Working Memory in Patients with Varying Degree of Hepatic Encephalopathy (HE): A Pilot EEG-fNIRS Study

Abstract

It is known that patients with covert hepatic encephalopathy (CHE) exhibit working memory abnormalities, but to date there is no study comparing patients with cirrhosis with/without CHE and controls with both electrophysiological and hemodynamic data collected at the same time.Here we collected behavioral [accuracy and reaction times (RTs), electrophysiological (evoked potentials) and hemodynamic (oxygenated and deoxygenated haemoglobin) correlates of an n-back task [formed by a control (0-back) condition, a low (1-back) and a high (2-back) working memory load conditions] in patients with cirrhosis with/without CHE: (1) at baseline (n = 21, males = 15, 58±8 yrs), and by comparison with controls (n = 21, males = 15, 57±11 yrs) and (2) after a 3-month course of rifaximin (n = 18, males = 12, 61±11 yrs), and by comparison to baseline.All patients showed slower RTs (p < 0.0001) and lower P2 amplitude compared with controls (p = 0.018); moreover, patients with CHE showed reduced accuracy (p < 0.0001) compared with controls, and patients without CHE showed higher oxygenated haemoglobin in the central dorsolateral prefrontal cortex in the 2-back compared with patients with CHE. Post-rifaximin, oxygenated haemoglobin increased in the central frontopolar cortex. In addition, in patients without CHE the RTs of the 2-back became comparable to those of the 0-back and P3 showed higher amplitude.In conclusion, the presence of cirrhosis seemed to have more effects than CHE on working memory at baseline. A course of treatment with rifaximin was more beneficial to patients without CHE, who probably had more room for improvement in this complex task.

Keywords: EEG; Hepatic encephalopathy; Rifaximin; Working memory; fNIRS.