PubMed was periodically searched with the terms “brain stimulation”, “vagal nerve stimulation”, “VNS”, “transcranial magnetic stimulation”, “TMS”, “thalamic stimulation”, “caudate”, “cerebellum”, “subthalamic nucleus”, “anterior thalamic nucleus”, and “substantia nigra”, combined with the term “epilepsy”. The last search was completed in August 2003. Other papers for inclusion were identified from the personal files of the authors, from previous reviews of the subject, either by these
ReviewBrain stimulation for epilepsy
Section snippets
Vagus-nerve stimulation (VNS)
VNS is the only approved stimulation therapy for epilepsy and was reviewed recently in this journal.6 It is thought to be effective on the basis of clinical trials.7 However, as with other therapeutic modalities, VNS has been applied to seizure types that were not assessed in the controlled clinical trials.
The vagus nerve has diffuse and widespread projections to the thalamus, amygdala, and forebrain through the nucleus tractus solitarius and to other cortical areas via the medullary reticular
Clinical efficacy
Several controlled trials have assessed the efficacy of VNS in patients age 12 years and older with complex partial and secondarily generalised seizures. Because patients can sense when VNS is active, traditional placebo-controlled trials have not been practical. Instead, the trials sought to show effectiveness by a “high” dose (0·25–3·50 mA, 30 s on, 5 min off, 30 Hz, 500 μs pulse duration) versus “low” dose (0·25–3·5 mA, 30 s on, 180 min off, 1 Hz, 130 μs pulse duration) design, with a null
Side-effects
VNS is generally well tolerated.26 3–6% of patients have postoperative infections. Common side-effects most evident during stimulation are (in about 15–20% of patients) cough, hoarseness, dyspnoea, pain, paraesthesia, headaches, and (in about 50–60%) voice alteration. These side-effects respond to changes of the stimulation settings, particularly lowering of pulse width.27 Left-vocal-cord paralysis, lower facial weakness, sternocleidomastoid spasm, and transient bradycardia or asystole during
Deep brain stimulation (DBS)
DBS electrodes are implanted by use of stereotaxic methods similar to those used to implant stimulators to treat movement disorders.41 In the case of diencephalic stimulation, electrodes are typically placed bilaterally into the anterior principal nucleus, the centromedian nucleus, or the subthalamic region. Stereotaxic and MRI targeting, recording of extracellular unit activity, and electroencephalographic monitoring of stimulation effects are used to monitor the accuracy of electrode
Cerebellar
Cerebellar outflow is inhibitory in nearly all patients, so stimulation of the cerebellum is attractive for the treatment of seizures.43 Nevertheless, cerebellar stimulation has variable effects in animal models.44 Anterior stimulation decreased hippocampal formation discharges; inconsistent results may have been related to variable stimulation parameters. Cerebellar hemisphere stimulation had little effect. Seizures worsened in some studies.
An uncontrolled study of 115 patients reported that
Subcortical
Interest in the possibility of subcortical stimulation arose from recognition of widespread non-specific anterior and intralaminar nuclear connections to mesial frontal, temporal, and limbic structures, and progressive recruitment of substantia nigra, subthalamic nucleus, and midline thalamic nuclei in animal models of epilepsy. Stimulation of the subthalamic nucleus, anterior thalamus, and substantia nigra in animal models inhibits limbic seizures.48, 49, 50 Seizures are not believed to
Caudate nucleus
In animal models, stimulation of the caudate nucleus can inhibit a rhinencephalic seizure focus53 and cortical54 and hippocampal55 penicillin-induced seizure foci. Effects of stimulation of the caudate nucleus on an aluminium-hydroxide seizure focus in the motor cortex of a non-human primate depended on frequency of stimulation: 10–100 Hz stimulation was inhibitory, but 100 Hz stimulation increased seizure frequency.56 If we presume that low-frequency stimulation is excitatory and
Thalamic
The thalamus is an attractive target for stimulation to treat seizures because it has widespread connections with the cortex. As long ago as 1942, Dempsey and Morison58 defined a cortical EEG “recruiting response” and “augmenting response” resulting from thalamic stimulation. Wilder and Schmidt59 showed that thalamic stimulation could end seizures in a primate epilepsy model.
