Poster No:
11
Submission Type:
Abstract Submission
Authors:
Yejin Ann1, Young-Min Sohn2, Seok-Jun Hong1
Institutions:
1Center for Neuroscience Imaging Research, Institute for Basic Science, Suwon, South Korea, 2Samsung Medical Center, Seoul, South Korea
First Author:
Yejin Ann
Center for Neuroscience Imaging Research, Institute for Basic Science
Suwon, South Korea
Co-Author(s):
Seok-Jun Hong
Center for Neuroscience Imaging Research, Institute for Basic Science
Suwon, South Korea
Introduction:
The anterior and centromedian thalamic nuclei (ATN/CM) are the two commonly target structures in deep brain stimulation (DBS) for epilepsy. To optimize this surgical planning, several MRI-based thalamic atlases have been introduced, but they were often unsuccessful for precise stimulation targeting[1]. While the efficacy of DBS has been assessed in many previous studies, a comprehensive investigation based on fully quantitative approaches is still lacking, especially in elucidating the relationship between varying rates of seizure reduction (SR) and DBS patterns or relating their pathological effects to an underlying anatomy. Here, we addressed these issues by 1) validating existing MRI thalamic atlases based on comparisons with histological data (to select the best reference atlas), 2) mapping the volume of DBS-activated tissue (VTA) on the thalamic nuclei of selected atlases (to assess how important precisely stimulating the targeted nucleus is to achieve SR), and 3) comparing the morphology of the thalamus between the control and patient groups.
Methods:
Through the literature survey, we have found 5 publicly released thalamic atlases (Allen[2], Freesurfer[3], Ilinsky[4], Distal[1] and Thomas[5]). We quantitatively validated these atlases by comparing to the thalamic nuclei from postmortem brain data6. For this, we manually segmented ATN and CM in histology (Fig1A) and calculated a dice index (DI) between our segmentations and the borders of nuclei in MRI atlases to choose the best reference atlas.
Analysis 1: We analyzed structural MRI of in-total 50 patients (31 focal [ATN] and 19 generalized epilepsy [CM]). We computed DI between the thalamic nucleus of the selected atlas and VTA at each subject. A receiver operating characteristic (ROC) curve was then plotted to evaluate the accuracy of classifying responder and non-responder groups (RG/NRG) across multiple DI values.
Analysis 2: Next, we conducted a shape analysis to investigate the pathological effect on the thalamic morphology across the two patient groups (ATN/CM) compared to the control using a SPHARM-MAT algorithm[7]. After modeling the thalamic structure's surface, two sample t-test was employed to quantify their morphological differences. False discovery rate (FDR) corrected for the Type-I error.
Results:
Our atlas validation demonstrated that ATN and CM boundaries in histology match to different MRI atlases, respectively. Indeed, the ATN showed the largest overlap with Thomas5, while in CM, the Freesurfer3 showed the best fit (Fig1B). The ROC (based on the DI between the nucleus atlas and VTA) for the accuracy of distinguishing RG and NRG revealed 0.75 of AUC for ATN and 0.489 for CM (Fig1C). It showed that the higher the overlap is with the atlas, the better seizure outcome is, in only ATN.
Similarly, the thalamic shape analysis also showed the significant group changes only in the ATN-targeted patients compared to controls (Fig2) but not in CM patients. In the ATN group, overall trends of findings showed an inward deformation across multiple nuclei (Fig2). Notably, this atrophy was specifically noted in the right ATN. In the CM-targeted group, however, there were no group differences between patients and controls.

·Figure 1. A) Histology images (left), Overlay masked thalamic nuclei (right). B) Dice index between masked nucleus and atlas. C) ROC curve based on the dice index (left – ATN, right – CM).

·Figure 2. Results of the shape analysis between the control and ATN patients (Left thalamus (top), Right thalamus (bottom), T-value map (left), Significant P-value map (right).
Conclusions:
Our study validated the existing MRI-based thalamic atlases based on postmortem histology. The series of analyses demonstrated that the efficacy of DBS may be different, depending on the targeted thalamic nucleus. Indeed, while the DBS targeting ATN seems moderately efficacious in terms of seizure reduction, the CM-DBS cases failed to show comparable therapeutic effects. In the shape analysis, the individual heterogeneity in the CM group may influence the observed negative finding. Further research is required for the individual-level phenotyping and also the whole-brain analysis that targets beyond the thalamic nuclei (e.g., whole-brain connectivity) to comprehensively investigate potential factors of brain substrates that can affect the DBS outcome.
Brain Stimulation:
Deep Brain Stimulation 1
Disorders of the Nervous System:
Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s) 2
Keywords:
Epilepsy
MRI
Thalamus
Other - Shape Analysis
1|2Indicates the priority used for review
Provide references using author date format
1. Ewert, S., Plettig, P., Li, N., Chakravarty, M. M., Collins, D. L., Herrington, T. M., ... & Horn, A. (2018). Toward defining deep brain stimulation targets in MNI space: a subcortical atlas based on multimodal MRI, histology and structural connectivity. Neuroimage, 170, 271-282.
2. Ding, S. L., Royall, J. J., Sunkin, S. M., Ng, L., Facer, B. A., Lesnar, P., ... & Lein, E. S. (2016). Comprehensive cellular‐resolution atlas of the adult human brain. Journal of comparative neurology, 524(16), 3127-3481.
3. Iglesias, J. E., Insausti, R., Lerma-Usabiaga, G., Bocchetta, M., Van Leemput, K., Greve, D. N., ... & Alzheimer's Disease Neuroimaging Initiative. (2018). A probabilistic atlas of the human thalamic nuclei combining ex vivo MRI and histology. Neuroimage, 183, 314-326.
4. Ilinsky, I., Horn, A., Paul-Gilloteaux, P., Gressens, P., Verney, C., & Kultas-Ilinsky, K. (2018). Human motor thalamus reconstructed in 3D from continuous sagittal sections with identified subcortical afferent territories. eneuro, 5(3).
5. Su, J. H., Thomas, F. T., Kasoff, W. S., Tourdias, T., Choi, E. Y., Rutt, B. K., & Saranathan, M. (2019). Thalamus Optimized Multi Atlas Segmentation (THOMAS): fast, fully automated segmentation of thalamic nuclei from structural MRI. Neuroimage, 194, 272-282.
6. Alkemade, A., Bazin, P. L., Balesar, R., Pine, K., Kirilina, E., Möller, H. E., ... & Forstmann, B. U. (2022). A unified 3D map of microscopic architecture and MRI of the human brain. Science advances, 8(17), eabj7892.
7. Shen, L. (2010). SPHARM-MAT v1. 0.0 documentation.