Abstracts

Rationale: Deep brain stimulation (DBS) is a rapidly expanding therapy for drug-resistant epilepsies, including Lennox-Gastaut syndrome (LGS). A common DBS target for LGS is the thalamic centromedian nucleus (CM). However, the CM covers a relatively large area of the thalamus and includes multiple sub-regions with distinct functions and connectivity. We aimed to determine the therapeutically optimal CM sub-region and connectivity profile for DBS in LGS.

Methods: 20 patients were recruited for the Electrical Stimulation of Thalamus for Epilepsy of Lennox-Gastaut Phenotype (ESTEL) study, a randomized controlled trial of CM-DBS for LGS. Patients underwent bilateral CM-DBS implantation (Medtronic 3389 leads; 4 contacts per lead), with 19 randomized. Clinical outcome was evaluated as % seizure reduction (in diaries) after 3 months of DBS relative to a 3-month pre-implantation baseline. Lead positions were determined by aligning post-surgical CT with pre-surgical MRI. Volumes of tissue activation (VTAs; 1 per hemisphere), representing the location and influence of DBS on tissue surrounding the lead contacts selected for stimulation (Fig. 1A), were estimated using a finite element method in Lead-DBS software (Neuroimage 2019;184:293-316). VTAs were projected into a common template space and weighted by patients’ magnitude of seizure reduction. Brain connectivity associated with clinical improvement was assessed by: (i) determining white matter fiber tracts involving patients’ VTAs, using normative diffusion MRI data; and (ii) performing a ‘discriminative fiber tract analysis’ in Lead-DBS, whereby each tract is assigned a t-score reflecting the degree to which its VTA involvement yielded higher/lower seizure reduction.

Results: Seizure reduction ranged from +4% to -89%. Greater seizure reduction was found in VTAs residing in the anterior/inferolateral CM, bordering ventrolateral and ventral posteromedial thalamic nuclei; in contrast, seizure reduction was lower in VTAs located more posteriorly and superomedially (Fig. 1B). Fiber tracts associated with greater seizure reduction were concentrated in known projection areas of the reticular activating system (RAS), including connections between thalamus, brainstem, and frontal cortex (Fig. 2A). Areas showing the strongest positive relationship between connectivity strength (i.e., number of VTA-connected tracts) and seizure reduction were in posterior frontal cortex, including premotor regions of superior and middle frontal gyri (Fig. 2B).

Conclusions: We define a functional ‘sweet-spot’ for CM-DBS in LGS, located in the anterior/inferolateral CM (histologically termed the parvocellular CM). Stimulation of fiber tracts within the RAS is associated with seizure reduction, and may contribute to subjective reports in the literature of ‘increased alertness’ during CM-DBS, which was also common in the ESTEL patients. These findings have important implications for DBS in LGS, both for planning future trials and for optimizing stimulation settings (e.g., choice of contacts) in patients with existing implanted devices.

Funding: Please list any funding that was received in support of this abstract.: National Health and Medical Research Council (NHMRC) Project Grant #1108881; Lennox-Gastaut syndrome (LGS) Foundation Post-doctoral Fellowship.