SEARCH

SEARCH BY CITATION

Keywords:

  • Deep brain stimulation;
  • epilepsy;
  • neurostimulation;
  • review article;
  • stimulation

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Deep Brain Stimulation of the Epilepsy Control Systems
  5. Direct Stimulation of a Seizure Focus
  6. Closed-Loop Stimulation
  7. Conclusions
  8. Conflict of Interest
  9. References
  10. Comments

Introduction. There is renewed interest in the use of electrical stimulation to control seizures in patients with medically refractory epilepsy. The evidence indicates that multiple nuclei are involved in the onset, spread, or termination of seizures. Establishing electrical stimulation parameters tailored to these nuclei that best control seizures is ongoing. Methods. The aim of this article is to review the published literature on electrical stimulation of the brain for epilepsy in animals and humans. Results. Animal and human research efforts have focused primarily on the study of the cerebellum, anterior thalamus, centromedian thalamus, substania nigra, caudate nucleus, subthalamic nucleus, and amygdalo-hippocampal complex. Electrical stimulation of deep brain nuclei has in some instances controlled seizures and epilepsy. The advent of seizure detection devices used in closed-loop studies has in part redefined the strategy to prevent seizure occurrence and limit spread. Discussion. A number of studies in animals and humans indicate that electrical stimulation may be an alternative treatment for some patients with medically intractable epilepsy who are not candidates for conventional surgical options. Conclusion. The reduction in the number and/or severity of seizures found in some studies supports further investigation into the effects of electrical stimulation on the brain and the continuation of testing in animals and humans.


Introduction

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Deep Brain Stimulation of the Epilepsy Control Systems
  5. Direct Stimulation of a Seizure Focus
  6. Closed-Loop Stimulation
  7. Conclusions
  8. Conflict of Interest
  9. References
  10. Comments

Pharmacotherapy is the cornerstone of treatment in the majority of patients diagnosed with epilepsy. Of the estimated 36% of patients with medically intractable epilepsy, surgery is often indicated (1). Still, this leaves many patients with medically intractable epilepsy without a surgical option, justifying the continued effort to find alternatives. Interest in deep brain stimulation (DBS) for epilepsy has swelled following the extraordinary success of DBS for movement disorders (2). Paralleling this interest in DBS for epilepsy has been the excitement surrounding the use of acute electrical stimulation (AES) for seizures. AES for seizures involves the activation of electrical stimulation following an electrographic change consistent with the onset of ictal activity. This is often described as “closed loop stimulation.” Penfield and Jasper recognized the potential for AES to arrest an ongoing seizure more than 50 years ago (3). However, the therapeutic potential of AES has only achieved clinical relevance now that automated seizure detection is available. At present, the use of AES has mostly been limited to controlling definable seizure foci.

On the other hand, DBS for epilepsy involves the chronic delivery of stimulation without reference to ongoing electroencephalogram (EEG) activity. This is often described as “open loop stimulation.” In patients with epilepsy, the brain oscillates between a functionally normal state and an abnormal, ictal one. This implies that the brain in patients with epilepsy is bistable or multistable (4). DBS for epilepsy hypothetically will establish or reinforce the functionally normal brain state. Alternatively, chronic stimulation of deep structures might suppress conditions or circuits that favor the emergence of seizures. Many protocols that have investigated DBS for epilepsy started with stimulation parameters derived from the literature on DBS for movement disorders. Establishing the optimum protocol for controlling epilepsy will not necessarily follow similar stimulation settings.

There are two conceptually diverging methods for selecting a stimulation site to “control” seizures. With the “direct control” method, the presumed source of seizures is the site for stimulation. The “remote control” method refers to the stimulation of a site that is remote from the source but presumed to control seizures (5). In general, DBS for epilepsy has been used to test remote control and AES has been used to test direct control. However, both techniques have been studied on different target types. Therefore, there are currently four paradigms that are used to test the effect of electrical stimulation on seizures (Fig. 1).

image

Figure 1. Schematic representation depicting the four paradigms used to study the effect of electrical stimulation on seizures. The open-loop paradigm applies chronic, intermittent or continuous, variable frequency stimulation to suppress epileptic activity. The closed-loop paradigm uses an implanted seizure detection device before activating electrical stimulation. In the direct control method, the electrode is implanted at the presumed source of epileptic activity. In the remote control method, the electrode is implanted in a site involved in the control of seizures.

