Advertisement

Mesial Prefrontal Cortex and Alcohol Misuse: Dissociating Cross-sectional and Longitudinal Relationships in UK Biobank

Open AccessPublished:March 20, 2022DOI:https://doi.org/10.1016/j.biopsych.2022.03.008

      Abstract

      Background

      Alcohol misuse is a major global public health issue. The disorder is characterized by aberrant neural networks interacting with environment and genetics. Dissecting the neural substrates and functional networks that relate to longitudinal changes in alcohol use from those that relate to alcohol misuse cross-sectionally is important to elucidate therapeutic approaches.

      Methods

      To assess how neuroimaging data, including T1, resting-state functional magnetic resonance imaging, and diffusion-weighted imaging, relate to alcohol misuse cross-sectionally and longitudinally in the UK Biobank, this study analyzed range of alcohol misuse in a population-based normative sample of 24,784 participants, ages 45 to 81 years old, in a cross-sectional analysis and a sample of 3070 participants in a longitudinal analysis 2 years later.

      Results

      Cross-sectional analysis showed that alcohol use is associated with a reduction in dorsal anterior cingulate cortex and dorsomedial prefrontal cortex gray matter concentration and functional resting-state connectivity (nodal degree: t24,422 = −12.99, p < 1 × 10−17). Reduced dorsal anterior cingulate cortex/dorsomedial prefrontal cortex functional connections to the ventrolateral prefrontal cortex, amygdala, and striatum relate to greater alcohol use. In a longitudinal analysis, higher resting-state nodal degree (t3036 = −3.27, p = .0011) and T1 gray matter concentration in the ventromedial prefrontal cortex relate to reduced alcohol intake frequency 2 years later. Higher ventromedial prefrontal cortex and frontoparietal executive network functional connectivity is associated with lower subsequent drinking longitudinally.

      Conclusions

      Dorsal versus ventromedial prefrontal regions are differentially related to alcohol misuse cross-sectionally or longitudinally in a large UK Biobank normative dataset. Our study provides a comprehensive understanding of the neurobiological substrates of alcohol use as a state or prospectively, thereby providing potential targets for clinical treatment.

