Posterior temporal epilepsy: Electroclinical features

Posterior Temporal Epilepsy: JAectrocbmcal Peatures M. Duchowny, MD,*-I‘P. Jayakar, MD, PhD,* T. Resnick, MD,*t B. Levin, PhD,t$ and L. Alvarez, MD*t I n the course of evaluating children with posterior temporal lobe epilepsy with subdural electrodes, we observed that their seizures commonly arose from basal rather than convexity foci and that they followed a stereotyped clinical sequence. Seizures characteristically began with behavioral arrest that coincided with basal temporal seizure discharges and progressed to motor signs as the seizure activity spread to the ipsilateral cortical convexity. Behavioral automatisms were observed in approximately half the patients, but were never the first or most prominent ictal manifestation. Focal lesions were identified preoperatively in 7 patients. We performed tailored temporal lobe resections in 14 patients, 10 (71%) of whom were seizure free (N = 9) or had occasional auras (N = 1) at a mean follow-up of 2 years. These findings suggest that in childhood, posterior temporal seizures frequently arise from basal cortex and have a consistent and recognizable ictal and electrographic semiology. In medically refractory patients, tailored temporal resection is an effective therapy. Duchowny M, Jayakar P, Resnick T, Levin B, Alvarez L. Posterior temporal epilepsy: electroclinical features. Ann Neurol 1994,35.427-411

Studies of patients with temporal lobe epilepsy provide few descriptions of seizures that arise in the posterior temporal region. While this form of epilepsy is less common than anterior temporal lobe seizures, it still affects a considerable number of patients, many of whom are medically resistant [l}. T h e lack of clear guidelines for recognizing and evaluating posterior temporal seizures has led to diverse approaches in their surgical management. While anterior temporal lobectomy prevents seizure spread to temporolimbic sites [Z, 31, it has not been uniformly successful in all patients 141. Corticectomy under electrocorticographic (ECoG) guidance effectively eliminates posterior convexity foci [ S ] , but is of little benefit when seizures begin in deeper regions. In our experience, temporal lobe seizures in childhood are often posterior in location [6]. Since posterior foci are typically outside the standard plane of anterior temporal resection, we implant subdural electrodes more posteriorly in order to define surgical boundaries. In the course of recording from subdural arrays, we observed that posterior temporal seizure foci commonly began in basal rather than convexity cortex and followed a stereotyped clinical sequence. Since these features have important implications for diagnosis and surgical therapy, we describe the electroclinical manifestations of subdurally confirmed seizures of posterior temporal origin.

From the ‘Seizure Unit and Department of Neuroscience, Miami Children’s Hospital, and the tDepartment of Neurology and iDivision of Neuropsychology, University of Miami School of Medicine, Miami, FL.

Subjects and Methods Fourteen patients ( 9 males, 6 females) who underwent surgery for medically resistant posterior temporal seizures form the basis of rhis report (Table). One patient was included in a prior study of tailored temporal resections in very young patients r61. These patients were selected from a total of 53 patients who underwent temporal lobectomy at Miami Children’sHospital between 1987 and 1991.To be included, patients had to have an interictal spike focus that was maximal at the posterior temporal (T5,T6) scalp electrode or a focal lesion in the posterior temporal region o n neuroimaging studies in association with a regional seizure focus. Tumoral patients were included only if seizure control was the primary goal of surgery. Patients have been followed postoperatively for up to 5 years, with mean follow-up of 2 years. Patients ranged in age from 3 to 22 years (mean, 12.4 years) and had a mean seizure duration of 8.6 years (range, 3- 19 years). Ten were experiencing daily seizures, and 4 had weekly seizures. One patient (Patient 5 ) had undergone an unsuccessful anterior temporal lobectomy 2 years earlier. Besides multiple outpatient electroencephalographic (EEG) evaluations, all patients underwent long-term video/ EEG monitoring using 32/64 channel systems (Telefactor). Closely spaced electrode montages and special electrodes (sphenoidal, anterior temporal, supraorbital, etc.) were employed routinely. Patients evaluated after 1990 also had offline digital EEG montage reformatting. Extraoperative intracranial monitoring further defined the epileptogenic region in 11 patients. The clinical seizure data, surface EEG, and neuroimaging studies determined the choice and placement

Received May 26, 1993, and in revised form Aug 12. Accepted for publication Sep 14, 1993.