Download figure to PowerPoint

In this article, we will review the evidence that supports or refutes the role of DBS for epilepsy and AES for seizures in both remote and direct control conditions. In early animal and human studies, these considerations were not taken into account. This confounds the interpretation and application of this research to current studies. Second, little attention was given to standardizing the type of seizure or model selected for study using electrical stimulation. It is now recognized that the onset and propagation of focal epilepsy and generalized epilepsy vary greatly along with the response to stimulation.

Deep Brain Stimulation of the Epilepsy Control Systems

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Deep Brain Stimulation of the Epilepsy Control Systems
  5. Direct Stimulation of a Seizure Focus
  6. Closed-Loop Stimulation
  7. Conclusions
  8. Conflict of Interest
  9. References
  10. Comments

General Concepts

Grouping related neuroanatomic structures with a proven role in seizure generation or modulation into definable epilepsy control systems serves as a framework both for understanding the rationale of choosing a target and for organizing areas of future study. For the purpose of this review, we have selected the control systems and structures that have been most frequently investigated using the remote control model. This includes the cerebellum, substania nigra (SNR) and related nuclei, anterior thalamus and the non-specific thalamus, specifically the centromedian nuclei (CM). Monitoring animals for an extended time is labor intensive and costly. Most studies therefore run a short course of stimulation prior to evoking an acute seizure through electroshock or chemoconvulsants and then grading severity. This is more consistent with AES for seizures and applicable to a closed-loop model. On the other hand, nearly all human trials studied open-loop DBS for epilepsy. The end point of these studies was a change in seizure frequency or severity over time with chronic, continuous, or intermittent stimulation.

Rationale for Cerebellar Stimulation

Irving Cooper initially hypothesized that chronic cerebellar stimulation capitalizes on the widespread Purkinje cell inhibitory output to contain pathologic, excitatory networks believed responsible for seizures (he later questioned this hypothesis, postulating that the effects were linked to activation of the reticular formation and inhibition of the thalamus) (6). This central tenant, crystallized in his work, was derived from both anatomic and physiologic studies that confirmed the functional importance of purkinje cells (7). Purkinje cells have a role in regulating seizures, controlling rigidity and related motor phenomena, and were recognized in early studies as the crucial cells in cerebellar electrical stimulation (8–12). Other studies suggested that the role of the cerebellum in seizures might be overestimated. For example, in the pentylenetetrazole (PTZ) model of generalized seizures in the rat, acute cerebellectomy did not influence seizure manifestations (13).

Animal Studies: Electrical Stimulation of the Cerebellum

While the direct involvement of the cerebellum in controlling seizures is still unsettled, numerous animal studies dating back to the mid 20th century have concluded that the cerebellum is a valid target for the study of electrical stimulation in seizures and epilepsy models(see Krauss & Koubeissi for review) (14). Many of these early studies were primarily descriptive in nature, lacked adequate controls or sufficient sample size, and routinely made inferences derived from what knowledge was available at the time. It is difficult to conclude based on these studies that cerebellar stimulation either reduces seizure frequency or markedly weakens its severity. Myers et al. tested electrical stimulation of the paleocerebellum of varying frequencies in four acute experimental cat epilepsy models without detecting any significant electrographic or behavioral changes (15). Acute, electrically induced focal seizure activity in the monkey did not respond to cerebellar hemispheric or dentate nucleus stimulation. Furthermore, no significant electrographic differences were observed at the sensorimotor cortex during stimulation (16). On the other hand, cerebellar stimulation increased seizure threshold in the rabbit following PTZ administration but not in the electrical stimulation induced seizure model (17). In a chronic epilepsy monkey model, anterior cerebellar vermis stimulation using parameters analogous to Cooper worsened seizure frequency (18). Reimer and colleagues described the effects of cerebellar stimulation on cobalt experimental epilepsy in the cat (19). Cerebellar stimulation neither prevented seizure onset nor shortened seizure duration but rather, in some instances increased seizure duration. Similarly, detectable electrocorticography activity was unchanged during fastigial and dentate chronic cerebellar stimulation in the cobalt epilepsy model in the squirrel monkey (20). However, in the hippocampal cobalt cat epilepsy model, Babb et al. reported that fastigio-bulbar stimulation during the ictus shortened duration and cut off clonic phase seizures (21). Of note, stimulation parameters were not standardized in these experiments. A recent publication supports the antiepileptic effect of high frequency (100–300 Hz) paleocerebellar stimulation in rats treated with benzylpenecillin sodium; an effect that was exacerbated using low-frequency (10–12 Hz) stimulation (22). The relatively few new animal studies published over the last decade indicate that interest in cerebellar stimulation for epilepsy is fading (see Myers et al. for summary of cerebellum stimulation studies 1941–1974) (15).

