Transcallosal resection of hypothalamic hamartomas in patients with intractable epilepsy

ARTICLE

Auteur(s) : A. Simon Harvey^1,2,4,7, Jeremy L. Freeman^1,2,3,4,7, Samuel F. Berkovic^1,5,7, Jeffrey V. Rosenfeld^1,6

1. Children’s Epilepsy Program,
2. Department of Neurology and
3. Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Victoria, Australia;
4. Departments of Paediatrics and
5. Medicine, University of Melbourne, Victoria, Australia;
6. Departments of Neurosurgery and Surgery, The Alfred Hospital and Monash University, Prahran, Victoria, Australia; and
7. Epilepsy Research Collaborative Centre, Austin and Repatriation Medical Centre, Heidelberg, Victoria, Australia.

Presented at the International Symposium on Hypothalamic Hamartoma and Epilepsy, Montreal Neurological Institute, Montreal, Canada, November 29^th 2001.

Resections for hypothalamic hamartoma (HH) in patients with intractable epilepsy have generally employed trans-sylvian or subfrontal approaches to the lesion, with a significant number of patients having residual lesion, ongoing seizures or neurovascular complications (see review and table in Rosenfeld et al., 2001) [1]. A recently published surgical series of 13 patients collected from four centres highlights the potential complications associated with these approaches [2]. Minimally invasive endoscopic, stereotactic [3] and radiosurgical [4, 5] approaches to HH offer promise, but all experience is limited. HH that cause epilepsy usually have a significant third ventricular component [6], with attachment to the mammillary body on one or both sides. This intrahypothalamic component may be important in the generation of seizures. We previously reported our early experience with resection of HH via a transcallosal interforniceal approach to the third ventricle [1, 7], the advantages of this approach being the clear view gained of the HH in the third ventricle, the ability to resect the HH using anatomical landmarks, minimal cortical disruption accessing the HH, and avoidance of important neurovascular structures which lie beneath the HH. To date, the neurosurgeon author (JVR) has performed transcallosal resection of HH in 44 children and adults with refractory gelastic epilepsy. Preoperative and postoperative findings are presented here for the series of 29 consecutive patients investigated and operated at the Royal Children’s Hospital, Melbourne until June 2002, for whom minimum postoperative follow-up is 12 months.

Patients and methods

Twenty-nine patients with HH and intractable epilepsy underwent surgery at the Children’s Epilepsy Program of the Royal Children’s Hospital, Melbourne; 24 patients underwent surgery at approximately monthly intervals over the last two years. Patient referrals were from the United States of America in 12, Australia in 7, United Kingdom in 5, Hong Kong in 3, Germany in 1 and Tanzania in 1.

Preoperative and postoperative evaluations

All patients underwent detailed preoperative and postoperative clinical, EEG, imaging, psychological, endocrine and ophthalmological assessments. The protocol for endocrine investigation is presented elsewhere in this issue (Freeman et al.). Neurological examinations and ophthalmological assessments were performed preoperatively and postoperatively in all cases, with visual perimetry performed in cooperative patients.
Conventional volumetric MRI acquisitions at 1.5 T were obtained in all patients, with thin, reformatted slices. High-resolution, high-contrast, T2-weighted images in coronal, axial and sagittal planes were also obtained and were found to be most helpful in showing the HH, the normal hypothalamic grey nuclei, and the heavily myelinated fornices, mammillary bodies and mammillothalamic tracts. MRI was repeated postoperatively using an identical protocol and the percentage of HH resected was estimated.
Detailed seizure histories were obtained and past medical notes and EEG reports were reviewed where available. Prolonged video-EEG monitoring was performed in 28/29 patients. Ictal or early postictal SPECT with Tc^99m-HMPAO was combined with video-EEG monitoring in 24 patients, and interictal SPECT was performed in 19. EEG was repeated after surgery and any change in spike-wave activity between preoperative and postoperative EEGs was calculated.
Neuropsychological assessment of cognitive abilities, behavioural problems and the impact of epilepsy was performed in all cases. The test batteries administered depended on each patient’s age and developmental abilities. Where possible, detailed assessment of memory was undertaken. Neuropsychological assessment was repeated postoperatively.
Due to the large proportion of patients referred from overseas, preoperative testing was generally performed in the week prior to surgery. Similarly, postoperative neurological, EEG, MRI, endocrine, visual and neuropsychological evaluations were performed between 2-3 weeks following surgery in most patients. Long-term follow-up of seizure outcome, medication usage, development, behaviour, weight and biochemical testing was in person for the seven Australian patients and via correspondence with the patient’s family and treating doctors in the 22 overseas patients. Analysis of seizure outcome with respect to patient and HH features was done using Fisher’s exact test for categorical data and the Mann-Whitney test for continuous age data.