Address correspondence to Dr Duchowny, Deparunenr of Neuroscience, Klein Pavilion-302, Miami Children’s Hospital, 3200 S.W. 60th Court, Miami, FL 33155.

Copyright 0 1994 by the American Neurological Association 427

Clinrcal Features of Patzerits uvth Posterior l’emporal Serzurn Patient No.

Age (yr):Sex

Seizure Duration (yr)

Interictal Seizure Frequency

1

l0lM

8

I -S lwk

2 3

l0iM liF

3 3

I-4fday 2-3lwk

4

9iP

8

2-ilday

5

12 10

?

19lM 12iM 14fF 14iF 20IM

10

6

liday 3-4lday 2-3 day

11

3-4lday

6

4-5lmo

22lM

l?

1 -8lday

11

1l l F

5

1-2fday

12

13iF

10

2-3fday

13 14

1l i M 10iM

9

I-2iday

s

1-2lwk

6 7 8

Spike FocusfField

Imaging Left parahippocampal gyrus lesion k f t pdrdhippocampal gyrus lesion Normal Left parahippocampal gyrus lesion Right poscerior inferior temporal hypodensity Normal Nonspecific atrophy Righr occipirotemporal lesion Nonspecific atrophy Left temporal volume reduction R g h t lateral occipitotemporal gyms cystic lesion Left parahippocampal gyrus lesion h l ar g ed right temporal horn Nornid

of electrodes. Combinations of depth, subdural strip, and grid electrodes were configured to monitor the basal, mesial, and convexity temporooccipital regions. Coverage of the temporal and occipital floor was obtained by passing multiple subdural strips or a 20-contact grid along the base of the middle cranial fossa over the tentorium cerebelli. The mesial temporal region was monitored by subdurd strip or depth electrodes. The lateral temporal convexity was moiiitorcd by a 20- or 32-contact subdural grid. All recordings were referenced to a subgaleal electrode placed at a distance from the focus. Intracranial EEG recording was terminated as soon as adequate information was obtained. All implanted patients underwent functional mapping of receptive language cortex via stimulation of subdural electrodes. Intracarotid amobarbital (amyta1)-testing was performed in the nonimplanted patient and in 2 patients with intracranial electrodes. Tailored excisional procedures were carried out under general endotracheal anesthesia. Regions of ictal onset and prominent interictal spiking were removed in all patients. Basal resections were restricted to corticectomies to minimize visual field deficits and included the inferior temporal and parahippocampal gyri. The vein of Labbe was undercut whenever it became necessary t o extend the resection posteriorly. Except in the patient who had a previous anterior temporal resection (Patient 5 ) , anterior temporal lobectomies were performed in all patients because of independent spiking or prominent secondary activation during seizures. Structural lesions were excised fully except in 1 patient who had a large mass that extended from the midtemporal region to the occipital pole (Patient 4).Following resection of the epilepto-

428

Annals of Neurology

Vol 35 No 4

April 1994

Outcome

Patholog)

No seizures

“Norind’

No seizures 9W( reduction

Ga~igliocytoma Ectopic iie u ro iis

Aura onlv

Oligodendroglinma

N o seizures

Gliosis

No improvement Ectopic neurons

No seizures

Hamartoma Cortical dysplasia Ectopic neurons Ectopic iieuroiis, gliosis, mesial rcrnporal sclerosis

No seizure^

Dysernbyroplastic tieuropathclioma

N o seizures No seizures No ~ e i t u r e s

No seizures 90C; reduction

00% reduction

Cortical dysplasia Microdysgencsis ectopic neurons “Normal”

genic region, rim ECoG was performed under light anesthesia, and the resection was extended to include residudl spikes if they were judged to be epileptogen~c and if eloquent cortex was not disrupted Seizure outcome was classihed into four caregories according to the criteria of Engrl [ 7} ( 1) seizure free, (2)90% reduction, (3) more than 50% reduction, and ( 4 )no improvement