Human Studies: Electrical Stimulation of the Cerebellum

Despite the shortfall in animal studies supporting cerebellar stimulation, human studies were performed (9,23). Davis and Emmonds reported improved seizure frequency in some patients with epilepsy who underwent bilateral cerebellar stimulation (superior medial cerebellar cortex) for the treatment of spasticity (24–26). These unanticipated clinical results prompted two small controlled studies. Both studies failed to support the findings of the uncontrolled studies. Although, the statistical analysis of first controlled study is in question (25,27,28). In addition, other, smaller, uncontrolled studies were inconclusive at best (29).

Recently, a double-blinded, randomized controlled pilot study of bilateral cerebellar stimulation on five patients with medically uncontrolled motor seizures was published (30). Included in the findings was a reduction in generalized tonic-clonic seizures during a six-month stimulation period. The stimulator was adjusted to deliver a charge density of 2.0 microC/cm(2)/phase at 10 pps with a fixed pulse width of 0.45 msec (ON/OFF cycles every 4 min). These rather contradictory effects, the lack of agreed upon target within the cerebellum and the absence of well-defined stimulation parameters have dampened the initial excitement about cerebellar stimulation in the treatment of epilepsy.

Rationale for Stimulation of the Nigral Control System

Since Iadorola and Gale first described the nigral control of epilepsy system in 1982, our knowledge of this system has evolved to understand the contributions of the striatum and subthalamus (31,32). Multiple studies have concluded that the basal ganglia and related structures are subcortical modulatory systems of absence seizure frequency and duration. The modulatory influence of the cortico-striato-nigral and cortico-subthalamo-nigral circuits on both normal and abnormal thalamo-cortical oscillations has been well characterized during the last decade. Electrical stimulation trials in animals, aimed at one or multiple nuclei in this circuit might serve to identify a site responsible for altering seizure threshold in humans. At this time, several trials in animals and fewer in humans have been carried out with mixed results.

Animal Studies: Electrical Stimulation of the Substania Nigra

In amygdala kindled rats, bilateral HFS (130 Hz, 60 µs, 100–200 µA) of the SNR prevented seizure recurrence (33). In the GAERS model, seizures were suppressed with 60 Hz, 60 µs stimulation of the SNR bilaterally (34). However, repeated stimulations lowered seizure threshold. Further animal studies are warranted to evaluate SNR as a target. There are no reported human cases of SNR stimulation for epilepsy in the literature.

