вход по аккаунту


Anterior temporal language areas in patients with early onset of temporal lobe epilepsy.

код для вставкиСкачать
Anterior Temporal Ianguage
Areas in Patients with Early Onset of
Temporal Lobe Epilepsy
Orrin Devinsky, MD, Kenneth Perrine, PhD, Rafael Llinas, BA, Daniel J. Luciano, MD,
and Michael Dogali, M D
Eighteen consecutive patients undergoing dominant temporal lobectomy underwent preoperative cortical stimulation
for language localization. Patients with naming deficits on anterior (4.5 cm from the temporal pole) temporal lobe
stimulation had earlier seizure onset vs those without such deficits (5.8 yr vs 12.9 yr; p < 0.04). There was a similar
trend for reading errors,(6.3 yr vs 12.4 yr; p < 0.052). Resections always spared at least 1 cm anterior to any langauge
area. There was no significant difference in postoperative neuropsychological tests between patients with and without
anterior language representation. Early onset of dominant temporal lobe seizure foci leads to a more widespread or
atypical distribution of language areas. Individual variability should be considered in epilepsy surgery to reduce
postoperative language deficits.
Devinsky 0, Perrine K, Llinas R, Luciano DJ, Dogali M. Anterior temporal language areas in
patients with early onset of temporal lobe epilepsy. Ann Neurol 1993;34:727-732
Penfield and colleagues [l, 21 first mapped language
cortex with electrical stimulation during epilepsy surgery. They demonstrated that speech could be disrupted with stimulation of Broca’s and Wernicke’s
areas as well as by stimulation of numerous dominant
hemisphere peri-Sylvian areas and the supplementary
motor area. Lesser and co-workers 131 investigated
speech arrest during cortical stimulation, finding positive effects (Le., contraction of pharyngeal muscles)
with stimulation of primary and supplementary motor
areas and negative effects (i.e., paralysis of pharyngeal
muscles) with stimulation of the inferior frontal gyrus.
Another area subserving language functions was described in the basal temporal region with electrical
stimulation of chronically implanted subdural electrodes {4}. Stimulation of the fusiform and basal inferior temporal gyri resulted in anomia, alexia, and shortterm verbal memory deficits.
Ojemann and colleagues { 5-91 used stimulation to
identify separate sites mediating naming, reading, phoneme identification, sequential orofacial movements,
and short-term verbal memory. They found individual
variability in language localization. Stimulation in areas
that cause speech arrest or naming errors in some patients cause no langauge dysfunction in other patients.
They also found the temporal language area either
more anterior or posterior to the expected location of
Wernicke’s area.
This study reports our findings on temporal lobe
language areas in a consecutive series of patients,
studying speech arrest, naming, and reading. We evaluated the hypotheses that (1) early age of seizure onset
is associated with more anterior displacement of temporal language areas and that (2) if anterior temporal
language areas are spared, postoperative language function would not be significantly different from those
with more posterior language areas.
From the Departments of Neurology and Neurosurgery, NYU
School of Medicine and Hospital for Joint Diseases, New York,
Address correspondence to Dr Devinsky, Department of Neurology, Hospital for Joint Diseases, 301 East 17th Street, New York,
NY 10003.
We studied a consecutive series of 18 patients undergoing
dominant temporal lobectomy. All subjects had a left (dominant, by the Intracarotid Sodium Amobarbital Test) hemisphere seizure focus. There were 8 males and 10 females;
ages ranged from 12 to 41 years (mean, 27 yr). Seizures
began between ages 1 month and 25 years (mean, 9.3 yr)
with an average seizure duration of 18.8 years. Educational
level ranged from 6 to 26 years of schooling (mean, 13 yr).
Table 1 summarizes the baseline neuropsychological test results.
Neuropathology abnormalities of the resected tissues included mesial temporal sclerosis (gliosis; n = 5 ) , neoplasms
(astrocytoma,oligodendroglioma,neurocytoma; n = 4 ) , vas-
Received Mar 24, 1993, and in revised form Jun 3. Accepted for
publication Jun 3, 1993.
