Abnormalities of voxel-based whole-brain functional connectivity patterns predict the progression of hepatic encephalopathy.

Resting state functional magnetic resonance imaging (fMRI) is an important tool for understanding the functional reorganization of the brain in cirrhotic patients. Previous studies revealed that functional integration failure were observed in the whole brain. However, the whole-brain functional connectivity analysis methods used in these studies have the limitation that the result relied on a priori definition of network nodes. Moreover, the utility of resting state functional connectivity in the diagnosis and prediction of hepatic encephalopathy (HE) is not well examined. In this study, we recruited 87 subjects consisting of patients without HE, with HE, and healthy controls. We employed a voxel-based, unbiased functional connectivity analysis and the functional connectivity density (FCD) metric to precisely study abnormalities in the intrinsic functional connectivity patterns of cirrhotic patients. FCD analyses showed that hub regions in the brain were less topologically important in cirrhotic patients, whereas non-hub regions became topologically important in the disease state. This trend was more apparent with the progression of cirrhosis severity. Most FCD abnormalities were associated with deficits in psychomotor function, executive control, or visual-spatial abilities (p < 0.05, AlphaSim corrected). FCD alterations in the left inferior parietal lobe and the right hippocampal gyrus/parahippocampal gyrus were significantly correlated with cognitive ability and blood ammonia level (p < 0.05, AlphaSim corrected). A pattern classification analysis indicated that whole-brain FCD differences distinguished cirrhotic patients from healthy controls and predicted disease severity with high accuracies. These findings suggest that voxel-based FCD analysis may be clinically important for the diagnosis and prediction of HE.