Animal Studies: Electrical Stimulation of the Caudate Nucleus (CN)

The potential anticonvulsant effect of electrical stimulation on the CN has not been sufficiently tested in either acute or chronic animal models of seizures. Mutani and Fariello showed brief (0.5 Hz, 1 msec pulse width) electrical stimulation of the caudate, with an interstimulus interval less than 2.5–3 sec, prevented the onset of epileptic activity in the cobalt epilepsy model in the cat (35). However, caudate stimulation was unsuccessful in stopping a seizure after it did start. Wagner et al. suppressed “weak to moderate” focal epileptic activity in the acute feline penicillin model using electrical stimulation of the caudate (36). Following the application of penicillin to the cat hippocampus, 10 Hz (0.5–1.0 msec, 0.1–0.5 mA) stimulation of the CN suppressed hippocampal spiking activity to a greater extent then 25 Hz stimulation when used for up to 180 sec (37). Oakley and Ojemann conducted a well-designed trial investigating caudate stimulation in monkeys (38). They tested the effect of chronic, electrical stimulation on the CN ipsilateral to a preexisting alumina cream seizure focus. Interestingly, they found that continuous, HFS (100 Hz, 1–4.0 mA and 1 msec pulse width) decreased seizure frequency in contrast to an alternating 10 min on/10 min off paradigm or in animals with caudate lesions. Unfortunately, follow-up studies were not conducted despite these encouraging results. Active research in the area of caudate stimulation in the animal model has largely fallen off.

Human Studies: Electrical Stimulation of the Caudate Nucleus

The initial trials that studied the effect of caudate stimulation on humans with known seizures were not systematic and yielded only marginal results at best (39). In recent uncontrolled human trials, interictal epileptic activity as well as seizure propagation was suppressed with low-frequency (4–8 Hz), short-duration (2–5 sec) stimulation of the CN head (40). Stimulation protocols were individually tailored to reflect the daily occurrence of seizures. Intensity was calibrated based on the patient's awareness of the stimulus during the ON phase. High-frequency stimulation of CN (50–100 Hz) intensified epileptic activity in the ipsilateral hippocampus and amygdala as well as scalp EEG (41).

Cortico-Subthalamo-Nigral Circuit

Interest in the subthalamic nucleus (STN) as a possible player in seizure modulation was a direct outgrowth of the research supporting the influence of the cortico-striato-nigral circuit on absence seizures in conjunction with independent experiments of the STN. Synchronized, rhythmic high-frequency bursts of action potentials have been recorded from the STN during cortical spike-wave discharges (SWD) (42). Increased excitatory activity of subthalamonigral neurons co-occurring with decreased striato-nigral GABAergic output could sustain SWD through the physiologic imbalance generated. In a convulsive model, micro-infusions of muscimol, a GABAA agonist, injected into the rat STN (unilateral or bilateral) significantly increased the threshold for clonic seizures in an acute fluorethyl seizure model (43). This effect was corroborated by Dybdal and Gale who in 2000 demonstrated that muscimol, injected bilaterally into STN, antagonized the effect of biculculline either systemically (IV) administered or focally applied to area tempestas (44).

Animal Studies: Electrical Stimulation of the Subthalamic Nucleus

Several animal studies using electrical stimulation have sought to explore the pathways outlined above, much of it using the GAERS model. Acute, HFS (130 Hz, 60 µs, intensity titrated to behavioral side-effects) of STN suppressed SWD in GAERS (45). This effect did not persist during continuous stimulation but was specific to bilateral STN stimulation. It is worth noting that an exogenous stimulus of sufficient intensity is capable of interfering with many EEG rhythms. Therefore, less intense epileptic events, in this case SWD, may be abolished by stimulation. Habituation could account for the failure of the stimulus when applied chronically. And furthermore, it is unclear if this effect is transferable to other models of more intense seizure with motor manifestations.

In the fluorethyl acute seizure model, clonic but not tonic-clonic seizure threshold was elevated compared with controls with acute bilateral STN stimulation at 130 Hz (an effect not reproducible at 260 and 800 Hz) (46). In a small study (six rats/group), HFS (130 Hz) of STN inhibited seizure generalization in a kainic acid (KA) model in the rat compared with controls and SNR stimulation (47). In a series of negative experimental trials testing muscimol injections, excitotoxic lesions and HFS (130 Hz) of STN, Shehab et al. were unable to block tonic seizures in an electroshock model of brainstem seizures in the rat (48). Our group tested the effect of HFS of the STN bilaterally in the KA model of acute status epilepticus: bilateral HFS of STN delayed EEG seizure onset, decreased the duration of the generalized EEG component of the status but did not change the total duration of EEG seizures (that included hippocampal focal seizures). Further experiments suggested that the lack of effect of hippocampal seizures may be due to lack of functional connectivity between the STN and the hippocampal formations in the rats (Boongird et al., Unpublished data).

