J Neurosurg 49:344-356, 1978
A method for surgical management of focal epilepsy, especially as it relates to children SIDNEY GOLDRING, M . D .
Department of Neurology and Neurological Surgery (Neurological Surgery), Washington University School of Medicine, St. Louis, Missouri ~" A method of surgical management for intractable epilepsy is described. The essential features are: 1) all surgical manipulation is carried out under general, rather than local, anesthesia; 2) the sensorimotor region is readily identified in the anesthetized patient by recording cortical sensory evoked responses; and 3) the epileptogenic focus is localized by extraoperative electrocorticography via indwelling epidural electrode arrays, localization deriving from recordings made during spontaneously occurring clinical seizures. Cases are presented to demonstrate that: 1) in some instances, recording of sensory evoked responses is the only means of sensorimotor localization in both the awake and anesthetized patient, and 2) spontaneous and electrically induced electroencephalographic seizure activity may provide false localization of the focus, the correct localization requiring recordings made during spontaneous clinical seizures. The outcome of surgery and the various epileptogenic lesions encountered are described. A good result has been achieved in 61% of patients followed 1 to 10 years. When the results obtained in children are analyzed alone, 70% have benefited from surgery. KEY WORDS 9 epilepsy 9 surgical management 9 sensory evoked response 9 extraoperative electrocorticography
HE method used most commonly in the surgical management of focal epilepsy requires that surgery be carried out under local anesthesia/-11 General anesthesia is avoided because it suppresses both spontaneously and electrically induced electroencephalographic (EEG) seizure activity. It also suppresses the motor responses to electrical stimulation of the cortical surface, information that is used to identify the sensorimotor region. Not all patients, however, can undergo surgery under local anesthesia with the necessary degree of cooperation, and this is especially true in children. Also, it has become increasingly clear that the most reliable evidence for localizing an epilep-
T
togenic focus is that which is obtained in electrical recordings made during a spontaneous convulsion. 1,2,',Is,1' During surgery, abnormal interictal electrical discharges may be unreliable, whether they occur spontaneously or are induced by electrical stimulation (afterdischarge). Even an after-discharge associated with the partial seizure from which the clinical convulsion develops sometimes provides erroneous localization. Furthermore, the partial seizure may spread rapidly into a generalized one, an undesirable event during surgery. A consequence of these considerations has been the development of a method that permits a more precise localization of the epilep-
344
J. Neurosurg. / Volume 49 / September, 1978
Surgery in epilepsy togenic focus and makes possible definitive treatment of patients, especially children, who heretofore would not have been considered for surgery. The important features of the procedure are: 1) all surgical manipulation is carried out under anesthesia; 2) the sensorimotor region is readily identified in the anesthetized patient by recording cortical sensory evoked responses; and 3) the epileptogenic focus is localized by extraoperative electrocorticography (ECG) via indwelling epidural and/or depth electrode arrays, with the localization derived from recordings made during spontaneously occurring seizures. The purpose of this paper is to review our experience in which indwelling epidural electrode arrays were used to localize the epileptogenic focus.
TABLE 1
Surgical management of intractable focal epilepsy: lO-year experience*
Method of Epileptogenie Focus Localization extraoperative ECG using epidural electrode arrays extraoperative ECG using depth electrodes intraoperative ECG or scalp EEG and x-ray film total *ECG = electrocorticography; EEG = encephalography.
No. of Cases 29 46 17 92 electro-
A craniotomy is performed under general endotracheal anesthesia, and the brain exposed and inspected. The dural incision is made around the entire circumference of the exposed dura, producing a free dural graft. This is done to minimize the development of
epidural clot which could attenuate the recorded electrical activity. Cortical sensory evoked responses are then recorded to identify the sensorimotor region. Previously, we5,7,12have shown that, under the conditions of recording to be described, a somatosensory evoked response is recorded only from the somatosensory and motor gyri. This was true in both the anesthetized and awake state. The somatosensory area always shows a response, but the appearance of a response in the motor cortex is variable. The observation that a response appears only in the sensorimotor region is the basis for using sensory evoked responses for sensorimotor localization. A Silastic template holding nine linearly oriented electrodes, each separated by a distance of 1.5 cm, is placed on the cortical surface in a plane parallel to the midline (Fig. 1). Simultaneous records are made from each adjacent pair of electrodes (that is, 1-2, 2-3, 3-4, etc) while the contralateral median nerve is electrically pulsed transcutaneously at the wrist. A special-purpose computer is used to quickly expose the sensory evoked response. Without computer processing, human cortical electrical responses to somatosensory stimulation are barely visible above the background waves of the spontaneous ECG. The computer stores and adds a brief epoch (100 to 200 msec) of electrical activity following each shock for approximately 25 stimuli. The spontaneous random brain waves do not add significantly, but, because the evoked responses have consistent polarity and latency, they summate and become clearly exposed above the background. The face area is similarly identified using a light tap to the upper or lower lip as
J. Neurosurg. / Volume 49 / September, 1978
345
Clinical Material and Methods
Patient Population During the years 1968 to 1978, 92 epileptic patients have come under our care. Of these, 29 patients were evaluated with extraoperative ECG by means of indwelling epidural electrode arrays, 46 were studied with extraoperative ECG using indwelling depth electrodes, and in 17 the focus was localized by intraoperative ECG or by scalp and radiological imaging procedures (Table 1). All but four of the 46 patients who had depth electrode ECG had temporal lobe foci. Those 29 patients undergoing extraoperative ECG with epidural electrode arrays were suspected of having a cortical focus in areas other than the medial temporal lobe. Seventeen were children (ages 2 to 14 years) who were having many seizures (10 to 50) daily, in spite of intensive medical management which pushed all combinations of anticonvulsant medication to toxic levels, and included a ketogenic diet. Seizures in the adults were equally intractable, although in two patients they only occurred several times monthly. Method
S. Goldring
FIG. 1. Identification of somatosensory area by recording cortical sensory evoked responses. Linear electrode array overlying cerebral cortex is shown (lower). Recordings are made simultaneously from each adjacent pair of electrodes (i.e., 1-2, 2-3, etc.). The numbers identifying each trace correspond to the numbers on the electrode array. The response profile establishes the area underlying Electrode 5 as the somatosensory hand area because: 1) the only traces that show a significant response are 4-5 and 5-6; 2) only Electrode 5 is common to both electrode combinations; and 3) the polarity of the two responses is reversed. Calibration pulse is 12.5 ~V and 18 msec; positive at the first electrode of each pair (G1) is up.
the sensory stimulus. We have used this method routinely for the past 10 years, even in patients under local anesthesia, because it has proved to be a quicker method of identifying the sensorimotor region than the use of electrical stimulation of the cortical surface to produce movement. After the sensorimotor region is identified, the dura is closed and a quasi-rectangular Silastic template holding 24 electrodes is sewn in epidurally (Fig. 2). The electrodes are arranged in rows with inter-electrode dis-
tances of 1.5 cm and 2 cm in the sagittal and coronal plane, respectively. The electrode template is placed so that it spans the suspected area of epileptogenicity, its surrounding area, and the sensorimotor region. Care is taken to assure that one of the electrodes immediately overlies the motor or sensory hand area. When this condition is met, the number and spacing of the electrodes is such that it is then possible to record both sensory evoked responses to the face and hand, and also to elicit motor responses in the
346
J. Neurosurg. / Volume 49 / September, 1978
S u r g e r y in e p i l e p s y
Fro. 2. Epidural electrode array sewn in place.