      Keywords

      Alcohol misuse/abuse is a major global public health issue. Alcohol use disorder (AUD) is among the most prevalent mental disorders worldwide (
      • Hasin D.S.
      • Stinson F.S.
      • Ogburn E.
      • Grant B.F.
      Prevalence, correlates, disability, and comorbidity of DSM-IV alcohol abuse and dependence in the United States: Results from the National Epidemiologic Survey on Alcohol and Related Conditions.
      ,
      World Health Organization
      Global Status Report on Alcohol and Health 2018.
      ). In the United Kingdom and globally, alcohol misuse is the greatest risk factor for death, ill health, and disability among 15- to 49-year-olds, with UK societal costs of £21 to £52 billion (
      • Burton R.
      • Henn C.
      • Lavoie D.
      • Wolff A.
      • Marsden J.
      • Sheron N.
      • et al.
      The Public Health Burden of Alcohol and the Effectiveness and Cost-Effectiveness of Alcohol Control Policies: An Evidence Review.
      ). AUD has been associated with aberrant neurocircuitry interacting with environmental, social, and genetic influences (
      • Kwako L.E.
      • Momenan R.
      • Litten R.Z.
      • Koob G.F.
      • Goldman D.
      Addictions Neuroclinical Assessment: A neuroscience-based framework for addictive disorders.
      ,
      • Voon V.
      • Grodin E.
      • Mandali A.
      • Morris L.
      • Doñamayor N.
      • Weidacker K.
      • et al.
      Addictions NeuroImaging Assessment (ANIA): Towards an integrative framework for alcohol use disorder.
      ). The neural substrates mapping to the cognitive processes underlying core addiction theoretical mechanisms of incentive salience, negative emotionality, and transition to compulsive behaviors have been systematically reviewed in the Addictions Neuroclinical Imaging Assessment (
      • Kwako L.E.
      • Momenan R.
      • Litten R.Z.
      • Koob G.F.
      • Goldman D.
      Addictions Neuroclinical Assessment: A neuroscience-based framework for addictive disorders.
      ,
      • Voon V.
      • Grodin E.
      • Mandali A.
      • Morris L.
      • Doñamayor N.
      • Weidacker K.
      • et al.
      Addictions NeuroImaging Assessment (ANIA): Towards an integrative framework for alcohol use disorder.
      ), in which extensive cortical and subcortical brain structures and networks were found relevant to alcohol misuse.
      While it is important to identify the neural mechanism and negative impacts of alcohol use, a more important target is to identify biomarkers that relate to change or successful reduction of alcohol use and, therefore, facilitate developing theoretical support for clinical interventions. In the recent decade, more efforts have been given to identify such biomarkers with task-based functional magnetic resonance imaging (fMRI) and gray matter volume/concentration associating with prospective alcohol use initiation or change (
      • Baranger D.A.A.
      • Demers C.H.
      • Elsayed N.M.
      • Knodt A.R.
      • Radtke S.R.
      • Desmarais A.
      • et al.
      Convergent evidence for predispositional effects of brain gray matter volume on alcohol consumption.
      ,
      • Whelan R.
      • Watts R.
      • Orr C.A.
      • Althoff R.R.
      • Artiges E.
      • Banaschewski T.
      • et al.
      Neuropsychosocial profiles of current and future adolescent alcohol misusers.
      ,
      • Squeglia L.M.
      • Ball T.M.
      • Jacobus J.
      • Brumback T.
      • McKenna B.S.
      • Nguyen-Louie T.T.
      • et al.
      Neural predictors of initiating alcohol use during adolescence [published correction appears in Am J Psychiatry 2017; 174:80].
      ). For example, in a recent large-scale analysis (N = 2423), smaller dorsolateral and insular gray matter volume was found to be significantly related to the predisposition and initiation of alcohol use spanning from adolescents to middle-age adults (
      • Baranger D.A.A.
      • Demers C.H.
      • Elsayed N.M.
      • Knodt A.R.
      • Radtke S.R.
      • Desmarais A.
      • et al.
      Convergent evidence for predispositional effects of brain gray matter volume on alcohol consumption.
      ). In an adolescent study, gray matter volume and activation of the bilateral superior frontal gyrus and right precentral gyrus in task-based fMRI at 14 years old was associated with alcohol misuse at age 16 (
      • Whelan R.
      • Watts R.
      • Orr C.A.
      • Althoff R.R.
      • Artiges E.
      • Banaschewski T.
      • et al.
      Neuropsychosocial profiles of current and future adolescent alcohol misusers.
      ). Similar adolescent studies identified reduced cortical thickness and/or activation of the rostral anterior cingulate and right superior frontal and frontal pole in 12- to 14-year-olds as related to moderate to heavy drinking by age 18 (
      • Squeglia L.M.
      • Ball T.M.
      • Jacobus J.
      • Brumback T.
      • McKenna B.S.
      • Nguyen-Louie T.T.
      • et al.
      Neural predictors of initiating alcohol use during adolescence [published correction appears in Am J Psychiatry 2017; 174:80].