Human Studies: Electrical Stimulation of the Subthalamic Nucleus

Preliminary human trials testing STN stimulation have yielded beneficial results in some cases warranting additional human research (49–51). Charbades et al., in an uncontrolled trial, reported seizure reductions in four out of five patients with severe epilepsy compared with pre-operative status (49). The stimulation parameters mirrored those used for movement disorders (130 Hz, 60–90 µs, intensity titrated to side-effects). Another clinical trial was performed at Cleveland Clinic Epilepsy Center using the same parameters as above failed to show long-lasting significant seizure reduction and control in none of the five patients enrolled in the study (52).

Rationale for Stimulation of the Non-Specific Thalamic System

The non-specific thalamic system includes the intralaminar, paralaminar, and midline thalamic nuclei. Most epilepsy stimulation studies in humans have concentrated on the intralaminar nuclei, specifically the centromedian nucleus (53–63). A number of animal studies have investigated in great detail the corticothalamic network (see Destexhe and Sejnowski 2003 for review) that is responsible for SWD but animal experiments testing DBS of the non-specific thalamic nuclei are scarce (64).

Human Studies: Electrical Stimulation of the Centromedian Nuclei

Velasco et al. have investigated electrical stimulation of the CM is patients with multiple cortical foci, bilateral symmetric foci, seizures arising from eloquent cortex and nonlesional epilepsy who are otherwise not resective surgery candidates (55). Much of their work, which began after bilateral implantation of CM in a 12-year-old in with Lennox-Gastaut sydrome (LGS) in 1984, has been derived from original reports on electrical stimulation of the nonspecific thalamic nuclei by Morrison and Dempsey (65). Several human trials have evaluated the effect of electrical stimulation on the CM primarily in patients diagnosed with LGS. Response rate was not as robust in epileptic patients who did not have LGS although stimulation tended to impede secondary generalization (55). In a recent report on CM stimulation, Velasco et al. reported an 87% reduction in seizures in patients followed for an average of 46 months. Stimulation settings were maintained at 130 Hz, 0.45 msec, 400–600 µA alternating in left/right 1-min trains with a 4-min quiescent interval phase. Furthermore, they have reported a residual effect lasting months after stimulation is off (54,60). In a placebo controlled pilot study, Fisher et al. did not find a statistically significant reduction in seizures when the stimulator was turned on (53).

Rationale for the Stimulation of the Anterior Nucleus of the Thalamus (AN)

The AN is a crucial structure in the propagation of limbic epilepsy. It is uniquely situated to project to the cingulate cortex, amygdala, hippocampus, orbito-frontal cortex and caudate and receives inputs from the mamillary bodies. The AN has been extensively investigated both in animal models of epilepsy and in human trials. Unlike other thalamic nuclei that have been the focus of DBS studies, it is not under the control of the reticular thalamus. A microinjection of PTZ into AN of the guinea pig exacerbated seizures (66). Electrolytic lesions of the mamillothalamic tracts protected against PTZ induced seizures (67). Ongoing seizure activity was abolished after microinjections of muscimol into the AN (66). A GABA transaminase inhibitor, injected into the AN, blocks seizures following intraperitoneal PTZ but not maximal electroshock seizures (66).