Copyright 0 1993 by the American Neurological Association 727
Table I . Baseline Nruropsycbological Test Re.rult.i
Verbal IQ
Performance IQ
Full Scale IQ
Boston Naming Test
Token Test
Reading (Wide Range
Achievement Test)
Controlled Oral Word
Association Test
BDAE RepetitionLow Probability
BDAE = Boston Diagnostic Aphasia Examination.
cular abnormalities (cavernous angioma; n
1), dysplasias
(hamartoma and heterotopias; n = 6), and inflammation
(n = 2). Preoperative magnetic resonance imaging (MRI)
was normal in 6 patients, with other scans revealing neoplasms (n = 4 ) ,increased signal o n T2 in the mesial temporal
region (n = 3), increased signal on T2 with atrophy in the
mesial temporal region (n = 2), lobar atrophy (n = l), decreased gray-white differentiation (n = 1 ), and bilateral
frontal hypointense lesions consistent with old hemorrhage
(n = I). Neurological examination was normal in 4 patients
with abnormalities in remaining patients shown for shortterm memory (n = lo), naming (n = 6), verbal fluency (n
= 2 ) , repetition (n = 11, decreased right nasal-labial fold
(n = l),diminished right arm swing (n = I), psychomotor
slowing ( n = l), dysarthria (n = l), and global cognitive
functioning ( n = 1). Interictal surface electroencephalogram
(EEG) revealed abnormal activity in all patients, including
localization to the left anterior mesial temporal lobe (n =
111, left temporal lobe (n = l),left posterior quadrant (n =
l), frontotemporal derivations (n = 2), bilateral anterior mesial temporal lobe with left predominance (n = l),and bilateral temporal regions wirh left predominance (n = 2).
Ten patients had intraoperative mapping, 7 had subdural
mapping, and 1 had both subdural mapping followed bj intraoperative mapping. Both intraoperative and extraoperative
results are included for the last patient, but statistical analyses
were based on a sample of 18, collapsing the intra- and extraoperative data for this patient.
The neuropsychological measures reported in the current
study were part of a comprehensive presurgical evaluation
performed during inpatient admissions for video-EEG monitoring. Tests were not administered following a seizure until
all postictal symptoms had cleared. Selected language tests
were readministered 6 months to 1 year following surgery to
determine postoperative language functions. The Wechsler
Adult Intelligence Scale-Revised (WAIS-R) was administered to determine intellectual ability lOJ. Confrontation
naming was measured with the Boston Naming Test Ell].
The Token Test assessed auditory verbal comprehension of
propositional language 112). Verbal fluency in a word genera-
728 Annals of Neurology
Vol 34
tion task was measured with the Controlled Oral Word Association test (“FAS”)11 21. The Reading subtest from the Wide
Range Achievement Test-Revised, Level 2 assessed oral
reading abilities [ 13J. Verbal repetition was assessed by the
Reading Words and Phrases: Low Probability scale from the
Boston Diagnostic Aphasia Examination (BDAE) [ 111.
Language testing was adapted from Ojemann and Whitaker’s
studies 15-71. Multiple trials of three phases each were presented on a Macintosh IIci computer. The first phase displayed a line drawing of a common object, which the patient
named aloud using the carrier phrase “This is a -”.
format was utilized to differentiate speech arrest (no response), which could be due to contraction of orofacial musculature or deficits in initiation, from anomia (word finding
difficulty). The second phase displayed an incomplete sentence with a word missing at the end. The patient read the
sentence aloud and provided the missing word, which was
contextually related to the body of the sentence (e.g., “We
The third and final
went to the airport to catch -”).
phase displayed a printed cue of “It was -”
as a prompt
for the patient to recall the line drawing that appeared on
the screen during the first phase. The naming and recall
phases were displayed for an average duration of 5 seconds
and the reading phase for an average duration of 7 seconds,
with alteration of these intervals if a patient was particularly
slow. Stimulation was applied during one of the three phases,
with at least 20 seconds between successive stimulations.
Testing with all stimuli was conducted on most patients, and
stimuli (line drawings and sentences) that could not be identified were eliminated. This stimulus selection preceded testing
for baseline rates, which occurred during the stimulation sessions.
Most patients were stimulated during all three phases;
however, 2 patients received stimulation only during naming
and reading but not recall, although recall was obtained and
will be reported. The entire three-phase trial was not used
at all sites for the other patients due to time constraints but
was used at the majority of sites.
For intraoperative cases, a 9-in. video monitor was
mounted on a mechanical arm 30 cm from the patient’s face
on the nonsterile side of the drapes. The arm was mounted
to a cart holding the computer, cortical stimulator, and a
separate video display for the examiner to view. For extraoperative cases, the identical setup was utilized at bedside, with
the mechanical arm mounted 30 cm from the patient’s face
with the patient in a sitting position on the bed.
A Grass Instruments S-88 cortical stimulator with a separate SIU-9 constant currentistimulus isolation unit was used
in earlier studies. A Grass Instruments $12 stimulator with
built-in constant current/stimulus isolation circuitry was utilized in later studies. The S-88 delivered a monophasic
square wave, while the S-12 delivered a biphasic square wave.
Parameters were set at 5 0 - H ~frequency, 0.3 msec pulse duration, and amperage ranging from 2 to 15 mA. Afterdischarges were monitored with concomitant electrocorticography utilizing separate subdural grids for intraoperative cases
and other portions of the subdural grid for extraoperative
mapping. Afterdischarges were determined with gradual increases in amperage in the earlier intraoperative studies but
No 5 November 1773
discontinued in later studies due to excessive time. These
later studies utilized 12 mA, which was decreased if afterdischarges were elicited. An amperage setting of 12 mA is sufficient to produce functional changes and approaches the 15
mA upper limit regarded as safe for human cortical sdmulation. Amperages for extraoperative stimulation were set 1
mA below the afterdischarge threshold, which was determined separately for each site. None of the results reported
below contain data from trials where afterdischarges were
For extraoperative mapping, stimulation was applied between adjacent pairs of electrodes on a chronically implanted
subdural grid. Earlier cases utilized PMT subdural grids with
5.0-mm electrodes imbedded in Silastic, while later cases utilized Ad-Tech subdural grids with 4.0-mm electrodes imbedded in plastic. Various combinations of 64- and 32-contact
subdural grids were utilized. The subdural grids were implanted during a first craniotomy, with the wires channeled
through the scalp slightly distal to the craniotomy flap. In
earlier cases the bone flap was replaced, while in later cases
the bone flap was not replaced and only a skin flap closed.
The patients were returned to a neurosurgical special care
unit, where electrocorticography and the language mapping
with cortical stimulation were performed. The patients returned for a second craniotomy for removal of the subdural
grid and neurosurgical resection 7 to 14 days following the
first craniotomy.
For intraoperative cases, the craniotomy and mapping were
performed under local anesthesia. Supraorbital blocks were
administered following positioning of the patient in order to
diminish discomfort during scalp infusion. The scalp was then
injected with bupivicaine around the perimeter of the craniotomy flap. The patients were administered alfentanil or propofol during removal of the bone flap to diminish discomfort.
Electrocorticography was then performed, initially using a
Grass electrocorticography apparatus with cotton wick electrodes in the first 8 cases and with 20-contact or 32-contact
subdural grids in subsequent cases. Stimulation was conducted with a monopolar probe refcrenced to a ground needle placed in the temporalis muscle in some cases and between adjacent pairs of electrodes on the subdural grid
placed over the cortex in other cases. When rare stimulationevoked seizures occurred during intraoperative and extraoperative mapping, further mapping was conducted following
return to baseline mental status.
Stimulation sites were selected in anticipation of the area
to be resected. The number of sites stimulated averaged 7
per patient for intraoperative cases, and 11 per patient for
subdural cases. For the purposes of localization, sites were
divided into 12 zones of interest on the temporal lobe. Four
quadrants per temporal gyrus (superior, middle, and inferior)
were identified on the basis of the distance from the temporal
pole as follows: anterior (0-2.9 cm), middle (3-4.4 cmj, posterior (4.5-5.9 cm), and far posterior ( G + cm). The middle
and posterior quadrants were most closely spaced because
the margins of the resections generally fell within these
boundaries and necessitated more sites within these regions.
Baseline resting of each phase without stimulation occurred during test sessions for each patient. Sites were considered to be positive for language function if the error rate
obtained with stimulation exceeded the baseline error rate
without stirnulation at a significant level ( p < 0.05) as determined by the binomial single sample test. The number of
patients demonstrating positive findings at each of the 12
sites was tallied.
For comparisons, we grouped responses occurring within
4.5 cm of the temporal pole as anterior, and responses
greater than 4.5 cm from the temporal pole as posterior.
Surgical excision of the dominant temporal lobe always
spared an area of at least 1 cm anterior to any sites at which
stimulation induced lanbwage dysfunction along that gyrus.
This 1-cm margin was retained for patients with anterior language areas as well as those with more common posterior
language representation.
Stutisticul AnuiysiJGroup comparisons of continuous data (e.g., neuropsychological test scores, age) between patients with and without
anterior language representation were tested with the independent sample Student’s t test. Group comparisons of noncontinuous data (e.g., sex, lesion type, history of febrile seizure, seizure, and EEG data) were analyzed with x2.
Preoperative to postoperative comparison of neuropsychological measure were analyzed with the paired sample Student’s t test.
We summarized percentages of naming (Table 2) and
reading (Table 3 ) errors during electrical stimulation
of t h e 12 lateral temporal lobe regions.
Patients with naming deficits o n stimulation of the
anterior temporal lobe had an earlier mean onset of
seizures (5.8 yr) compared with those without naming
deficits o n stimulation of t h e anterior temporal lobe
(12.0 yr; fi < 0.04). This finding was most significant
w h e n considering stimulation of the superior temporal
gyrus. Patients with naming errors with stimulation of
t h e anterior 4.5 c m of the superior temporal gyrus had
an earlier m e a n onset of seizures (3.2 yr) compared
with those without naming errors (13.0 yr; p < 0.003).
Patients with reading errors during stimulation of
the anterior portion of t h e middle temporal gyrus had
an earlier mean age of seizure onset (4.3 yr) compared
with those without such errors (12.4 yr; p < 0.008).
There was a trend for patients with reading errors during stimulation of the anterior temporal lobe t o have
an earlier average age of seizure onset (6.3 yr) compared with those without such errors (13.1 yr; p <
T h e r e were n o differences between patients with
and without naming or reading errors during stimulation of t h e anterior temporal lobe for the following
variables: lesion type, history of febrile seizure, sex,
contralateral epileptiform activity on EEG recordings,
secondary generalized tonic-clonic seizures, or type of
partial seizure (simple versus complex).
Patients with naming errors on stimulation of the
anterior portion of t h e middle temporal gyrus had
Uevinsky et al: Language Areas and Temporal Lobectomy
Table 2. Naming Ewors with Dominant Temporal Lobe Electrical Stimulation
Section of the Gyrus
Far Posterior
1112 (8.3)
7144 (15.9)
7117 (41.2)
17192 (18.5)
6112 (50)
17172 (23.6)
416 (66.7)
11/52 (21.1)
619 (6.7)
20158 (34.4)
2/15 (13.3)
4190 (4.4)
4115 (26.6)
1 1 / 1 1 (9.9)
218 ( 2 5 )
4 / 2 9 (13.8)
2/10 (20)
9/32 (28.1)
Patients = number of patients with positive responses; Stimuli = number of stimuli with positive responsesistimuli performed at that site;
numbers in parentheses = percentages of positive responses; N D = not done.
Tuble 3. Reading Ewors with Dominant Temporal Lobe Ehctkcal Stimulation
Section of the Gyrus
Far Posterior
3110 (30)
9/40 (22.5)
5114 (35.7)
24182 (29.2)
519 (55.5)
25/73 (34.2)
315 ( 6 0 )
16/31 (51.6)
216 (33.3)
4128 (14.2)
4112 (30.8)
16186 (18.6)
1112 (8.3)
8180 ( 1 0 )
616 ( 7 5 )
28/58 (48.3)
219 (22.2)
6128 (21.4)
218 ( 2 5 )
5133 ( 1 5 . 1 )
Patients = number of paticnts with positive responses; Stimuli = number of stimuli with positive responses/stimuli performed at that site;
numbers in parentheses = percentages of positive responses; ND = not done.
lower baseline scores on word fluency tests than those
without naming errors in this region ( p < 0.03). There
were no other differences between patients with and
without naming or reading errors during stimulation of
the anterior temporal lobe for any of the neuropsychological baseline test results.
There was a significant decline in scores on the Boston Naming Test postoperatively (preoperative mean,
46; postoperative mean, 39.2; p < 0.03). There were
no significant postoperative changes in verbal fluency
(Controlled Oral Word Association Test), comprehension (Token Test), reading (WRAT-R), or repetition
(BDAE-low probability). There were no significant differences in postoperative language functions between
patients with anterior and posterior temporal language
Early onset of dominant temporal lobe epilepsy is associated with more anteriorly distributed temporal lan-
guage representation. Both naming and reading functions were more often identified in the anterior
temporal lobe of patients with early onset of epilepsy.
Surgical resections sparing these anterior language areas were associated with similar postoperative language
functions as in patients with more posterior temporal
language areas. We postulate that seizure foci in the
dominant temporal lobe during language development
produces a more widespread or diverse pattern of localization.
Ojemann and colleagues [5-91 found stimulationinduced naming errors over a wide area of the left
lateral cortex, extending beyond the traditional anatomical limits of Broca’s and Wernicke’s areas. The
posterior language areas of the parietal and temporal
lobes showed a greater degree of anatomical variability
than frontal language areas [7}. Ojemann [57 found
anterior temporal language sites in I of 10 patients.
Emphasizing the distinction between areas with naming
errors during only some of the stimulation trials from
730 Annals of Neurology Vol 34 ,No 5 November 1993
those with naming errors during all stimulation trials,
he found that sites in which stimulation only intermittently evoked naming errors could be removed without
postoperative dysnomia. In contrast, when resection
extended to within 1 cm of a superior temporal gyrus
site that was associated with consistent naming deficits,
a postoperative dysnomia was observed.
Ojemann {5] and Ojemann and Whitaker { 7 } found
parietal representation of language was more extensive
in patients with seizure foci extending to the posterior
temporal lobe and patients with low verbal IQs. Paradoxically, none of 10 patients had sites with consistent
stimulation-induced naming errors on the posterior
half of the superior temporal gyrus, the classical location of Wernicke’s area E71.
Early onset of left hemisphere seizure foci is associated with altered language lateralization, with an increased incidence of right hemisphere dominance 112,
131. Rasmussen and Milner 1141 suggested that early
left frontal or parietal, but not temporal, injury can
displace language functions partially or completely to
the right hemisphere. Right hemisphere language displacement is more likely if the person is pathologically
left-handed, without a family history of sinistrality { 151.
In such cases, damage to the sensorimotor areas of the
left frontoparietal region displaces motor dominance
to the right hemisphere (left hand). Language lateralization may thus be determined by both the direct effects
of left frontoparietal injury as well as a tendency for
motor aspects of language to develop in the same hemisphere dominant for other motor functions. This hypothesis is consistent with the observation that among
patients with refractory epilepsy and right hemisphere
language functions, there is often a dissociation between fluency and comprehension; fluency has a
stronger localization on the right side, naming and
comprehension has a stronger localization on the left
side { 161.
The effects of early onset of temporal lobe seizure
foci on language development may therefore differ
from those in the frontoparietal regions. If early seizure foci are restricted to the left temporal lobe, motor
dominance will not shift to the right hemisphere because the left sensorimotor areas are intact. With left
hemisphere motor dominance, language functions almost always remain on the left side. The effects of left
temporal seizure foci during early development appear
to produce a displacement of language functions either
anteriorly, as demonstrated in our study, or into parietal regions, as demonstrated by Ojemann and colleagues {5-91. While they found that only early onset
posterior temporal lobe seizure foci displaced language
functions into parietal regions, we did not find a localized area within the temporal lobe in which early onset
seizures were associated with anteriorly displaced larguage functions.
One of the difficulties in interpreting functional localization based on stimulation studies in epilepsy patients is the lack of “normative” data. The best normative data come from patients with seizures or brain
tumors developing after age 10 years, when language
localization has been determined. However, even in
these cases, there may have been electrographic discharges preceding clinical seizures or neuropathological abnormalities present since early life. Therefore,
the variability of stimulation-defined language localization in the normal population remains unknown.
We found a mild decline in naming abilities following dominant temporal lobectomy. This decline was
similar in patients with or without anterior temporal
language functions. However, the resection always
spared at least 1 cm anterior to sites in which stimulation consistently evoked speech arrest, naming or reading errors, for patients with anterior language sites and
those with more normally represented posterior language sites. Therefore, it appears postoperative language functions are related to preservation of language
areas, as previously demonstrated [ S ] .
Additional studies are needed to define normative
patterns of language localization and the developmental factors that modify these patterns. The most
challenging cases for the future are those in which the
lateral temporal seizure focus and language areas overlap. New approaches such as multiple subpial transections may be effective in such cases 1171 and warrant
further study.
1. Penfield W, Jasper H. Epilepsy and the functional anatomy of
the human brain. Boston: Little, Brown, 1054
2. Penfield W, Roberts L. Speech and brain mechanisms.
Princeton, NJ: Princeton University Press, 1959
3. Lesser RP, Lueders H, Dinner DS, et al. The location of speech
and writing functions in the frontal language area. Brain 1784;
1 0 7 ~ 2 5-291
4. Luders H, Lesser RP, Hahn J, et al. Basal temporal language
area demonstrated by electrical stimulation. Neurology 1986;
5. Ojemann GA. Individual variabiliry in cortical locahzation of
language. J Neurosurg 1977;50:164-169
6. Whitaker H, Ojemann GA. Graded localisation of naming from
electrical stimulation mapping of left cerebral cortex. Nature
1977;270:50-5 I
7. Ojemann GA, Whitaker HA. Language localization and variability. Brain Lang 1978;6:239-260
8. Ojemann GA. Interrelationships in the localization of language,
memory, and motor mechanisms in human cortex and thalamus.
In: Thompson RA, Green JR, eds. New perspectives in cerebral
localization. New York: Raven Press, 1982:157-175
9. Ojemann GA, Dodrill CB. Verbal memory after left temporal
lobectomy for epilepsy: mechanism and inrraoperative prediction. J Nrurosurg 1985;62:101-107
10. Wechsler, D. Wechsler adult intelligence scale-revised. New
York: Psychological Corporation, 1987
11. Goodglass H , Kaplan E. Assessment of aphasia and related disorders. Philadelphia: Lea and Febiger, 1972
12. Benton AL, Hamsher KDS, Varney NR, Spreen 0. Contribu-
Devinsky et al: Language Areas and Temporal Lobectomy
tions to neuropsychological assessment. New York: Oxford
University Press, 1983
13. Jastak S, Wilkinson GS. Wide range achievement test-revised.
Wilmingcon, DE: Jastak Associates Inc, 1984
14. Rasmussen T. Milner B. The role of early left-brain injut). in
determining the lateralization of cerebral speech functions. Ann
N Y Acad Sci 1977;299:355-169
15. Rausch R, Walsh GO. Right-hemisphere language dominance
in right-handed epileptic patients. Arch Neurol 1984;41: 10771080
16. Woods RP, Dodrill CB, Ojemann GA. Brain injury, handedness, and speech lateralization in a series of amobarbital studies.
Ann Neurol 1988;23:510-518
17. Morrell F, Whsler WW, BleckTF’. Multiple subpid transection:
a new approach to the surgical treatment of epilepsy. J Neurosurg 1989;70:231-239
732 Annals of Neurology Vol 34 No 5 November 1993
Без категории
Размер файла
548 Кб
patients, language, area, temporal, anterior, early, epilepsy, onset, lobel
Пожаловаться на содержимое документа