Large-Scale Neuronal Network Dysfunction in Diabetic Retinopathy

• Published in: Journal of neural transplantation & plasticity 2020, Vol.2020 (2020), p.1-13
• Main Authors: Deng, Qin-Qin, Dan, Han-Dong, Qi, Chen-Xing, Tong, Yan, Huang, Xin, Shen, Yin
• Format: Article
• Language: English
• Subjects: Adult; Brain; Brain - physiopathology; Brain Mapping; Cognitive ability; Computational linguistics; Diabetes; Diabetic neuropathy; Diabetic retinopathy; Diabetic Retinopathy - physiopathology; Female; Glaucoma; Glucose; Glycosylated hemoglobin; Humans; Image Processing, Computer-Assisted; Language processing; Magnetic Resonance Imaging; Male; Medical imaging; Medical research; Medicine, Experimental; Middle Aged; Myopia; Natural language interfaces; Neural Pathways - physiopathology; Neurons; Neurophysiology; Plasma; Principal components analysis; Retinal detachment; ROC Curve; Studies; Type 2 diabetes
• ISSN: 2090-5904; 0792-8483; 1687-5443
• DOI: 10.1155/2020/6872508

Diabetic retinopathy (DR) patients are at an increased risk of cognitive decline and dementia. There is accumulating evidence that specific functional and structural architecture changes in the brain are related to cognitive impairment in DR patients. However, little is known regarding whether the functional architecture of resting-state networks (RSNs) changes in DR patients. The purpose of this study was to investigate the intranetwork functional connectivity (FC) and functional network connectivity (FNC) of RSN changes in DR patients using independent component analysis (ICA). Thirty-four DR patients (18 men and 16 women; mean age, 53.53±8.67 years) and 38 nondiabetic healthy controls (HCs) (15 men and 23 women; mean age, 48.63±11.83 years), closely matched for age, sex, and education, underwent resting-state magnetic resonance imaging scans. ICA was applied to extract the nine RSNs. Then, two-sample t-tests were conducted to investigate different intranetwork FCs within nine RSNs between the two groups. The FNC toolbox was used to assess interactions among RSNs. Pearson correlation analysis was conducted to explore the relationship between intranetwork FCs and clinical variables in the DR group. A receiver operating characteristic (ROC) curve was conducted to assess the ability of the intranetwork FCs of RSNs in discriminating between the two groups. Compared to the HC group, DR patients showed significant decreased intranetwork FCs within the basal ganglia network (BGN), visual network (VN), ventral default mode network (vDMN), right executive control network (rECN), salience network (SN), left executive control network (lECN), auditory network (AN), and dorsal default mode network (dDMN). In addition, FNC analysis showed increased VN-BGN, VN-vDMN, VN-dDMN, vDMN-lECN, SN-BGN, lECN-dDMN, and AN-BGN FNCs in the DR group, relative to the HC group. Furthermore, altered intranetwork FCs of RSNs were significantly correlated with the glycosylated hemoglobin (HbA1c) level in DR patients. A ROC curve showed that these specific intranetwork FCs of RSNs discriminated between the two groups with a high degree of sensitivity and specificity. Our study highlighted that DR patients had widespread deficits in both low-level perceptual and higher-order cognitive networks. Our results offer important insights into the neural mechanisms of visual loss and cognitive decline in DR patients.