face and hand by stimulating through the same electrodes. The remainder of the closure is completed and recordings are begun on the following day. Both the patient's behavior and the ECG are continuously recorded and displayed side by side on the split screen of a television monitor. The taping of these recordings permits the additional opportunity for replays of any seizure that might occur. This is necessary to determine accurately which cortical area shows the first sign of abnormal activity in the moments preceding the onset of clinical convulsion, essential information for localizing the focus. Because the patient is awake, it is now possible to record motor responses to transdural cortical stimulation as well as sensory evoked responses, identifying the relationship between the sensorimotor region and the focus. If the recordings are from the dominant hemisphere,
speech arrest due to electrical stimulation is also used to determine the boundaries of the speech area. Thus, all of the techniques used in localizing an epileptogenic focus during the usual surgical procedure under local anesthesia can be performed by recording and stimulating epidurally with indwelling epidural electrode arrays. The recordings are performed in a leisurely manner, free from the usual stresses of the operating room, and the opportunity of recording the ECG during a spontaneous clinical seizure, or, if necessary, an activated one, is markedly enhanced. Recording with the epidural electrode arrays usually continues for only 24 to 48 hours, since most of these patients have at least several seizures a day. If the observations identify the boundaries of a single epileptogenic focus, then the brain tissue within the epileptogenic zone is excised at a
J. Neurosurg. / Volume 49 / September, 1978
347
S. G o l d r i n g
The following two cases illustrate the usefulness of sensory evoked responses and show that in some instances they are the only means of obtaining sensorimotor localization in the anesthetized or awake state. Case Report: A 4-year-ol'd gfff (C. ~ : ) fiad had intractable seizures since 6 months of age, progressive mental retardation with loss of all language function, and left hemiparesis which waxed and waned in severity. Seizures started with left adversive movement of the head and eyes, and were occurring 10 to 12 times a day at the time of surgery. Complete radiological work-up revealed only blunting of the right frontal horn on computerized tomography (CT) scan. Upon exposure, the brain appeared normal. Under general anesthesia
the sensorimotor region could b e identified only by recording sensory evoked responses; the cortex was electrically inexcitable for producing movement. An epidural electrode array was sewn in place with Electrode 19 overlying the sensory hand area (Fig. 3). Recording of evoked responses in the awake state again identified the sensorimotor region, but cortical stimulation through all electrodes (monopolar and bipolar combinations) still failed to produce movement. The resting ECG showed spontaneous spiking in the somatosensory hand area, and the first two recorded seizures began in that gyrus (Fig. 4). At the onset of seizure activity the left hand clenched, and there was then stiffening of the arm and adversion of head and eyes to the left. The subsequent five seizures, however, began in the right prefrontal cortex with a left adversive clinical onset (Figs. 5 and 6). On the basis of these findings, a right prefrontal cortical excision spared the sensorimotor region; the pathological diagnosis was focal cortical dysplasia. The seizure frequency has decreased significantly, the hemiparesis has disappeared, and the child is developing a vocabulary. Comment. fn tills case we ffem,onstrated that seizures occasionally could begin in the sensorimotor gyri, the child had exhibited a progressive hemiparesis preoperati, vely, and electrical stimulation of the cortex failed to produce any movement; all thq~se data suggested that the sensorimotor ~area was non-functional. However, the recording of sensory evoked responses not only :identified the sensorimotor region, but also showed that at least the somatosensory area was physiologically functional. Had we not leaLrned this we might have included the sem~ory and motor gyri in the resection and risk ed a permanent hemiparesis. This case also demonstrates the importance of reoording as many spontaneous seizures as possJible. We have encountered four other children in whom electrical stimulation of cerebral cortex failed to produce any movement, and the sensorimotor region was identified only by recording sensory evoked responses (Table 2). l'wo of these children also had a preoperative hemiparesis, which disappeared after surgery. The progressive hemiparetic disabilities in all of these children obviously represented sustained postictal deficit due to frequent focal seizures.
348
J. Neurosurg. / Volume 49 / September, 1978
FIG. 3. Photograph of the exposed brain (C.W.): anterior is to the left and midline is at the bottom. The numbered tickets overlie the precise areas from which recordings were made with the epidural electrode array. The area underlying No. 19 was identified as the somatosensory handarea by recording responses from electrode combinations 20-19 and 19-18 shown at bottom; note polarity reversals of response. The area underlying No. 14 was identified as the somatosensory face area. The black line between 22-21, 16-15 etc. is the posterior limit of the frontal lobe resection.
second craniotomy; if not, the surface electrode array is removed and the operative wound closed. Results
Use of Sensory Evoked Responses for Localizing Cerebral Function
Surgery in epilepsy
FIG. 4. Seizure begins in sensorimotor region at electrode 19 (see Fig. 3). Observe the appearance of seizure activity at electrode sites 14-15 which is also in sensorimotor domain. The initial high voltage burst of activity (arrows) ushered in both seizures that began in the sensorimotor region. Note recordings from left cortical surface. These were made from epidural electrodes introduced through a small (1 cm) trephine opening. Such recordings on the side contralateral to the hemisphere of predominant interest have been made in several patients, especially those having adversive seizures
Case Report. This 13-year-old boy (T.L.) had suffered focal seizures since the age of 6 years. The ictus began with a feeling of numbness in the left face and progressed to facial twitching that spread to involve the left arm; the seizure frequently became generalized. A CT scan suggested a hypodense area in the right posterior frontal region; the remainder of the radiological workup showed no abnormality. During surgery the sensorimotor region was localized by recording sensory evoked responses to electrical stimulatiort of the median nerve and tactile stimulation of the face. When the level of anesthesia was decreased, electrical stimulation of cerebral cortex elicited only
*Motor cortex was electrically inexcitable; no movement could be obtained by electrical stimulation in both the anesthetized and awake state.
J. Neurosurg. / Volume 49 / September, 1978
349
TABLE 2 Children in whom sensorimotor identification was made possible only by recording sensory evoked responses*
Case Age No. (yrs) I 2 3 4 5
Diagnosis
Excision
3 2
focal cortical dysplasia frontal tuberous sclerosis frontal & temporal 4 tuberous sclerosis frontal 2~/2 heterotopia occipitoparietal 2~A ganglioglioma posterior parietal
S. G o l d r i n g
FIG. 5. Spontaneous clinical seizure first evident electrographicallyin pre-frontal cortex (arrows, and refer to Fig. 3) as record becomes iso-electric. The spiking in sensorimotor leads occurred periodicallyfor many minutes in the resting record, but was never associated with a clinical seizure. Record continues in Fig. 6.
hand and arm movements; no facial motor responses could be produced. The lower half of the motor gyrus was mildly swollen and showed a 5-mm raised and discolored area at its upper end. The dura was closed and an epidural electrode array sewn in place, with Electrode 21 overlying the motor hand area and Electrode 14 situated just posterior and superior to the abnormal area (Fig. 7). With the patient awake, stimulation via the epidural electrodes again produced only hand and arm movement, but no facial motor response. By contrast, evoked response recording identified the boundary between the hand and face area; stimulation of both the upper lip and median nerve evoked a response
from a combination of Electrodes 14-15 (Fig. 7). Medial to this area only the nerve stimulation evoked a response. Recordings were made during six spontaneous seizures, all of which began at the site of Electrode 7 (Fig. 8). On the basis of this observation and the findings of sensorimotor localization, a resection of about the lower half of the motor gyrus was carried out. The medial limit of resection was at the boundary of face and upper extremity representation (Fig. 9). The subcortical tissue had a brownish discoloration, was abnormally soft, and did not appear to extend beyond the limits of the resection; pathological diagnosis was mixed glioma. Immediately after operation the patient ex-
350
J. Neurosurg. / Volume 49 / September, 1978
Surgery in epilepsy
FIG. 6. Continuation of spontaneous seizure shown in Fig. 5. The seizure discharge ultimately spreads to the other hemisphere.