      ). Activation of prefrontal regions relating to alcohol use change was also frequently reported in other task-based fMRI studies with relevant alcohol-related cognitive processes such as response inhibition (
      • Norman A.L.
      • Pulido C.
      • Squeglia L.M.
      • Spadoni A.D.
      • Paulus M.P.
      • Tapert S.F.
      Neural activation during inhibition predicts initiation of substance use in adolescence.
      ,
      • Wetherill R.R.
      • Squeglia L.M.
      • Yang T.T.
      • Tapert S.F.
      A longitudinal examination of adolescent response inhibition: Neural differences before and after the initiation of heavy drinking.
      ,
      • Heitzeg M.M.
      • Nigg J.T.
      • Hardee J.E.
      • Soules M.
      • Steinberg D.
      • Zubieta J.K.
      • Zucker R.A.
      Left middle frontal gyrus response to inhibitory errors in children prospectively predicts early problem substance use.
      ) and alcohol cue reactivity (
      • Nguyen-Louie T.T.
      • Courtney K.E.
      • Squeglia L.M.
      • Bagot K.
      • Eberson S.
      • Migliorini R.
      • et al.
      Prospective changes in neural alcohol cue reactivity in at-risk adolescents.
      ), albeit with smaller sample sizes. In summary, previous studies point toward a reduced frontal-centered structural and functional network relating to executive, control, and inhibitory function, which is engaged in alcohol misuse initiation and/or change longitudinally.
      Based on previous findings, we hypothesize that the brain networks that relate to prospective longitudinal changes in alcohol use can be dissected from those related to alcohol use as a state in the frontal networks engaging cognitive control and inhibitory processes. We address the hypothesis within the UK Biobank data, a normative population with a range of alcohol consumption patterns. The Biobank brain imaging dataset contains nearly 40,000 participants aged 45 to 81 years old scanned with multimodal MRI, including T1-weighted structural data, resting-state fMRI (rfMRI), and diffusion-weighted imaging (DWI) (
      • Sudlow C.
      • Gallacher J.
      • Allen N.
      • Beral V.
      • Burton P.
      • Danesh J.
      • et al.
      UK Biobank: An open access resource for identifying the causes of a wide range of complex diseases of middle and old age.
      ). The extensively validated Alcohol Use Disorders Identification Test (AUDIT) (
      • Saunders J.B.
      • Aasland O.G.
      • Babor T.F.
      • de la Fuente J.R.
      • Grant M.
      Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO Collaborative Project on Early Detection of Persons With Harmful Alcohol Consumption–II.
      ), which is a sensitive index of problem alcohol misuse, was assessed at baseline. Longitudinally, alcohol intake frequency was assessed in a large subset (more than 3000) of participants. Thus, the Biobank dataset offers a unique combination of cross-sectional and longitudinal data to identify vulnerable neural substrates and functional network-based biomarkers.
      Here, we analyzed the Biobank dataset to first ask how brain structure and functional connectivity (FC) in an adult normative population relate to alcohol use in cross-sectional analysis and, more importantly, associate with alcohol use perspective changes in longitudinal analysis (Figure 1). We used a network approach, investigating the resting-state FC networks over and above the gray and white matter structure, so that we can identify biomarkers based on connectivity and functional hemodynamics. The hemodynamic-based biomarker is potentially more useful as a target for brain stimulation (
      • Allen E.A.
      • Pasley B.N.
      • Duong T.
      • Freeman R.D.
      Transcranial magnetic stimulation elicits coupled neural and hemodynamic consequences.
      ). In addition, we also systematically estimated the convergence evidence between rfMRI and structural MRI. To our knowledge, our study is the first large-sample, multimodal MRI (particularly FC-based), prospective study for alcohol use.
      Figure thumbnail gr1
      Figure 1Analysis flowchart. Multimodal brain imaging data, including T1 gray matter concentration, resting-state functional magnetic resonance imaging (fMRI) functional connectivity and nodal degree, and diffusion-weighted imaging (DWI) fractional anisotropy (FA), were related to alcohol use in different regression models. Support vector regression with split-half cross-validation was used to test how well the alcohol use–brain relationship can be cross-validated and replicated. Smoking behavior was analyzed in the same pipeline as a comparison to alcohol use and to identify common substance misuse effects on the brain. Age, gender, intracranial volume, and participants’ data collection site were controlled for in all models. In addition, age was analyzed in the same pipeline as alcohol use, so that we could compare the alcohol use effect with normal aging. MNI, Montreal Neurological Institute; VBM, voxel-based morphometry.

