Diverse perspectives on developments in epilepsy surgery

Abstract

The objective of this article is to review the dramatic changes that have occurred in the field of epilepsy surgery since the founding of Epilepsy Action in 1950. We have chosen to consider these advances from the biomedical perspective (the physician and basic scientist), and the behavioral perspective (the psychologist and the patient). Both of these viewpoints are equally important in understanding the evolution of epilepsy surgery over the past 60 years, but may not always be well synchronized.

1. Introduction

To put this review in historical perspective, the modern era of surgical treatment for epilepsy was some 70 years old by the time of the founding of Epilepsy Action, with initial seminal work having been already performed in the United Kingdom (1). In 1879, William Macewen (2) of Glasgow reported the first resection of an “invisible” lesion to treat epilepsy based on localization derived from the clinical seizure observations of John Hughlings Jackson (3), followed by a report on a series of cases in 1881 (4). Five years later, Victor Horsley published his seminal paper in Brain (5), combining clinical-pathological correlations by Jackson (3) and electrical stimulation studies carried out on monkeys by Ferrier (6) to confirm that localization of an epileptogenic region could be determined by the signs and symptoms characterizing the onset of the habitual clinical seizures. Using this approach, localization was initially suspected by clinical observation of the seizure onset, and confirmed intraoperatively by identification of a structural abnormality, often accompanied by direct brain stimulation.

Subsequent work in the early 20^th century, particularly from Germany (7,8) and Canada (9) further confirmed the efficacy of surgery as a treatment for epilepsy, but these procedures were limited to a small number of patients with focal seizures for whom a structural lesion could be demonstrated. Use of lesion-directed surgery was greatly enhanced by development of pneumoencephalography in 1919 (10) and cerebral angiography in 1934 (11). Also at this time, a revolutionary approach to monitoring both normal and abnormal brain function was being developed (12,13). The electroencephalogram (EEG) allowed patterns associated with epilepsy to be recognized (14), and the ability of EEG interictal spikes to localize epileptogenic tissue, including identification of mesial temporal structures as the site of onset of psychomotor seizures (15). Yet despite EEG later playing a critical role in epilepsy surgery, by 1950 it played no role in the standard lesion-based approach to epilepsy surgery. There were very few centers capable of performing surgery as a treatment for epilepsy, and relatively few patients underwent surgery at those centers; for many the risks were high and the benefits relatively low compared to today’s practice.

2. The 1950s

2.1. The biomedical perspective

The decade following the founding of Epilepsy Action was a period of seminal advances in understanding the patho-anatomical basis of focal epilepsy and applying this information to surgical treatment. Despite active research on EEG in epilepsy, epilepsy surgery continued to be lesion-directed based predominantly on evidence from angiography and pneumoencephalography of structural brain abnormalities. This changed with the landmark publication of Bailey and Gibbs (Figs. 1 and 2) (16). They were the first to report a series of patients who underwent resection localized exclusively by interictal EEG. These patients all had temporal lobe resections for psychomotor seizures. A year earlier, Penfield (Fig. 3) and Flanigan (17) had published a larger series of patients who underwent temporal lobe resection for epilepsy based on lesion identification, and the following year Jasper (Fig. 3) et al. (18) published the complete EEG findings on these patients. These three papers confirmed that psychomotor seizures originated in the temporal lobe, and that temporal lobe resections were beneficial in alleviating disabling seizures. As a result, there was a virtual explosion of surgical activity worldwide, almost exclusively focused on temporal lobe resections (1). Interestingly, these initial resections were corticectomies for fear of possible deleterious consequences from hippocampal removal, and only approximately one-third of patients achieved seizure freedom. Surgical results improved greatly once surgeons realized that mesial temporal structures could be safely resected (19–21). It might be argued, therefore, that initial results largely represented a placebo effect and that temporal lobe surgery might never have been pursued had it not been for this phenomenon.

Fig. 1.

Percival Bailey (1892–1973).

Fig. 2.

Erma L. (1904–1988) and Frederick A. Gibbs (1903–1992).

Fig. 3.