Patient characteristics

The mean age at transcallosal surgery was 10 years (range 4-23 years). Prior treatment failures included anticonvulsant drugs in 29 patients, a ketogenic diet in six and vagal nerve stimulation in four. One patient had undergone prior temporal lobectomy and one a frontotemporal corticectomy. Eight patients had undergone prior HH surgery including resection via a subfrontal or trans-sylvian approachin seven, radiofrequency thermocoagulation in three and stereotactic radiosurgery in one.
The HH varied in size on MRI from a small spheroid approximately 7mm in diameter to a large irregular lesion with diameters of 28 × 34 × 42 mm in three planes. The proportion of the lesion lying above the normal level of the floor of the third ventricle (intraventricular component) varied from 100% to less than 10%. The smallest lesions encountered were attached to the hypothalamus above the mammillary body on one side and lay entirely within the third ventricle. Most HH extended into the interpeduncular cistern, and in four cases extended to the pontine level. The mammillary region of the hypothalamus was involved in all patients. Unilateral or predominantly unilateral attachment of the HH was seen in 6 and 12 cases respectively.
Gelastic seizures were present in all patients at some stage. In addition, 27 patients had complex partial seizures, 20 had tonic seizures (11 with drop attacks) and six had tonic-clonic seizures. Average seizure frequency was 10 per day. All but one patient had an abnormal interictal EEG; most patients had frequent generalised or lateralised spike-wave discharges which increased markedly in sleep. Nineteen patients had the electroclinical phenotype of symptomatic generalised epilepsy at the time of surgery, defined as the presence of (i) tonic seizures or epileptic spasms, and (ii) abundant spike-wave activity, with or without electrodecrements. Ictal EEG patterns were variable with many patients having either no ictal EEG change (typically in gelastic seizures), widespread low voltage generalised fast activity (typically in tonic seizures) or lateralised or asymmetric frontotemporal fast and/or spike-wave patterns (typically in partial seizures). Temporal or frontal pseudolocalisation’ was seen in many patients. Confirmation of HH seizure origin with specific modalities was not a requirement for surgery; rather, an electroclinical picture consistent with HH seizure origin, with or without subsequent seizure spread, and absence of a cortical lesion or incongruous electroclinical picture, were the bases for proceeding with surgery. A variable degree of HH ictal hyperperfusion was present in many patients (data analysis in progress) (figure 1); in those patients early in our series, HH hyperperfusion was a major impetus to for resection of the lesion. Intracranial EEG monitoring was not employed in any of our patients.
Thirteen patients had central precocious puberty, 21 had intellectual disability and 18 had behavioural problems including outbursts of rage and aggression.

Operative technique

A small frontoparietal craniotomy was performed and the right hemisphere was gently retracted. Using the operating microscope, a 1.5-2.0 cm opening was made in the anterior portion of the body of the corpus callosum just behind the genu (figure 2a), protecting the pericallosal vessels. The third ventricle was entered between the leaves of the septum pellucidum and the most anterior segment of the fornices, where the fornices part along the anterior border of the foramina of Monro to pass inferiorly as the columns (figure 2b). The trajectory of this anterior approach was generally towards the centre or posterior part of the HH in the third ventricle, corresponding to the tuberal and mammillary regions of the hypothalamus.
The HH was usually seen in the third ventricle as a smooth mass, often with a cleft between it and the ventricular wall. Large HH sometimes filled and distorted the third ventricle, obscuring the lateral walls. The HH was removed using an ultrasonic aspirator, using the walls of the third ventricle as the lateral limit of the resection and the translucent pial floor of the third ventricle as the inferior limit of the resection, at least in the mid and anterior portions. Posteriorly, the HH was invariably attached to the mammillary bodies and attempts were made to preserve these and their connections. The HH was usually rubbery and gliotic in consistency, making it relatively distinguishable from the surrounding soft normal cerebral tissue around it in most cases.

Results

Comparison of preoperative and postoperative MRI scans revealed that greater than 95% resection of the HH was achieved in 18/29 patients (figure 3), with 100% resection seemingly achieved in 12 patients. Between 75% and 95% resection was achieved in seven patients, four of whom had complete or near-complete disconnection of the residual hamartoma lying anteriorly, laterally or inferiorly. Less than 50% of the HH was resected in four patients. Incomplete resection was likely if the HH was attached anteriorly to the pituitary stalk and optic chiasm, if the HH extended inferiorly in the interpeduncular recess to the level of the pons, if the HH had a broad base on the lateral hypothalamus, or if the HH was so large that it filled the third ventricle and distorted the lateral and posterior margins of the third ventricle. Histopathology was consistent with a hamartoma in all cases.
A small, unilateral, anterior thalamic-capsular infarct occurred in two patients, both of whom had diathermy in the resection bed to arrest bleeding from perforating vessels. Subtle hemiparesis and an incomplete unilateral third nerve paresis resolved completely by the end of the first postoperative week in one of these two patients. Mild-moderate hemiparesis resolved completely over several months in the other patient. No postoperative visual defects occurred in any patient.
Early short-term memory disturbance, manifesting typically by repeated questioning, was noted in 14 patients. In most cases this resolved over several days or weeks, but in four cases, memory impairment persisted to a significant degree; one patient had previously undergone a left temporal lobectomy with resultant severe verbal memory impairment, and another patient had previously undergone resections and radiofrequency ablations of the hamartoma, complicated by middle cerebral artery territory stroke. From parent reports, there was the impression of improving memory function over several months in these affected patients.
Postoperative follow-up for the 29 patients ranged from one to nearly six years (mean 30 months, range 12-70 months). Seizure evolution and postoperative outcome are shown schematically in figure 4. Seizures occurred in the first postoperative week in 16 (59%) patients. These tended to lack the gelastic component of the preoperative seizures, occurred primarily from sleep, were predominantly tonic in nature, and in most patients decreased in frequency or ceased over subsequent months. In terms of current seizure status, 15 patients are seizure-free (patients 4, 5, 9, 11-18, 20, 21, 26, 29), seven have seizures at > 90% reduced frequency (patients 1, 2, 3, 8, 25, 28 and patient 22 at time of death), three have seizures at 55-80% reduced frequency (patients 7, 23, 27), one has seizures at 40% reduced frequency (patient 10), and three have seizures at the same frequency as prior to surgery (patients 6, 19, 24). Six of the seizure-free patients had seizures in the early postoperative period that subsequently remitted. Persistent seizures for many patients were auras or brief gelastic events in which the parents noted a sudden behavioural change, maybe with a smile, but with no laughter or impairment of consciousness. Only four patients had seizure recurrence after a year or more of seizure-freedom, these being infrequent auras or mild partial seizures at a much reduced frequency in three (patients 1, 2, 3), and seizures of the same severity and magnitude as preoperatively in one (patient 6). No patient had worsening of seizure frequency or severity following surgery. Most patients are on reduced medication and nine seizure-free patients are on no antiepileptic medication.
Seizure-freedom following surgery was not associated with (i) preoperative clinical findings such as intellectual disability (P = 1.0), behaviour disorder (P = 0.6) or precocious puberty (P = 0.6), (ii) HH characteristics such as maximum diameter (P = 0.5) or proportion of HH being intraventricular in location (P = 0.4), (iii) the proportion of HH resected, analysed for both > 95% (P = 0.6) and 100% (P = 0.6) resection, or (iv) patient age at seizure onset (P = 0.1) or age at surgery (P = 0.2). Similarly, seizure-freedom was not associated with the absence of symptomatic generalised epilepsy (P = 0.2), the age-at-onset of tonic seizures specifically (P = 0.8) or the duration of tonic seizures prior to surgery (P = 0.3).
Paralleling the resolution of seizures in these patients was a gradual reduction in slow spike-wave activity on interictal scalp EEG, most notable in the SGE patient group with abundant spike-waves. Twelve of the 19 patients with SGE were the subject of a detailed perioperative EEG analysis reported elsewhere [8]. The mean percentage of EEG containing spike-wave activity in these 12 patients decreased from 17% to 4% in awake EEGs and from 48% to 19% in asleep EEGs, with spike wave being abolished in the awake EEG in seven patients.
There were early postoperative improvements in behaviour, attentiveness and speech output observed in many children. Three autistic, non-verbal children began to speak soon after surgery. These early postoperative behavioural improvements were maintained, with loss of aggression and cessation of rage episodes in all patients with favourable seizure outcomes and in one of the patients with no seizure improvement. Mood disturbance occurred in three seizure-free patients and required antidepressant medication. Two non-seizure-free patients developed psychotic symptoms during periods of postoperative seizure fluctuation, both of whom had prior psychiatric disturbances.
One patient, the oldest in the series, died suddenly and unexpectedly in his sleep seven months after surgery. Seizures were reduced by greater than 95% and an initial postoperative exacerbation of a preoperative psychosis had resolved. No cause of death was found at postmortem, and therefore death was regarded as SUDEP.