      Methods and Materials

      Participants

      Imaging data from 39,679 participants from the UK Biobank (Project ID 64044) were filtered for alcohol use, smoking, age, gender, total intracranial volume, and data quality in T1, rfMRI, and DWI. Demographics and usable participant numbers are reported in Tables S1 and S2.

      Behaviors

      Participants’ alcohol use scores are based on AUDIT (
      • Saunders J.B.
      • Aasland O.G.
      • Babor T.F.
      • de la Fuente J.R.
      • Grant M.
      Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO Collaborative Project on Early Detection of Persons With Harmful Alcohol Consumption–II.
      ), with alcohol intake frequency used for longitudinal change (reversing the original coding so that higher coding corresponds to more frequent intake: 6 = daily drinking; 5 = 3−4 times/week; 4 = 1−2 times/week; 3 = 1−3 times/month; 2 = special occasion; 1 = never; alcohol intake frequency data are in the same direction as AUDIT). We added participants’ smoking data into the analysis as a comparison with alcohol. Smoking is based on pack-years (daily cigarettes smoked used for longitudinal change).
      Age, gender, imaging data collection site, and total intracranial volume were included as control variables. Other potentially confounding variables, including depression, anxiety, body mass index, and blood pressure were assessed and are included in Table S1 and Figure S11.

      Brain Imaging Data Processing

      The preprocessing steps are described in the Biobank documentation (https://biobank.ctsu.ox.ac.uk/crystal/ukb/docs/brain_mri.pdf). Based on the preprocessed data, we calculated voxelwise gray matter concentration from T1, nodal degree and FC from rfMRI, and fractional anisotropy (FA) from DWI. More details are given in the Supplement.

      General Linear Models Relating Brain Measures to Behaviors

      The general linear models were built on each voxel with all available participants for each imaging modality and behaviors of interest.
      For simple linear models, we first regressed out age, gender, total intracranial volume, and data collection site from the voxel signal intensity, e.g., voxel gray matter concentration, nodal degree, or FC. We then related the residual values to the AUDIT scores (alcohol intake frequency change used for longitudinal analysis). The simple linear models were built with MATLAB (R2020b; The MathWorks, Inc.) “fitlm” function, and effect size (Cohen’s f2) was reported. Multiple linear models are given in detail in the Supplement.
      We further built five interaction models between AUDIT scores and age/gender, pack-years and age/gender, and AUDIT and pack-years, using MATLAB “fitlm” with interaction terms.
      We conducted similar analyses for smoking as a comparison with alcohol to demonstrate the potential common effects of substance misuse. Aging was analyzed in the cross-sectional analysis to show the normal aging brain pattern as a comparison with alcohol use and smoking.
      Multiple comparisons corrections are given in the Supplement. Brain statistical results were mapped onto brain maps using MATLAB package DPABI (
      • Yan C.G.
      • Wang X.D.
      • Zuo X.N.
      • Zang Y.F.
      DPABI: Data processing & analysis for (resting-state) brain imaging.
      ) and BrainNet Viewer (
      • Xia M.
      • Wang J.
      • He Y.
      BrainNet Viewer: A network visualization tool for human brain connectomics.
      ).

      Support Vector Regression Models

      Given the advantage of the Biobank’s large sample, we built support vector regression (SVR) models on the cross-sectional gray matter and rfMRI nodal degree data. The SVR model with split-half cross-validation was used to test if the brain-behavior relationships are stable and replicable. The SVR models were built with the ‘fitrlinear’ function in MATLAB with the ‘svm’ learner. We used a brain atlas with 625 similarly sized regions of interest, respecting the region boundaries of the Automated Anatomical Labeling 90 atlas (
      • Wang J.
      • Wang X.
      • Xia M.
      • Liao X.
      • Evans A.
      • He Y.
      GRETNA: A graph theoretical network analysis toolbox for imaging connectomics [published correction appears in Front Hum Neurosci 2015; 9:458].
      ). Pearson correlation between predicted and real behavioral residual scores was used to measure accuracy of the SVR split-half cross-validation model. We give an example of the SVR model settings in the Supplement.

      Results

      Our analytic flow chart is shown in Figure 1. We report the behavioral scores and their relationships with T1 gray matter concentration, rfMRI nodal degree, and FC. DWI FA results are reported in the Supplement. We focus on the simple linear regression results between variables of interest and brain measure residual scores, controlling for age, gender, intracranial volume, and data collection site from gray matter signal intensity. The model accuracy of the SVR was reported following the univariate regression analyses in each section.
      The multiple linear regression, where all variables were analyzed in one regression model, demonstrated the same pattern of results as the simple linear regression. The interaction models did not show significant interaction effects, which means that alcohol use (smoking) and age contributed to brain functional and structural changes independently. Alcohol use (smoking) has no interaction with gender on the brain measures.

