A Longitudinal Study of Neurocognitive Biomarkers in Predicting Drinking Patterns Over Time - an EEG/ERP study

Alcohol consumption is a common practice across the world. In the United States, approximately 65% - 88% of the population has reported drinking at least once in their lifetime (Delker et al., 2016). In contrast, the lifetime prevalence of alcohol use disorder (AUD) has been estimated at around 29.1% (Grant et al., 2015), and the one-year incidence rate of alcohol abuse at around 1.02% (Delker et al., 2016). Such drastic discrepancies highlight the need for effective tools that help identify who may be at greater risk of developing AUD over time. Much research in this area has been focused on genetics. Although research synthesis efforts suggest that AUD is 50% heritable (Verhulst et al., 2015), identifying specific genetic variants to predict AUD risk has proven challenging and yielded mixed results (Tawa et al., 2016). Decades of research on the genetic basis of AUD and their mixed findings also highlight the complexity and heterogeneity of the etiological trajectory of AUD. As a result, researchers have begun searching for AUD endophenotypes indexed by neurocognitive biomarkers that could potentially provide deeper insights into the etiological development and prognosis of such psychopathology (Iacono, 2018). Specifically, recent research suggests that neurophysiology markers obtained through electroencephalogram (EEG) or event-related potential (ERP) measures are promising candidates as endophenotypes for AUD (Iacono, 2018; Kinreich et al., 2021; Salvatore et al., 2015). Despite preliminary evidence (Kinreich et al., 2021), research is still scarce in understanding the longitudinal relationship between biomarkers generated from EEG/ERP measures and drinking outcomes over time. The P3 is an extensively researched ERP component that can be reliably elicited during the "Oddball" paradigm, in which participants respond to targets that are subjectively deemed as less likely and relevant to the task (Donchin & Coles, 1988). While multiple theories exist regarding the psychological significance of the P3 component, it is generally believed to be associated with cognitive processes of working memory, attention allocation, and context updating (Donchin & Coles, 1988). The amplitude of the P3 component is considered to reflect CNS inhibition with a larger amplitude associated with a greater inhibition (Porjesz et al., 2005). Given the reproducibility and ubiquity of the P3 component, it has been widely utilized in psychiatric research and subsequently proposed as an endophenotype marker for AUD (Porjesz et al., 2005; Salvatore et al., 2015). Two competing models exist regarding the potential use of the P3 component as a biomarker for predicting AUD risk. The first is the “deficit model”, which is evidenced by research that found reduced P3 amplitude among AUD patients or high-risk populations, compared to healthy controls (for review, see Porjesz et al., 2005). This suggests a potential impairment in inhibitory control among individuals with AUD or at high risk. Research also indicates that the “deficits model” extends to the social domain, where reduced P3 amplitude to facial expressions was found in the AUD group, compared to the control group (Hoffman et al., 2019). This is also in line with the general social deficits of individuals with AUD observed using behavioral measures (Bora & Zorlu, 2017). The second model is referred to as the “sensitivity model”, where the increased risk for drinking problems is characterized by enhanced sensitivity or hyperexcitability towards alcohol cues or social rewards (Fairbairn et al., 2015; Kang et al., 2021; Martins et al., 2019). Research supporting the “sensitivity model” found enhanced P3 amplitude towards alcohol-related pictures in alcohol-administration studies or among individuals with low sensitivity to acute alcohol effects (Cofresí et al., 2022; Kang et al., 2021; Martins et al., 2019). In addition, increased activity in “reward-related” brain regions or behaviors toward positive social stimuli (e.g., smiling facial expressions) has been theorized to predict increased risk for AUD (Fairbairn et al., 2015; Foulkes & Blakemore, 2016). A major limitation of prior research is that experiments are typically conducted under a cross-sectional framework combined with a between-group design (e.g., AUD patients vs. healthy controls). Therefore, it is unclear whether altered P3 amplitude is a trait-like vulnerability or a result of chronic use of alcohol. Further, the predictive power of P3 amplitude as an endophenotype for AUD is less known. In light of the above information, here we aim to examine the longitudinal association between P3 amplitude measured at baseline from two Oddball tasks and problematic drinking at two different time points over time. One task is focused on alcohol cue salience (“alcohol cue salience task”), where participants responded to alcohol or non-alcohol pictures as targets (for details of the task, see Kang et al., 2021). The second task aims to evaluate P3 responses to neutral or positive social stimuli, specifically neutral vs smiling faces ("facial emotion task"). It's worth noting that this study utilizes a unique group-drinking design in which participants are paired as strangers or friends dyads and consume either alcohol or control beverages together at the beginning of the study. Afterward, during the facial emotion task, participants were asked to respond to faces from both the group-drinking members and non-group-drinking members as targets.