Abstracts

Rationale: Modulation of the anterior nucleus of thalamus (ANT) has emerged as a potential therapy option in multifocal epilepsies where widespread network dysfunction precludes surgical resection. Critical to the patient selection and optimization of this circuit-based neuromodulation therapy is the knowledge of recruitment of cortical circuits (target engagement) at different stimulation parameters(1). Using low and high-frequency stimulation trials, we mapped the cortical responses with high precision using stereo-EEG (sEEG). We hypothesize that cortical de/activation will vary depending on low and high-frequency stimulation. Methods: An IRB approved prospective study, where consenting adults (N=7) with suspected TLE undergoing SEEG (PMT electrodes) were recruited for additional research-recording from the ANT. Post implant CT was merged with pre-operative MRI for precise electrode localization. Trials with 10/50/100Hz biphasic stimulation of ANT were performed using Nicolet cortical stimulator (pulsewidth: 300/400uS, train-duration: 5s, amplitude: 3mA). For seizure detection, EEG channels with epileptogencity index (EI) > 0.2 were subjected to line length (LL) detection and the channels that survived thresholding were parsed as seizure network (SN). HFA was estimated (50-150 Hz, increment of 10Hz) on bipolar derivation following 60Hz artifact removal of the short inter-stimulus intervals and the final HFA estimated was normalized to the baseline HFA power to exclude the effect of non-biological artifacts (2). Increase in broadband gamma power (high frequency activity 50-150 Hz) in field potentials index local cortical activation.  Results: Electrical stimulation of ANT with low and high frequency produced opposite effects in the cortex. Low frequency stimulation resulted in cortical activation where as high-frequency stimulation resulted in deactivation. With both stimulation parameters, there was no temporal change in the evoked HFA response in the cortex between early stimulation and late stimulation (Fig1).  When the gamma band was dissected into smaller frequency bands of 10Hz increments, we noted that low frequency stimulation produced maximal HFA activation between 50-60Hz (Fig2) while high frequency stimulation was not significantly different. Lastly, both low and high frequency stimulations showed no difference in HFA coactivation-deactivation between SN and other cortical electrodes. Conclusions: Stimulation of ANT at different frequencies produces distinct patterns of cortical coactivation and deactivation. The lack of variation in response between seizure network (mesial temporal regions) and other cortical channels indicates that the recruitment of cortical circuits with ANT stimulation is driven by structural connectivity. This knowledge will aid in optimizing patient selection, as seizure network is likely to be modulated if it is within connectivity of ANT network.  Funding: No funding