT) [ l l , 22, 261, pertechnetate
brain scanning [2, 10, 12, 401, angiography [35], pathology [24], and regional cerebral blood flow [ 3 7 ] .
We studied five stable aphasic patients by F D G
PECT and CT to investigate the relationships between the clinical syndrome, CT, and LCMRglc.

From the *UCLA School of Medicine, Los Angeles, CA 90024,
and the +Veterans Administration Medical Center, Sepulveda, CA

Address reprint requests to Dr Metter, Department of Neurology,
Veterans Administration Medical Center, 16111 Plummer St,
Sepulveda, CA 91343.

91343.

Methods
Five patients were selected from the stroke rehabilitation
unit at the VA Medical Center, Sepulveda, CA. In each
case, aphasia had resulted from what appeared to be a
single ischemic event, diagnosed clinically and by laboratory investigation including, where appropriate, lumbar
puncture, electroencephalogram (EEG), and isotope brain
scan. The interval between stroke and the study varied
from six weeks to fourteen years, and the patients' clinical
deficits were stable. Each patient was clinically evaluated by
one of us (E. J. M.). Patient selection was based o n availability, cooperativeness, and apparent expressivelreceptive
dichotomy. Testing consisted of a full neurological exam-

Received Aug 18, 1980, and in revised form Dec 1. Accepted for
publication Dec 19, 1980.

0364-51 34/81/080173-11$01.25 @ 1981 by the American Neurological Association

173

ination and the Boston Diagnostic Aphasia Examination
(BDAE) [GI. Each patient also had a C T scan, with and
without contrast, performed at 15 degrees to the canthomeatal line. Informed consent was obtained from the
patient or family before the study.
Methods of radionuclide preparation, positron emission
computed tomography (PECT), and data analysis are described in detail elsewhere [18, 301. PECT was performed
with the ECAT positron tomograph (ORTEC, Inc., Life
Sciences, Oakridge, T N ) [28,291. T h e spatial resolution
employed was 1.7 cm in the image plane and 1.8 cm in the
axial direction.
Five to 10 mCi of FDG was injected intravenously, and
blood samples were obtained over time from a hand vein to
determine FDG and glucose concentrations [301. Six brain
levels were scanned sequentially, parallel to the canthomeatal plane, from upper cerebellum to above the cerebral ventricles starting 4 0 minutes after injection. A normal
scan is presented in Figure 1; i t represents the radioactivity
measured in each area (i.e., IHFtissue concentrations). Calculation of local cerebral metabolic rate for glucose
(LCMRglc), in milligrams per 100 mg of tissue per minute,
required measurement of the time course of plasma '*F activity, glucose concentration, and local IXFtissue concentration. LCMRglc was calculated using the model of
Huang, Phelps, and associates with the rate constants determined from normal young adults [8, 18, 301. Regions of
interest were selected on the tomograph display screen,
local values of I8F tissue concentration or LCMRglc were
determined, and comparisons were made between selected
zones and corresponding contralateral zones. T h e final
value was expressed as percentage difference between the
corresponding areas.
Areas of abnormality were determined from F D G and
CT. These lesions were then plotted on templates prepared
from the atlas of Matsui and Hirano 12 11. The templates for
C T utilized cross sections at 15 degrees to the canthomeatal line, while those for F D G were at zero degrees
to the line. These templates represent our best estimate of
lesion extent as determined from the original scans. For
selected areas from the F D G tomographs, differences were
calculated for LCMRglc.

Results
Patient 1
A 48-year-old man with a previous history of hypertension and adult-onset diabetes mellitus had sudden onset of
right-handed weakness and aphasia. When studied six
weeks later, he demonstrated mild right central facial
weakness, mild weakness and incoordination of his right
hand, normal sensation to pain, touch, vibration, and position, and mild astereognosis on the right. He had marked
difficulty in following pictorially presented directions and
in constructional tasks. He had no neglect, right-left
disorientation, o r finger agnosia. His hypertension was
adequately controlled with a diuretic and clonidine. His
diabetes was being treated by diet alone. An EEG demonstrated bifrontal intermittent delta activity, more prominent on the left, suggestive of a deep lesion. Pertechnetate
brain scan, skull roentgenogram, electrocardiogram, and
routine laboratory evaluation were normal.
174