Results Clinicrtl Seizures Two distinct clinical presentations correlated with location of seizure onset. Thirteen patients (Patients 1-1 3) who had posterior basal temporal onset initially exhibited arrested behavior lasting between 5 and 30 seconds during which time all goal-directed activity ceased. Similar behavioral changes could not be clearly identified in 1 patient whose seizures began exclusively out of sleep (Patient 4). Two patients (Patients 8 and 10) experienced visual hallucinations in association with their behavioral arrest. Arrest of behavior was invariably followed by motor seizures in all but 1 patient (Patient 11). Motor seizure activity generally lasted less than 5 minutes and consisted primarily of forced contralateral head turning in 10 patients, 7 of whom also had tonic stiffening of the contralateral arm. Clonic movements of the face, arm, and leg appeared late in the motor sequence, usually

toward seizure termination. O n e patient (Patient 10) uttered a brief cry. Automatisms were observed in 7 patients but were never the first sign, and with one exception (Patient 11), always occurred late in the ictal sequence, after the convulsive movements had ceased. Four patients (Patients 1, 2, 5 , and 11) exhibited oroalimentary automatisms, 2 (Patients 6 and 12), had gestural automatisms, and 1 patient (Patient 9) had both. Two patients (Patients 13 and 14) with lateral posterior temporal seizures (Patient 13 had independent foci in the basal and lateral temporal regions) initially manifested complex auras (deja vu, formed visual hallucination) or behavioral arrest with complex oroalimentary automatisms (Patient 14). Motor manifestations were not part of the seizure sequence of either patient. There were no reports of ictal blindness.

A

Elect Yoenrephalogruphy O n the scalp EEG, interictal spikes were maximal in amplitude at the T,/T, electrode, but typically exhibited a wider field that included either the occipital, central, temporal, or posterior frontal region. In 2 lesional patients, spiking localized to the middle temporal region, but the sphenoidal and anterior temporal (T,/TL)electrodes were never maximally involved. One patient (Patient 6) exhibited independent spiking from the ipsilateral frontal lobe. Slowing of posterior background occurred in 3 patients (Patients 2, 7,and 11). Interictal and ictal EEG data were convergent in 13 patients so that lateralization of the seizure focus was never difficult. Following seizure onset from the posterior temporal electrode, epileptic activity consistently spread to the ipsilateral hemisphere. Three patients (Patients 3, 4, and 9) demonstrated regionalized parietal lobe involvement prior to wider hemispheric activation; contralateral hemispheric activation was associated with generalized convulsions. Intracranial EEG recording localized the seizure focus to the basal posterior temporal region (i.e., > 4 cm posterior to the anterior temporal tip) in 12 patients. One patient (Patient 14) had seizures from the lateral posterior temporal convexity and 1 patient (Patient 13) had independent basal and lateral convexity foci. Seizures that began in the temporal base subsequently spread to the ipsilateral parietal and frontal convexity in 11 of 12 patients (Fig). In 2 patients (Patients 2 and lo), the lateral convexity did not activate until 30 seconds after basal temporal seizure onset, but convexity spread then proceeded rapidly. The patient with no motor seizure activity (Patient 11) had a region of seizure onset restricted to 2 cm of basolateral temporal cortex. Seizures subsequently appeared in the anterior temporal lobe approximately 1 second later. Interictal involvement of basal cortex in this patient

B ~~~~

~~

Patierit 4. (A)Typical seizure pattern recorded on sralp and sirbdiiral efectmrrtcepha~ogram ( E E G ) .Seizure onset (closed arrow) is .first ohsewed in the midposterior tpmporal region of the basal temporooccipital grid (Pi. After approximately 4 secowds. seizure actiility appears in the lateral temporal conz,exicy (G, at which time changes are first obserwd o n the scalp EEG. Note also the attewuation of basal J.eizure disc-bargingcoincident t o actiit~tionof the corii,exity.(B) Axial TI-uveighted magnetic re~mria~tr image reilealing an irregularly bordered region of inc-vedsed .rigrial in the lejl parahippocampal gyrus. The les.iot. J hoii:; heterogeneous signal intensity and absetic-e of mass effrct.

I h c h o w n y et al: Electroclinical Features in PTE

429

was always widespread and consisted of multifocal spiking and sharp wave discharges. Two patients (Patients 3 and 14) had independent occipital spikes. Neuroimrlging FiBdiings S i x patients had focal lesions of the posterior parahippocampal or occipitotemporal gyrus (see Fig 1) and 1 patient had a focal lesion of the posterior inferior temporal region. There was no mass effect and no patient had multiple lesions. Two patients exhibited generalized cortical atrophy, 1 had enlargement of the contralateral temporal horn, and 1 showed reduced ipsilateral temporal lobe volume. Three patients had normal findings.