In our total experience of extraoperative recording with both epidural electrode arrays and depth electrodes, we have encountered seven cases in which electrically induced
seizures provided false localization of the epileptogenic focus. This is not to say that electrical stimulation of cerebral cortex was successful in identifying the epileptogenic focus in all of the others. In the majority of patients, electrical stimulation (at an intensity that produced hand movement from the motor hand area) either failed to produce after-discharge or clinical seizure, or produced only electrical after-discharge without any clinical concomitant. The following case is an example of false localization by an electrically induced seizure, Case Report. This 20-year-old man (J.R.) suffered intractable seizures occurring three to five times a month for 11 years. The ictus was ushered in by a visual aura: vision "goes out" with a flickering traveling from left to right, head and eyes turn to the right usually followed by a tonic-clonic convulsion.
J. Neurosurg. / Volume 49 / September, 1978
35]
hibited only a mild left facial weakness. The next day he developed a dense paresis of the left upper extremity and face that cleared in 72 hours. Ten months postoperatively he is without neurological deficit (Fig. 10) and free of seizures. Comment. In addition to demonstrating the usefulness of sensory evoked responses for sensorimotor localization in epilepsy surgery, this case also illustrates its value for safe and precise excision of certain cerebral glial tumors and arteriovenous malformations.
Importance of Recording during Spontaneously Occurring Seizures
S. Goldring preceded by electrical seizure activity from electrodes overlying occipital cortex, sites where electrical stimulation failed to generate after-discharge or clinical seizure! The resected occipital lobe tissue showed reactive gliosis. The patient is seizure-free 22 months after surgery. Comment. In addition to demonstrating false localization by cortical stimulation, this case also shows that by using extraoperative ECG one can safely activate seizures, even generalized ones, in order to localize the true epileptogenic focus. Surgical Outcome
During 2 days of extraoperative recording, the patient failed to have a seizure. Electrical stimulation via each of the epidural recording electrodes at a stimulus strength just adequate for producing a motor hand response from the motor hand area failed to elicit the patient's clinical seizure. However, from two adjacent electrodes overlying the posterior temporal region, seizure discharge was readily produced accompanied by receptive dysphasia, both lasting 4.5 minutes. Thus, a seizure was produced, but it was not the patient's clinical seizure and he had never experienced such an episode before. Pentylenetetrazol activation was then carried out. After a total of 200 mg (50 mg/30 sec) the patient reported his visual aura, his head and eyes turned to the right, and he underwent a generalized seizure. This episode was
Twenty-four of the 29 patients who had undergone implantation of an epidural electrode array had either excision of an epileptogenic focus (21 patients) or a biopsy of cortical tissue (three patients). Eighteen of the patients who had excision have been followed for 1 to 10 years and, of these, 11 (61%)have had a good result. A good result is defined as: 1) no seizures; 2) a reduction in seizure frequency that permits educability or employment, neither of which was possible before; or 3) a reduction in seizure frequency that prevents institutionalization. There have been no infections and no hematomas. Indeed, the epidural electrode array acts as an epidural drain and, since it is removed in 24 to 48 hours, it should differ little from the usual epidural drain as a potential source of infection. In one of the first patients in whom an epidural electrode array was used, the bone flap was removed and steam-autoclaved before its insertion at the second procedure; however, it underwent ase0tic necrosis and had to be removed. Because all surgical manipulation can be accomplished under anesthesia and prolonged ECG can be carried out painlessly in an anxiety-free atmosphere, the method is uniquely suited for the management of children. Thus, 17 of the 29 patients managed with epidural electrode arrays were between the ages of 2 and 14 years. In 13 of these children, extraoperative recording led to excision of an epileptogenic focus. In the other four, a single dominant focus could not be identified; two of these had a cortical biopsy, and in two no tissue was removed. Ten of the patients who had excision of an epileptogenic focus have been followed for 1 year or more (Table 3). Seven (70%) have had a good
352
J. Neurosurg. / Volume 49 / September, 1978
Fie. 7. Photograph of the exposed brain (T.L.): anterior is to the left and midline is at bottom. Traces show sensory evoked responses. Observe that from Electrode pair 14-15 a response could be evoked from either the face or hand, identifying this area as the boundary between face and hand representation. Black line identifies central fissure. Electrical stimulation of cortical surface anterior to this line through epidural Electrodes 14, 20, and 21 produced hand and arm movements. No facial motor responses were produced at Electrode 14 or from any of the other electrode sites. Note small abnormal raised area just anterior to the site of Electrode 14 (see text for details).
S u r g e r y in e p i l e p s y
FIG. 8. Same case as in Fig. 7. Seizure activity is first seen in electrode combinations 6-7 and 7-8 indicating that seizure begins at Electrode 7. It is noteworthy that electrical stimulation at this site did not evoke a seizure, but did so at the site of Electrode 3 (see Fig. 7), an example of false localization of epileptogenic foci occasionally seen with electrical stimulation.
The variety of lesions encountered in our series is similar to that reported for the focal epilepsies treated at the Montreal Neurological Institute between 1961 and 1970. 8 Of the eight adults who had focal excision, histological examination showed nonspecific gliosis in three, gliosis and neuronal loss in
one, posttraumatic meningocerebral cicatrix in one, residuum of cerebral abscess in one, and changes of Sturge-Weber disease in one. In only one case was no pathological change defined. In all of our children, the excised tissue showed pathological change (Table 4). In addition to the entities encountered in the adults, there were lesions of tuberous sclerosis, focal cortical dysplasia, chronic encephalitis, ganglioglioma, mixed glioma, arteriovenous malformation, and heterotopia. In five patients, the diagnosis was known and there was radiological localization of a lesion before surgery. In the remainder, the diagnosis was unknown and
J. Neurosurg. / Volume 49 / September, 1978
353
result. Cases 8 and 9 had a moderate reduction in seizure frequency, but their quality of life has not changed significantly, and therefore they, together with Case 10 are considered as not benefiting from surgery.
Pathological Nature of the Epileptogenic Lesions
S. Goldring
F~c. 9. Same case as in Figs. 7 and 8. Operative photograph showing area of resection. Anterior is left and midline is at bottom. Ticket 14 identifies the boundary of facial and hand motor representation (see Fig. 7). Black line is central fissure.