      Behavioral Scores

      The behavioral data from the time point 1 alcohol (n = 24,784; male participants: n = 11,156) T1 voxel-based morphometry analysis is reported (mean [SD]: age = 63.46 [7.47] years old; AUDIT score = 5.19 [4.16]) (Figure S1). In the smoking analyses, there were n = 30,778 participants (pack-years 5.08 [11.69]) and n = 8157 smokers only (19.17 [15.67]). Pearson correlation between AUDIT scores and alcohol intake frequency at the first timepoint is r = 0.58 (p < 1 ×10−17, n = 24,631).
      We analyzed longitudinal alcohol intake frequency change from the second versus first imaging visits (n = 3070). The difference between the second and first imaging visits was an average of 2.35 years (SD = 0.71, range = 1–7 years, 71% with 2-year difference and 26% with 3-year difference). The majority (68%) of participants’ drinking frequency was unchanged, 18% reduced (negative scores), and 13% increased (positive scores). For the frequent daily drinkers (code 6, n = 457), 27% (n = 124) reduced drinking.

      Cross-sectional Gray Matter Voxel-Based Morphometry

      Alcohol Use

      Greater alcohol use was related to reduced gray matter concentration in the dorsal anterior cingulate (dACC)/dorsomedial prefrontal cortex (dmPFC) (Bonferroni-corrected voxel p < .001, cluster size > 400 mm3; peak coordinates: 0, 41, 29; t = −7.37, p = 1 × 10−13, effect size = 0.0022; cluster size = 916 mm3) (Figure 2). The same gray matter correlates can be seen when related to first timepoint alcohol intake frequency (Figure S2).
      Figure thumbnail gr2
      Figure 2T1 voxel-based morphometry cross-sectional analysis. The first row shows the brain correlates from simple linear regression and the second row shows the average brain maps (training sets) from the support vector regression with split-half cross-validation. In the first row, for alcohol use, smoking, and aging, the dorsal anterior cingulate and dorsomedial prefrontal cortex are significant for three behaviors in a negative direction, which means more alcohol use, smoking, and aging were related to smaller gray matter concentration (for alcohol use and smoking, Bonferroni-corrected voxel p < .001 and cluster size > 400 mm3; for aging, a higher threshold, t < −40, was applied to identify the peak relationships). The alcohol effect is very localized, while for smoking, related regions also extend to the dorsolateral prefrontal cortex, insula, and striatum. For aging, the peak related regions are mainly surrounding the ventricles and large sulcus, including the insula and striatum. The brain correlates in the simple linear regression overlap with the peak regions identified in the support vector regression. L, left; R, right.
      The peak cluster can also be seen in the SVR model (model accuracy: r = 0.12, range = 0.05–0.15, 10,000 times cross-validation; p < .00001, permutation test 10,000 times) (Figure 2).

      Smoking

      Greater smoking (in number of packs per year) was also related to reduced gray matter concentration in the dACC/dmPFC, thalamus, caudate, and bilateral insula (Bonferroni-corrected voxel p < .001, cluster size > 400 mm3) (Figure 2). The prefrontal cluster includes the dACC/dmPFC, which overlaps with alcohol use gray matter correlates (peak coordinates: 0, 42, 27; t = −13.45, p < 1 × 10−17, effect size = 0.0059; cluster size = 177,346 mm3).
      The above map was similar when only including smokers. Owing to the non-normal distribution of smoking pack-years, we only included smokers in the SVR analysis (model accuracy: r = 0.16, range = 0.05–0.21, 10,000 times cross-validation; p < .00001, permutation test 10,000 times) (Figure 2). The dACC/dmPFC region is significant in both simple linear regression and SVR.
      The effects of age and gray matter are reported in the Supplement.

      Cross-sectional rfMRI Nodal Degree

      Voxel nodal degree reflects connectivity of a voxel to other voxels, assessed as the total number (weight) of connections of a given voxel with all other voxels that passed a connectivity threshold (Pearson correlation r > 0.2). With rfMRI, we first analyzed how nodal degree relates to alcohol use, smoking, and aging to understand how these variables related to brain function from a network perspective.

