Abstracts

Rationale: The anterior and centromedian thalamic nuclei (ATN, CM) are commonly targeted regions in deep brain stimulation (DBS) for epilepsy. The motivation for this target strategy is related to the circuits they are involved in. For example, the ATN, which plays a vital role in seizure propagation of focal epilepsy (especially with a mesial temporal focus), is a node of the Papez circuit linking overall limbic structures. The CM, often associated with generalized epilepsy, is also the core of the cerebello-thalamic-cortical (CTC) circuit characterized by diffuse cortical projections. While the efficacy of targeting these nuclei has been previously established, the biological substrates that may affect its outcomes remain poorly understood. Here, we addressed this question by assessing i) the relation between DBS target accuracy and seizure reduction (SR) and ii) the pathological effects on thalamic nuclei and their white matter bundles based on the morphological analysis and track density imaging (TDI).


Methods: Our study consists of three analyses: 1) To examine the effect of DBS target accuracy on SR, we analyzed structural MRIs of 53 patients (34 focal [ATN], 19 generalized epilepsy [CM]) and computed the dice index (DI) between the thalamic nucleus and volume of tissue activated (VTA). A receiver operating characteristic (ROC) curve was plotted to evaluate the classification accuracy for responder and non-responder groups (RG/NRG) (Fig 1A). 2) Next, we delineated the Papez and CTC circuits in 27 patients (14 ATN, 13 CM) by diffusion tractography (Fig 1B) and compared their track density between the groups using the voxel-wise TDI analyses. 3) We also investigated the morphological changes of the thalamus across the patient groups using SPHARM-MAT, an established shape analytical tool.


Results: The ROC curve for classifying RG/NRG revealed an AUC of 0.78 for ATN and 0.417 for CM (Fig 1A). It suggests that in the ATN (but not CM) group, the higher the overlap between VTA and the thalamic nucleus, the better the SR is. This ATN-specific effect is also found in the TDI analysis, where it demonstrated excessively increased tract density of the Papez circuit in the ATN group (Fig 2A) but no difference in the CTC circuit of the CM group (although there was a tendency for diffuse increases across their tracts; Fig 2B). The thalamic shape analysis also corroborated this pattern: the ATN group showed a severe inward deformation in bilateral thalami, whereas the CM group presented much lesser morphological anomalies only in the right thalamus (Fig 2C,D).


Conclusions: Our study demonstrated that DBS efficacy varies according to the target accuracy and pathological effects. While targeting ATN seems moderately efficacious, the CM stimulation failed to show comparable therapeutic effects. This may relate to the significantly increased tract density and severe atrophic thalamic nucleus in the ATN group, motivating why the ATN indeed should be targeted in this group. In contrast, the mild changes in the CM group suggest investigating more distributed areas connected to this nucleus, beyond the CM itself and its CTC circuit.


Funding: NA