Discussion

A variety of surgical approaches exist to treat intractable epilepsy associated with HH. Resection of HH from classical subfrontal and trans-sylvian routes, with or without cortical resection for improved access, have been associated with significant risks of stroke, cranial nerve palsy, endocrinopathy and residual lesion with postoperative seizures [1, 2]. These approaches provide limited access to the HH, especially the intrahypothalamic component, when the third ventricle is entered from beneath. Greater access to the intrahypothalamic component may be obtained by entering the third ventricle via a trans-lamina terminalis approach, various transcallosal approaches, a transcortical trans-foraminal (Monro) approach, or a supratentorial or infratentorial approach from behind [9, 10]. Although the trans-lamina terminalis approach is favoured by many neurosurgeons, there is retraction of both frontal lobes to gain access and potential for injury to the olfactory nerves, optic chiasm and the anterior cerebral arteries; furthermore, there is restricted access to the inferior and posterior margins of the HH in the third ventricle, rendering it more difficult to obtain resection or disconnection of the HH attached to the mammillary bodies.

Three transcallosal approaches to the third ventricle are described: (i) the midline interforniceal approach [12], (ii) the transforaminal approach through the lateral ventricle and the foramen of Monro [10, 12, 13], and (iii) the transchoroidal or subchoroidal approach traversing the choroidal fissure, the velum interpositum or tela choroidea between the fornix and the thalamus [12, 14, 15]. In our modified, transcallosal, interforniceal approach, the third ventricle is entered more anterior to that described by Apuzzo and Amar [11]. We favour this transcallosal approach to HH with an intraventricular attachment, even when there is a large interpeduncular component, because of the better view of the HH in the third ventricle, the ability to perform microsurgical resection from within the lesion, the surgical landmarks provided by the lateral and inferior margins of the third ventricle, and the separation from critical neurovascular structures beneath the HH.

Small subcortical stroke with transient hemiparesis occurred in two of our 29 patients, most probably as a result of diathermy of perforating vessels between the HH and the floor of the third ventricle. Permanent cranial nerve injury did not occur in our surgical series. Postoperative endocrine disturbance was minimal, with transient hypernatraemia, mild thyroxine deficiency and weight gain being the only endocrine consequences of surgery in some patients (see Freeman et al. in this issue). This minimal endocrinological and neurological morbidity contrasts with that reported when subfrontal or trans-sylvian operative approaches to HH are employed [1, 2], most likely due to the limited traction of the frontal lobe, the separation from major neurovascular structures and the pituitary stalk, the resection being performed with gentle aspiration from entirely within the lesion, and the avoidance of diathermy where possible.