Wilder Penfield (1891–1976) and Herbert H. Jasper (1906–1999).

The 1950s were also a time of rapid advances in diagnostic EEG, particularly as applied to localization for surgery. Various noninvasive approaches using nasopharyngeal, tympanic, and sphenoidal electrodes were already in use by 1950, but direct intraoperative EEG recording, electrocorticography (ECoG), was felt to be most important for localization of the epileptogenic region. Although Bickford and Kairns performed the first chronic depth electrode recordings by inserting multistranded insulated wires into a cerebral bullet track at Oxford in 1944 (22), the modern approach to depth electrode recording was pioneered by Jean Talairach (Fig. 4) and Jean Bancaud (Fig. 5) in Paris in the late 1950s (23). The value of stereo EEG for three-dimensional localization of epileptogenic tissue was rapidly appreciated; however, French law at that time permitted electrodes to remain in place for only several hours, so analysis was based on interictal spike activity, as well as ictal discharges induced by electrical stimulation and convulsant drugs.

Fig. 4.

Jean Talairach (1911–2007).

Fig. 5.

Jean Bancaud (1921–1994).

An important contribution to research on the pathological basis of temporal lobe epilepsy occurred in 1953. Murray Falconer (Fig. 6) introduced a standardized en bloc anterior temporal lobectomy procedure in London, which provided large intact tissue specimens for pathological examination (24). As a result, clinical pathological correlations revealed that a high percentage of patients with temporal lobe epilepsy had hippocampal sclerosis. This development was essential for later detailed pathophysiological research on epileptogenic hippocampus, and elucidation of the prognostic importance of hippocampal sclerosis.

Fig. 6.

Murray A. Falconer (1910–1977).

2.2 The behavioral perspective

In the 1950s, the practice of epilepsy surgery was not only being advanced by significant electrophysiological and patho-anatomical discoveries, but also by seminal contributions from those committed to understanding brain-behavior relationships, such as Brenda Milner (Fig. 7) and Juhn Wada (Fig. 8). A striking illustration is the now famous case of H.M., who in 1954 underwent a bilateral temporal lobectomy for medically intractable seizures that produced a dense postoperative amnesia (25). One year later, Milner and Penfield (26) reported impaired recent memory function after unilateral temporal excision in the presence of bilateral mesial temporal pathology, again involving the hippocampi. While H.M. was not the first case to undergo a bilateral procedure (27), Milner’s systematic and scholarly evaluations of the surgical patients of Penfield and Scoville are widely recognized for identifying the critical role of the mesial temporal region in recent memory function (28). Her work also demonstrated the important dissociation between episodic and procedural memory, and gave rise to the modern day material-specific model of episodic memory that ascribed learning and retention of verbal material to the left temporal lobe (29,30).

Fig. 7.

Brenda Milner (1918-).

Fig. 8.

Juhn Wada (1924-).

Wilder Penfield at the Montreal Neurological Institute (MNI) first promoted the use of a multidisciplinary approach, including routine neuropsychological examination, for the surgical evaluation of epilepsy patients (31). Pre-surgery, neuropsychology played an increasing diagnostic role to (i) aid determination of the location of cerebral abnormality, (ii) identify the risk for iatrogenic cognitive impairments, particularly in language and memory (including intra-operative mapping), and (iii) assess the presence of cortical reorganization of cognitive functions, notably right hemisphere dominance for speech in left-handed patients (32). Relevant to this third issue, Juhn Wada pioneered the intracarotid amobarbital procedure (also known as the ‘Wada test’) that became the method for determining hemispheric specialization for speech and language functions pre-surgery (33,34). Introduction of the technique to the MNI in 1955 demonstrated that early onset of seizures was associated with greater prevalence of atypical language organisation, and it was subsequently used to assess lateralized contributions of mesiotemporal structures to memory function (35,36).

