Synopsis

Different types of GTCS could be explained by the different brain network. Cortex-subcortex structural covariate connectivity might help us to understand the mechanism in the new insight. Thalamus was the important regions in classification the two types of GTCS.

Purpose

Generalized Epilepsy is the most serious type of central nervous system diseases. It could be separated into two types according to the seizure symptom: the Idiopathic generalized epilepsy with generalized tonic-clonic seizures (IGE-GTCS) and Secondary GTCS (S-GTCS). It is difficult to classification only according the clinical symptoms while they need different therapeutic strategy. Former studies showed that the subcortex nucleus (such as thalamus, striatum, cerebellum et al.) played important roles during the seizure generalization and propagation, and there might be different models for the two GTCS epilepsy. The abnormal cortex-subcortex connectivity might be used to provide some evidences to help us understand the mechanisms of the two GTCS types and help to classify them in future.

Brain network is the useful tool to describe the pathway of seizure propagation, and the structural covariate connectivity could depict the long-range effect of diseases. In this work, we used structural covaraite connectivity and winner-take-all(WTA) strategy to build up the cortex-subcortex network of I-GTCS and S-GTCS, and to help us find out some evidences of different seizure types.

Material and methods

One hundred and eleven IGE-GTCS patients, 111 S-GTCS patients and 111 age- and gender-match healthy controls were included in this study. The diagnosis was made by two experienced neurologists in our hospital according the ILAE 1989, Our study was approved by local ethical standards committee, and the participants or legal guardians provided written informed consent. All Participants went on the high-resolution Structural MRI scans (3.0T, Trio Tim, Siemens).

Data transformation was done by mricron (https://www.nitrc.org/projects/mricron), and the future process was done by CAT12(http://www.neuro.uni-jena.de/cat/) . The moderated gray matter volume map and total brain tissue volume was used for the later analysis.

According the Zhang et al.' work, we separate the cortex into 5 ROIs, and gained the stratum, thalamus and cerebellum from AAL template. We used our in-house software ASBC (http://www.jlradiology.com/File.do) to do the cortex-subcortex structural connectivity network and build up the sub-regions of subcortex according the WTA strategy. Then, the structural covarate networks between sub-regions of subcortex were constructed.

The 5000 times permutation method was used to find out the differences between two groups of participants.

Results

Compared to the healthy controls, the patients showed increased connectivity numbers in frontal-subcortex and motor-subcortex network, and decreased in somatosensory-subcortex connectivity. Separately, the S-GTCS showed additional decrease in partial&occipital-subcortex connectivity.

Compared to the S-GTCS, the IGE-GTCS showed increased connectivity of motor-subcortex, partial&occipital-subcortex and temporal-subcortex, while decreased ones of frontal-subcortex and somatosensory-subcortex.

There were also some differences in the comparison of network of sub-regions of subcortex. All participants showed strong connectivity intra-brain regions (stratum, thalamus and cerebellum). Compared to the healthy controls, the two patient group showed increased connectivity strengths among the intra-connectivity in brain regions and inter-connectivity between stratum and cerebellum. The S-GTCS showed wider abnormal, and more enhanced stratum-cerebellum connectivity was found.

Compared to the S-GTCS, the IGE-GTCS showed decreased covariance in intra-connectivity in thalamus and the inter-connectivity between stratum and thalamus. Meanwhile, there were enhanced covariance in intra-connectivity in stratum and inter-connectivity between thalamus and cerebellum.

Conclusion

The cortex-subcortex connectivity pattern comparison results suggest that the structural covariate connectivity could be a useful tool to separate the subcortex, and this separation would be affected by the diseases. GTCS would significantly affect the covarate change trends of cortex-subcortex.

Also, the abnormal sub-region subcortex network comparison results demonstrated that the disease widely affected the covariance in the whole subcortex. Different types of GTCS showed similar pattern of damage pattern. The difference between two patient groups mainly pointed to the thalamus, and the I-GTCS showed more heterogeneity than S-GTCS. It might be some indirect evidence for the different propagation networks of different GTCS types.

Acknowledgements

References

1. Zhang D, Snyder A Z, Fox M D, et al. Intrinsic functional relations between human cerebral cortex and thalamus[J]. Journal of neurophysiology, 2008, 100(4): 1740-1748.

2. Ji G J, Zhang Z, Xu Q, et al. Identifying corticothalamic network epicenters in patients with idiopathic generalized epilepsy[J]. American Journal of Neuroradiology, 2015.

3. Zhang Z, Liao W, Xu Q, et al. Hippocampus‐associated causal network of structural covariance measuring structural damage progression in temporal lobe epilepsy[J]. Human brain mapping, 2017, 38(2): 753-766.