Annals of Neurology

Vol 10 N o 2

August 1981

F i g 1 . Normal FDG PECT. The scans were obtained at zero
degrees to the canthomeatal line in a normal individual (as
described under Methods) and depict tissue "FDG concentrations. The templates on the left are traced drawings of
zero-degree sections from Matsui and Hirano [21]with their
labeling changed t o letters. These templates were used for all
FDG PECT. (Templates derived from Matsui T , Hirano A:
An Atlas of the Human Brain for Computerized Tomography.
Tokyo, Igaku-Shoin, 1978.)
For this and .tubsequent figures, structures are labeled as
follows: A, lateral occipital gyrus; B, middle temporal gyrus;
C, superior temporal gyrus; D, sylvian jissure; E, supramarginal gyms: F, postcentral gyrus: G, calcarine sulcus: H, precentral gyrus: I, inferiorfrontal gyrus; J, middle frontal gyms:
K, superior frontal gyrus; L, cingulate gyrus: M, caudate nucleus: N, precuneus; 0,thalamus; P, splenium; Q globus pallidus; R, inferior temporal gyrus; s, angular g.yrus; T, lingual
gyrus: U, fusiform gyrus; V, pars opercularis; W, pars triangularis; X, cuneus; Y , body of corpus callosum: Z, internal
capsule; AA, putamen; BB, hippocampal gyrus; CC, occipital
lobe, DD, insula; EE, Ammon's horn.

Language evaluation demonstrated good comprehension
of conversational speech and single words but difficulties in
understanding complex messages and relational terms.
Repetition of single words and phrases was good. Reading
comprehension was good, but when reading aloud he
would omit or substitute words. Writing was adequate for
one-syllable words, but he misspelled more complex
words. He spoke in sentences of adequate length and complexity but had word-findings difficulties, occasionally perseverated, and often omitted the ends of words. O n the
BDAE (Table), he had moderately severe aphasia but did
not fit any specific pattern.
C T demonstrated two punctate areas of decreased density in the left caudate nucleus and possibly the left internal capsule, and one punctate area in the right caudate (Fig
2). The overlying cortex appeared normal. FDG PECT (Fig
3) demonstrated a 24% decrease in LCMRglc in the left

Individual Scores on the Boston Diagnostic
Aphasia Examination
Patient No.
Determination

Ranee

1
~

Severity rating
Articulatory rating
Phrase length
Verbal agility
Word discrimination
Body part
identification
Commands
Complex material
Responsive nami ng
Confrontation
naming
Animal naming
Body part naming
Word reading
Sentence reading
Repetition (words)
Repeti tion-high
probability
Repetirion-low
probability
Neologisms
Literal paraphasias
Verbal paraphasias
Extended
paraphasias
Reading-s ymbols
Word recognition
Comprehension of
oral spelling
Word-picture
matching
Reading sentences
and paragraphs

2

3

4

5
4
6
6
11
30

~

0-5
1-7
1-7
0-14
0-72

3
2
7
7
70

6
52

1
9
72

1
3
2
6
45

0-20

19

6

19

19

16

0-15
0-12
0-30

12
8
26

3
1
11

14
6
18

6
5
16

8
6
19

0-105

97

74

89

60

82

0-23
0-30
0-30
0-10
0-10
0-8

5
26
30
7
10
7

4
15
20
0
8
0

6

3

19
23
3
8
2

6
27
4
9

9
21
30
4
9

6

5

0-8

4

0

1

5

2

12
3
2

2
3
17
0

1
16
1
0

1
2
8
0

0-12
0-16
0-24
0-16

2

2

6
6

3

0-10
0-8
0-8

10
7
5

8
6
1

10
8
6

9
5
2

10
8
3

0-10

10

6

10

2

10

0-10

7

2

7

3

6

F i g 2. (Patient 1 ) CT demonstrates lucent areas i n both internal capsules and the ldt caudate. All C T scans were obtained at
15 degrees to the canthomeatal line. The templates are directly
from Matsui and Hirano [211with labeling changed t o letters
as described i n Figure I . These templates will be used for all
C T . (Templates derivedfrom Matsui T , Hirano A : A n Atlas
of the Human Brain for Computerized Tomography. Tokyo,
Igaku-Shoin, 1978.)

thalamus and major depression in the left striatum. In addition, there was a 15% decrease in the left middle and
inferior temporal gyri.

This patient demonstrated many of
the characteristics of a thalamic aphasia, with anomia, perseveration, and sparing of comprehension.
Thalamic lesions have been associated with several
clinical presentations of aphasia [ 5 , 19, 25, 34, 361.
Acutely, most of these patients demonstrate reduction of spontaneous speech, anomia, perseveration,
trailing off of words, decreased voice volume, and
paraphasias. Repetition is intact, and variable changes
are noted in comprehension. This syndrome resolves
over weeks to months, with improvement in spontaneous speech and frequently clearing of the entire
syndrome. Thalamic aphasia has many features in
common with transcortical [ 51 or anomic aphasia. Lesions in cortical gray or subcortical white matter are
believed responsible for transcortical and anomic
aphasias, while most reported cases of thalamic
aphasia have been secondary to hemorrhagic thalamic
lesions with no evidence of cortical involvement [5,
11, 25, 34, 361.
COMMENTS.

