Structural Covariance Network of Cortical Gyrification in Benign Childhood Epilepsy with Centrotemporal Spikes

Abstract

Benign childhood epilepsy with centrotemporal spikes (BECTS) is associated with cognitive and language problems. According to recent studies, disruptions in brain structure and function in children with BECTS are beyond a Rolandic focus, suggesting atypical cortical development. However, previous studies utilizing surface-based metrics (e.g., cortical gyrification) and their structural covariance networks at high resolution in children with BECTS are limited. Twenty-six children with BECTS (15 males/11 females; 10.35 ± 2.91 years) and 26 demographically matched controls (15 males/11 females; 11.35 ± 2.51 years) were included in this study and subjected to high-resolution structural brain MRI scans. The gyrification index was calculated, and structural brain networks were reconstructed based on the covariance of the cortical folding. In the BECTS group, significantly increased gyrification was observed in the bilateral Sylvain fissures and the left pars triangularis, temporal, rostral middle frontal, lateral orbitofrontal, and supramarginal areas (cluster-corrected p < 0.05). Global brain network measures were not significantly different between the groups; however, the nodal alterations were most pronounced in the insular, frontal, temporal, and occipital lobes (FDR corrected, p < 0.05). In children with BECTS, brain hubs increased in number and tended to shift to sensorimotor and temporal areas. Furthermore, we observed significantly positive relationships between the gyrification index and age (vertex p < 0.001, cluster-level correction) as well as duration of epilepsy (vertex p < 0.001, cluster-level correction). Our results suggest that BECTS may be a condition that features abnormal over-folding of the Sylvian fissures and uncoordinated development of structural wiring, disrupted nodal profiles of centrality, and shifted hub distribution, which potentially represents a neuroanatomical hallmark of BECTS in the developing brain.

Keywords: MRI; Rolandic epilepsy; connectome; cortical development; hub.

Figure 1

Significant cortical gyrification clusters projected onto the inflated surfaces of the bilateral hemispheres. The significant clusters show increased gyrification in the BECTS group compared with their peer controls and surviving cluster correction (p < 0.05). Detailed information for these clusters is shown in Table 2.

Figure 2

Within-group global network measures, between-group differences in these measures across a range of network densities, and area under the curve (AUC) results. Normalized clustering (A), normalized path length (B), small-world index (C), global efficiency (D), local efficiency (E), and mean node betweenness (F) of the BECTS and healthy control (HC) networks. The red inverted triangle indicates the difference between the two groups.

Figure 3

Between-group differences of regional measures (normalized regional degree, betweenness, and clustering) across a range of network densities [i.e., area under the curve (AUC) results]. The red inverted triangle indicates the difference between the two groups. All regions survived following FDR correction (p < 0.05).

Figure 4

Network hubs. The red color indicates hubs specific to healthy controls (HCs), the green color highlights hubs specific to children with BECTS, and the yellow color represents hubs that are common in both groups.

Figure 5

Random and targeted attack analysis. Alterations in the size of the largest preserving component of the network are shown as a function of a fraction of randomly (left panel) and targetedly (right panel) removed nodes. In many proportions of the deleted nodes and the area under the curve (AUC) results, there was no significant difference between the two groups (p > 0.05) of the resilience of the networks to both targeted and random attacks. In general, network of BECTS showed less robust to the targeted attack as compared with the controls, but the statistical significance only appears at a few fractions of deleted nodes (p < 0.05, the red star indicates a significant difference across different network densities between the two groups).

Figure 6

Relationship between cortical gyrification and chronological age in children with BECTS. In children with BECTS, we did not observe significant age effects on the foci of Sylvian fissures, but we did detect frontal association cortices with significant positive relationships between cortical gyrification and age (vertex p < 0.001, cluster-level correction).

Figure 7

Relationship between cortical gyrification and chronological age in typical developmental controls. Similar to children with BECTS, we also observed significant positive relationships between cortical gyrification and age (vertex p < 0.001, cluster-level correction).

Figure 8

Correlation between cortical gyrification and duration (months) of epilepsy in children with BECTS. We observed a positive correlation between cortical gyrification in children with BECTS and the duration of epilepsy in the frontal cortices (vertex p < 0.001, cluster-level correction).