Animal Studies: Electrical Stimulation of the Anterior Thalamic Nucleus

Mirski et al. demonstrated that HFS of the AN elevated the seizure threshold during continuous PTZ infusion compared with controls and LFS (8 Hz) which proved to be proconvulsant in this setting (68). The protective effect of HFS on the AN in the PTZ model has been reproduced in other rat seizure models, notably in the pilocarpine model (69,70). In this model, bilateral HFS (but not unilateral) delayed the onset of behavioral seizure manifestations. However, HFS did not prevent eventual SE in the majority of animals at this dose (320 mg/kg i.p.) similar to unstimulated groups; in contrast bilateral lesioning of the AN forestalled the effects of the pilocarpine. In a follow-up study which addressed optimum stimulus settings, 500 µA significantly increased latency to seizure onset and SE (70). Hourly motor seizure frequency was reduced with unilateral HFS of AN or unilateral lesioning of AN and blocked with bilateral HFS of AN or bilateral lesioning in an acute focal seizure model injection of KA into the sensorimotor cortex (71). In a related study using the same subgroups, seizure severity and frequency were reduced compared with controls following intramygdaloid KA injection (72). The authors concluded that the absence of convulsions in rats following HFS of AN in either model was potentially the result of a microthalamotomy effect. On the other hand, Lado observed an increase in daily seizure frequency during chronic bilateral HFS of AN in the KA model of chronic epilepsy (73).

Human Studies: Electrical Stimulation of the Anterior Thalamic Nucleus

In 1986, Upton et al. reported improvements in seizure frequency in four out of six patients with chronic bilateral stimulation (60–70 Hz, 3.5 V, 300 µs pulses) of the AN (74). Lee et al. reported a 75.4% reduction in seizure frequency in three patients with bilateral AN implants (75). In the same study, 49.1% had a reduction in seizure frequency with STN stimulation. Hodai et al. reported a mean reduction in seizure frequency of >50% after HFS (100 Hz) of the AN. This reduction was noted after implantation, prior to the stimulator being activated (76). Preliminary studies addressing the safety and in some cases efficacy of HFS of the AN have paved the way for controlled studies (77,78).

In a trial of four patients with inoperable mesial temporal lobe epilepsy (TLE), Osorio et al. reported a 75.6% mean reduction in seizure frequency after cyclic 1 min ON/5 min OFF, 145–170 Hz, 60–90 µs and 1.5–5 V stimulation of the AN (79). The protocol included a six-month baseline followed by 36-month treatment period. A phase III, randomized, double blind trial, SANTE—Stimulation of the Anterior Nucleus of the Thalamus for Epilepsy, is currently testing safety and efficacy of stimulation on the AN in patients with partial onset epilepsy.

Direct Stimulation of a Seizure Focus

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Deep Brain Stimulation of the Epilepsy Control Systems
  5. Direct Stimulation of a Seizure Focus
  6. Closed-Loop Stimulation
  7. Conclusions
  8. Conflict of Interest
  9. References
  10. Comments

General Concepts

Pollo and Villemure designated chronic electrical stimulation of a suspected source for seizures as “direct control” in clear contrast to the methods described as “remote control” previously elaborated on in this review (5). (Cortical stimulation, an example of direct control when applied to a cortically based seizure focus, is not discussed in this review because it is not a deep brain structure). The preponderance of in vivo and in vitro animal studies investigating the etiology and mechanism of focal epilepsy are centered around the mesial temporal lobe leaving little doubt that this region of the brain is an ideal site to test direct control. The number of patients with TLE who are not candidates for surgical resection is small. Nevertheless, this number is not insignificant given the prevalence of TLE in the population overall and warrants the development of minimally invasive surgical alternatives. Furthermore, patients with proven focal TLE will not necessarily benefit from DBS at distant structures. This was observed in patients with complex partial seizures who underwent DBS of the CM nucleus (56).

Animal Studies: Direct Stimulation of the Hippocampus or Amygdala

There are a limited number of experiments in animals that have specifically tested focal DBS in models of hippocampal epilepsy. Despite this shortage, human studies continue to accumulate in the literature. There is in vivo evidence supporting the disruption of epileptogenesis by electrical stimulation. Low-frequency stimulation (1 Hz for 15 min) or quenching, suppresses the expression of amygdala kindled seizures in the adult rat (80). A companion study corroborated the effect of LFS on the basolateral amygdala in immature fast kindled rats (81). The authors postulated that LFS induced long-term depression or depotentiation. In another study of TLE in rats, continuous, HFS (130 Hz) delivered over one week to the hippocampal region increased afterdischarge threshold, afterdischarge latency and shortened its duration (82).