Metter et al: LCMRglc in Aphasia

175

Fig 3. (Patient I ) FDG PECT thows arear of decreared FDG
concentration in the left thalamur and left temporul lobe, par-

ticularly the middle and inferior temporal gyri

Our patient demonstrated the features of a
thalamic aphasia except for the expected reduction of
spontaneous speech and decreased voice volume.
This difference may be due to the more limited extent of the lesion than usually occurs in hemorrhage.
The more prominent metabolic suppression in the
thalamus and striatum as compared to cortex suggested that these deeper regions were the site of the
primary lesion. While an independent cortical lesion
cannot be ruled out, we suggest that the decrease in
cortical metabolic rate may be secondary and may indicate a relationship between these subcortical and
cortical areas in the genesis of thalamic aphasia. This
is consistent with the suggestion of Cappa and Vignolo [ S ] that “the linguistic role of the thalamus is
related to that of the marginal language areas” rather
than the “central language zones,” which are perisylvian in location.
Pdtient 2
A 57-year-old, right-handed, bilingual Mexican-American
man had sudden onset of right hemiparesis and aphasia. He
was competent in English, had been educated in the United
States, and before his illness worked as an engineer, past
medical history included hypertension, gout, and medullary sponge kidney. Examination ten months after onset
demonstrated an incomplete right homonymous hemian-

176 Annals of Neurology

Vol 10 NO 2

August 1981

opia, right central facial weakness, minimal drift of his
right hand, hyperactive reflexes on the right with no
Babinski sign, a mild right hemisensory loss, and normal
blood pressurc
Language evaluation in English demonstrated difficulty
in understanding more than simple spoken statements and
simple conversation. H e was able to carry out one-part instructions but was unable to follow more complex ones. He
could read single letters and match printed words to pictures but was unable to follow written directions. H e was
unable to write words except by copying. He spoke English
fluently in complete, complex sentences; however, much of
the speech lacked content and meaning. Paraphasia was
only occasionally noted and there were few, if any,
neologisms. Word-finding difficulties were frequent. He
was able to ask and answer simple questions and could
carry on a limited conversation. H e had an orofacial
apraxia and early in his course was unable to repeat syllables or words modeled for him. The BDAE (Table) demonstrated a pattern consistent with severe Wernicke’s
aphasia. The patient was noted to have similar problems in
Spanish.
C T showed a large perisylvian infarction involving much
of the territory of the middle cerebral artery (Fig 4 ) . The
precentral gyrus appeared to be partially spared, while
damage was present in the pars triangularis, Broca’s area,
the temporal and parietal areas, and the underlying white
matter. Higher and lower sections showed damage to most
of the parietal and much of the middle and inferior temporal gyri. FDG PECT demonstrated a 79% reduction in
LCMRglc in the left parietal and posterior temporal regions
(Fig 5 ) . A t the level of Rroca’s area the frontal lobes
showed 65% reduction of LCMRglc. At levels higher than
the pars triangularis the frontal lobe appeared normal. Depression of FDG was noted also in the left occipital area
and basal ganglia. The left thalamus showed 44p, reduction
in LCMRglc but appeared intact by CT.
COMMENTS. Based on the extensive structural and
metabolic changes throughout the major language
areas, including Broca’s, this patient would be expected to have a mixed aphasia with marked expressive difficulties, yet his speech was fluent. His fluency
might be explained by preservation of function in the
island of tissue in the vicinity of the precentral gyrus
noted on CT and FDG PECT, but one wonders how
this area could be connected to other cortical regions
since it appears to be entirely surrounded by infarcted tissue. Alternatively, fluency could be based
on altered cortical organization, with greater right
hemisphere involvement, related to his bilinguality
[I].
Decreased metabolic activity was present in the
basal ganglia and thalamus (intact by CT), reflecting
either reduction in the activity of thalamocortical cirCuitS O r retrograde degeneration of thalamic neurons.
Cortical areas adjacent to the infarction also demonstrated some depression of LCMRglc, suggesting loss
of corticocortical interactions.

F i g 4. (Patient 2) C T demonstrates a large left perisylvian
infarct. The bifrontal hypodensity noted o n the lowest scun ZJ
artifactual. (Templates derived from Matsui T , Hirano A: A n
Atlas of the Human Brain for Computerized Tomography,
Tokyo, Igaku-Shoin, 1978.)