      Alcohol Use

      Greater alcohol use was related to reduced nodal degree in whole-brain gray matter. The peak region (at threshold: t < −10, cluster size > 400 mm3) is in the dACC/dmPFC (peak coordinates: 0, 40, 24, t = −12.99, p < 1 × 10−17, effect size = 0.0069, cluster size = 8288 mm3) (Figure 3). The finding converged with alcohol use gray matter correlates, including when controlling for gray matter concentration. Same nodal degree brain correlates can be seen when related to first time point alcohol intake frequency (Figure S2). Large extent regions are significant (Figures S8 and S9) with more liberal t values. We focus on the dACC/dmPFC because it is the peak cluster in both T1 and rfMRI results.
      Figure thumbnail gr3
      Figure 3Relating resting-state functional magnetic resonance imaging nodal degree to behaviors cross-sectional analysis. In simple linear regression, for alcohol use, smoking, and aging, the dorsal anterior cingulate cortex/dorsomedial prefrontal cortex are significant for all three behaviors in a negative direction, which means more alcohol use, smoking, and aging were related to smaller nodal degree (for alcohol use and smoking, t < −10, p < 1 × 10−17, and cluster size > 400 mm3; for aging, a higher threshold, t < −40, was applied to identify the peak relationships). For smoking, the bilateral superior and middle frontal gyri are also related. For aging, the peak relationships are surrounding the ventricles and sulcus, including the bilateral insula and parietal lobule. The brain correlates in the simple linear regression overlap with the peak regions identified in the support vector regression. L, left; R, right.
      The dACC/dmPFC cluster can also been seen in the SVR model (model accuracy: r = 0.11, range = 0.07–0.14, 10,000 times cross-validation; p < .00001, permutation test 10,000 times) (Figure 3).

      Smoking

      Greater smoking pack-years was related to reduced nodal degree in the dACC/dmPFC and bilateral superior and middle frontal gyri (t < −10, cluster size > 400 mm3) (Figure 3). The dACC/dmPFC cluster is significantly related to smoking (peak coordinates: 2, 40, 26, t = −11.38, p < 1 × 10−17, effect size = 0.0043, cluster size = 1104 mm3) and was evident when controlling for alcohol use and gray matter concentration.
      The above map is similar when only including smokers. The dACC/dmPFC and bilateral middle and superior frontal gyri are significant in both simple linear regression and SVR (model accuracy: r = 0.13; range = 0.08–0.18, 10,000 times cross-validation; p < .00001, permutation test 10,000 times).
      The effects of age and nodal degree are reported in the Supplement.

      Cross-sectional Resting-State FC

      Given the dACC/dmPFC nodal degree overlap with alcohol use, smoking, and age, we further conducted dACC/dmPFC FC analyses to investigate dissociable patterns of connectivity. The dACC/dmPFC peak cluster from nodal degree (t < −11), which overlaps with alcohol gray matter voxel-based morphometry (Bonferroni-corrected voxel p < .001), was used as a region of interest in the seed-based whole-brain FC analysis.

      Alcohol Use and Smoking

      Both greater alcohol use and smoking were associated with reduced FC between the dACC/dmPFC and other brain regions, with no increased FC observed (Figure 4; significant region reports in Table S3 for alcohol use and Table S4 for smoking). Reduced connectivity with the bilateral inferior frontal, dorsolateral prefrontal, and lateral orbitofrontal cortices was specific to alcohol use and unrelated to smoking or aging, whereas reduced connectivity with bilateral superior frontal gyri and medial orbitofrontal cortices was specific to smoking. In contrast, decreased subcortical connectivity between the dACC/dmPFC and the putamen, caudate, thalamus, and amygdala was observed for both alcohol and smoking but was unrelated to aging, highlighting common subcortical effects across substance misuse.
      Figure thumbnail gr4
      Figure 4Relating resting-state functional magnetic resonance imaging functional connectivity to behaviors cross-sectional analysis. The first row shows the brain correlates in the lateral and medial views (Bonferroni-corrected voxel p < .01), and the second row shows the relationships in the subcortical regions (false discovery rate–corrected voxel p < .001). Reduced connectivity with the bilateral inferior frontal, dorsolateral prefrontal, and lateral orbitofrontal cortices was specific to alcohol use and unrelated to smoking or aging, whereas reduced connectivity with the bilateral superior frontal gyri and medial orbitofrontal cortices was specific to smoking. In contrast, dorsal anterior cingulate cortex/dorsomedial prefrontal cortex decreased subcortical connectivity with the putamen, caudate, thalamus, and amygdala were observed for both alcohol and smoking but unrelated to aging, highlighting common subcortical effects across substance misuse. L, left; R, right.
      The FA results reported in the Supplement show negative correlations for both alcohol and smoking in white matter tracts surrounding the dACC/dmPFC, including the corpus callosum and cingulum.

