Functional network connectivity imprint in febrile seizures

Abstract

Complex febrile seizures (CFS), a subset of paediatric febrile seizures (FS), have been studied for their prognosis, epileptogenic potential and neurocognitive outcome. We evaluated their functional connectivity differences with simple febrile seizures (SFS) in children with recent-onset FS. Resting-state fMRI (rs-fMRI) datasets of 24 children with recently diagnosed FS (SFS-n = 11; CFS-n = 13) were analysed. Functional connectivity (FC) was estimated using time series correlation of seed region-to-whole-brain-voxels and network topology was assessed using graph theory measures. Regional connectivity differences were correlated with clinical characteristics (FDR corrected p < 0.05). CFS patients demonstrated increased FC of the bilateral middle temporal pole (MTP), and bilateral thalami when compared to SFS. Network topology study revealed increased clustering coefficient and decreased participation coefficient in basal ganglia and thalamus suggesting an inefficient-unbalanced network topology in patients with CFS. The number of seizure recurrences negatively correlated with the integration of Left Thalamus (r = - 0.58) and FC of Left MTP to 'Right Supplementary Motor and left Precentral' gyrus (r = - 0.53). The FC of Right MTP to Left Amygdala, Putamen, Parahippocampal, and Orbital Frontal Cortex (r = 0.61) and FC of Left Thalamus to left Putamen, Pallidum, Caudate, Thalamus Hippocampus and Insula (r 0.55) showed a positive correlation to the duration of the longest seizure. The findings of the current study report altered connectivity in children with CFS proportional to the seizure recurrence and duration. Regardless of the causal/consequential nature, such observations demonstrate the imprint of these disease-defining variables of febrile seizures on the developing brain.

Figure 1

Multi-slice view of significant (FDR corrected; p < 0.0.5) seed to voxel-based functional connectivity for Complex (CFS) vs Simple febrile seizure (SFS) of (a) left middle temporal pole (b) right middle temporal pole (c) bilateral thalamus. Red to yellow colour indicates increased connectivity and blue to green colour indicates decreased connectivity in CSF in comparison with SFS. The colour bar indicates “FC t-values” of the connectivity voxels and the dark green cluster in the connectivity map shows the seed region.

Figure 2

Whole brain network topological properties average across all ROIs for (a) network segregation (b) network integration (c) small worldness (d) global efficiency for complex febrile seizure (CFS) (red line graph) and simple febrile seizure (blue line graph). The error bar in the graph represents a 95% confidence interval and the star represents a significant difference between complex and simple febrile seizure with correction for multiple comparisons; FDR < 0.05 for the number of sparsity (N = 17).

Figure 3

Glass brain view of group difference between CSF and SFS for regional network segregation and integration (a) brain regions showed higher segregation/local connectivity (i.e., nodal clustering coefficient) in CFS compared to SFS and (b) brain regions showed decreased integration (i.e., nodal participation coefficient) in CFS compared to SFS with multiple comparisons correction of FDR < 0.05 for no of ROIs (N = 90). Red to yellow color indicates increased network segregation and blue to green color indicates decreased network integration in CSF in comparison with SFS. The color bar indicates the “t-values” of statistical difference between CFS and SFS. The glass brain view were constructed using BrainNet Viewer.

Figure 4

Correlation of FC and Network topology with clinical variables of all CFS and SFS subjects. The FC effect size is the mean t-values of every voxel in the corresponding cluster.