TABLE 3
Children followed for I or more years Case No. 1" 2 3 4 5 6 7 8 9 10
Diagnosis
Result
arteriovenous malformation chronic encephalitis cicatrix and porencephalic cyst (birth injury) ganglioglioma tuberous sclerosis tuberous sclerosis cicatrix and porencephalic cyst (birth injury) ganglioglioma focal cortical dysplasia posttraumatic cicatrix & porencephalic cyst
no seizures no seizures no seizures no seizures marked reduction in seizure frequency marked reduction in seizure frequency marked reduction in seizure frequency moderate reduction in seizure frequenc moderate reduction in seizure frequenc no reduction in seizure frequency
*This patient did not have preoperative recordings with indwelling electrodes, but is included because localization of his epileptogenic focus was uniquely aided by recording of sensory evoked responses. He had continuous partial seizures involving the left hand. During surgery, under anesthesia, exploration of the central region with the stimulating electrode produced, from a wide area, movement that always started in the face. Identification Of the hand area (the site which related to the body part involved in the partial seizure) could only be accomplished by recording cortical responses to median nerve stimulation. This localizatign along with electrocorticography enabled the identification and correct excision of the epileptogenic focus. 354
J. Neurosurg. / Volume 49 / September, 1978
Surgery in epilepsy all radiological procedures were either normal, or only partially localized the extent of the lesion.
Discussion We found electrical inexcitability of the motor cortex in some of the children who showed a somatosensory evoked response. It is doubtful that this observation indicates only a pathological process that affects the motor cortex, but spares the somatosensory gyrus (for instance, a predominantly posterior frontal lesion). In Cases 4 and 5 (Table 2), the epileptogenic lesions were located predominantly in the occipitoparietal and posterior parietal regions, respectively. Yet a sensory evoked response was readily recorded, while electrical stimulation of the motor cortex failed to elicit movement. One possible explanation is that sensory evoked responses are less sensitive to the disease process than are cortically induced movements - - a differential susceptibility similar to that observed under general anesthesia. An alternative explanation, but probably less likely, may relate to a disparity in ontogenic development between these two responses. In the newborn rabbit a single sensory stimulus is capable of evoking a cortical response, although its configuration differs from that observed in the mature animal2 Repetitive cortical stimulation with brief shocks however, fails to recruit the neuronal population underlying the stimulating electrode to repetitive firing? a condition that is normally a prerequisite for the production of movement. Since the five children in whom we found the cortex electrically inexcitable were also the youngest, it is possible that at least in some (Cases 4 and 5, Table 2) the failure of cortical stimulation to produce movement relates to immaturity of the motor cortex. The use of extraoperative ECG in the surgical management of epilepsy has limitations. We have encountered instances where it would have been desirable to record or stimulate in areas between electrodes, but have been unable to do so because of the fixed electrode positions. We hope to resolve this problem by doubling the number of electrodes in the recording arrays from 24 to 48. This will result in inter-electrode distances of 1 cm rather than 1.5 and 2 cm as is the case now. The new arrays, coupled with an efJ. Neurosurg. / Volume 49 / September, 1978
FIG. 10. The patient (T.L.) performing fine finger movements. There were no motor deficits following resection shown in Fig. 9. See text for details.
ficient switching matrix under computer control, will permit facile scanning of the numerous recording sites so that changes can be made easily and quickly. The system should provide better spatial resolution of the epileptogenic boundary.
TABLE 4 Pathology of epileptogenic lesions in children
Case No.
Preop Diagnosis & Radiological Localization
1
yes
Pathology
posttraumatic cicatrix & porencephalic cyst 2 yes cicatrix & porencephalic cyst (birth injury) 3 yes cicatrix & porencephalic cyst (birth injury) 4 yes cicatrix & porencephalic cyst (birth injury) 5 yes tuberous sclerosis 6 partial radiological ganglioglioma localization only 7 partial radiological mixed glioma localization only 8 partial radiological heterotopia localization only 9 partial radiological focal cortical dysplasia localization only 10 neither ganglioglioma 11" neither reactive gliosis (cause unknown) 12" neither chronic encephalitis (suspect) 13 neither chronic encephalitis 14 neither tuberous sclerosis 15 neither arteriovenous malformation *Biopsy only. 355
S. Goldring 5. Goldring S, Aras E, Weber PC: Comparative study of sensory input to motor cortex in animals and man. EEG Clin Neurophysiol 29:537-550, 1970 6. Hunt WE, Goldring S: Maturation of evoked response of the visual cortex in the postnatal rabbit. EEG Ciin Neurophysiol 3:465--471, 1951 7. Kelly DL, Goldring S, O'Leary JL: Averaged evoked somatosensory responses from exposed cortex of man. Arch Neurol 13:1-9, 1965 8. Mathieson G: Pathologic aspects of epilepsy with special reference to the surgical pathology of focal cerebral seizures, in Purpura DP, Penry JK, Walter RD (eds): Neurosurgical Management of the Epilepsies. Advances in Neurology. New York: Raven Press, 1975, Vol. 8, pp 107-138 9. Penfield W, Jasper H: Epilepsy and the Functional Anatomy of the Human Brain. Boston: Little, Brown and Co, 1954, p 1 10. Purpura DP, Penry JK, Walter RD (eds): Neurosurgical Management of the Epilepsies. Advances in Neurology. New York: Raven Press, 1975, Vol. 8, 356 pp Acknowledgments 11. Rasmussen T: Cortical resection in the treatment of focal epilepsy, in Purpura DP, Penry This work was made possible by the technical JK, Walter RD (eds): Nenrosurgical Manageassistance of Isaac Edwards, Carl Pieper, and ment of the Epilepsies. Advances in Neurology. Lloyd Simpson. New York: Raven Press, 1975, Vol. 8, pp 139-154 References 12. Stohr PE, Goldring S: Origin of somatosen1. Crandall PH: Developments in direct recordsory evoked scalp responses in man. J ings from epileptogenic regions in the surgical Neurosurg 31:117-127, 1969 treatment of partial epilepsies, in Brazier 13. Talairach J, Bancaud J: Lesion, "irrative" MAB (ed): Epilepsy: Its Phenomena in Man. zone and epileptogenic focus. Confin Neurol New York/San Francisco/London: Academic 27:91-94, 1966 Press, 1973, pp 287-309 14. Van Buren JM, Ajmone Marsan C, Mutsuga 2. Crandall PH: Postoperative management and N: Temporal lobe seizures with additional foci criteria for evaluation, in Purpura DP, Penry treated by resection. J Neurosurg 43:596-607, JK, Walter RD (eds): Neurosurgical Manage1975 ment of the Epilepsies. Advances in Neurology. New York: Raven Press, 1975, Vol. 8, pp 265-280 The work was aided by U.S. Public Health Ser3. Do Carmo R J: Direct cortical and recruiting responses in postnatal rabbit. J Neurophysiol vice Grants NS 06947, NS 04513, and NS 05580. Address reprint requests to: Sidney Goldring, 23:496-504, 1960 4. Dodson WE, Prensky AL, DeVivo DC, et al: M.D. Department of Neurology and Neurological Management of seizure disorders: Selected Surgery (Neurological Surgery), Barnes Hospital aspects. Part 1. J Pediatr 89:527-540, 1976 Plaza, St. Louis, Missouri 63110.
Another limitation is inherent in the rationale for localizing epileptogenic foci with extraoperative ECG, namely, identification of the focus on the basis of recordings made during spontaneously occurring seizures. This limits the procedure predominantly to those patients who experience frequent daily seizures. Finally, this method necessitates a secondary craniotomy. This disadvantage is outweighed, however, by the following considerations: 1) we have, thus far, not experienced a ~'ound infection; 2) the long, arduous operation (for both patient and surgeon) using intraoperative ECG under local anesthesia is avoided; with extraoperative ECG, each stage under anesthesia requires only several hours; 3) the procedure permits management oY both children and those adults who carnot undergo surgery under local anesthesia; and 4) localization of the epileptogenic focus is more precise.
356
J. Neurosurg. / Volume 49 / September, 1978
A method for surgical management of focal epilepsy, especially as it relates to children SIDNEY GOLDRING, M . D .
Department of Neurology and Neurological Surgery (Neurological Surgery), Washington University School of Medicine, St. Louis, Missouri ~" A method of surgical management for intractable epilepsy is described. The essential features are: 1) all surgical manipulation is carried out under general, rather than local, anesthesia; 2) the sensorimotor region is readily identified in the anesthetized patient by recording cortical sensory evoked responses; and 3) the epileptogenic focus is localized by extraoperative electrocorticography via indwelling epidural electrode arrays, localization deriving from recordings made during spontaneously occurring clinical seizures. Cases are presented to demonstrate that: 1) in some instances, recording of sensory evoked responses is the only means of sensorimotor localization in both the awake and anesthetized patient, and 2) spontaneous and electrically induced electroencephalographic seizure activity may provide false localization of the focus, the correct localization requiring recordings made during spontaneous clinical seizures. The outcome of surgery and the various epileptogenic lesions encountered are described. A good result has been achieved in 61% of patients followed 1 to 10 years. When the results obtained in children are analyzed alone, 70% have benefited from surgery. KEY WORDS 9 epilepsy 9 surgical management 9 sensory evoked response 9 extraoperative electrocorticography
HE method used most commonly in the surgical management of focal epilepsy requires that surgery be carried out under local anesthesia/-11 General anesthesia is avoided because it suppresses both spontaneously and electrically induced electroencephalographic (EEG) seizure activity. It also suppresses the motor responses to electrical stimulation of the cortical surface, information that is used to identify the sensorimotor region. Not all patients, however, can undergo surgery under local anesthesia with the necessary degree of cooperation, and this is especially true in children. Also, it has become increasingly clear that the most reliable evidence for localizing an epilep-
T
togenic focus is that which is obtained in electrical recordings made during a spontaneous convulsion. 1,2,',Is,1' During surgery, abnormal interictal electrical discharges may be unreliable, whether they occur spontaneously or are induced by electrical stimulation (afterdischarge). Even an after-discharge associated with the partial seizure from which the clinical convulsion develops sometimes provides erroneous localization. Furthermore, the partial seizure may spread rapidly into a generalized one, an undesirable event during surgery. A consequence of these considerations has been the development of a method that permits a more precise localization of the epilep-
344
J. Neurosurg. / Volume 49 / September, 1978
Surgery in epilepsy togenic focus and makes possible definitive treatment of patients, especially children, who heretofore would not have been considered for surgery. The important features of the procedure are: 1) all surgical manipulation is carried out under anesthesia; 2) the sensorimotor region is readily identified in the anesthetized patient by recording cortical sensory evoked responses; and 3) the epileptogenic focus is localized by extraoperative electrocorticography (ECG) via indwelling epidural and/or depth electrode arrays, with the localization derived from recordings made during spontaneously occurring seizures. The purpose of this paper is to review our experience in which indwelling epidural electrode arrays were used to localize the epileptogenic focus.
TABLE 1
Surgical management of intractable focal epilepsy: lO-year experience*
Method of Epileptogenie Focus Localization extraoperative ECG using epidural electrode arrays extraoperative ECG using depth electrodes intraoperative ECG or scalp EEG and x-ray film total *ECG = electrocorticography; EEG = encephalography.
No. of Cases 29 46 17 92 electro-
A craniotomy is performed under general endotracheal anesthesia, and the brain exposed and inspected. The dural incision is made around the entire circumference of the exposed dura, producing a free dural graft. This is done to minimize the development of
epidural clot which could attenuate the recorded electrical activity. Cortical sensory evoked responses are then recorded to identify the sensorimotor region. Previously, we5,7,12have shown that, under the conditions of recording to be described, a somatosensory evoked response is recorded only from the somatosensory and motor gyri. This was true in both the anesthetized and awake state. The somatosensory area always shows a response, but the appearance of a response in the motor cortex is variable. The observation that a response appears only in the sensorimotor region is the basis for using sensory evoked responses for sensorimotor localization. A Silastic template holding nine linearly oriented electrodes, each separated by a distance of 1.5 cm, is placed on the cortical surface in a plane parallel to the midline (Fig. 1). Simultaneous records are made from each adjacent pair of electrodes (that is, 1-2, 2-3, 3-4, etc) while the contralateral median nerve is electrically pulsed transcutaneously at the wrist. A special-purpose computer is used to quickly expose the sensory evoked response. Without computer processing, human cortical electrical responses to somatosensory stimulation are barely visible above the background waves of the spontaneous ECG. The computer stores and adds a brief epoch (100 to 200 msec) of electrical activity following each shock for approximately 25 stimuli. The spontaneous random brain waves do not add significantly, but, because the evoked responses have consistent polarity and latency, they summate and become clearly exposed above the background. The face area is similarly identified using a light tap to the upper or lower lip as
J. Neurosurg. / Volume 49 / September, 1978
345
Clinical Material and Methods
Patient Population During the years 1968 to 1978, 92 epileptic patients have come under our care. Of these, 29 patients were evaluated with extraoperative ECG by means of indwelling epidural electrode arrays, 46 were studied with extraoperative ECG using indwelling depth electrodes, and in 17 the focus was localized by intraoperative ECG or by scalp and radiological imaging procedures (Table 1). All but four of the 46 patients who had depth electrode ECG had temporal lobe foci. Those 29 patients undergoing extraoperative ECG with epidural electrode arrays were suspected of having a cortical focus in areas other than the medial temporal lobe. Seventeen were children (ages 2 to 14 years) who were having many seizures (10 to 50) daily, in spite of intensive medical management which pushed all combinations of anticonvulsant medication to toxic levels, and included a ketogenic diet. Seizures in the adults were equally intractable, although in two patients they only occurred several times monthly. Method
S. Goldring
FIG. 1. Identification of somatosensory area by recording cortical sensory evoked responses. Linear electrode array overlying cerebral cortex is shown (lower). Recordings are made simultaneously from each adjacent pair of electrodes (i.e., 1-2, 2-3, etc.). The numbers identifying each trace correspond to the numbers on the electrode array. The response profile establishes the area underlying Electrode 5 as the somatosensory hand area because: 1) the only traces that show a significant response are 4-5 and 5-6; 2) only Electrode 5 is common to both electrode combinations; and 3) the polarity of the two responses is reversed. Calibration pulse is 12.5 ~V and 18 msec; positive at the first electrode of each pair (G1) is up.
the sensory stimulus. We have used this method routinely for the past 10 years, even in patients under local anesthesia, because it has proved to be a quicker method of identifying the sensorimotor region than the use of electrical stimulation of the cortical surface to produce movement. After the sensorimotor region is identified, the dura is closed and a quasi-rectangular Silastic template holding 24 electrodes is sewn in epidurally (Fig. 2). The electrodes are arranged in rows with inter-electrode dis-
tances of 1.5 cm and 2 cm in the sagittal and coronal plane, respectively. The electrode template is placed so that it spans the suspected area of epileptogenicity, its surrounding area, and the sensorimotor region. Care is taken to assure that one of the electrodes immediately overlies the motor or sensory hand area. When this condition is met, the number and spacing of the electrodes is such that it is then possible to record both sensory evoked responses to the face and hand, and also to elicit motor responses in the
346
J. Neurosurg. / Volume 49 / September, 1978
S u r g e r y in e p i l e p s y
Fro. 2. Epidural electrode array sewn in place.