      Relating Brain Measures to Longitudinal Behavioral Change

      We repeated the multimodal brain-behavior univariate analyses using the longitudinal behavioral data, including alcohol intake frequency change and daily smoking change. We used scores at the second imaging visit minus scores at the first visit and related these to brain measures from the first imaging visit investigating biomarkers of perspective behavioral change. We found nodal degree and gray matter concentration significantly related to alcohol intake frequency change but not smoking change.
      With all available participants, we found that greater nodal degree (resting-state) and gray matter concentration (T1) was related to reduced drinking frequency at approximately 2 years of follow-up (Figure 5). The ventromedial prefrontal cortex (vmPFC) in particular was significant in both resting-state nodal degree and T1 gray matter analyses (nodal degree: Gaussian random field [GRF]–corrected voxel p < .01 and cluster p < .05, convergent regional peak coordinates −4, 56, −20, t = −3.27, p = .0011, effect size = 0.0035, cluster size = 1376 mm3; T1: GRF-corrected voxel p < .02 and cluster p < .05, convergent regional peak coordinates −1, 39, −29, t = −3.24, p = .0012, effect size = 0.0034, cluster size = 1445 mm3). In addition, the ventral striatum and amygdala also showed significance in the gray matter analysis.
      Figure thumbnail gr5
      Figure 5Relating brain to longitudinal alcohol use change. The first row brain maps show the regions that related to alcohol use change in resting-state nodal degree (Gaussian random field [GRF]–corrected voxel p < .01 and cluster p < .05) and gray matter (GM) concentration (GRF-corrected voxel p < .02 and cluster p < .05) analyses. The ventromedial prefrontal cortex (vmPFC) shows significance in both maps. We further separated the participants into six groups based on their drinking frequency at the first visit (group 1 = never drinks, group 2 = special occasions only, group 3 = 1–3 times a month, group 4 = once or twice a week, group 5 = three or four times a week, group 6 = daily or almost daily) and then related the brain measures to alcohol intake frequency change in each group. The result shows that only group 6, the frequent drinkers group, drives the nodal degree relationships, as shown in the first brain map in the second row (GRF-corrected voxel p < .01 and cluster p < .05). The significant vmPFC cluster (in red circle) is used as a seed to calculate its whole-brain functional connectivity. The vmPFC functional connections that relate to reduced drinking frequency are shown in the second brain map in the second row (GRF-corrected voxel p < .01 and cluster p < .05). The connected regions match to the frontoparietal central executive network, including the bilateral superior and middle frontal gyri and the inferior parietal lobule. L, left; R, right; ROI, region of interest.
      We further separated the participants into six groups based on their drinking frequency at the first imaging visit. We then related brain nodal degree to the drinking frequency change and only considered participants who changed. The frequent drinkers (group 6) drove the nodal degree relationship in the vmPFC region (GRF-corrected voxel p < .01 and cluster p < .05) (Figure 5), which converges with the whole group nodal degree findings (convergent region peak coordinates: 10, 48, −22, t = −3.06, p = .0027, effect size = 0.0768, cluster size = 1224 mm3). Greater nodal degree of bilateral subcortical regions in the caudate, putamen, pallidum, and thalamus also related to lower subsequent alcohol consumption (left hemisphere peak coordinates: −8, −26, 12, t = −3.47, p = .0007, effect size = 0.0988, cluster size = 4672 mm3; right hemisphere peak coordinates: 22, 14, −4, t = −3.41, p = .0009, effect size = 0.0951, cluster size = 5440 mm3).
      Because the vmPFC had a convergent significant effect on alcohol use change, we further analyzed which functional connections from the vmPFC relate to alcohol use change (Figure 5). Greater vmPFC connectivity to the bilateral frontoparietal central executive network (
      • Menon V.
      Large-scale brain networks and psychopathology: A unifying triple network model.
      ) related to subsequent lower alcohol use change (GRF-corrected voxel p < .01 and cluster p < .05). Regions and statistics are reported in the Supplement.