Alongside these advances, Ward Halstead developed a battery of neuropsychological measures that when applied to epilepsy surgery, showed that surgery did not produce generalized cognitive decline, and that post-operative improvement was observable on some tasks (37). In London, Meyer and Yates also reported minimal impact of temporal lobectomy on IQ scores, but selectively identified the task of paired-associate learning as having increased sensitivity to hippocampal dysfunction (38). Their work supported the material-specific model of memory, and distinguished recovery of language disturbance from auditory-verbal learning impairments post-surgery. This was an important precursor to contemporary theories of the role of the hippocampus in ‘binding’ associations (30), and later identification of arbitrary relational learning as a neurocognitive marker of mesial temporal epileptogenesis (39).

From the patient’s perspective, epilepsy surgery is an elective procedure meaning that understanding its effects on cognition and behavior are central to making an informed decision (40). In the 1950s, outcome studies initially measured the efficacy of surgery in terms of postoperative seizure frequency (16,17,19), however soon after psychological functioning was considered mainly due to the referral of patients from psychiatric sources and the then current opinion associating temporal lobe epilepsy with psychological problems (41,42). The results generally indicated that psychiatric improvement followed seizure relief or improvement, despite a time lag of 12–24 months between the two (43–46). During this period, a transient worsening in patient mood or behavior could occur, that Ferguson and Rayport (47) attributed to a process of psychological adjustment. This process stemmed from the patient’s pre-operative psychosocial situation, where ‘illness’ behaviors and enmeshed family dynamics often engendered a need for the patient to learn to become ‘well’ after surgery (47–49). In recognition of this, Taylor (Fig. 9) spearheaded a series of studies in London that promoted a comprehensive review of patient change pre- to post-surgery across a range of domains, including psychological, psychiatric, family, socio-economic, and sexual functioning (49–51).

Fig. 9.

David C. Taylor (1933-).

3. The latter part of the 20^th century

3.1 The biomedical perspective

Bouchet and Cazauvieilh first described hippocampal sclerosis in the brains of patients with epilepsy in the early nineteenth century (52), but by the 1960s it was still being debated whether this was a result, or a cause, of epilepsy. Falconer’s en bloc temporal resection made it possible to demonstrate that the presence of hippocampal sclerosis predicted an excellent postoperative outcome (46,53), indicating that this pathological abnormality was epileptogenic. Not only was this conclusion ultimately important for selection of surgical candidates and determining prognosis, it identified brain tissue that would become valuable for basic research into the fundamental mechanisms of epilepsy.

In 1963, Paul Crandall (Fig. 10) and his colleagues at UCLA were the first to report the use of chronic stereotactically implanted depth electrodes to record EEG changes occurring at the onset of spontaneous seizures (54). Crandall then combined these chronic electrophysiological recordings with Falconer’s standardized en bloc anterior temporal resection, which permitted detailed electroclinical-pathological correlations. This approach became the basis for innovative multidisciplinary collaborative investigations into fundamental mechanisms of human epilepsy that are now pursued at epilepsy surgery centers worldwide (55–57).

Fig. 10.

Paul Crandall (1923-).

At the same time, Ross Adey at UCLA was devising EEG telemetry technology for the National Aeronautics and Space Administration to record from chimpanzees orbiting the earth (58). Crandall and colleagues enlisted the collaboration of Adey to develop the first epilepsy EEG telemetry unit (59), which permitted artifact-free continuous EEG recordings from depth electrodes over long periods of time. During the 1970s, this advance was accompanied by simultaneous cinemagraphic, and then video monitoring, of ictal behavior, in order to obtain second-by-second electroclinical correlations of ictal onset and propagation. Over the next decade, EEG artifact rejection became sufficiently sophisticated to permit long-term video and EEG monitoring using scalp and sphenoidal electrodes (60,61).

The advent of X-ray computed tomography (CT) in the 1970s made it possible to visualize the entire brain in three dimensions, to identify localized structural lesions in many patients with focal epilepsy whose seizures had been diagnosed as cryptogenic (62). This marked the beginning of modern neuroimaging, which has gradually moved us back to a lesion-directed approach to epilepsy surgery. Neuroimaging has not replaced EEG, however, which remains essential for demonstrating that an identified lesion is epileptogenic. With the later advent of magnetic resonance imaging (MRI), which has higher resolution than CT, MRI rapidly replaced CT for localizing structural lesions in epilepsy.