face and hand by stimulating through the same electrodes. The remainder of the closure is completed and recordings are begun on the following day. Both the patient's behavior and the ECG are continuously recorded and displayed side by side on the split screen of a television monitor. The taping of these recordings permits the additional opportunity for replays of any seizure that might occur. This is necessary to determine accurately which cortical area shows the first sign of abnormal activity in the moments preceding the onset of clinical convulsion, essential information for localizing the focus. Because the patient is awake, it is now possible to record motor responses to transdural cortical stimulation as well as sensory evoked responses, identifying the relationship between the sensorimotor region and the focus. If the recordings are from the dominant hemisphere,
speech arrest due to electrical stimulation is also used to determine the boundaries of the speech area. Thus, all of the techniques used in localizing an epileptogenic focus during the usual surgical procedure under local anesthesia can be performed by recording and stimulating epidurally with indwelling epidural electrode arrays. The recordings are performed in a leisurely manner, free from the usual stresses of the operating room, and the opportunity of recording the ECG during a spontaneous clinical seizure, or, if necessary, an activated one, is markedly enhanced. Recording with the epidural electrode arrays usually continues for only 24 to 48 hours, since most of these patients have at least several seizures a day. If the observations identify the boundaries of a single epileptogenic focus, then the brain tissue within the epileptogenic zone is excised at a
J. Neurosurg. / Volume 49 / September, 1978
347
S. G o l d r i n g
The following two cases illustrate the usefulness of sensory evoked responses and show that in some instances they are the only means of obtaining sensorimotor localization in the anesthetized or awake state. Case Report: A 4-year-ol'd gfff (C. ~ : ) fiad had intractable seizures since 6 months of age, progressive mental retardation with loss of all language function, and left hemiparesis which waxed and waned in severity. Seizures started with left adversive movement of the head and eyes, and were occurring 10 to 12 times a day at the time of surgery. Complete radiological work-up revealed only blunting of the right frontal horn on computerized tomography (CT) scan. Upon exposure, the brain appeared normal. Under general anesthesia
the sensorimotor region could b e identified only by recording sensory evoked responses; the cortex was electrically inexcitable for producing movement. An epidural electrode array was sewn in place with Electrode 19 overlying the sensory hand area (Fig. 3). Recording of evoked responses in the awake state again identified the sensorimotor region, but cortical stimulation through all electrodes (monopolar and bipolar combinations) still failed to produce movement. The resting ECG showed spontaneous spiking in the somatosensory hand area, and the first two recorded seizures began in that gyrus (Fig. 4). At the onset of seizure activity the left hand clenched, and there was then stiffening of the arm and adversion of head and eyes to the left. The subsequent five seizures, however, began in the right prefrontal cortex with a left adversive clinical onset (Figs. 5 and 6). On the basis of these findings, a right prefrontal cortical excision spared the sensorimotor region; the pathological diagnosis was focal cortical dysplasia. The seizure frequency has decreased significantly, the hemiparesis has disappeared, and the child is developing a vocabulary. Comment. fn tills case we ffem,onstrated that seizures occasionally could begin in the sensorimotor gyri, the child had exhibited a progressive hemiparesis preoperati, vely, and electrical stimulation of the cortex failed to produce any movement; all thq~se data suggested that the sensorimotor ~area was non-functional. However, the recording of sensory evoked responses not only :identified the sensorimotor region, but also showed that at least the somatosensory area was physiologically functional. Had we not leaLrned this we might have included the sem~ory and motor gyri in the resection and risk ed a permanent hemiparesis. This case also demonstrates the importance of reoording as many spontaneous seizures as possJible. We have encountered four other children in whom electrical stimulation of cerebral cortex failed to produce any movement, and the sensorimotor region was identified only by recording sensory evoked responses (Table 2). l'wo of these children also had a preoperative hemiparesis, which disappeared after surgery. The progressive hemiparetic disabilities in all of these children obviously represented sustained postictal deficit due to frequent focal seizures.
348
J. Neurosurg. / Volume 49 / September, 1978
FIG. 3. Photograph of the exposed brain (C.W.): anterior is to the left and midline is at the bottom. The numbered tickets overlie the precise areas from which recordings were made with the epidural electrode array. The area underlying No. 19 was identified as the somatosensory handarea by recording responses from electrode combinations 20-19 and 19-18 shown at bottom; note polarity reversals of response. The area underlying No. 14 was identified as the somatosensory face area. The black line between 22-21, 16-15 etc. is the posterior limit of the frontal lobe resection.
second craniotomy; if not, the surface electrode array is removed and the operative wound closed. Results
Use of Sensory Evoked Responses for Localizing Cerebral Function
Surgery in epilepsy
FIG. 4. Seizure begins in sensorimotor region at electrode 19 (see Fig. 3). Observe the appearance of seizure activity at electrode sites 14-15 which is also in sensorimotor domain. The initial high voltage burst of activity (arrows) ushered in both seizures that began in the sensorimotor region. Note recordings from left cortical surface. These were made from epidural electrodes introduced through a small (1 cm) trephine opening. Such recordings on the side contralateral to the hemisphere of predominant interest have been made in several patients, especially those having adversive seizures
Case Report. This 13-year-old boy (T.L.) had suffered focal seizures since the age of 6 years. The ictus began with a feeling of numbness in the left face and progressed to facial twitching that spread to involve the left arm; the seizure frequently became generalized. A CT scan suggested a hypodense area in the right posterior frontal region; the remainder of the radiological workup showed no abnormality. During surgery the sensorimotor region was localized by recording sensory evoked responses to electrical stimulatiort of the median nerve and tactile stimulation of the face. When the level of anesthesia was decreased, electrical stimulation of cerebral cortex elicited only
*Motor cortex was electrically inexcitable; no movement could be obtained by electrical stimulation in both the anesthetized and awake state.
J. Neurosurg. / Volume 49 / September, 1978
349
TABLE 2 Children in whom sensorimotor identification was made possible only by recording sensory evoked responses*
Case Age No. (yrs) I 2 3 4 5
Diagnosis
Excision
3 2
focal cortical dysplasia frontal tuberous sclerosis frontal & temporal 4 tuberous sclerosis frontal 2~/2 heterotopia occipitoparietal 2~A ganglioglioma posterior parietal
S. G o l d r i n g
FIG. 5. Spontaneous clinical seizure first evident electrographicallyin pre-frontal cortex (arrows, and refer to Fig. 3) as record becomes iso-electric. The spiking in sensorimotor leads occurred periodicallyfor many minutes in the resting record, but was never associated with a clinical seizure. Record continues in Fig. 6.
hand and arm movements; no facial motor responses could be produced. The lower half of the motor gyrus was mildly swollen and showed a 5-mm raised and discolored area at its upper end. The dura was closed and an epidural electrode array sewn in place, with Electrode 21 overlying the motor hand area and Electrode 14 situated just posterior and superior to the abnormal area (Fig. 7). With the patient awake, stimulation via the epidural electrodes again produced only hand and arm movement, but no facial motor response. By contrast, evoked response recording identified the boundary between the hand and face area; stimulation of both the upper lip and median nerve evoked a response
from a combination of Electrodes 14-15 (Fig. 7). Medial to this area only the nerve stimulation evoked a response. Recordings were made during six spontaneous seizures, all of which began at the site of Electrode 7 (Fig. 8). On the basis of this observation and the findings of sensorimotor localization, a resection of about the lower half of the motor gyrus was carried out. The medial limit of resection was at the boundary of face and upper extremity representation (Fig. 9). The subcortical tissue had a brownish discoloration, was abnormally soft, and did not appear to extend beyond the limits of the resection; pathological diagnosis was mixed glioma. Immediately after operation the patient ex-
350
J. Neurosurg. / Volume 49 / September, 1978
Surgery in epilepsy
FIG. 6. Continuation of spontaneous seizure shown in Fig. 5. The seizure discharge ultimately spreads to the other hemisphere.