      Discussion

      We highlight a multimodal dissociation of structural and functional brain networks in cross-sectional and longitudinal relationships with alcohol use in a large normative older adult cohort. The dACC/dmPFC region cross-sectionally relates to alcohol misuse (in the past 1 year, measured by AUDIT) and smoking severity (measured by pack-years) convergently across gray matter concentration, rfMRI nodal degree, and FC and white matter FA using both linear regression and machine learning methods. Smoking severity shows a similar but independent effect in the dACC/dmPFC but with differing FC patterns. Both forms of substance use appear to show similar effects as aging in this region but also show effects independent of aging. In contrast, ventral and subcortical regions, such as the vmPFC, ventral striatal, and amygdala, show integrity across both gray matter concentration and nodal degree, and connectivity with the frontoparietal central executive network relates to resilience and vulnerability, specifically in alcohol use frequency at 2 years. Thus, we highlight dissociable roles of the mesial prefrontal cortex with alcohol misuse on dorsal structures of the dmPFC/dACC implicated in impaired top-down cognitive control; in contrast, we highlight ventral and subcortical structures of the vmPFC, ventral striatum, and amygdala, in relation to prospective resilience and vulnerability implicated in incentive motivation, negative emotionality, and habit theories of addictions.

      Cross-sectional Effect of dACC/dmPFC

      The dACC/dmPFC region appears to relate to the cumulative effects of substance misuse. Within gray matter, smoking seems to have a stronger effect than alcohol use, which might be because pack-years considers lifetime nicotine use, while AUDIT only considers the past year. The dACC is activated in both alcohol cue reactivity and craving behaviors (
      • Schacht J.P.
      • Anton R.F.
      • Myrick H.
      Functional neuroimaging studies of alcohol cue reactivity: A quantitative meta-analysis and systematic review.
      ). This hub region is implicated in cognitive interference and control tasks in healthy control subjects and other substance misuse, such as opiates (
      • Yücel M.
      • Lubman D.I.
      • Harrison B.J.
      • Fornito A.
      • Allen N.B.
      • Wellard R.M.
      • et al.
      A combined spectroscopic and functional MRI investigation of the dorsal anterior cingulate region in opiate addiction.
      ) and cocaine (
      • Kaufman J.N.
      • Ross T.J.
      • Stein E.A.
      • Garavan H.
      Cingulate hypoactivity in cocaine users during a GO-NOGO task as revealed by event-related functional magnetic resonance imaging.
      ). dACC glutamate is further negatively correlated with alcohol use symptom severity (
      • Grecco G.G.
      • Chumin E.J.
      • Dzemidzic M.
      • Cheng H.
      • Finn P.
      • Newman S.
      • et al.
      Anterior cingulate cortex metabolites and white matter microstructure: A multimodal study of emergent alcohol use disorder.
      ) and is reduced in opiate-dependent subjects (
      • Yücel M.
      • Lubman D.I.
      • Harrison B.J.
      • Fornito A.
      • Allen N.B.
      • Wellard R.M.
      • et al.
      A combined spectroscopic and functional MRI investigation of the dorsal anterior cingulate region in opiate addiction.
      ), along with impairments in neuronal density and functional viability (
      • Moffett J.R.
      • Ross B.
      • Arun P.
      • Madhavarao C.N.
      • Namboodiri A.M.A.
      N-acetylaspartate in the CNS: From neurodiagnostics to neurobiology.
      ) in substance misuse.
      Similar to our findings, previous studies also show general whole-brain gray matter and white matter reductions in alcohol misuse (
      • Pfefferbaum A.
      • Sullivan E.V.
      • Rosenbloom M.J.
      • