Aberrant brain intra- and internetwork functional connectivity in children with Prader-Willi syndrome

Abstract

Purpose: Prader-Willi syndrome (PWS) suffers from brain functional reorganization and developmental delays during childhood, but the underlying neurodevelopmental mechanism is unclear. This paper aims to investigate the intra- and internetwork functional connectivity (FC) changes, and their relationships with developmental delays in PWS children.

Methods: Resting-state functional magnetic resonance imaging datasets of PWS children and healthy controls (HCs) were acquired. Independent component analysis was used to acquire core resting-state networks (RSNs). The intra- and internetwork FC patterns were then investigated.

Results: In terms of intranetwork FC, children with PWS had lower FC in the dorsal attention network, the auditory network, the medial visual network (VN) and the sensorimotor network (SMN) than HCs (FWE-corrected, p < 0.05). In terms of internetwork FC, PWS children had decreased FC between the following pairs of regions: posterior default mode network (DMN) and anterior DMN; posterior DMN and SMN; SMN and posterior VN and salience network and medial VN (FDR-corrected, p < 0.05). Partial correlation analyses revealed that the intranetwork FC patterns were positively correlated with developmental quotients in PWS children, while the internetwork FC patterns were completely opposite (p < 0.05). Intranetwork FC patterns showed an area under the receiver operating characteristic curve of 0.947, with a sensitivity of 96.15% and a specificity of 81.25% for differentiating between PWS and HCs.

Conclusion: Impaired intra- and internetwork FC patterns in PWS children are associated with developmental delays, which may result from neural pathway dysfunctions. Intranetwork FC reorganization patterns can discriminate PWS children from HCs.

Registration number on the chinese clinical trail registry: ChiCTR2100046551.

Keywords: Functional connectivity; Independent component analysis; Magnetic resonance imaging; Prader-Willi syndrome.

Fig. 1

The 12 core RSNs selected by the independent component analysis. aDMN, anterior default mode network; AN, auditory network; DAN, dorsal attention network; lFPN, left frontoparietal network; lVN, lateral visual network; mVN, medial visual network; pDMN, posterior default mode network; pVN, posterior visual network; rFPN, right frontoparietal network; RSNs, resting-state networks; SAN, salience network; SMN, sensorimotor network; VAN, ventral attention network

Fig. 2

PWS group showed a decrease in intranetwork functional connectivity compared to HCs by two-sample t-test (FWE-corrected p < 0.05 and a cluster size > 10 voxels). A Decreased core regions in DAN. B Decreased core regions in AN. C Decreased core region in mVN; and D Decreased core region in SMN. AN, auditory network; DAN, dorsal attention network; FWE, the familywise error; HCs, healthy controls; L, left; mVN, medial visual network; PWS, Prader-Willi syndrome; R, right; SMN, sensorimotor network

Fig. 3

PWS group showed a decrease in internetwork functional connectivity compared to HCs by two-sample t-test (FDR-corrected p < 0.05). A Inter-network functional connectivity matrix; B and C Significant differences in intra-network functional connectivity between PWS compared with HCs. aDMN, anterior default mode network; AN, auditory network; DAN, dorsal attention network; FDR, the false discovery rate; HCs, healthy controls; lFPN, left frontoparietal network; lVN, lateral visual network; mVN, medial visual network; pDMN, posterior default mode network; pVN, posterior visual network; rFPN, right frontoparietal network; PWS, Prader-Willi syndrome; SAN, salience network; SMN, sensorimotor network; VAN, ventral attention network

Fig. 4

Significant correlations between intranetwork and internetwork functional connectivity and developmental variables in children with PWS. A Pearson correlation shows a increase between intra-network FC (right IPL of DAN and right INS of AN) and developmental quotients. B Pearson correlation shows a decrease between inter-network FC (pDMN-aDMN) and developmental quotients. aDMN, anterior default mode network; FC, functional connectivity; INS, insula; IPL, inferior parietal lobule; pDMN, posterior default mode network

Fig. 5

ROC curve for intranetwork FC alterations as indicators. AUC, area under the curve; CI, confidence interval; HES, heschil gyrus; IFGoperc, opercular part of inferior frontal gyrus; INS, insula; IPL, inferior parietal lobule; LING, lingual gyrus; PreCG, precentral gyrus; ROC, receiver operating characteristics; SE, standard error