Short-term memory disturbance due to septal, forniceal or mammillary body injury is the major disincentive to third ventricular surgery. Transient short-term memory impairment was reported by Apuzzo and Amar in 30% of patients in their series who underwent classical interforniceal resection of third ventricular lesions; in 75% of patients the memory disturbance resolved in the week after surgery and by 3 months, all had returned to preoperative status [11]. In our transcallosal approach, the fornices are retracted laterally to some degree, more on the side of entry (right side) than the other side, but probably less overall than with a more posterior approach. We were usually able to preserve the mammillary bodies and their connections (as seen on postoperative T2-weighted MRI scans in three planes), even if the HH distorted these structures. This is because the fornices are displaced around the edge of the lesion and the mammillary bodies are usually displaced to the inferiolateral edge of the HH. When the HH was attached unilaterally, surgical trauma was presumably limited to memory structures on the one side and may not have been clinically apparent. Short-term memory disturbance may have been undetected in non-verbal or intellectually disabled patients in our series, or alternatively, may have recovered promptly or been adapted in young children, due to cerebral plasticity.
Our postoperative seizure results, with 76% of patients being seizure-free or having a 90% or greater reduction in seizures (mean follow-up 2.5 years), are comparable to those reported with anterior temporal lobectomy in patients with mesial temporal lobe epilepsy [16]. The results are somewhat surprising, considering the long-standing and sometimes progressive nature of the patients’ epilepsies and associated neurobehavioural problems. Interestingly, no patient or lesion characteristics were predictive of seizure freedom following surgery. This suggests that young patients of any age, with HH of any size and intraventricular location, who have intractable epilepsy of any duration and severity may become free of seizures with transcallosal surgery.
Most striking was the neurological improvement in patients with SGE, in whom the majority showed marked improvement or remission of gelastic, partial and tonic seizures, reduction in spike-wave activity or EEG normalisation, acceleration in development, and improvement in behaviour. The evolution of SGE from gelastic seizures is well described [17], and has been proposed as a model for both progressive epilepsy [18] and pervasive developmental disorders of childhood [19]. In our series, the decrease in interictal spike-wave activity was not immediate following resection of the HH, and the resolution of tonic seizures took up to six months in some cases. Delayed seizure remission is described following other surgical approaches to HH [20], and spike wave discharges on the scalp EEG of patients with HH have been shown to not necessarily correlate with discharges originating in the HH [8, 21]. We have proposed that the slow spike-wave activity and tonic seizures that develop over time in some patients with HH, are extralesional phenomena related to secondary epileptogenesis [8]. The electroclinical remission in our SGE patients is similar to that described by Morrell for patients he classified as having the intermediate form of secondary epileptogenesis [22]. One potential anatomical conduit for the propagation of seizures from the HH and the remote generation of generalised spike-wave is the mammillothalamic tract, whose fibers ascend to the anterior thalamic nucleus and then project to the anterior cingulate gyrus [23].
Our surgical series of HH patients is the largest reported, with all patients operated upon with a standardised perioperative protocol and a single operative approach over a relatively short period at a single institution. We have shown that it is feasible and safe to resect HH via a transcallosal interforniceal approach to the third ventricle with microsurgical removal and stereotactic guidance. We believe that the remarkable seizure improvement is related to removal or disconnection of the intraventricular, intrahypothalamic component of the HH. The low frequency and benign nature of complications are most likely due to being able to adequately visualise and operate entirely within the HH, limit the resection to the margins of the third ventricular walls and pial floor, minimise traction on the frontal lobes and fornices, and avoid diathermy during resection. Based on published surgical series, we believe that the transcallosal approach is superior to conventional trans-sylvian and subfrontal approaches to HH, with respect to completeness of resection, seizure outcome and complications. The challenges ahead for clinicians managing these complex cases are (i) comparing long-term efficacy and morbidity of HH resections with radiosurgical approaches, (ii) establishing criteria for patient selection and timing of surgery, (iii) refining surgical techniques to further minimise risks to memory and endocrine function, and (iv) understanding why apparently complete HH resection does not render some patients seizure-free.

Acknowledgements

The authors thank Ms Catherine Bailey (EEG Technologist), Ms Elizabeth Grimmer (Nurse Specialist), Mr Michael Kean (MR Technologist), Ms Jacqueline Wrennall (Clinical Neuropsychologist), Dr Margaret Zacharin and her endocrinologist colleagues, the neuroscience nursing staff, and the anaesthetic and operating suite staff for their important contributions to the management of this patient group. We are also grateful to the patients, their families and the Hypothalamic Hamartoma Uncontrolled Gelastic Seizures (HHUGS; http://www.geocities.com/hhugs2001) support group, who had faith and trust in our epilepsy program and assisted us with collection of accurate clinical information and follow-up; their efforts have definitely advanced knowledge and led to treatment advances in this rare and frequently devastating condition. n

References

1. Rosenfeld JV, Harvey AS, Wrennall J et al. Transcallosal resection of hypothalamic hamartomas, with control of seizures, in children with gelastic epilepsy. Neurosurgery 2001; 48: 108-18.

2. Palmini A, Chandler C, Andermann F et al. Resection of the lesion in patients with hypothalamic hamartomas and catastrophic epilepsy. Neurology 2002; 58: 1338-47.