In two animal studies that explicitly studied the effect of seizure focus stimulation, neither one reported a significant improvement in seizure frequency. Daily two-hour high (50 Hz) or low (1 Hz) frequency stimulation of the right hippocampus and right perforant path failed to suppress interictal and spontaneous seizures in rats after intrahippocampal KA injection (83). Tanaka et al. tested continuous high (130 Hz, 100–200 µA) and low (10 Hz, 50–100 µA) frequency stimulation in rats with KA limbic seizures: HFS aggravated seizure frequency (84). In rats treated with LFS, seizure frequency was unchanged or inconsistently decreased.

Human Studies: Direct Stimulation of the Hippocampus or Amygdala

While animal studies have not made any significant inroads in decreasing seizure frequency with focal hippocampal stimulation, human studies have demonstrated success in some cases. In a feasibility study, subacute, hippocampal HFS up to 21 days in patients invasively monitored significantly reduced daily seizures and interictal spikes after a brief bump in seizure frequency once antiepileptic medications were withdrawn. Patients in the sham group continued to exhibit seizures (85,86). In a pilot study of 12 patients with mesial TLE, continuous, HFS of the hippocampus was safe and efficacious in the majority of patients enrolled (87–89). However, only one patient was reported to be seizure free. By comparison, 58% of patients were seizure free in a randomized controlled trial between medical and surgical therapy for TLE (90). In a double-blinded study of hippocampal HFS in nine patients, five patients with normal MRIs had a seizure reduction of 95%, four of whom became seizure free. The remaining four patients with MRI confirmed hippocampal sclerosis had a seizure reduction of 50–70% (91). In four of the patients, the stimulation was turned on immediately and the remaining five patients had a one month off period preceding the stimulation. In a double-blinded, cross over study which sought to examine seizure frequency after unilateral (190 Hz, 90 µs, sub-threshold stimulation) hippocampal stimulation in four patients with refractory mesial TLE, a median reduction of 15% was reported (92). At four-year follow-up, one patient achieved sustained improvement in seizure frequency. Pollo and Villemure reported a 50–95% reduction in seizure frequency with amygdalo-hippocampal complex stimulation in five patients followed from five months to three years (parameters not reported) (5). None of these studies addressed LFS.

Closed-Loop Stimulation

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Deep Brain Stimulation of the Epilepsy Control Systems
  5. Direct Stimulation of a Seizure Focus
  6. Closed-Loop Stimulation
  7. Conclusions
  8. Conflict of Interest
  9. References
  10. Comments

In contrast to the independent, continuous effect of open-loop stimulation, closed-loop stimulation relies on ultra-short stimulation periods activated in response to epileptic events. Most of the preliminary research stems from cortical stimulation studies, although deep targets have also been investigated (93). Early trials evaluated the safety and tolerance to closed-loop stimulation through an externalized “responsive” neurostimulator. Kossoff et al. described four patients who were tested with the externalized device during chronic invasive monitoring prior to resective surgery (94). Osorio published the results of a pilot study that tested closed-loop stimulation on local targets (local closed loop, LCL) and deep targets (remote closed loop, RCL) (95). Eight patients were enrolled in the pilot study divided evenly between the two groups. High-frequency stimulation (>100 Hz) was delivered to the epileptogenic zone in the LCL group and to the AN in the RCL group. The results of this pilot study were promising with responders in both groups.

Based on the safety and efficacy of the externalized, closed-loop system, an implantable closed-loop system was soon developed. The Responsive Neurostimulation System (RNS) (Neuropace Inc, Mountainview, CA, USA) includes a programmable pulse generator/electrographic monitor that sits atop the dura connected to either intracranial depth or strip electrodes. The neurostimulator is capable of detecting and parsing electrocortigraphic signals consistent with pre-epileptic or epileptiform activity through either of the intracranial electrodes after a post-implantation programming phase. When an event is detected, a biphasic pulse burst is administered through the depth or surface strip electrode. The site of stimulation is the suspected epileptogenic zone or zones. This site and the parameters selected are at the discretion of the epileptologist. An invasive evaluation is not a necessary prerequisite for device implantation. Nevertheless, the development of implantable hardware with accurate seizure detection and interrogation and programming features will be an invaluable asset in future models of closed-loop DBS. For example, DC fields, potentially hazardous to neural tissue when used in an open-loop model, may have a role when coupled to a seizure detection device.

