Wilder Penfield's development of surgical methods for treating focal cerebral seizures, beginning with his early work in Montreal in 1928, is reviewed. The reliance on seizure pattern and focal brain pathology was enormously enhanced by the advent of EEG and direct electrical recording from the cortex during surgery. The recognition in the early 1950's of mesial temporal structures in the pathogenesis of temporal lobe epilepsy with automatism lead to more rational and successful surgical treatment. Positron emission tomography and especially magnetic resonance imaging have recently added to the essential diagnostic information of focal epilepsy provided by EEG.

Preoperative EEG investigations of patients with temporal lobe seizures include extracranial interictal and ictal recordings during wakefulness and sleep, including long-term EEG and video-monitoring. Interictal epileptiform discharges when evaluated conservatively and in conjunction with other EEG and non-EEG localizing information, provide valuable guidance for the identification of the area to be resected, as do ictal recordings. When extracranial EEG features in conjunction with non-EEG data provide conflicting localizing information, intracranial recordings with stereotaxically implanted depth and epidural electrodes are used. Intracranial recordings must be designed to avoid biasing the exploration strategy in favor of one's preferred localizing hypothesis. Patients with evidence for bitemporal epileptogenic dysfunction in extracranial EEG recordings are suitable candidates for intracranial recordings. The majority of the patients explored in this manner show that all or more than 80% of their seizures arise from one temporal lobe. Excision of that lobe yields satisfactory results in a fair proportion of these patients. The number of satisfactory outcomes is, however, still somewhat less than in patients with unilateral temporal foci in extracranial EEG recordings.

The first section of this article deals with specific anatomic and pathophysiologic factors which contribute to a poor EEG localization of the interictal epileptic abnormality and to the unreliable seizure onset localization commonly reported in patients with frontal lobe epilepsy. The localizing effectiveness of long term EEG monitoring was reviewed in four different groups of frontal lobe epileptic patients who underwent preoperative EEG investigation with extracranial and intracranial electrodes. The results of this study reveal a continuum distribution of interictal epileptic disturbances, ranging from focal abnormalities to lobar or multi-lobar epileptogenesis. A frontal lobe localization of the seizure generator based on ictal recordings obtained with extracranial electrodes is rather poor and much more reliable results can be obtained by depth-electroencephalography.

The concept of epileptogenic zone is defined as a large area of cortical neurons arranged in concentric circles of variable degrees of epileptogenicity. This is particularly so in frontal lobe epilepsy since the interictal and ictal epileptic abnormalities are poorly localized, often absent and at times misleading in terms of localization. In temporal lobe epilepsy, the epileptogenic zones may be more restricted.

An introduction to the clinical study of postrolandic epileptic seizures. We reviewed, in 145 epilectics studied with SEEG, 800 clinical and electrographic seizures originating from the post-rolandic areas (590 spontaneous Sz and 260 induced by stimulation). The intra-cranial electrodes were implanted using a technique described by Talairach et al. (1974). Seizure onsets were recorded in the centro-parietal region (64 patients) in the partietal49 and in the occipital region.15 Additionally in 15 patients, mixed Sz onset were recorded. One hundred and eight (108) patients underwent surgical removal of their epileptic focus. (69 on the right, 39 on left.) 65% were cured (Sz free or occas Sz, f/up 3 years). The main ictal cal features are discussed. Emphasis is placed on the role of the operculo-insular cortex in the functional organization in man, based on Sz arising from the supra-temporal (fronto-pariental) cortex. The study of the pattern of onset and spread of seizures originating in the post-rolandic areas and of their clinical correlates allow a topographic differential diagnosis.

Electroencephalographic and neuro-imaging data were correlated with the lobe of cortical resective surgery among 48 patients age 16 years or less with medically intractable partial seizures. Decisions about surgery were determined by clinical judgement and therefore our results reflect the weight of these factors in clinical practise. Neuroimaging disclosed a lesion over the operated lobe in 32 patients (67%). Multiple EEGs demonstrated a most active spike focus over the operated lobe in 32 patients (67%). Although only 34 patients had a recorded seizure, this clearly arose over the operated lobe in 23 (68% of the 34 patients; 48% of series). Persistent focal delta activity appeared in 20 of the 48 patients (42%). No instance was encountered in which false lateralization of regional or focal interictal EEG spikes or delta when normal apiculate phenomena were excluded. Thus, the scalp EEG remains a valuable guide to epileptogenesis, even among young patients, and should continue to play a major role in determining the region of cortical resection. Effectiveness of cortical resection for seizure control was clearly better among patients with a single major lesion and among those with normal intelligence.