3. Kuzniecky R, Guthrie B, Mountz J et al. Intrinsic epileptogenesis of hypothalamic hamartomas in gelastic epilepsy. Ann Neurol 1997; 42: 60-7.

4. Regis J, Bartolomei F, de Toffol B et al. Gamma knife surgery for epilepsy related to hypothalamic hamartomas. Neurosurgery 2000; 47: 1343-51.

5. Dunoyer C, Ragheb J, Resnick T et al. The use of stereotactic radiosurgery to treat intractable childhood partial epilepsy. Epilepsia 2002; 43: 292-300.

6. Arita K, Ikawa F, Kurisu K et al. The relationship between magnetic resonance imaging findings and clinical manifestations of hypothalamic hamartoma. J Neurosurg 1999;91:212-20.

7. Harvey AS, Freeman JL, Rosenfeld JV et al. Postoperative course and seizure outcome following transcallosal resection of hypothalamic hamartoma in seventeen children with intractable gelastic epilepsy. Epilepsia 2001; 42 (suppl 7): S216.

8. Freeman JL, Harvey AS, Rosenfeld JV et al. Generalized epilepsy in hypothalamic hamartoma: Evolution and postoperative resolution. Neurology 2003; 60: 762-7.

9. Apuzzo MLJ. Surgery of the third ventricle. 2nd ed. Baltimore: Williams and Wilkins, 1998.

10. Yasargil MG. Intraventricular tumours. Microneurosurgery. Volume IVB: Microsurgery of CNS Tumors. New York: Thieme, 1996: 313-38.

11. Apuzzo MLJ, Amar AP. Transcallosal interforniceal approach. In: Apuzzo MLJ. Surgery of the third ventricle. Baltimore: Williams and Wilkins, 1998: 421-52.

12. Shucart W. Removal of tumors in the third ventricle. In: Wilson CB. Neurosurgical Procedures: Personal Approaches to Classic Operations. Baltimore: Williams and Wilkins, 1992: 52-62.

13. Ehni G, Ehni BL. Considerations on transforaminal entry. In: Apuzzo MLJ. Surgery of the third ventricle. Baltimore: Williams and Wilkins, 1998: 391-419.

14. Wen HT, Rhoton AL, Jr., de Oliveira E. Transchoroidal approach to the third ventricle: an anatomic study of the choroidal fissure and its clinical application. Neurosurgery 1998; 42: 1205-17.

15. Patterson RH, Lavyne MH. Two approaches to the anterior third ventricle. I. Via the velum interpositum. II. Via the lamina terminals and sphenoid sinus. In: Samii M. Surgery in and around the brain stem and the third ventricle. New York: Springer-Verlag, 1986: 329-33.

16. McIntosh AM, Wilson SJ, Berkovic SF. Seizure outcome after temporal lobectomy: current research practice and findings. Epilepsia 2001; 42: 1288-307.

17. Berkovic SF, Andermann F, Melanson D et al. Hypothalamic hamartomas and ictal laughter: evolution of a characteristic epileptic syndrome and diagnostic value of magnetic resonance imaging. Ann Neurol 1988; 23: 429-39.

18. Berkovic SF, Kuzniecky RI, Andermann F. Human epileptogenesis and hypothalamic hamartomas: new lessons from an experiment of nature. Epilepsia 1997; 38: 1-3.

19. Deonna T, Ziegler AL. Hypothalamic hamartoma, precocious puberty and gelastic seizures: a special model of “epileptic” developmental disorder. Epileptic Disord 2000; 2: 33-7.

20. Nishio S, Morioka T, Fukui M et al. Surgical treatment of intractable seizures due to hypothalamic hamartoma. Epilepsia 1994; 35: 514-9.

21. Kahane P, Munari C, Minotti L et al. The role of the hypothalamic hamartoma in the genesis of gelastic and dacrystic seizures. In: Tuxhorn I, Hothausen H, Boenigk H. Paediatric epilepsy syndromes and their surgical treatment. London: John Libbey, 1997: 447-61.

22. Morrell F. Varieties of human secondary epileptogenesis. J Clin Neurophysiol 1989; 6: 227-75.

23. Truex RC, Carpenter MB. The diencephalon and corpus striatum. In: Truex RC, Carpenter MB. Human Neuroanatomy. Baltimore: Williams and Wilkins, 1964: 379-442.