Patient 3
A 61-year-old, right-handed man with a history of rheumatic heart disease and mitral stenosis had sudden onset of
a right hemiplegia and aphasia in 1964. At the time of examination in 1978 he had a spastic right hemiplegia, increased stretch reflexes, Babinski sign, and decreased sensation to pain, touch, position, and vibration on the right.
He walked with a cane and was independent in activities of
daily living.
Language evaluation demonstrated mild impairment in
auditory comprehension. Reading and writing were moderately impaired. His speech was nonfluent and telegraphic
and was characterized by word-finding difficulties, paraphasias, and moderate orofacial apraxia. The BDAE
(Table) demonstrated Broca’s aphasia of marked severity.
CT (Fig 6) showed a large area of decreased density in-

Fzg 5. (Patient 2) F D G PECT demonstrates extensive decrease
i n F D G concentration i n much of the left hemi@here, including major tubcortical structures.

volving the left temporal, parietal, and frontal regions and
the basal ganglia. The involvement in the temporal lobes
seemed to spare much of the middle and inferior temporal
gyri, while extensive damage to the inferior parietal lobule
extended through all higher sections (not shown). The
frontal involvement included much of the posterior aspect
of the inferior frontal gyrus but spared the middle and
superior frontal gyri. F D G PECT (Fig 7) correlated well
with CT, showing 735% reduction of LCMRglc in the frontotemporal area. Decreased LCMRglc was noted in the left
thalamus (635% reduction), where no structural lesion was
observed by CT.

CT and F D G PECT demonstrated a
large lesion in the left perisylvian area. Given the
extent of cortical damage, it is surprising that this paCOMMENTS.

Metter et al: LCMRglc in Aphasia

177

F i g 6. (Patient .?) C T revealc a left perisylvian lesion that is
not as extensiue or a.c posteriorly located as the one i n Patient 2.
There appears to be a relatizdy preserved cortical strip i n the
middLe and lower figures. (Templates derived from Matsui T ,
Hirano A: A n Atlac of the Human Brain for Computerized
Tomography. Tokyo, Igaku-Shoin, 1978.)

tient had only mild impairment in auditory comprehension. This patient’s C T and FDG scans showed
lesions similar to those of Patient 2, yet the two men
presented very different abnormalities of language
and speech, one with Wernicke’s and one with
Broca’s aphasia. The major difference on CT and
FDG PECT between the patients was sparing of the
middle and inferior temporal gyri in Patient 3.
Whether these two gyri are part of Wernicke’s area is
unclear 141, but they are generally considered to be a
secondary language area.
Patient 4
A 7 1-year-old, right-handed man experienced sudden
onset of right hemiparesis and aphasia in 1973. He had a

178

Annals of Neurology

Vol 10 N o 2

August 1981

F i g 7 . (Patient 3) F D G PECT shows severe metabolic depression in the left temporal and parietal regions extending beyond
the area of structural damage noted in Figure 6 . Marked depression is noted i n the relatirely pre.rrwed cortical rim.
history of prior myocardial infarction. Since the stroke,
he suffered occasional generalized seizures for which he
took phenytoin. Examination at the time of this study in
1978 demonstrated an incomplete right homonymous
hemianopia, right central facial weakness, spastic right
hemiparesis with minimal voluntary movement, hyperactive reflexes on the right with a right Babinski sign, decreased sensation to pain, touch, and stereognosis on the
right, and orofacial apraxia. In addition, he had paratonic
resistance on the left and incoordination of the left leg. He
was mildly to moderately demented and had severe memory dysfunction in addition to his language problem.
Language evaluation demonstrated moderate impairment in auditory comprehension. Naming, repetition, and
reading comprehension were poor, and he was unable to

F i g 8. (Patient 4) C T shows no structural abnormalities in
either Wernicke’s or Broca’s areas but a subcortical lesion at
levels higher than the lateral ventricles. Dilatation i.c prominent i n the l d t lateral ventricle. (Templates derived from
Matsui T , Hirano A : A n Atlas of the Human Brain for
Computerized Tomography. Tokyo, Igaku-Shoin, 1978.)

write. He could not protrude his tongue voluntarily or repeat vowels. His speech was agrammatical and telegraphic.
He spoke in one- and two-word phrases. The BDAE
(Table) demonstrated severe aphasia that included all the
features of a Broca’s aphasia plus some deficit in auditory
comprehension.
CT (Fig 8) showed an enlarged left ventricular system
and a region of decreased density high in the left parietooccipital area (mainly above the level of the ventricles) extending into the left thalamus. Broca’s and Wernicke’s
areas appeared intact. F D G PECT (Fig 9) demonstrated
more extensive abnormalities than C T had shown. A 479%
reduction of LCMRglc was noted in the entire cortex of the
left hemisphere at all levels higher than the lateral
ventricles. Reduced FDG concentration was also seen in
the posterior temporal lobe (involving the superior, mid-

