• Gaoyan Zhang, Yue Cheng, Wen Shen, Baolin Liu, Lixiang Huang, and Shuangshuang Xie.
• Tianjin Key Laboratory of Cognitive Computing and Application, School of Computer Science and Technology, Tianjin University, Tianjin 300350, China.
• Korean J Radiol. 2018 May 1; 19 (3): 452-462.

ObjectiveTo investigate brain regional homogeneity (ReHo) changes of multiple sub-frequency bands in cirrhotic patients with or without hepatic encephalopathy using resting-state functional MRI.Materials And MethodsThis study recruited 46 cirrhotic patients without clinical hepatic encephalopathy (noHE), 38 cirrhotic patients with clinical hepatic encephalopathy (HE), and 37 healthy volunteers. ReHo differences were analyzed in slow-5 (0.010-0.027 Hz), slow-4 (0.027-0.073 Hz), and slow-3 (0.073-0.198 Hz) bands. Routine analysis of (0.010-0.080 Hz) band was used as a benchmark. Associations of abnormal ReHo values in each frequency band with neuropsychological scores and blood ammonia level were analyzed. Pattern classification analyses were conducted to determine whether ReHo differences in each band could differentiate the three groups of subjects (patients with or without hepatic encephalopathy and healthy controls).ResultsCompared to routine analysis, more differences between HE and noHE were observed in slow-5 and slow-4 bands (p < 0.005, cluster > 12, overall corrected p < 0.05). Sub-frequency band analysis also showed that ReHo abnormalities were frequency-dependent (overall corrected p < 0.05). In addition, ReHo abnormalities in each sub-band were correlated with blood ammonia level and neuropsychological scores, especially in the left inferior parietal lobe (overall corrected p < 0.05 for all frequency bands). Pattern classification analysis demonstrated that ReHo differences in lower slow-5 and slow-4 bands (both p < 0.05) and higher slow-3 band could differentiate the three groups (p < 0.05). Compared to routine analysis, ReHo features in slow-4 band obtained better classification accuracy (89%).ConclusionCirrhotic patients showed frequency-dependent changes in ReHo. Sub-frequency band analysis is important for understanding HE and clinical monitoring.

22 keywords...

hide…