Epilepsy in the dog is a common and frequently debilitating condition that often leads to a shortened lifespan, death, and major distress for owners of affected dogs.[1-4] The disease in dogs occurs at a frequency similar to epilepsy in people, with seizures occurring in 0.5–5% of all dogs and with idiopathic epilepsy being the most common underlying cause.[5, 6] Although antiepileptic drugs (AEDs) continue to form the cornerstone of epilepsy treatment in this species, up to 25% of epileptic dogs remain refractory to medication. This compares to a similar proportion in the human epileptic population, among whom approximately one third continue to suffer seizures despite medication. For a subset of human patients with hippocampal sclerosis, resection of the mesial temporal lobe will result in a reduction in seizure frequency of 80–90%, but for many patients this is inappropriate because of the location of the seizure focus, unacceptable adverse effects such as memory impairment, or multiple seizure types. As a result, other potential treatments are actively being investigated.[9, 10]
The success of vagal nerve stimulation as a form of neurostimulation capable of improving seizure control has led to research into deep brain stimulation (DBS) as a treatment for epilepsy in recent years.[9, 11-20] Targets proposed for stimulation have included the anterior nucleus of the thalamus (ANT), the centromedian nucleus of the thalamus (CMT), the subthalamic nucleus, and the hippocampus. Much of this work has been performed in rodent models and has provided valuable insight into the pathogenesis of seizures. However, rodent models suffer from disadvantages related to their size and dissimilar anatomy as compared with people. For this reason, large animal models such as the sheep and minipig are being developed that will allow better translation of DBS strategies into people. Recently, one such study has investigated the feasibility of implanting leads into the ANT and hippocampus of a sheep model as part of the evaluation of a novel device capable of both stimulating and recording from the brain.[22, 23]
Vagal nerve stimulation as a treatment for epilepsy in dogs has shown some promise, suggesting that electrical stimulation of the nervous system in this species also may be capable of improving seizure control. With the advent of the Brainsight system and other frameless stereotaxy devices, it has recently become possible to access deep regions of the brain in dogs and potentially to stereotactically implant DBS electrodes.[25, 26] Given the potential for epileptic dogs to provide a spontaneous model in which to test thalamic stimulation, therefore, the aim of this study was to investigate the feasibility and safety of placement of stimulating and recording electrodes in the ANT and hippocampus of normal dogs using the Brainsight system, as well as to test a novel device capable of providing DBS in conjunction with simultaneous EEG recording.
- Top of page
- Materials and Methods
This study showed that it is feasible to place DBS and recording electrodes in the thalamus and hippocampus of dogs with accuracy and with minimal morbidity. Using the Brainsight system, an overall mean accuracy of 4.6-mm placement error was achieved in the dogs with successful fixation. Given the size and spacing of the electrodes, this accuracy allowed the positioning of at least 1 electrode in the target in all dogs. In all cases, the dogs recovered without adverse neurologic effects and showed no outward sign of the effects of the implantation or stimulation. In addition, stimulation and recording from the electrodes showed that stimulation of the ANT can produce evoked responses identical to those obtained in other large animal models in which similar studies have been performed. The shape and latency of the EPs was consistent with activation of the circuit of Papez, suggesting that the dog is a suitable model for the testing of DBS strategies in the treatment of epilepsy.
To the authors' knowledge, this is the first report to describe the implantation of indwelling electrodes in the brain of dogs for the purposes of recording and stimulating simultaneously. Other studies in the field of veterinary medicine have examined the use of stereotaxy for needle placement, but predominantly in the setting of lesion biopsy.[28-31] In addition, studies have examined image-guided freehand or directed biopsy techniques, both to biopsy lesions and to implant indwelling catheters for delivering brachytherapy. Overall, the morbidity and mortality described in most studies in dogs is low, ranging from 12 to 29% and 6 to 9%, respectively.[28, 30, 32] This is higher than the 3–12% morbidity rate and 0–1% mortality rate reported in human patients.[33, 34] The limited numbers in our study do not allow for direct comparison, but the use of the Brainsight system to implant electrodes did not result in any neurologic deficits in any of the dogs involved. It is likely that the implantation of these electrodes is a safer procedure because there is no suction applied to the brain parenchyma and therefore the risk of intracranial hemorrhage is smaller than with needle biopsy. In addition, the dogs in this study had no structural pathology, as compared with dogs undergoing needle biopsy to diagnose lesions, in which the presence of additional blood vessels within the lesions would increase the likelihood of hemorrhage, the biggest cause of postbiopsy morbidity.