On the basis of the idea that the thalamus is a pacemaker for the cortex,60 the American neurosurgeon Irving Cooper
Centromedian nucleus
The centromedian nucleus has been thought of as part of the nonspecific thalamus, but its output has stronger relations with the caudate nucleus than with the cortex.62 Despite the absence of major monosynaptic connections between the centromedian nucleus and cortex, cellular firing patterns in the centromedian nucleus and cortex are strongly related in different stages of human sleep.63 Stimulation of the centromedian nucleus at three-stimulations per second by depth electrodes can produce
Anterior thalamic nucleus
Metabolic mapping studies have shown the anterior thalamus to be substantially involved in generalised seizures. Radiolabelled deoxyglucose mapping studies have highlighted the anterior thalamus in pentylenetetrazol-induced seizures in guinea pigs.72
The anterior thalamus is in the circuit of Papez, which goes from the hippocampus via the fornix to the mammillary bodies of the hypothalamus, to the anterior thalamus, the cingulate gyrus, and then—via the cingulum bundle—to the entorhinal cortex
Subthalamic nucleus
The experience from stimulation of this nucleus for the treatment of movement disorders makes it an appealing target. At least nine patients have been implanted in uncontrolled studies so far.52, 77, 78, 79 Up to 80% reduction in seizure frequency was reported in six; the others did not respond. Mild facial twitching and numbness in the extremities responded to adjustment of the stimulation settings.78 Sharp waves, thought to be an expression of direct glutamatergic pathways that could modulate
Direct stimulation of the epileptic focus
Homosynaptic long-term depression and long-term potentiation are persistent changes in synaptic strength that can be induced by electrical stimulation. Generally, low-frequency stimulation is inhibitory and high-frequency stimulation is facilitatory. Long-term depression can be induced by low frequency stimulation in hippocampus.81
Several preclinical studies found potential effects of brain stimulation on cortical excitability in epilepsy models. The nature of the effect may depend on the order
Hippocampus
Bilateral, depth, hippocampal or unilateral, subdural, basal temporal electrodes were implanted in ten patients before a temporal lobectomy.89 Antiepileptic drugs were discontinued from 48 h to 72 h before 2–3 weeks of continuous 130 Hz electrical stimulation, delivered 23 h per day. In seven patients whose stimulation electrode contacts were placed within the hippocampal formation or gyrus and who experienced no interruption in the stimulation programme, stimulation stopped clinical seizures
Implantable seizure detection
Implantable devices incorporating seizure detection algorithms have entered clinical trials;91 these devices should deliver stimulation to epileptogenic zones when the onset of ictal activity is detected. The strategy is distinct from other stimulation approaches, in that its aim is to block seizures acutely, rather than alter, cortical excitability on a chronic basis (patients with VNS implants can increase stimulation with an external magnet if they feel a seizure is about to occur).
Transcranial magnetic stimulation
TMS with either a hand-held magnet or a frame that can be aligned to predetermined coordinates is used widely in clinical neurophysiology to measure motor-cortex excitability.92 Patients can experience sensations such as “skin crawling” during stimulation.
Motor threshold, the main measurement used to judge cortical excitability, probably reflects cortical and spinal neuronal membrane excitability, as well as intracortical synaptic and corticospinal connections.93 The currents generated by TMS,
Selection of patients
Except for VNS, none of the brain-stimulation therapies has been shown to be effective in controlled trials; VNS is generally more complex and more expensive (at least in the short run [5–10 years]) than antiepileptic drug treatment. VNS is a useful adjunctive therapy for patients with localisation-related epilepsy characterised by complex partial and secondary generalised seizures that do not respond to antiepileptic drugs, or for those who do not tolerate drug therapy and are not candidates
Conclusions
Although stimulation for uncontrolled epilepsy with electrodes implanted in the brain itself has been done for many years, the initial claims for therapeutic success have not been confirmed by the very few controlled studies that have been done so far. The best structures to stimulate and the most effective stimuli to use are unknown. It is not clear why high-frequency and low-frequency stimulation have contrasting effects in different seizure models and structures. Many patients who have
Search strategy and selection criteria
References (109)
Vagus nerve stimulation for the treatment of epilepsy
Lancet Neurol
(2002)- et al.
MRI of the brain is safe in patients implanted with the vagus nerve stimulator
Seizure
(2001) Left vagal nerve stimulation in children with medically refractory epilepsy
J Pediatr
(1999)- et al.
Vagal nerve stimulation in tuberous sclerosis complex patients
Pediatr Neurol
(2001) - et al.
Left vagal nerve stimulation in six patients with hypothalamic hamartomas
Pediatr Neurol
(2000) - et al.
Adverse events in children receiving intermittent left vagal nerve stimulation
Pediatr Neurol
(1998) - et al.
Anticonvulsant effect of anterior thalamic high frequency electrical stimulation in the rat
Epilepsy Res
(1997) - et al.
Electrical stimulation of substantia nigra pars reticulata is anticonvulsant in adult and young male rats
Exp Neurol
(2002) - et al.
High-frequency stimulation of the subthalamic nucleus suppresses absence seizures in the rat: comparison with neurotoxic lesions
Epilepsy Res
(1998) - et al.
Effects of chronic stimulation of the caudate nucleus on a preexisting alumina seizure focus
Exp Neurol
(1982)