Hepatic encephalopathy (HE) is a spectrum of neuropsychiatric abnormalities associated with liver dysfunction, which is a significant complication in most patients with liver cirrhosis. It encompasses a wide range of neurological symptoms that exhibit varying degrees of severity and impacts a range of functions, including psychomotor abilities, intellectual faculties, cognitive processes, emotional responses, behavioral patterns, fine motor skills, and more.

HE can be divided into five stages based on clinical symptoms or etiology, ranging from grade 0 to grade 4. The minimal hepatic encephalopathy (MHE) or grade 0 HE is located at the beginning of this spectrum and was formerly known as subclinical or latent hepatic encephalopathy. The risk of progression from grade 0 (MHE) to grade 4 is significant, with increased mortality and hospitalization rates for patients. Accurate assessment of HE in its initial stage (i.e., MHE) holds the potential to enable the implementation of suitable intervention strategies, which helps prevent the progression of HE to more severe stages. Such proactive measures could have substantial implications for enhancing the quality and efficacy of treatment, thereby contributing to improved outcomes for affected individuals.

The increasing evidence have shown that the presence of alpha oscillation disturbance in the electroencephalogram (EEG) is a hallmark of MHE (Amodio and Montagnese, 2015). In recent study, Nikhilkumar et al. explored the correlation between MHE and EEG changes, in which all cases were compared with an equal number of healthy controls (Patel et al., 2019). They found that the mean frequency of alpha waves in MHE patients is statistically significant lower than that in healthy controls. This situation is named as “slowing of alpha waves” in their work, which can be used as supportive evidence and for monitoring of MHE (Garnier et al., 2016b). Han et al. found that lower alpha relative power in patients with MHE than in patients without MHE (Han et al., 2020). In the work of Zhang et al., they also found significant decreasing spectral power densities of the alpha oscillations in the HE sub-group (Zhang et al., 2022). However, the neural correlates of such abnormal EEG phenomenon are poorly understood. The further studies on the disturbance of alpha waves, especially regarding pathological mechanisms, are significant for deep understanding the early stages of HE (i.e., MHE) and preventing its progression and deterioration.

The pathogenesis of HE remains unclear, but it is widely accepted that the accumulation of ammonia plays a pivotal role in its pathophysiology. Recent studies have shed light on the central role of astrocytes in the pathogenesis of HE, which are critical components of the blood-brain barrier as they communicate directly with neurons (Häussinger et al., 2021; Prakash and Mullen, 2010; Ciećko-Michalska et al., 2012). The alterations related to astrocytes due to HE-relevant factors impair the “astrocytic-neuronal communication” which disturbs oscillatory networks in brain as reflected by the symptoms of HE (May, 2012; Häussinger and Görg, 2019). And in work of (Butz et al., 2013), it is pointed that the emergence of disturbances of oscillatory brain activity precedes the development of HE symptoms after a cascade of complex events, where an important one is disturbances of astrocyte-neuronal communication. However, it is not yet understood how this mechanism relates to the observed disturbance of alpha waves in the early stage of HE. Therefore, the primary objective of this study is to delve into the mechanistic details associated with astrocyte-neuronal communication concerning the disruption of alpha waves (specifically, the reduction in frequency and power spectral density) observed in MHE.

Neural computational modeling is a rapidly growing research area that has proven to be effective in the biomedical field. It allows for the analysis and comprehension of observed EEG phenomena, testing of mechanistic hypotheses, and prediction of system behaviors over time and/or space (Shayegh et al., 2013; Bhattacharya et al., 2011; Li et al., 2020). The microscopic and mesoscopic modeling approaches are developed according to the level of biological organization to be modeled. The former approach mainly focuses on characterizing detailed activity for each individual neuron, while the latter approach focuses more on characterizing average activity for a population of neurons (or neural mass). Compared to microscopic models, mesoscopic models are an essential tool for comprehending the behavior of neural systems at a level of abstraction that is both biologically plausible and computationally feasible. As a result of this recognition, there has been a growing interest in studying mesoscopic models to investigate physiological and pathological phenomena. Here, the typical one is “Liley model”, which is proposed by Liley in 1997 to clarify the genesis of alpha rhythms in the mammalian brain (Liley, 1997). Following this, a series of modifications to the Liley model were proposed and applied in an attempt to explain the underlying mechanisms of various physio-pathological states. For example, Bojak et al. incorporated a slow process of synaptic resource into the Liley model and explored the emergence of spatially heterogeneous burst suppression during deep anesthesia (Bojak et al., 2015). Hartoyo et al. inferred the mechanisms underlying alpha blocking by fitting the Liley model to EEG spectra from 82 different individuals (Hartoyo et al., 2020).

In our recent work, we integrated three known mechanisms underlying acute HE (AHE) into the Liley model to construct a new computational model NCM-AHE, which is applied to explain the emergency of triphasic waves (TPWs) in AHE EEGs (Song et al., 2019). Inspired by our previous work, the NCM-AHE is taken as the fundamental model in this work to investigate the mechanistic details of alpha wave disturbance (AWD) related to astrocyte-neuronal communication in the MHE. A novel mathematical model MHE-AWD-NCM that encompasses the communication dynamics between astrocyte and neuronal populations is constructed firstly, building upon our previously established NCM-AHE model. It comprises two pathways: feedforward communication originating from the cortical neuron population (CNP) and targeting the astrocyte population (AP), as well as feedback communication from the AP back to the CNP. In addition, we introduce the concepts of peak power density and peak frequency within the alpha band as quantitative measures of AWD. These metrics enable us to investigate the impact of key model parameters on the disturbance of alpha waves in MHE. Subsequently, sensitivity analysis is performed to assess the relative significance of various parameters in influencing alpha wave disruptions in the context of MHE.