Outcome At follow-up of up to 5 years (mean, 2 years), 10 (7'1%) patients achieved a class 1 seizure outcome (9 were completely seizure free, 1 with occasional auras), 3 patients obtained a 90% reduction (class 2), and 1 was unimproved (class 4 ) (see Table). All patients with postoperative seizures were nonlesional and had proved independent eyileptogenic foci. Two of the patients with a 90% reduction had occipital foci, and l had a temporal convexity focus located near Wernicke's region. The unimproved patient had a documented independent frontal spike focus which remained active after surgery. This patient had been functionally disabled by the posterior temporal seizures, and his family therefore viewed his outcome favorably. Neuropathology Pathological diagnoses of the surgical specimens are given in the Table. Developmental lesions were most common and included ectopic neurons, gliosis, and dysplasia in over half the specimens. The remainder showed benign or slow-growing tumors. Two patients had no pathological abnormalities. One patient (Patient 10) exhibited mesial temporal sclerosis in conjunction with other pathological findings. This patient was 22 years old and had been experiencing seizures since age 3.

Discussion Two consistent findings emerged from our study of posterior temporal epilepsy in childhood. We observed that posterior temporal seizures are likely to arise from basal rather than lateral convexity neocortex, and to have a reliably stereotyped ictal semiology. In the awake state, posterior temporal seizures begin with sudden behavioral arrest followed by secondary motor convulsive activity, usually in the form of tonic contraversive movement. During sleep, the initial change in behavioral state cannot be appreciated, and motor symptomatology predominates from the outset. 430 Annals of Neurology Vol 35 No 4 April 1994

Posterior temporal seizures differ clinically from anterior temporal lobe seizures in that automatisms are always late in the sequence and remain relatively minor, in contrast to their prominent motor symptomatology. While automatisms are not rare in patients with posterior temporal seizures, they tend to be inconspicuous and late, and in our experience are never the earliest sign or the sole seizure pattern. Patients with posterior temporal seizures may experience visual auras suggesting involvement of the occipital cortex, but visceral and autonomic sensations similar to those described by patients with anterior temporal foci were not observed. i t is often stated that temporal lobe seizure manifestations overlap with those of frontal seizures and are therefore not localizing [S). While we agree that the anatomical implications of clinical seizure data must be interpreted cautiously, our observations are fully consistent with previous studies showing that prominent and early automatisms are more typical of anterior temporal lobe epilepsy {9, lo), especially with limbic involvement 110, 11), and rare in posterior temporal seizures. Our data also reveal that epileptic dysfunction confined to basal temporal cortex is capable of arresting ongoing behavior. Language disturbances have been observed during stimulation of one temporal base but this effect is found only during dominant temporal lobe stimulation 112). We do not know whether the behavioral arrest that we observed was partially aphasic in origin since we did not perform language testing during seizure episodes. However, the fact that similar behavior changes accompanied nondominant basal temporal seizures strongly implicates impairment of functions other than language. The prominence of motor convulsions in posterior temporal seizures emphasizes the close anatomical connectivity of the posterior temporal base to the frontal lobe. Occipital seizure foci also propagate to frontal cortex [ 131 and it is reasonable to speculate that posterior temporal foci access similar convergent fascicular pathways. Contraversive tonic movements result from premotor cortical activation but their late appearance in the ictal sequence of posterior temporal epilepsy is more consistent with extrafrontal seizure origin. By contrast, primary frontal lobc seizures produce head version almost immediately at seizure onset {14, 151. The electrographic manifestations of basal posterior temporal seizures are also stereotypic. The posterior temporal elecrrode is most active on scalp recording, while the anterior temporal and sphenoidal electrodes generally remain silent, a scalp EEG topography rarely seen in anterior temporal lobe seizures. The wide epileptic spike field is more consistent with deep seizure origin, but precise localization is not possible from the scalp-recorded EEG. Depth electrodes are of proven