The next neuroimaging technique to play a critical role in presurgical evaluation was functional, not structural. In the early 1980s, positron emission tomography (PET) with 18F-fluorodoxyglucose (FDG) revealed unilateral temporal hypometabolism in patients with hippocampal sclerosis and other mesial temporal lesions not visible on CT or MRI (60,63). Combining scalp and sphenoidal video-EEG monitoring with FDG-PET and neuropsychological testing obviated the need for invasive electrophysiological investigations in many patients (60). Subsequently, FDG-PET became useful in identifying large unilateral areas of cortical dysplasia in infants and small children with secondary generalized epilepsy (64) who were candidates for hemispherectomy and multilobar resections.

By 1990, improvement in MRI resolution permitted convincing demonstration of hippocampal sclerosis (65), and is now able to demonstrate cortical dysplasia (66) and a variety of other subtle lesions in patients with epilepsy who were previously diagnosed as cryptogenic. Consequently, a significant percentage of patients who undergo surgery have hippocampal sclerosis and cortical dysplasia that is identified noninvasively, many of whom might not have been considered surgical candidates a decade or so earlier. Additional PET tracers, such as flumazenil (a benzodiazepine receptor ligand), and α-methyl-tryptophane (AMT), have also become useful in identifying potentially epileptogenic abnormalities in patients with normal MRI (67).

During this time, single photon emission computed tomography (SPECT) was applied to reveal decreased cerebral perfusion in epileptogenic temporal lobes but with a somewhat lower spatial resolution than PET. SPECT became particularly useful, however, when interictal hypoperfused regions became hyperperfused ictally (68,69). Consequently, by the end of the 20^th century, a variety of functional and structural neuroimaging approaches had become indispensable in presurgical evaluation. These technologies also made it possible for neuroscientists interested in basic mechanisms of epilepsy to noninvasively identify the neuroanatomical substrates of a variety of ictal behaviors and to begin to develop concepts of neuronal networks underlying different epilepsy syndromes.

The last decade of the 20^th century was a time of major conceptual change in the practice of epilepsy surgery. In the 1980s, most epilepsy surgery teams pursued only one particular surgical approach, depending on where they were trained (1). The first Palm Desert Conference on Surgical Treatment of the Epilepsies was held in California in 1986, where virtually all of the epilepsy surgery programs in the world at the time convened to compare strategies and outcomes (70). Prior to this there had been only one textbook on epilepsy surgery (71), and very few centers published their work, so this was a novel opportunity for sharing experiences and developing standardized approaches. Over the next five years most epilepsy surgery programs tried approaches used at other centers, and adopted different approaches for different types of epilepsy. A follow-up Palm Desert conference in 1992 confirmed that most centers were now performing depth electrode and subdural grid recordings selectively, operating on many patients based on noninvasive data, and carrying out a variety of surgical procedures for different epilepsy conditions (72). During the intervening six years there was also a tremendous increase in publications on epilepsy surgery, a doubling of epilepsy surgery centers worldwide, and improvement in seizure outcomes of surgical treatment specifically for anterior temporal lobe resections (Table 1).

Table 1.

Outcomes before 1985 and from 1986 to 1990

	Seizure-free	Number of patients (%)		Total
		Improved	Not Improved
Limbic resections
Before 1985	1296 (55.5)	648 (27.7)	392 (16.8)	2336 (100)
1986–1990 ATL	2429 (67.9)	860 (24.0)	290 (8.1)	3579 (100)
AH	284 (68.8)	92 (22.3)	37 (9.0)	413 (100)
Neocortical resections
Before 1985	356 (43.2)	229 (27.8)	240 (29.1)	825 (100)
1986–1990 ETR	363 (45.1)	283 (35.2)	159 (19.8)	805 (100)
L	195 (66.6)	63 (21.5)	35 (11.9)	293 (100)
Hemispherectomies
Before 1985	68 (77.3)	16 (18.2)	4 (4.5)	88 (100)
1986–1990 H	128 (67.4)	40 (21.1)	22 (11.6)	190 (100)
MR	75 (45.2)	59 (35.5)	32 (19.3)	166 (100)
Corpus callosotomies
Before 1985	10 (5.0)	140 (71.0)	47 (23.9)	197 (100)
1986–1990	43 (7.6)	343 (60.9)	177 (31.4)	563 (100)

ATL, anterior temporal lobectomy; AH, amygdalohippocampectomy; ETR, extratemporal resection; L, lesionectomy; H, hemispherectomy; MR, large multilobar resection.

