The impact of alcohol consumption on brain networks: Investigating differential effects in bipolar disorder

Background: Widespread alterations to cortical regions particularly in areas related to reward processes have been described following non-dependent alcohol use and independently in bipolar disorder (BD). Additionally, alterations in the microstructural organisation of white matter are associated with non-dependent alcohol consumption and within BD. Moreover, alcohol use in the disorder impacts functional networks related to emotion, cognition, and introspection, these alterations are associated with mood lability and negative illness trajectory. Identifying the interactive and network level effects of alcohol use and diagnosis on the brain may elucidate the impact of alcohol on reward and affective circuitry and its contribution to relapse in BD. This thesis uses a range of MRI modalities and theoretical developments in network neuroscience to uncover previously unknown knowledge about the impact of alcohol on the brain and its compound effects for people with a diagnosis of BD. Methods: For all studies alcohol use was assessed using the Alcohol Use Disorders Identification Test, Consumption sub score (AUDIT-C). This test validly and reliably assesses alcohol consumption across a range of levels and a number of differing testing contexts. Study One: Forty-six psychiatrically healthy and 40 BD (DSM-V-TR) participants underwent T1-weighted (MPRAGE) MRI scanning at 3T and the AUDIT-C. Cortical regions of interest were parcellated based on the Desikan-Killiany atlas (Freesurfer v.5.3.0) and included the anterior cingulate (ACC), dorsolateral prefrontal (dlPFC), and orbitofrontal cortices (OFC), and insula. Cortical thickness was examined for an effect of alcohol use and compared between BD and control groups covarying for age, sex and diagnosis. Study Two: Thirty-four BD-I (DSM-IV-TR) and 38 psychiatrically healthy controls underwent T1 and diffusion-weighted MRI scanning, and the AUDIT-C. Connectomes comprising 34 cortical and nine subcortical nodes bilaterally (Freesurfer v5.3) connected by fractional anisotropy-weighted edges derived from non-tensor based deterministic constrained spherical deconvolution tractography (ExploreDTI v4.8.6) underwent permutation-based topological analysis (NBS v1.2) and were examined for effects of alcohol use and diagnosis-by-alcohol use accounting for age, sex and diagnosis. Study Three: Forty BD-I (DSM-IV-TR) and 46 psychiatrically-healthy controls underwent T1 and resting state functional MRI scanning, and the AUDIT-C. Functional images were decomposed using spatial independent component analysis, into 14 resting state networks (RSNs), which were examined for effect of alcohol use and diagnosis-by-alcohol use accounting for age, sex and diagnosis. Results: There was no difference in alcohol use scores between BD or control participants in all of the studies. Study One: For all participants, alcohol use was associated with reduced cortical thickness of the left ACC (T=-2.984, pFDR=0.016), left OFC (T=-2.508, pFDR=0.025), and left insula (T=-2.385, pFDR=0.025). For BD participants only, there was an association between alcohol use and cortical thickness in the left dlPFC (T=-2.237, p=0.032). Study Two: Alcohol was significantly related to a subnetwork, encompassing connections between fronto-limbic, basal ganglia and temporal nodes (Frange=5-8.4, p=0.031). A portion of this network (18%), involving cortico-limbic and basal ganglia connections, was differentially impacted by alcohol in the BD relative to the control group (Frange=5-8.8, p=0.033), Study Three: For BD participants greater alcohol use was associated with increased connectivity of the paracingulate gyrus within a default mode network (DMN) and reduced connectivity within an executive control network (ECN) relative to controls. Independently greater alcohol use was associated with increased connectivity within an ECN, and reduced connectivity within a DMN. Conclusions: Taken together these papers suggest that alcohol is associated with structural alterations to specific reward related structures, is found to impact subnetwork connectivity, and that these alterations are further reflected in aberrant functional connectivity of networks subserving introspective and affective processes. Moreover, for those with a diagnosis of BD, these results suggest that pre-existing structural and functional alterations may be placed at additional vulnerability to the impact of alcohol use, which is reflected in alterations to network patterning between structures subserving affective and cognitive processes. This altered pattern of connectivity may impact on reward expectancies and emotion processing, thus influencing emotional lability within the disorder. These results indicate that clinical guidelines should reflect the impact that alcohol use may have on illness trajectory for people with a diagnosis of BD. 2023-10-14