In our total experience of extraoperative recording with both epidural electrode arrays and depth electrodes, we have encountered seven cases in which electrically induced
seizures provided false localization of the epileptogenic focus. This is not to say that electrical stimulation of cerebral cortex was successful in identifying the epileptogenic focus in all of the others. In the majority of patients, electrical stimulation (at an intensity that produced hand movement from the motor hand area) either failed to produce after-discharge or clinical seizure, or produced only electrical after-discharge without any clinical concomitant. The following case is an example of false localization by an electrically induced seizure, Case Report. This 20-year-old man (J.R.) suffered intractable seizures occurring three to five times a month for 11 years. The ictus was ushered in by a visual aura: vision "goes out" with a flickering traveling from left to right, head and eyes turn to the right usually followed by a tonic-clonic convulsion.
J. Neurosurg. / Volume 49 / September, 1978
35]
hibited only a mild left facial weakness. The next day he developed a dense paresis of the left upper extremity and face that cleared in 72 hours. Ten months postoperatively he is without neurological deficit (Fig. 10) and free of seizures. Comment. In addition to demonstrating the usefulness of sensory evoked responses for sensorimotor localization in epilepsy surgery, this case also illustrates its value for safe and precise excision of certain cerebral glial tumors and arteriovenous malformations.
Importance of Recording during Spontaneously Occurring Seizures
S. Goldring preceded by electrical seizure activity from electrodes overlying occipital cortex, sites where electrical stimulation failed to generate after-discharge or clinical seizure! The resected occipital lobe tissue showed reactive gliosis. The patient is seizure-free 22 months after surgery. Comment. In addition to demonstrating false localization by cortical stimulation, this case also shows that by using extraoperative ECG one can safely activate seizures, even generalized ones, in order to localize the true epileptogenic focus. Surgical Outcome
During 2 days of extraoperative recording, the patient failed to have a seizure. Electrical stimulation via each of the epidural recording electrodes at a stimulus strength just adequate for producing a motor hand response from the motor hand area failed to elicit the patient's clinical seizure. However, from two adjacent electrodes overlying the posterior temporal region, seizure discharge was readily produced accompanied by receptive dysphasia, both lasting 4.5 minutes. Thus, a seizure was produced, but it was not the patient's clinical seizure and he had never experienced such an episode before. Pentylenetetrazol activation was then carried out. After a total of 200 mg (50 mg/30 sec) the patient reported his visual aura, his head and eyes turned to the right, and he underwent a generalized seizure. This episode was
Twenty-four of the 29 patients who had undergone implantation of an epidural electrode array had either excision of an epileptogenic focus (21 patients) or a biopsy of cortical tissue (three patients). Eighteen of the patients who had excision have been followed for 1 to 10 years and, of these, 11 (61%)have had a good result. A good result is defined as: 1) no seizures; 2) a reduction in seizure frequency that permits educability or employment, neither of which was possible before; or 3) a reduction in seizure frequency that prevents institutionalization. There have been no infections and no hematomas. Indeed, the epidural electrode array acts as an epidural drain and, since it is removed in 24 to 48 hours, it should differ little from the usual epidural drain as a potential source of infection. In one of the first patients in whom an epidural electrode array was used, the bone flap was removed and steam-autoclaved before its insertion at the second procedure; however, it underwent ase0tic necrosis and had to be removed. Because all surgical manipulation can be accomplished under anesthesia and prolonged ECG can be carried out painlessly in an anxiety-free atmosphere, the method is uniquely suited for the management of children. Thus, 17 of the 29 patients managed with epidural electrode arrays were between the ages of 2 and 14 years. In 13 of these children, extraoperative recording led to excision of an epileptogenic focus. In the other four, a single dominant focus could not be identified; two of these had a cortical biopsy, and in two no tissue was removed. Ten of the patients who had excision of an epileptogenic focus have been followed for 1 year or more (Table 3). Seven (70%) have had a good
352
J. Neurosurg. / Volume 49 / September, 1978
Fie. 7. Photograph of the exposed brain (T.L.): anterior is to the left and midline is at bottom. Traces show sensory evoked responses. Observe that from Electrode pair 14-15 a response could be evoked from either the face or hand, identifying this area as the boundary between face and hand representation. Black line identifies central fissure. Electrical stimulation of cortical surface anterior to this line through epidural Electrodes 14, 20, and 21 produced hand and arm movements. No facial motor responses were produced at Electrode 14 or from any of the other electrode sites. Note small abnormal raised area just anterior to the site of Electrode 14 (see text for details).
S u r g e r y in e p i l e p s y
FIG. 8. Same case as in Fig. 7. Seizure activity is first seen in electrode combinations 6-7 and 7-8 indicating that seizure begins at Electrode 7. It is noteworthy that electrical stimulation at this site did not evoke a seizure, but did so at the site of Electrode 3 (see Fig. 7), an example of false localization of epileptogenic foci occasionally seen with electrical stimulation.
The variety of lesions encountered in our series is similar to that reported for the focal epilepsies treated at the Montreal Neurological Institute between 1961 and 1970. 8 Of the eight adults who had focal excision, histological examination showed nonspecific gliosis in three, gliosis and neuronal loss in
one, posttraumatic meningocerebral cicatrix in one, residuum of cerebral abscess in one, and changes of Sturge-Weber disease in one. In only one case was no pathological change defined. In all of our children, the excised tissue showed pathological change (Table 4). In addition to the entities encountered in the adults, there were lesions of tuberous sclerosis, focal cortical dysplasia, chronic encephalitis, ganglioglioma, mixed glioma, arteriovenous malformation, and heterotopia. In five patients, the diagnosis was known and there was radiological localization of a lesion before surgery. In the remainder, the diagnosis was unknown and
J. Neurosurg. / Volume 49 / September, 1978
353
result. Cases 8 and 9 had a moderate reduction in seizure frequency, but their quality of life has not changed significantly, and therefore they, together with Case 10 are considered as not benefiting from surgery.
Pathological Nature of the Epileptogenic Lesions
S. Goldring
F~c. 9. Same case as in Figs. 7 and 8. Operative photograph showing area of resection. Anterior is left and midline is at bottom. Ticket 14 identifies the boundary of facial and hand motor representation (see Fig. 7). Black line is central fissure.