From Reference 113, with permission.

By the time of the second Palm Desert conference, the tremendous increase in available antiepileptic drugs had become an obstacle to timely surgical intervention because neurologists could almost always find another drug to try. In fact, the term “medically refractory epilepsy” lost its practical usefulness when it would literally take a lifetime to try every available antiepileptic drug in every conceivable combination in each individual patient. A seminal conceptual advance introduced at the Palm Desert conference, therefore, was the definition of surgically remediable syndromes as conditions with a known etiology and a predictable natural history of pharmacoresistance after failure of a few appropriate antiepileptic drugs at maximal doses (73). Surgical treatment for these conditions is cost-effective because presurgical evaluation can usually be performed noninvasively and a high percentage of patients, by definition, will become free of disabling seizures postoperatively. Mesial temporal lobe epilepsy with or without hippocampal sclerosis is the prototype of a surgically remediable syndrome, but patients with demonstrable structural lesions that can be surgically resected also have surgically remediable epilepsy, as do infants and small children with secondary generalized epilepsies due to diffuse structural lesions limited to one hemisphere. It is important to note that many patients with pharmacoresistant epilepsy who do not have a surgically remediable epilepsy syndrome as defined here, might still benefit from surgical treatment and could even become seizure free. Such patients, however, usually require invasive evaluation, and thus the risks are higher and the benefits lower than for patients with well-defined surgically remediable epilepsy syndromes. Reducing the risks and increasing the benefits of surgery for such patients remains a major challenge for the future.

3.2 The behavioral perspective

By the latter part of the 20^th century, the Wada test had become a key tool in the neuropsychologist’s diagnostic armamentarium in epilepsy surgery centers worldwide. This was despite a range of methodological concerns arising in the late 1960s and early 1970s about possible widespread diffusion of the barbituate, individual differences in arterial distribution, imprecise determination of the duration of maximal drug effect, and the use of different stimuli and criteria for assessing language and memory across centers. Initially progress to address these concerns was slow, likely reflecting that the majority of clinicians using the Wada test had been trained at, or followed the published protocols of one surgical center (32). Although these concerns remain relevant today, by the 1990s considerable progress had been made, with work by Jones-Gotman and colleagues at the MNI (74,75), and Loring and colleagues in Georgia (30) contributing to a range of studies focusing on the reliability and validity of the procedure, accompanied by efforts to define and standardize an optimal approach (76).

In 1990, the National Institutes of Health (NIH) ‘Consensus Conference on Surgery for Epilepsy’ declared neuropsychological examination a necessary diagnostic procedure in the surgical evaluation of epilepsy patients (77). The material-specific model had become the cornerstone of the field, guiding both presurgical decision making and dominating the approach to the assessment and interpretation of cognitive outcomes (39,40). Rausch and colleagues at UCLA had demonstrated that subfield neuronal loss in the left hippocampus specifically predicted impaired learning of arbitrarily related words (the ‘hard’ pairs) of the paired-associate learning task, but not semantically-rich prose (78). Around the same time, Saling and colleagues in Melbourne reported mildly impaired story recall in patients with well characterised left or right mesial temporal foci, but selectively impaired learning of arbitrarily related words in patients with left mesial temporal foci (79). Notably, this led them to consider task-specificity, rather than material-specificity, as the more relevant factor in verbal memory outcomes, that distinguishes medial versus lateral specialisation (or arbitrary versus semantic forms of learning) within the temporal lobe (38).