F i g 9. (Patient 4) F D G PECT demonstrates an extensive decrease i n F D G concentration at levels above the left lateral
ventricles. Metabolic depression is also present in the thalamus
and posterior temporal region, where no lesion wac zirihle on
C T (see F i g 81.

dle, and inferior gyri), the left thalamus, the supramarginal
gyrus, the angular gyrus, and possibly Wernicke’s area.
COMMENTS. This patient showed all the features of
Broca’s aphasia plus poor auditory comprehension.
Based on the deficit in comprehension, he was
classified as a mixed aphasic. The most impressive aspect of this patient’s clinical syndrome was his severe
expressive difficulties, yet there was no evidence of
damage to Broca’s area on CT or FDG PECT. Mohr
et a1 [24]have argued that a large perisylvian lesion,
which includes damage to Broca’s area but extends
beyond it, is required for Broca’s aphasia. Luria and

Metter et al: LCMRglc in Aphasia

179

Hutton [20] described two forms of motor aphasia:
efferent motor aphasia, a kinetic breakdown characterized by inability to make the transition from one
articuleme to another; and afferent motor aphasia, a
kinesthetic breakdown causing inability to form articulemes and associated with an orofacial apraxia.
According to Luria and Hutton, this “afferent syndrome” results from a lesion in the postcentral gyrus.
Perhaps our patient’s expressive difficulties are based
on an afferent, kinesthetic abnormality related to his
parietal lobe damage.

Patient 5
A 57-year-old, right-handed man had sudden onset of
aphasia. One month earlier he had had an episode of mild
right-sided weakness, dysarthria, and swallowing difficulties that lasted less than a day. Past medical history was
otherwise unremarkable. Examination demonstrated a mild
drift of the right upper extremity. The remainder of his
neurological examination was normal, including stereognosis, graphesthesia, and tactile localization.
Language evaluation demonstrated that he was able to
understand simple, but not complex, directions and conversation. Repetition was poor even for single words. He
was incapable of reading sentences but could match printed
words to pictures. H e was able to write only a few letters to
dictation. His spontaneous speech was fluent, but he used
neologsms and jargon in syntactically and grammatically
complete sentences that gave no information. The BDAE
results (Table) can be interpreted as showing severe Wernicke’s aphasia.
C T demonstrated an area of decreased density on the left
side involving the superior and middle temporal, angular,
and supramarginal gyri and extending deep to involve
periventricular structures by the posterior horn of the lateral ventricle. The left ventricular system appeared t o be
enlarged compared to the right, with mild posterior displacement (Fig 10). F D G PECT demonstrated more extensive metabolic damage (Fig 11). In areas where the C T
showed destructive lesions there was a 67% reduction of
LCMRglc. In addition, a mild reduction of FDG concentration was seen in the left basal ganglia and a 15% reduction of LCMRglc in the left thalamus.

The clinical syndrome in this patient
correlated well with C T and F D G PECT. He had
Wernicke’s aphasia with a major structural defect in
Wernicke’s area. The metabolic abnormality, as measured by FDG PECT, involved a larger area of brain
than the abnormalities in tissue x-ray density noted
on CT. It is not known whether the ventricular enlargement on C T reflected more diffuse loss of
hemispheral brain substance. Several studies have
noted similar ventricular enlargement following
cerebral infarction [9, 13, 251. In the late stages
following infarction, dilation of the ipsilateral
ventricular system may occur with displacement toward the infarction [7, 251. These observations d o
COMMENTS.

180

Annals of Neurology

Vol 10 N o 2

August 1981

Fig 10. (Patient 5 ) C T demonstrates u left temporal and
parietal le.rion with extensive dumage to Wernic-Re’sarea. There
is some dilatution of the left ventricular system. (Templates derived from Mutsui T , Hirano A: A n Atlas of the Human
Bruin for Computerized Tomogruphy. Tokyo, Iguku-Shoin,
1978.)

not eliminate the possibility of more diffuse damage,
but they demonstrate that such an explanation is not
necessary in our patient.
Discussion
At present there are limitations in the use and interpretation of F D G PECT. First, the determination
of LCMRglc with F D G is based on an operational
equation of a model of the biochemical behavior of
glucose and deoxyglucose in brain. T h e equation is
dependent on empirically derived rate constants and
a lumped constant. Those used in this study were
measured from normal young adults. LCMRglc has
been found to be quite insensitive to severalfold
changes in rate constants when determinations are

F i g 11. (Patient 5 ) FDG PECT demonstrates depression of
FDG concentration throughout the leji temporal lobe andpossibly in parts of the inferior frontal lobe. The middle figure
exemplzfies some of the problems in localization that result from
limitations in photographic reproduction. The lesion outlined
on the template encompasses the area of decreased FDG concentration observed on the original scan, which is produced on
x-ray film. After printing, the decrease i n FDG concentration
is no longer visible i n some temporal and posteriorfrontal
areas.

made in a near steady state, as was done in our patients [S, 361. With pathological states, large changes
may occur in the constants and lead to errors in calculated LCMRgIc. Hawkins et a1 (in preparation)
have estimated the rate and lumped constants for infarcted brain tissue following thromboembolism.
Major differences were found in the rate constants,
while the lumped constant was unaffected in patho-

logical tissue. A 20 to 40% difference in the estimate
of LCMRglc can occur when these constants are compared to those from normal young adults. The normal constants tend to overestimate the degree of depression, particularly in areas of very low LCMRglc.
These errors in calculated LCMRglc are important to
keep in mind but d o not detract from the use of tissue 18F concentration as an index of glucose metabolism.
A second limitation to the F D G PECT method is
that normative date for LCMRglc are currently available only for young adults. This makes it difficult to
compare our patients to a normal population. For this
reason, emphasis has been placed on right-left differences, and no inferences have been drawn with regard to the activity of the noninfarcted hemisphere.
Third, the spatial resolution of the positron tomographs was less than that obtained with CT, making
delineation less precise for anatomical structure. The
hatched areas of the templates are estimates of lesion
extent conservatively drawn from the original scans,
explaining the minor discrepancies between the illustrated scans and templates.
Despite these limitations, our study demonstrates
that F D G PECT can reveal the location of lesions not
detected by other methods and provides an improved
assessment of the degree and extent of the metabolic
deficit in aphasia. This was most evident in Patient 1,
whose metabolic scan revealed the location of a lesion that was not observed by CT and that explained
the clinical picture. In this case, had the study been
done with C T alone, the caudate lesions might have
been interpreted as causing the aphasia while the
metabolic suppression in the thalamus went undetected.
The metabolic abnormalities demonstrated by
PECT were more extensive than the C T evidence of
structural damage in all five patients. The finding of
suppression of metabolism in noninfarcted brain is
consistent with xenon 133 studies of regional cerebral blood flow, which have shown reduced flow
throughout the involved hemisphere both acutely
and chronically following ischemic stroke [231,
and with the observations of Kuhl et a1 [lS], who
found extensive cortical, striatal, and thalamic
hypometabolism in stroke patients by F D G PECT in
the absence of any CT abnormality. Reduced cerebral metabolism in noninfarcted tissue suggests cortical suppression or deactivation, which may be secondary to a functional decrease in metabolic activity of
intact cells, to neuron loss, to undercutting of cortical
fiber tracts with wallerian degeneration, or to persistent ischemia [181.
These metabolic data add a new dimension to the
study of anatomicoclinical correlations in aphasia.
Jnfarcted tissue may have widespread effects on

Metter et al: LCMRglc in Aphasia

181

structurally intact tissue, as demonstrated by decreased LCMRglc in areas synaptically connected to
the damaged area. These metabolic changes may have
functional importance. For example, the decreased
LCMRglc in the thalamus in patients whose primary
lesion appears cortical, and the reduced cortical activity in Patient 1, whose primary lesion appears
thalamic, suggest that thalamocortical circuits have
functional importance in language. Admittedly, the
small number of patients and lack of pathological correlation make it impossible to exclude multiple lesions accounting for these changes. However, the
consistency of the thalamocortical metabolic interaction suggests a true relationship, since none of our
patients showed normal LCMRglc in the left
thalamus when the cortex was extensively damaged
and since the involved cortical areas and thalamus
have a different vascular supply. Whether these
thalamocortical connections are involved in language
storage, retrieval, and conception or in some of the
motor or sensory feedback circuits can be determined only by further studies.
Our findings also suggest a possible relationship
between the left thalamus and posterior portions of
the middle and inferior temporal gyri in aphasia. Each
of our patients demonstrated abnormalities in
LCMRglc in these areas, with no other cortical areas
being consistently involved. In Patients 2 through 5
the reduction of LCMRglc was less in thalamic than
in cortical areas, suggesting that the thalamic changes
were secondary to cortical alterations.
In this study, clinicopathological correlations were
made in aphasia by classifying patients using the
BDAE, which allows for separation of aphasias in a
classification using traditional anatomical consideratioos. The assumption is that particular patterns
of language involvement are associated with damage
to specific cortical structures [3]. Patients 1 and 5
demonstrate the success of this approach in some patients. In Patient l, metabolic suppression was noted
in the middle and inferior temporal lobes, areas considered to be responsible for some cases of anomic
aphasia. Patients 2 , 3, and 4 demonstrate limitations
in our current understanding of the pathological
anatomy of aphasia. Each showed large lesions with
extensive suppression of LCMRglc, yet their clinical
syndromes could not be predicted from the anatomical or metabolic abnormalities. Other factors that
could affect the clinicoanatomical correlation include
bilinguality, dementia, and recovery.
Ou r investigation suggests that one reason for
the failure of clinicoanatomical correlations may be
technical, i.e., the result of limitations in present
methods. For example, Patient 1 failed to show
clinicoanatomical correlation on CT, but FDG PECT
resulted in excellent correlation, accurately showing

182

Annals of Neurology

Vol 10 No 2

August 1981

the lesion at the site predicted by the clinical picture.
In addition to structural anatomy, a chemical
anatomy exists in the brain. At present this chemical
anatomy is not accessible to imaging techniques, and
the current precision of PECT does not permit visualization of small structures. If the latter becomes
possible, improved correlation may provide a more
accurate theory of language.
Supported by Department of Energy Contract DE-AM03-76SF00012 and by US Public Health Service Research Grants
7R01-GM-24839 and 1P01-NS-15654.
Presented in part at the 104th Annual Meeting of the American
Neurological Association, St. Louis, MO, Oct 4-6, 1979.
The authors thank Dr D. Frank Benson and Dr Walter Riege for
reviewing the manuscript, Mrs Hermine Kavanau and Mrs Mary
Lee Griswold for preparing the photographs, Mrs Diane Webb for
preparing the manuscript, and Ms Joanne Miller for data acquisition.

References
I . Albert ML, Obler L K The Bilingual Brain: Neuropsychological and Neurolinguistic Aspects of Bilingualism. New York,
Academic, 1978
2. Benson DF: Fluency in aphasia: correlation with radioactive
scan localization. Cortex 3:373-394, 1967
3. Benson D F Aphasia, Alexia, Agraphia. New York,
Churchill/Livingstone, 1979
4. Bogen JE, Bogen GM: Wernicke’s region-where is it? Ann
N Y Acad Sci 280334-843, 1976
5. Cappa SF, Vignolo LA: Transcortical features of aphasia following left thalamic hemorrhage. Cortex 15:121-130, 1979
6. Goodglass H, Kaplan E: The Assessment of Aphasia and Related Disorders. Philadelphia, Lea & Febiger, 1972
7. Greenberg JH, Reivich M, Alavi A, et al: Mapping of functional neuronal activity in man with L8F-fluorodeoxyglucose.
Neurology 29:545, 1979
8. Huang SC, Phelps ME, Hoffman EJ, et al: Noninvasive determination of local cerebral metabolic rate of glucose in man.
Am J Physiol 238 (Endocrinol Metab 1):E69-E82, 1980
9. H u k W, Schiefer W: Computerized tomography (Siretom) of
acute cerebrovascular events. In Lanksch W, Kazner E (eds):
Cranial Computerized Tomography. New York, SpringerVerlag, 1976, pp 264-272
10. Karis R, Horenstein S: Localization of speech parameters by
brain scan. Neurology (Minneap) 26:226-230, 1976
11. Kertesz A, Harlock W, Coates R: Computer tomographic localization, lesion size and prognosis in aphasia and nonverbal
impairment. Brain Lang 8:34-50, 1979
12. Kertesz A, Lesk D, McCabe P: Isotope localization of infarcts
in aphasia. Arch Neurol 34:590-601, 1977
13. Kinkel WR, Jacobs L computerized axial transverse tomography in cerebrovascular disease. Neurology (Minneap)
26:924-930, 1976
14. Kuhl DE, Edwards RQ: Image separation radioisotope scanning. Radiology 80:653-662, 1963
15. Kuhl DE, Edwards RQ, Ricci AR, et al: Quantitative section
scanning using orthogonal tangent correction. J Nucl Med
14:196-200, 1973
16. Kuhl DE, Engel J Jr, Phelps ME, Selin C: Epileptic patterns
of local cerebral metabolism and perfusion in humans determined by emission computed tomography of I8FDG and
I3NH,. Ann Neurol8:348-360, 1980

17. Kuhl DE, Hoffman EJ, Phelps ME, et al: Design and application of the Mark IV scanning system for radionuclide tomography of the brain. In Medical Radionuclide Imaging: Proceedings of the International Atomic Energy Agency Symposium on Medical Radionuclide Imaging, Los Angeles, CA,
Oct 25-29, 1976. Vienna, IAEA, 1977, vol 1, pp 309320
18. Kuhl DE, Phelps ME, Kowell AP, et al: Effects of stroke on
local cerebral metabolism and perfusion: mapping by emission
computed tomography of **FDG and 13NH,. Ann Neurol
8:47-60, 1980
19. LoVerme SR, Alexander M: Language and neurobehavioral
deficits following thalamic and putaminal hemorrhages of the
left hemisphere. Neurology 29:599, 1979
20. Luria AR, Hutton J T A modern assessment of the basic forms
of aphasia. Brain Lang 4:129-151, 1977
21. Matsui T, Hirano A: An Atlas of the Human Brain for Computerized Tomography. Tokyo, Igaku-Shoin, 1978
22. Mazzocchi F, Vignolo LA: Localization of lesions in aphasia:
clinical-CT scan correlation in stroke patients. Cortex
151627-653, 1979
23. Meyer JS, Kanda T, Fucuuchi Y,et al: Clinical prognosis correlated with hemispheric blood flow in cerebral infarction.
Stroke 2:383-394, 1971
24. Mohr JP, Pessin MS, Finkelstein S, et al: Broca aphasia: pathological and clinical. Neurology 28:311-324, 1978
25. Mohr JP, Watters WC, Duncan GW: Thalamic hemorrhage
and aphasia. Brain Lang 2:3-17, 1975
26. Naeser MA, Hayward RW: Lesion localization in aphasia with
cranial computed tomography and the Boston Diagnostic
Aphasia Exam. Neurology 28:545-551, 1978
27. New PFJ, Scott WR: Computed Tomography of the Brain and
Orbit (EM1 Scanning). Baltimore, Williams & Wilkins, 1975,
pp 332-361
28. Phelps ME, Hoffman EJ, Huang SC, et al: ECAT A new
computerized tomographic imaging system for positronemitting radiopharmaceuticals. J Nucl Med 19:635-647,
1978
29. Phelps ME, Hoffman EJ, Mullani NA, et al: Application of

annihilation coincidence detection to transaxial reconstruction
tomography. J Nucl Med 16:210-224, 1975
30. Phelps ME, Huang SC, Hoffman EJ, et al: Tomographic measurement of local cerebral glucose metabolic rate in humans
with (F- 18)2-fluoro-2-deoxy-~-glucose:
validation of method.
Ann Neurol6:371-388, 1979
31. Phelps ME, Kuhl DE, Mazziotta JC: Metabolic mapping of the
brain’s response to visual stimulation: studies in man. Science
2 11:1445-1448, 1981
32. Reivich M, Kuhl DE, Wolf A, et al: Measurement of local
cerebral glucose metabolism in man with 18-F-2-deoxy-~glucose. In Ingvar DH, Lassen N (eds): Cerebral Function,
Metabolism and Circulation (A CBF symposium in Copenhagen, Denmark, June 1977). Acta Neurol Scand 56 [Suppl64]:
190-191, 1977
33. Reivich M, Kuhl DE, Wolf A, et al: The 18F-fluorodeoxyglucose method for the measurement of local cerebral
glucose utilization in man. Circ Res 44:127-137, 1979
34. Reynolds AF, Turner IT,Harris AB, et al: Left thalamic
hemorrhage with dysphasia: a report of five cases. Brain Lang
7162-73, 1979
35. Rosenfield DB, Goree JA: Angiographic localization of infarcts in aphasia. Arch Neurol 34:590-601, 1977
36. Samarel A, Wright TL, Sergay S, et al: Thalamic hemorrhage
with speech disorder. Trans Am Neurol Assoc 101:283-285,
1976
37. Soh K, Larsen B, Shinhoj E, et al: Regional cerebral blood
flow in aphasia. Arch Neurol 35:625-632, 1978
38. Sokoloff L Relationship between functional activity and energy metabolism in the CNS. Trans Am SOC Neurochem
11:171, 1980
39. Sokoloff L, Reivich M, Kennedy C, et al: The 14Cdeoxyglucose utilization: theory, procedure and normal values in conscious and anesthetized albino rat. J Neurochem 28:897-916,
1977
40. Verkas M, Schoutens A, Demo1 D, et al: Study in cerebrovascular disease: brain scanning with technetium 99m pertechnetate; clinical correlation. Neurology (Minneap) 2 5:553558, 1975

Metter et al: LCMRglc in Aphasia

183