benefit for anterior foci but rarely localize posterior temporal foci due to their limited trajectory and restricted sampling ability. By contrast, subdural XrdJ’S provide simultaneous coverage of convexity and basal regions, and are thus able to localize seizure origin and define intracortical seizure spread. Considering how difficult it is to document basal posterior temporal seizure foci, their favorable surgical outcome was not anticipated. We attribute some of the success to our goal of removing all of the epileptogenic region, and in lesional patients, to removing as much functionally abnormal tissue as possible. In lesional patients, we also employed intraoperative ultrasound to document the extent of the lesion and its relationship to the seizure focus [lbl. Since we always resected physiologically as well as anatomically abnormal tissue, we do not know whether selective lesionectomy would have produced similar results. We were reluctant to perform lesionectomy because in children, epilepsy often results from developmental abnormalities [ 17) that prove to be more extensive than is appreciated from the imaging studies. Wieser El81 has already shown that the anterior temporal base is highly epileptogenic and that its removal improves the outcome of anterior temporal lobectomy r191. Our experience with posterior temporal seizures reaffirms this concept and additionally suggests that some failures of anterior temporal lobectomy may result from posterior basal foci that are undetected using standard monitoring protocols. We therefore propose that the basal temporal region be monitored in patients of all ages with temporal seizure foci who exhibit behavioral arrest, prominent motor symptomatology, and late or absent automatisms.

References 1. King DW, Ajmone Mdsan C. Clinical events and ictal patterns in epileptic patients with EEG temporal lobe foci. Ann Neurol 1977;2:138-147 2. Olivier A, Gloor P, Andermann F, lves J. Occipirotemporal epilepsy studied with stereotaxically implanted depth electrodes

and successfully treated by temporal resection. Ann Neurol 1982;11:428-432 1. Sperling hlK, Cahan LD, Brown WJ, Relief of seizures from a predominantly posterior temporal tumor with anterior temporal lobectomy. Epilepsia 1989;30:559-563 4. Fish D, Andermann F, Olivier A. Complex partial seizures and small posterior temporal or extratemporal structural lesions: SUTgical management. Neurology 1991;4 1 : 178 1- 1784 5 , Blume WT, Whiting SE, Girvin JP. Epilepsy surgery in the posterior cortex. Ann Neurol 1991;29:638-645 6.Duchowny M, Levin B, Jayakar P, et al. Temporal lobectomy in early childhood. Epilepsia 1932;33:298-303 7. Engel J Jr. Outcome with respect to epileptic seizures. In: Engel J Jr, ed. Surgical treatment of the epilepsies. New York: Raven

1987;553-571 8. Ajmone-Marsan C. Commentary: clinical characteristics of partial seizures. In: Engel J Jr, ed. Surgical treatment of the epilepsies. New York: Raven, 1987:121-127 ‘9. Theodore WH, Porter RJ, Penrg JK. Complex panial seizures: clinical characteristics and differential diagnosis. Neurology (Cleve) 1983;33:1115-1 12 1 10. Quesney LF. Clinical and EEG features of complex partial seizures of temporal lobe origin. Epilepsia 1986;27(suppl2):27-45 11. Maldonado HM, Delgado-Escueta AV, Walsh GO, et al. Complex partial seizures of hippocampal and amygdalar origin. Epilepsia 1988;29:420-433 12. Luders H, Lesser RP, Hahn J, et al. Basal temporal language area. Brain 1991;114:743-7>4 13. Ludwig BI, Ajmone-Marsan C. Clinical ictal patterns in epileptic patients with occipital electroencephalographic foci. Neurology 1975;25:463-47 1 14. Wyllie E, Liiders H, Morris HH, et al. The laterallzing significance of versive head and eye movements during epileptic seizures. Neurology 1986;36:606-611 15. Jayakar P, Duchowny M, Resnick T, Alvarez L. Icral head deviation: lareralizing significance of the pattern of head movement. Neurology 1992;42:1989-1992 16. Altman N , Duchowny MS, Jayakar P, e t al. Placement of intracerebral depth electrodes during excisiond surgery for epilepsy: value of intraoperacive ultrasound. AJNK 1992;l 3:254-256 17. Vinters H V , Mah V, Shields WD. Neuropathologic correlates of pediatric epilepsy. In: Duchowny M, Trevor K, Alvarez L, eds. Pediatric epilepsy surgery. New York: Demos, 1990;227235 18. Wieser HG. Electroclinical features of the psychomotor seizures. Stuttgartihndon: Gustav FischeriButtemorths, 1983 19. Nayel MH, Awad IA, Liiders H. Extent of mesiobad resection decermines outcome after temporal lobectomy for intractable complex partial seizures. Neurosurgery 1991;29:55-61

Duchowny et al: Electroclinical Features in FTE

431