Consistent with this notion of task-specificity, Hermann and colleagues (80) had reported that retroactive interference in word list learning might provide the most ‘pure’ indicator of memory functioning of the left temporal lobe, as it is unrelated to language adequacy. Following this, Helmstaedter and colleagues showed that memory tasks with a semantic component decline from pre-operative levels following en bloc temporal resections, whereas this is not evident after selective amygdalohippocampectomy. This provided further support for the notion of specialization of memory functions within the temporal lobe, again fractionating the roles of mesial and lateral temporal cortical structures (81).

Perhaps most salient at the NIH Consensus Conference in 1990, a new way of assessing surgical outcome was conceived “…that would take advantage of validated and quantitative methods to assess the quality of life and health status of individuals.” (77) This was, in part, anticipated by Carl Dodrill’s work on quantifying the psychosocial aspects of living with epilepsy through development of the Washington Psychosocial Seizure Inventory (82). The health related quality of life approach (HRQOL), however, extended this work by utilizing a general measure, to allow comparison with other chronic conditions, supplemented by epilepsy specific items (77). While a large number of HRQOL measures have now been established for people with epilepsy across cultures, Vickrey and colleagues at UCLA first responded to the NIH recommendation by developing the Epilepsy Surgery Inventory-55 (83), followed by a measure established by Baker and colleagues in Liverpool (84,85). These subjective measures of quality of life were designed to reflect the patient’s perception of daily functioning and experience of health and well-being across a range of dimensions, including general, physical, mental, and social functioning.

Initial studies employing HRQOL measures demonstrated that improvements after surgery were linearly related to a reduction in seizure frequency (84,86). Soon after, the importance of other predictors of post-operative HRQOL was recognized, including mood (particularly depression), cognition (perceived memory), employment, driving, and anticonvulsant cessation (87,88). In the case of low mood, a strong negative linear relationship with HRQOL was described by Gilliam and colleagues (89), likely reflecting that a sense of well-being is a basic component of human existence maintained by adaptive psychological processes that are essential for species survival (90). Around the same time, work by Trimble and colleagues in London (91), Blumer and colleagues in Memphis (92), and subsequently Wilson and colleagues in Melbourne (93) demonstrated the greater risk for early mood disturbance following resection of the temporal lobe compared to resection outside this region, including increased risk for de novo depression (92,93). These findings support a broader, emerging idea that shared pathogenic mechanisms may underpin mesial temporal epileptogenesis and mood disturbance (94), and possibly impaired episodic associative memory function.

While HRQOL research has principally focused on improvements in well-being after surgery, by the last decade of the 20^th century progress had also been made in characterizing the nature of the adjustment process reported by patients, particularly within the first 24 months after efficacious surgery. Bladin and Wilson likened this process to being ‘burdened with normality’, as patients face new challenges and refashion their psychosocial milieu as they undergo a psychological and social transition from chronically ill to well (90,95). Characterization of this commonly reported adjustment process concurred with an increasing emphasis in the literature on the need for post-operative rehabilitation programs that routinely address all aspects of a patient’s functioning to maximize the ‘real life’ benefits of seizure freedom and decrease the significant risk of adjustment difficulties and mood disturbance post-surgery (96). In preparing patients for the vicissitudes of post-operative life, the importance of addressing pre-surgical expectations of both patients and family members was also recognized (97,98), including the impact of expectations on the patient’s view of the success of epilepsy surgery (99).

4. The 21^st century

The past decade has seen continued advances in diagnostic testing and microsurgical techniques. This has increased the accuracy of localization of the epileptogenic region and the safety of surgery, as well as expanding the number of patients considered surgical candidates to include those with no demonstrated lesions on MRI, or diffuse or multifocal structural abnormalities, only one of which might be responsible for generating habitual seizures. Technologies that were still considered experimental at the end of the 20^th century have now become available for general use, including magnetoencephalography (MEG), functional MRI (fMRI), and simultaneous EEG-fMRI recordings with the capability of localizing alterations in blood flow associated with interictal spikes and ictal EEG discharges. Additional useful approaches include MR spectroscopy (MRS), to localize metabolic alterations associated with the epileptogenic region, diffusion tensor imaging (DTI) to map alterations in fiber tracks that may indicate the presence of an epileptogenic lesion, and statistical parametric mapping (SPM) of structural MRI, which provides high-resolution definition of localized abnormalities that are not visible on standard MRI.