TABLE 3
Children followed for I or more years Case No. 1" 2 3 4 5 6 7 8 9 10
Diagnosis
Result
arteriovenous malformation chronic encephalitis cicatrix and porencephalic cyst (birth injury) ganglioglioma tuberous sclerosis tuberous sclerosis cicatrix and porencephalic cyst (birth injury) ganglioglioma focal cortical dysplasia posttraumatic cicatrix & porencephalic cyst
no seizures no seizures no seizures no seizures marked reduction in seizure frequency marked reduction in seizure frequency marked reduction in seizure frequency moderate reduction in seizure frequenc moderate reduction in seizure frequenc no reduction in seizure frequency
*This patient did not have preoperative recordings with indwelling electrodes, but is included because localization of his epileptogenic focus was uniquely aided by recording of sensory evoked responses. He had continuous partial seizures involving the left hand. During surgery, under anesthesia, exploration of the central region with the stimulating electrode produced, from a wide area, movement that always started in the face. Identification Of the hand area (the site which related to the body part involved in the partial seizure) could only be accomplished by recording cortical responses to median nerve stimulation. This localizatign along with electrocorticography enabled the identification and correct excision of the epileptogenic focus. 354
J. Neurosurg. / Volume 49 / September, 1978
Surgery in epilepsy all radiological procedures were either normal, or only partially localized the extent of the lesion.
Discussion We found electrical inexcitability of the motor cortex in some of the children who showed a somatosensory evoked response. It is doubtful that this observation indicates only a pathological process that affects the motor cortex, but spares the somatosensory gyrus (for instance, a predominantly posterior frontal lesion). In Cases 4 and 5 (Table 2), the epileptogenic lesions were located predominantly in the occipitoparietal and posterior parietal regions, respectively. Yet a sensory evoked response was readily recorded, while electrical stimulation of the motor cortex failed to elicit movement. One possible explanation is that sensory evoked responses are less sensitive to the disease process than are cortically induced movements - - a differential susceptibility similar to that observed under general anesthesia. An alternative explanation, but probably less likely, may relate to a disparity in ontogenic development between these two responses. In the newborn rabbit a single sensory stimulus is capable of evoking a cortical response, although its configuration differs from that observed in the mature animal2 Repetitive cortical stimulation with brief shocks however, fails to recruit the neuronal population underlying the stimulating electrode to repetitive firing? a condition that is normally a prerequisite for the production of movement. Since the five children in whom we found the cortex electrically inexcitable were also the youngest, it is possible that at least in some (Cases 4 and 5, Table 2) the failure of cortical stimulation to produce movement relates to immaturity of the motor cortex. The use of extraoperative ECG in the surgical management of epilepsy has limitations. We have encountered instances where it would have been desirable to record or stimulate in areas between electrodes, but have been unable to do so because of the fixed electrode positions. We hope to resolve this problem by doubling the number of electrodes in the recording arrays from 24 to 48. This will result in inter-electrode distances of 1 cm rather than 1.5 and 2 cm as is the case now. The new arrays, coupled with an efJ. Neurosurg. / Volume 49 / September, 1978
FIG. 10. The patient (T.L.) performing fine finger movements. There were no motor deficits following resection shown in Fig. 9. See text for details.
ficient switching matrix under computer control, will permit facile scanning of the numerous recording sites so that changes can be made easily and quickly. The system should provide better spatial resolution of the epileptogenic boundary.
TABLE 4 Pathology of epileptogenic lesions in children
Case No.
Preop Diagnosis & Radiological Localization
1
yes
Pathology
posttraumatic cicatrix & porencephalic cyst 2 yes cicatrix & porencephalic cyst (birth injury) 3 yes cicatrix & porencephalic cyst (birth injury) 4 yes cicatrix & porencephalic cyst (birth injury) 5 yes tuberous sclerosis 6 partial radiological ganglioglioma localization only 7 partial radiological mixed glioma localization only 8 partial radiological heterotopia localization only 9 partial radiological focal cortical dysplasia localization only 10 neither ganglioglioma 11" neither reactive gliosis (cause unknown) 12" neither chronic encephalitis (suspect) 13 neither chronic encephalitis 14 neither tuberous sclerosis 15 neither arteriovenous malformation *Biopsy only. 355
S. Goldring 5. Goldring S, Aras E, Weber PC: Comparative study of sensory input to motor cortex in animals and man. EEG Clin Neurophysiol 29:537-550, 1970 6. Hunt WE, Goldring S: Maturation of evoked response of the visual cortex in the postnatal rabbit. EEG Ciin Neurophysiol 3:465--471, 1951 7. Kelly DL, Goldring S, O'Leary JL: Averaged evoked somatosensory responses from exposed cortex of man. Arch Neurol 13:1-9, 1965 8. Mathieson G: Pathologic aspects of epilepsy with special reference to the surgical pathology of focal cerebral seizures, in Purpura DP, Penry JK, Walter RD (eds): Neurosurgical Management of the Epilepsies. Advances in Neurology. New York: Raven Press, 1975, Vol. 8, pp 107-138 9. Penfield W, Jasper H: Epilepsy and the Functional Anatomy of the Human Brain. Boston: Little, Brown and Co, 1954, p 1 10. Purpura DP, Penry JK, Walter RD (eds): Neurosurgical Management of the Epilepsies. Advances in Neurology. New York: Raven Press, 1975, Vol. 8, 356 pp Acknowledgments 11. Rasmussen T: Cortical resection in the treatment of focal epilepsy, in Purpura DP, Penry This work was made possible by the technical JK, Walter RD (eds): Nenrosurgical Manageassistance of Isaac Edwards, Carl Pieper, and ment of the Epilepsies. Advances in Neurology. Lloyd Simpson. New York: Raven Press, 1975, Vol. 8, pp 139-154 References 12. Stohr PE, Goldring S: Origin of somatosen1. Crandall PH: Developments in direct recordsory evoked scalp responses in man. J ings from epileptogenic regions in the surgical Neurosurg 31:117-127, 1969 treatment of partial epilepsies, in Brazier 13. Talairach J, Bancaud J: Lesion, "irrative" MAB (ed): Epilepsy: Its Phenomena in Man. zone and epileptogenic focus. Confin Neurol New York/San Francisco/London: Academic 27:91-94, 1966 Press, 1973, pp 287-309 14. Van Buren JM, Ajmone Marsan C, Mutsuga 2. Crandall PH: Postoperative management and N: Temporal lobe seizures with additional foci criteria for evaluation, in Purpura DP, Penry treated by resection. J Neurosurg 43:596-607, JK, Walter RD (eds): Neurosurgical Manage1975 ment of the Epilepsies. Advances in Neurology. New York: Raven Press, 1975, Vol. 8, pp 265-280 The work was aided by U.S. Public Health Ser3. Do Carmo R J: Direct cortical and recruiting responses in postnatal rabbit. J Neurophysiol vice Grants NS 06947, NS 04513, and NS 05580. Address reprint requests to: Sidney Goldring, 23:496-504, 1960 4. Dodson WE, Prensky AL, DeVivo DC, et al: M.D. Department of Neurology and Neurological Management of seizure disorders: Selected Surgery (Neurological Surgery), Barnes Hospital aspects. Part 1. J Pediatr 89:527-540, 1976 Plaza, St. Louis, Missouri 63110.
Another limitation is inherent in the rationale for localizing epileptogenic foci with extraoperative ECG, namely, identification of the focus on the basis of recordings made during spontaneously occurring seizures. This limits the procedure predominantly to those patients who experience frequent daily seizures. Finally, this method necessitates a secondary craniotomy. This disadvantage is outweighed, however, by the following considerations: 1) we have, thus far, not experienced a ~'ound infection; 2) the long, arduous operation (for both patient and surgeon) using intraoperative ECG under local anesthesia is avoided; with extraoperative ECG, each stage under anesthesia requires only several hours; 3) the procedure permits management oY both children and those adults who carnot undergo surgery under local anesthesia; and 4) localization of the epileptogenic focus is more precise.
356
J. Neurosurg. / Volume 49 / September, 1978