Epileptic seizures of focal origin often occur unpredictably as do interictal spikes. It is often assumed that spikes increase prior to seizures of focal origin and that antiepileptic medication affects spikes and seizures in a parallel fashion. We review evidence that this assumption is invalid and that there is a clear dissociation between spikes and seizures: increases in spiking before seizures have not been clearly documented; decreases in antiepileptic medication do not result directly in increased spiking; seizures are often followed by long-lasting increases in spiking; finally, seizures are no more likely when spikes are frequent than when spikes are rare. It therefore appears that spikes and seizures are two quite distinct phenomena, both originating in the epileptic focus but varying over time differently from what is most often believed.

In a consecutive series of 40 patients selected by EEG studies for surgical treatment of temporal lobe seizures, magnetic resonance imaging showed structural lesions in 25% and signal abnormalities, usually in the mesial temporal region, in another 35%. Pathological changes included structural lesions in over 30% of patients and sclerosis of the amygdala in almost half of the series. These findings further substantiate the implication of the amygdala in the pathogenesis of temporal lobe seizures with automatism and amnesia.

Computed tomography and magnetic resonance imaging enable the identification of neuronal migration disorders during life. Several specific syndromes have been identified and early diagnosis of previously unrecognized entities is now possible. We report 51 patients with imaging. Thirty-two had a single widespread cortical dysplastic lesion. Twenty-eight had focal corticectomies. From a pathological standpoint, these encompassed focal cortical dysplasia (14 cases) and forme fruste of tuberous sclerosis (10 cases). These two groups of patients were indistinguishable from the clinical and radiological standpoint. In only two was the MRI examination normal. In addition, there were 10 with bilateral perisylvian dysplasia, four with diffuse cortical dysplasia or the "double cortex" syndrome, three with hemimegalencephaly, one with megalencephaly, and one with nodular neuronal heterotopia. The electroclinical and imaging findings led to the development of specific surgical strategies for the alleviation of the intractable seizures in each of these radiologically-defined syndromes.

Many biologically active tracers are available for positron emission tomography (PET) investigations, but most studies of epilepsy have utilized l8F-fluorodeoxyglucose (FDG) to measure local cerebral metabolic rate for glucose. Over 70% of patients with medically refractory partial seizures demonstrate an interictal zone of hypometabolism corresponding to the epileptogenic region. This metabolic defect commonly involves the temporal lobe in patients with complex partial seizures of mesial temporal origin, and is encountered less consistently with seizures of extratemporal neocortical origin. Although false localization is less likely with FDG-PET than with EEG, the hypometabolic zone merely reflects a focal functional deficit and its epileptogenicity must still be demonstrated electrophysiologically. When hemispherectomy or large multilobar resections are planned in small children, FDG-PET also provides useful supporting evidence that the contralateral hemisphere is functioning normally. It is difficult to obtain FDG-PET scans and to interpret results during spontaneous partial seizures. Ictal scans can be more easily obtained with single photon emission computed tomography (SPECT), which may provide information for planning surgical resections.

A retrospective study of 383 patients who had undergone either left or right anterior temporal lobectomy (varying with respect to the amount of hippocampus excised) revealed that there were no significant differences in surgical outcome between those patients having a large and those patients having a small hippocampal removal. Although left temporal-lobe excisions were significantly smaller than right temporal-lobe excisions, both medially and laterally, patients with left temporal-lobe removals had an overall better surgical outcome. Patients in the successful surgicaloutcome group did better than the unsuccessful surgical-outcome group on a variety of cognitive measures, including tests of intelligence, delayed verbal memory, and verbal fluency. The results suggest that, for many patients with medically intractable epilepsy, an anterior temporal lobectomy including the amygdala may suffice to reduce their seizure frequency.

We have compared the relative localizing value of common tests in the presurgical evaluation of epilepsy in 73 patients with depth electrode (SEEG) confirmed epileptogenic foci and excellent surgical outcome. We found the tests of abnormality (CT, amytal, neuropsychology) to be frequently non-informative and often discordant with SEEG. The EEG was concordant with the SEEG in 38% of cases and provided lateralization in 78%. EEG was less localizing than ictal behavior in frontals but not in temporals. The need for SEEG is rather evident in these results but could become attenuated by the development of the localizing power of scalp EEG.