fMRI now also provides a noninvasive means of mapping essential neocortical functions, permitting the boundaries of localized corticectomies adjacent to eloquent cortex to be determined noninvasively prior to surgery. This has generally been accompanied by a shift to validate less invasive techniques for assessing language and memory functions, using the Wada test as a benchmark. In conjunction with a worldwide shortage of amobarbital, this has led to reevaluation of the clinical indications for the Wada test (76). Recent surveys of a large number of epilepsy surgery programs have indicated that the Wada test is no longer routinely used in every patient. This has prompted a call for the development of a set of guiding principles for use of the Wada, to ensure the risk-benefit ratio is justified (76).

The epilepsy surgery setting has provided fertile ground for basic scientists interested in understanding fundamental neuronal mechanisms of epilepsy (100). For several decades, hippocampal sclerosis has been the primary focus of attention, but in recent years other common epileptogenic lesions, particularly focal cortical dysplasia, have been the subject of investigation. Not only is resected epileptogenic tissue available for molecular biological, microanatomical, and electrophysiological in vitro investigations, but several epilepsy centers have developed the capability of chronic in vivo microelectrode techniques to record localized field potentials and single unit activity, as well as microdialysis for evaluating neurotransmitter release. While the ultimate objective of this research is to identify primary epileptogenic abnormalities that could be the target for novel pharmacotherapy and other treatments, it has also identified disturbances that could lead to the development of reliable biomarkers of epileptogenicity.

There is still no single diagnostic approach that can identify not only the location but the extent of the epileptogenic region, that is the area necessary and sufficient for generation of spontaneous seizures, and thus the minimal area that needs to be resected in order to produce a seizure free outcome. The location and extent of the epileptogenic region remains approximated based on a variety of diagnostic information, including EEG, sophisticated neuroimaging, and neuropsychological testing, as well as observation of seizure semiologies. The identification of a biomarker that would reliably localize uniquely to epileptogenic tissue would be a major advance for epilepsy surgery. Such a biomarker could greatly decrease the cost and risk of presurgical evaluation, improve surgical results, and increase the number of patients who might be considered surgical candidates. Basic research in the epilepsy surgery setting, along with parallel reiterative research using experimental animal models of human surgically remediable epilepsy syndromes, have revealed a number of disturbances that could become targets for effective biomarkers (Table 2). At present, there are two potential biomarkers under active investigation.

Table 2.

Target mechanisms for biomarkers of epileptogenicity and epileptogenesis

Cell loss

Axonal sprouting

Synaptic reorganization

Altered neuronal function (gene expression profiles, protein products)

Neurogenesis

Altered glial function and gliosis

Inflammatory changes

Angiogenesis

Altered excitability and synchrony

There is evidence that AMT, on PET, localizes to epileptogenic tissue, particularly when there are multiple potential epileptogenic lesions, as in the case of tuberous sclerosis (101). More recent interest, however, has been in pathological high-frequency oscillations (pHFOs), brief 100–600 Hz EEG events often associated with interictal spikes which are believed to be summated action potentials of the synchronously bursting neurons characteristic of epileptogenic tissue (102). Although reports from several centers suggest that pHFOs could more reliably identify epileptogenic tissue than interictal EEG spikes and even ictal onset (103), these events can only be recorded from electrodes within or on the surface of the brain. To be most useful, noninvasive approaches to recording pHFOs are needed, and perhaps will be provided by MEG or EEG-fMRI. If the accuracy of these putative biomarkers is confirmed, even when invasive recording is required this could greatly reduce the cost and increase the efficacy of resective surgical treatment. As pHFOs are frequent interictal events, the invasive recording could be completed in a few hours, without the need for long-term chronic recording to capture spontaneous seizures.

