A Systematic Review connecting Functional Connectivity and Alcohol Use Disorder with Relevance Maps

1772 

Abstract Submission 

Marco Bottino¹, Natálie Bočková¹, Nico Poller¹, Michael Smolka², Justin Böhmer³, Henrik Walter⁴, Michael Marxen¹

¹Technische Universität Dresden, Dresden, Saxony, ²Department of Psychiatry and Psychotherapy, Technische Universität Dresden, Dresden, Germany, ³Charité – Universitätsmedizin Berlin, Berlin, Brandenburg, ⁴Division of Mind and Brain Research, Department for Psychiatry, Charité–Universitätmedizin Berlin, Berlin, Germany

Marco Bottino  
Technische Universität Dresden
Dresden, Saxony

Natálie Bočková  
Technische Universität Dresden
Dresden, Saxony

Nico Poller  
Technische Universität Dresden
Dresden, Saxony

Michael Smolka  
Department of Psychiatry and Psychotherapy, Technische Universität Dresden
Dresden, Germany

Justin Böhmer  
Charité – Universitätsmedizin Berlin
Berlin, Brandenburg

Henrik Walter  
Division of Mind and Brain Research, Department for Psychiatry, Charité–Universitätmedizin Berlin
Berlin, Germany

Michael Marxen, Ph.D.  
Technische Universität Dresden
Dresden, Saxony

Alcohol use disorder (AUD) has been associated consistently with aberrations in brain functional connectivity. To provide an overview of current findings, we systematically reviewed the literature linking resting-state functional MRI (rs-fMRI) to AUD. However, studies and methodologies are very heterogeneous. We devised an algorithm to deal with this heterogeneity and create quantitative relevance maps for regions as well as connections that can be used to formulate hypotheses in future studies.

We identified 248 papers with a systematic search across PubMed, Scopus, and Web of Science using the query "alcohol" AND "connectivity" AND ("resting" OR "rest"). After screening and excluding studies, as shown in Fig. 1, 50 papers with 95 separate analyses remained for detailed analysis. Potential biases, especially in studies linking alcohol use disorder (AUD) to specific seeds, prompted the decision to prioritize whole-brain analyses (40/95) in creating relevance maps. The Automated Anatomical Atlas version 3.1 (AAL, Rolls et al., 2020) was employed to standardize results. All significant results concerning rs-fMRI and AUD were collected in a table, which we used to create two outputs: 1) a relevance matrix counting all studies that reported a correlation with AUD for a particular connection between regions, accounting for the direction of the effect; 2) a ranking of the regions, counting the number of whole-brain analyses citing a particular region, independently of the direction of the effect. We also categorized the most common methods and results in the field.

As a general overview, most of the studies have a sample size of less than 100 subjects and use a group comparison approach, where standard divisions are healthy controls vs. AUD, binge vs. light drinkers, and, in longitudinal studies, relapsers vs. abstainers. Methodologically, two main brain parcellations are employed: atlas-based (e.g., AAL or Harvard-Oxford from Desikan et al., 2006) or data-driven independent component analysis. The subsequent rs-fMRI connectivity analyses typically involve generating subject-specific statistical maps (e.g., through dual regression) or functional connectivity matrices through Pearson's correlation. Analyses fall into three main types: group statistical comparisons, correlation/regression models predicting addiction severity measures, and machine learning for binary classification of AUD against controls. Severity measures include structured clinical interviews and scores on different aspects of the disorder, e.g., AUDIT score (Saunders et al., 1993).
The relevance matrix in Figure 2a includes connection scores from -2 to 2: 8.4% of all scores are not zero. Noteworthy, positive associations include a triangle involving the insula, anterior cingulate cortex, and left superior temporal cortex. Negative associations include connections between the middle temporal cortex and precuneus, as well as between the superior temporal cortex and the right lingual gyrus. The region ranking (Fig. 2b) highlights the most cited brain regions in whole-brain analyses: dorsolateral prefrontal cortex (PFC), dorsal anterior cingulate cortex, dorsomedial PFC, insula, putamen, and left precuneus are on top of the list . Despite being a common seed, the amygdala has significant results in only one analysis.

While the literature linking rs-fMRI connectivity and AUD is substantial, it is also very heterogeneous and partially biased toward particular brain regions. Our methodological approach aimed to address potential biases and to ensure a comprehensive analysis of the diverse findings. The resulting relevance maps summarize our current knowledge of which regions and connections of the brain are associated with AUD. Additionally, they provide a basis to formulate more precise and unbiased hypotheses in future projects and interpret new data .

Psychiatric (eg. Depression, Anxiety, Schizophrenia) ²

fMRI Connectivity and Network Modeling ¹

Task-Independent and Resting-State Analysis

Cortical Anatomy and Brain Mapping

Brain Atlases

Addictions

Data analysis

Data Organization

DISORDERS

FUNCTIONAL MRI

Psychiatric Disorders

Other - Alcohol;