Altered topological organization of resting-state functional networks in children with infantile spasms

Abstract

Covering neuroimaging evidence has demonstrated that epileptic symptoms are associated with the disrupted topological architecture of the brain network. Infantile spasms (IS) as an age-specific epileptic encephalopathy also showed abnormal structural or functional connectivity in specific brain regions or specific networks. However, little is known about the topological alterations of whole-brain functional networks in patients with IS. To fill this gap, we used the graph theoretical analysis to investigate the topological properties (whole-brain small-world property and modular interaction) in 17 patients with IS and 34 age- and gender-matched healthy controls. The functional networks in both groups showed efficient small-world architecture over the sparsity range from 0.05 to 0.4. While patients with IS showed abnormal global properties characterized by significantly decreased normalized clustering coefficient, normalized path length, small-worldness, local efficiency, and significantly increased global efficiency, implying a shift toward a randomized network. Modular analysis revealed decreased intra-modular connectivity within the default mode network (DMN) and fronto-parietal network but increased inter-modular connectivity between the cingulo-opercular network and occipital network. Moreover, the decreased intra-modular connectivity in DMN was significantly negatively correlated with seizure frequency. The inter-modular connectivity between the cingulo-opercular and occipital network also showed a significant correlation with epilepsy frequency. Together, the current study revealed the disrupted topological organization of the whole-brain functional network, which greatly advances our understanding of neuronal architecture in IS and may contribute to predict the prognosis of IS as disease biomarkers.

Keywords: functional brain network; graph theory; infantile spasm; modularity; small-world.

FIGURE 1

Group differences in the small-world topological metrics between the IS and control groups. In the range of sparsity (0.05–0.4), the topologies of (A) λ, (B) γ, and (C) σ in both groups exhibited small-world property. Bar charts plot the significant differences of the AUC of λ, γ, and σ between the IS children and controls (p < 0.05, FDR corrected). ***Significant difference between the two groups. P, patient group; HC, healthy control group.

FIGURE 2

Group differences in the network efficiency of (A) E_glob and (B) Eloc between the IS and control groups in the range of sparsity (0.05–0.4). Bar charts plot the significant differences of the AUC of E_glob and E_loc between the IS children and controls (p < 0.05, FDR corrected). ***Significant difference between the two groups. P, patient group; HC, healthy control group.

FIGURE 3

Comparison results of the AUC of the averaged functional connectivity strength within and between the modules (p < 0.05, FDR corrected). *The connectivity can withstand the FDR correction.

FIGURE 4

The correlation (A) between the intra-modular connections within DMN and the epilepsy frequency, (B) between the inter-modular connections and the epilepsy frequency.