The results from early studies testing Neuropace are encouraging. Fountas et al. reported that in seven out of eight patients, seizure frequency was reduced more than 45% with an implantable closed-loop stimulation device after a mean follow-up of 9.2 months (96). In the majority of patients enrolled in this study, the seizure focus was located in the hippocampus. In a feasibility study, the interim analysis of 24 patients found that 43% of patients with complex partial seizures and 35% of patients with total disabling seizures, had at least a 50% reduction in seizure frequency (97). A multi-center clinical trial evaluating the efficacy of Neuropace in patients with simple partial seizures (motor or sensory) or complex partial seizures (motor) with or without secondary generalization is ongoing.(For a complete compendium of inclusion and exclusion criteria, see http://www.neuropace.com/trials/overview.html).

Conclusions

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Deep Brain Stimulation of the Epilepsy Control Systems
  5. Direct Stimulation of a Seizure Focus
  6. Closed-Loop Stimulation
  7. Conclusions
  8. Conflict of Interest
  9. References
  10. Comments

The stimulation parameters for epilepsy should be systematically titrated to influence the properties of the nucleus or site under study rather than duplicating the effects of a lesion. In vitro and in vivo experimental evidence has uncovered a number of deep nuclei that may modulate epileptogenic foci. In some cases, DBS at these sites has reduced seizure frequency and/or severity in both animal and human studies. Direct control of seizures, especially in the hippocampus also holds promise. The design of automated seizure detection devices and their clinical applications in closed-loop stimulation research has arguably transformed the field of AES. New research supporting the notion that changes in brain dynamics are detectable well before seizure onset should help refine closed-loop stimulation. Furthermore, improvements in the testing and analysis of dynamic epileptic networks will continue to shape future endeavors in this field.

References

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Deep Brain Stimulation of the Epilepsy Control Systems
  5. Direct Stimulation of a Seizure Focus
  6. Closed-Loop Stimulation
  7. Conclusions
  8. Conflict of Interest
  9. References
  10. Comments

Comments

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Deep Brain Stimulation of the Epilepsy Control Systems
  5. Direct Stimulation of a Seizure Focus
  6. Closed-Loop Stimulation
  7. Conclusions
  8. Conflict of Interest
  9. References
  10. Comments

Dr. Kendall Lee wrote an editorial in response to this manuscript, rather than writing a public “comment”. I will submit this guest editorial separately.—TS

As a specialist of epilepsy surgery, I welcomed the publication of this article the Journal of Neuromodulation. As the authors mentioned, a significant number of patients with intractable epilepsy still experience poor control of their condition despite modern aggressive medical management and/or resective surgery.

Among the modern technologies in medicine, such as gene therapy, stem cell therapy and neuromodulation, the neuromodulation approach is the most practical. Anterior thalamic nucleus is the only reliable target among the deep brain stimulation trials; however, the result is modest, though we do not know the appropriate candidates and they may have potential surgical complications. However, we should find better targets – maybe multiple targets or better stimulation. More long-term animal studies must be conducted.

Hippocampal /amygdala stimulation would be very interesting and easy to understand and would be applicable for candidates such as those with common mesial temporal epilepsy and bilateral temporal lobe epilepsy, although the result is not satisfactory. However, with more advanced tools, such as epilepsy electrodes from the device companies, we could have better results.

For closed-loop stimulation, so called RNS is very attractive tool and applicable to humans, although it has modest results. However, the result could be enhanced with better hardware.

Finally, as a specialist in epilepsy, it is important to read this article to know the concept of brain stimulation and results in general.

Jung-Kyo Lee, MD, PhD Professor, Department of Neurosurgery Asan Medical Center Ulsan University Seoul, South Korea