Abstract
BACKGROUND: Spastic cerebral palsy (SCP) is associated with extensive alterations in regional cortical morphology. However, the specific effects of SCP on the topological organization of morphological brain networks remain largely unknown. This study aimed to investigate these effects and explore their potential correlations with clinical manifestations in SCP children. METHODS: Structural magnetic resonance imaging and clinical data were collected from 31 children with SCP and 29 sex- and age-matched children with typical development. Single-subject morphological brain networks were constructed separately based on four different morphological indices [i.e., the cortical thickness (CT), fractal dimension, gyrification index, and sulcus depth], which were further characterized using graph-based network approaches. Permutation tests were used to examine between-group differences in regional morphology, interregional morphological connectivity (MC), and graph-based network properties. For magnetic resonance imaging (MRI)-based features showing significant between-group differences, Spearman partial correlations were used to examine their relationships with the clinical variables in the patients. RESULTS: Compared with the control group, the SCP group only showed alterations in the CT-based morphological brain networks. Specifically, the SCP group displayed an increased characteristic path length (t=3.909, P=4.0×10(-4)), which was negatively correlated with the verbal comprehension index (rho=-0.435, P=0.023), processing speed index (rho=-0.452, P=0.018), and full-scale intelligence quotient (rho=-0.471, P=0.013) of the SCP children, and positively correlated with the Gross Motor Function Classification System (rho=0.399, P=0.039) and Manual Ability Classification System (rho=0.459, P=0.016). Further, the SCP group showed decreased MC for 161 connections. These connections were mainly linked to the right area 25 (a part of the anterior cingulate cortex) at the nodal level and to regions in the default mode network at the subnetwork level. Among these, the MC between the right area 11l, part of the orbital and polar frontal cortex, and the right medial belt complex, part of the early auditory cortex, was positively correlated with the Communication Function Classification System in the SCP children (rho=0.662, P=1.7×10(-4)). These results remained unchanged after excluding preterm children from the SCP group. CONCLUSIONS: SCP is associated with abnormal morphological brain network topology, which may contribute to disturbances in motor and cognition in patients.