Although the ANT is a relatively large structure, one study suggests that the internal complexity of the thalamus may make effective targeting difficult (Osorio et al. 2007). They recorded in bilateral hippocampi during bilateral ANT stimulation in four patients, with highly variable responses across patients. They also confirmed placement with post-operative MRI and found that only three of the 16 stimulating contacts (two per nucleus in four patients) were located in the ANT proper. Curiously, the patient with contacts in the ANT had the most modest reduction in seizures with stimulation.

Since ANT lesions have themselves been shown to reduce seizures (Mullan et al. 1967), the question of a “lesion effect” from stimulator placement is also important. Hodaie et al. (2002) reported an immediate mean reduction in seizure frequency of 56.8 % after electrode implantation compared to baseline. They found no significant change in seizure reduction out to an average of 15 months when the stimulators were turned on, nor when they were turned off for a period of 2 months in a blinded fashion. These patients were subsequently followed for 2–7 years, during which time seizure frequency reductions over baseline were maintained, but did not improve (Andrade et al. 2006). In another case series, however, some individual patients were noted to suffer an increase in seizure frequency when the stimulators were unintentionally turned off (without awareness of the patient), with return of seizure improvement with resumption of stimulation (Kerrigan et al. 2004). These results only emphasized that a randomized, controlled, blinded study was needed.

In 2010 the anticipated results of the multi-center, double-blind SANTE trial (Fisher et al. 2010) were published. To be included in this trial patients had to have partial seizures, with or without secondary generalization, in line with the logic that ANT stimulation disrupts subcortical–cortical propagation of seizures along a distinct limbic pathway. Patients had failed at least three antiepileptic drugs and had no evidence of a structural brain lesion. Patients were implanted with bilateral ANT stimulating electrodes after completing a 3-month baseline period during which their antiepileptic drug regimen remained stable. If an electrode was not within the ANT on post-operative imaging, it was replaced. Patients were randomized to stimulation on or stimulation off groups 4 weeks after implantation. The blinded phase lasted 3 months and included 108 patients. Using a model that adjusted for repeated measurements, the authors reported a 29 % greater seizure frequency reduction in the stimulation group compared to the control group during month 3 of the blinded phase. It is instructive to look at the (unadjusted) median seizure frequency improvement in the blinded phase over time (Fig. 14.1). Both the stimulation and the control groups saw a comparable reduction in seizures in the month after implantation, before stimulation, which may capture a conjectured “lesional effect.” In the ensuing blinded months, however, the two groups progressively separate, the difference becoming significant in the third month. After the 3-month blinded phase, all patients entered an open trial of stimulation with restrictions on stimulation parameters. After month 13 stimulation parameters were allowed to vary freely. Median seizure frequency reduction at 13 months was 41 %, and increased to 56 % at 25 months. These results demonstrate a benefit of high-frequency bilateral ANT stimulation for refractory epilepsy. Further, more than having failed antiepileptic medication, 44.5 % of patients had failed vagal nerve stimulation, and 24.5 % had achieved no benefit from surgical resection, demonstrating that ANT DBS may be the therapy of choice in certain patients with partial epilepsy. Further trials need to be done in patients with primary generalized seizures, such as those with epilepsy syndromes (e.g. West syndrome, Lennox-Gastaut syndrome), who are often among the most difficult patients to treat (Sander 2003).