One hundred consecutive patients between 1961 and 1980, with temporal lobectomy including excision of half or more of the hippocampal formation, have been analyzed for degree of reduction of seizures. Thirty-nine percent have become and remained seizure free for follow-up period of 2 to 24 years (median 12 years). Another 24 patients achieved a marked reduction in seizure tendency as defined in the report and 25 showed a lesser degree of improvement. In fifteen patients no change in seizure frequency was recorded. Thus in this series of 100 patients, a complete or marked reduction of the seizure tendency was achieved in 63%.

The evidence for the role of the amygdala in temporal lobe seizures is supported by this follow-up (2-20 years) of 100 patients who were treated surgically by excision of the antero-lateral temporal cortex, most or all of the amygdala and minimal resection of the hippocampus. The findings showed 53 patients seizure-free or with rare or occasional seizures, 10 patients with marked seizure reduction and 37 with moderate or less reduction of seizures. The results are the same as in another 100 patients where, in addition, half or more of the hippocampus was resected.

Follow-up data of 2 to 49 years' duration are presented on 257 patients with cortical excisions for non-tumoral frontal lobe epilepsy. Twenty-six percent have become and remained seizure free. Another 30% have had a marked reduction of seizure tendency. Thus a total of 56% have had a long term complete or marked reduction of seizure tendency after the cortical resection. The cortical resections were classified into 5 regional categories. Forty-seven percent of patients with anterior frontal resections were in the seizure free category, as were 35% of patients with excisions limited to the convexity cortex and 32% of patients with maximal or near maximal frontal lobectomy. Only 18% of patients with frontal parasaggital excisions were in the seizure free category, and 10% of the patients who had excision of frontal cortex plus limited excision of adjacent central and/or temporal cortex.

Patients with epileptogenic lesions that are more or less restricted to the central, parietal or occipital regions represent 14% of our total epilepsy surgery series and consisted of 322 patients at the end of 1980. Tumoral or major vascular lesions were present in 117 patients (37%), but this report is limited to the remaining 203 patients (63%) with non-tumoral epileptogenic lesions. The presumed etiology was birth trauma or anoxia in 25%, postnatal brain trauma in 25%, and unknown in 24%, miscellaneous lesions were present in 16%, post-inflammatory gliosis or multiple factors were present in the remaining 10%. Follow-up data of 2-51 years' duration (median 18 years) are available for 186 of these 203 patients. Sixty-three patients (34%) remain seizure-free. Forty-two patients (23%) have a marked reduction of seizures. The remaining 81 patients (44%) have a lesser reduction. The variation in clinical seizure patterns is considerably greater than in temporal lobe epilepsy. Focal sensorimotor phenomena are common at the onset, whether the seizures begin in the central region or further posterior. The success following cortical resection is equally gratifying and the failures just as distressing as in surgery for epilepsy involving other brain regions.

Selective amygdalohippocampectomy (AHE) offers a real chance of cure only in patients with welldefined, precisely localized “epileptogenic area”, i.e. seizure focus. Therefore, a priori only a small proportion of all patients with epilepsy can meet the criteria for selective surgical interventions. From the evidence in patients meeting the criteria for AHE, we conclude that this technique is to be preferred to the “standard” anterior temporal lobectomy and represents a more selective but still effective surgical treatment of epilepsy.

We have briefly reviewed the experimental and clinical evidence for the importance of the amygdala and hippocampal formation in temporal lobe epilepsy. More specifically, we have analyzed our own experience in patients with temporal lobe epilepsy investigated with intracerebral stereotaxic electrodes and operated by various modalities of resection. Our results, in agreement with previous experimental and clinical work, provide further evidence for an overwhelming predominance of limbic participation in temporal lobe epilepsy. As a result, more and more selective procedures are being carried out involving the mesial structures. However, this shift has been slow and progressive because of the proven value of cortico-amygdalo-hippocampectomy which provides excellent results on seizure tendency with low morbidity.

In this review we summarize a number of technical and methodological approaches that have been used in our laboratory to study human brain slices maintained in vitro. The findings obtained in the course of these studies appear to be relevant in establishing he mechanisms that underlie physiological phenomena of the human brain such as synaptic plasticity or responses to neuroactive drugs. Moreover, these data are important for understanding certain fundamental mechanisms of epilepsy. In this respect, however, we aution that the mechanisms that apply to different forms of clinical epilepsy might be ifficult to find given the variability present in the pathogenesis of human epilepsy.