Given the small size of some of the targets involved, such as the ANT, obtaining accurate electrode placement is extremely important in a trial such as the one reported here. A mean placement error of 4.6 mm was acceptable, given that in all cases, this allowed for contact of at least 1 pair of contacts with the target, although further testing in more cases is necessary to definitively assess accuracy of placement. Several factors are likely to have contributed to the placement error. When using the Brainsight system, a registration to the animal's MRI scans is obtained by locating homologous markers on the animal and on the display using an optical position sensor. In people, frameless stereotaxy devices, such as the Brainlab and Stealth systems, use up to 9 registration points glued to the skin over the skull, as compared with the 5 registration points clustered around the front of the skull with the Brainsight system.[35, 36] The difficulties associated with gluing and maintaining registration points to the skin and the mobility of the skin in dogs precludes this possibility. As a result, the registration process involves some error, and the further away the subsequent trajectory is from the registration device located on the frontal bone, the more this error is exaggerated. Differences in skull shape and size among breeds are also likely to contribute to challenges in lead placement, because a smaller skull results in a greater distance between the frame and the target. In addition, the smaller brachycephalic skull shape offers less scope for an accurate lateral approach for lead placement. Finally, it is likely that movement of the electrodes occurred during or after stylette placement, which in some cases was >5 mm. Solutions to improve this currently are being examined, and in the future, different methods of maintaining the electrodes in place and of securing them to the calvarium will be utilized. In humans, special securing devices such as the Guardian burr hole cap are available that are placed in the burr hole and which then have a locking cap overlaid to secure the lead. It is likely that future improvements will involve a similar solution. Finally, the process of fusing the CT scans and MRI scans is imperfect and the hyperdense appearance of the electrode tips on CT scan makes the precise location of the electrode tip difficult to determine. Some studies have utilized the superior soft tissue imaging characteristics of MRI to image DBS electrodes during or after placement to establish the exact location of contact points.[37-39] However, the presence of ferromagnetic material in the electrodes raises the concern of thermal injuries when MRI is performed, as well as generating a signal void as a result of local distortion of the magnetic field. For these reasons, CT generally is regarded as a safer modality for postoperative imaging. The exaggerated appearance of the electrodes on CT has been characterized, with a 1.27-mm lead having been shown to generate an artifact of up to 3.4 mm. For this reason, fusing postoperative CT images with preoperative MRI is commonly used to combine the anatomical accuracy of MRI images with the safety of CT, and studies have shown the registration of the 2 involves a discrepancy of approximately 1.5 mm. Therefore, a substantial proportion of the placement error in this study is likely to have been arisen from the fusion of CT and MRI images. In the future, the use of improved electrodes may allow MRI scanning to replace CT scanning as the modality of choice for postoperative assessment. This would eliminate the error associated with fusion of the 2 images while simultaneously evaluating small amounts of hemorrhage or edema associated with the procedure.
The rationale for stimulation of the ANT as part of seizure therapy comes from the discovery that the ANT plays a crucial role in the propagation of limbic epilepsy as part of the Papez circuit (Fig 6). The Papez circuit was described in 1937 by James Papez as a circuit linking hippocampal output via the fornix and mammillary nucleus in the posterior hypothalamus to the ANT. The ANT projects to the cingulum bundle deep to the cingulate gyrus, which travels around the wall of the lateral ventricle to the parahippocampal cortex. This structure then completes the circuit by returning to the hippocampus. Atrophy or sclerosis of any structures within the circuit has been known to cause epilepsy, as occurs with mesial temporal sclerosis. Although mesial temporal sclerosis has yet to be described in dogs, the importance of the Papez circuit in the propagation of seizure activity suggests that it remains an important target for the treatment of epilepsy in dogs. Two other circuits are believed to be involved in the propagation of seizures: the corticothalamic circuit and the mammillary circuit. The corticothalamic circuit involves propagation of seizure activity from the motor cortex to the caudate, putamen, the globus pallidus and the thalamus, and unilateral, repeated stimulation of this circuit can result in continual partial seizures. The mammillary circuit also involves the ANT, from which seizure activity is propagated to the mammillary bodies to the brainstem.
Targeting of these circuits in the treatment of epilepsy is based on the theory that stimulation delivered to these pathways can prevent the generalization and propagation of seizures, although the exact mechanisms remain unclear. It is known that the unilateral stimulation of the anterior thalamus in amygdala-kindled seizures in rats decreases the incidence of generalized seizures. EEG and concurrent thalamic recording after surgery in human patients are consistent with a recruitment pattern that correlates with clinical improvement. Several studies of ANT stimulation have been performed, the largest of which is a multicenter, double-blind, randomized clinical trial entitled Stimulation of the Anterior Nucleus of the Thalamus for Epilepsy (SANTE). The treatment group of the 110 patients in the SANTE trial showed a median seizure frequency reduction of 29% at 3 months, 41% at 1 year, 56% at 2 years, and 68% by the end of 3 years as compared with baseline. Other, smaller, nonrandomized trials of ANT stimulation also have been performed with 4–6 patients per trial, with seizure frequency reduction of 14–75%.[9, 11, 16-19]
Perhaps the most exciting prospect raised by this trial is the possibility of closed loop recording and stimulation for the treatment of epilepsy. Long-term stimulation has been performed in human patients for a number of years, and long-term EEG recording recently has become available, but the device described here is one of the first that combines both capabilities simultaneously. This, in turn, raises the prospect of intelligent stimulation that can be instigated in response to seizure activity as detected in the EEG recording. Such devices would represent an important advance in brain stimulation for the treatment of epilepsy, and potentially could revolutionize the lives of human and canine patients with epilepsy.