The concept of identifying new markers of outcome has also emerged in current behavioral research (40). Moving away from the material-specific model, neurocognitive markers of mesial temporal dysfunction have been proposed, with arbitrary relational learning providing a candidate endophenotype of disruption of the rhinal-hippocampal interface (39). Progress has also been made in identifying neurobiological and psychological markers of risk for mood disturbance after surgery (104,105), with reduced contralateral hippocampal volume pre-surgery recently being identified as a marker for increased risk of depression post-surgery (106). This approach requires an understanding of the complex interaction of biological, cognitive, psychological, and social factors at play in a given individual, that combined, create the framework for an individual’s trajectory from seizure onset to chronicity and subsequent response to surgical treatment. For example, preliminary work linking the process of postoperative adjustment with HRQOL outcomes indicates that patients with early onset of seizures (before or during adolescence) are more likely to experience a greater sense of self-change after surgery, that while tumultuous, ultimately concurs more beneficial effects for HRQOL (107). In contrast, work linking HRQOL with cognitive outcomes shows that the “double whammy” of seizure recurrence and memory decline has a highly detrimental effect on HRQOL (108). Greater understanding of these interactions and their individual differences lies at the heart of identifying markers of risk for, and resilience to, poor post-operative outcomes. This is directly relevant to the ongoing development of models of post-surgical rehabilitation that can target treatment interventions to an individual’s specific phase and needs in post-operative recovery, to facilitate improvements in day-to-day functioning over the long-term.

Access to epilepsy surgery remains a major obstacle to be overcome. Epilepsy surgery is arguably the most underutilized of all therapeutic approaches that are generally accepted as effective. Perhaps only 1% of potential surgical candidates are ever referred to an epilepsy surgery center in industrialized countries, and at the beginning of this century very few individuals with epilepsy living in the developing world had access to surgical treatment. Several important advances have been made in the past decade in an attempt to improve this situation. Reluctance on the part of physicians and patient to consider surgical intervention has been attributed to several factors. Fear of surgery is understandable, although the risk of morbidity and mortality is much greater with uncontrolled disabling seizures than it is with surgery. We obviously have not done as good a job as we could in educating the general public and the medical community, despite the significant increase in publications since the first Palm Desert conference, that all support the safety and efficacy of surgical treatment for epilepsy.

One criticism at the end of the last century was that there had never been a randomized controlled trial (RCT) of epilepsy surgery that clearly confirmed the safety and efficacy that is reported in published uncontrolled series. This concern was overcome in 2001 with the publication of a RCT of epilepsy surgery at the University of Western Ontario (109) that clearly demonstrated the superiority of surgical treatment over continued pharmacotherapy in patients with temporal lobe epilepsy. Based on this, in 2003, the American Academy of Neurology, in association with the American Epilepsy Society and the American Association of Neurological Surgeons, published a practice parameter based on the RCT and a metanalysis, declaring surgical therapy the treatment of choice for pharmacoresistant temporal lobe epilepsy (110). Interestingly, two-thirds of patients were seizure free in the RCT and also in the metanalysis, indicating that the results of uncontrolled series are reliable. Nevertheless, a similar recommendation was not permitted for corticectomies based on metanalysis because no RCT had been performed.

The practice parameter also recommended that surgical treatment be carried out in a timely fashion in order to avoid the development of irreversible psychological and social consequences of disabling epileptic seizures. That same year, results of a multicenter study indicated that the average interval from onset of epilepsy to referral for surgery at seven centers in the United States was 22 years (111). An encouraging development over the past decade resulting from improvement in presurgical diagnostic approaches, has been the establishment of highly successful epilepsy surgery programs in emerging countries of Latin America, Asia, and the eastern Mediterranean. Recently, however, a study at UCLA revealed that the RCT and practice parameter has had no effect on the timing of surgical referral, which was exactly the same over two four-year periods taken in the 1990s, and towards the end of the current decade (112). Considerable work is ongoing to make epilepsy surgery safer, cheaper, and more effective, and tremendous advances have occurred in the 60 years since the founding of Epilepsy Action. Clearly a major effort is needed to communicate these advances to the general public and medical community in order to address the unacceptable treatment gap between potential surgical candidates and actual surgical treatment worldwide. In this regard, Epilepsy Action and the journal Seizure continue to play a vital role.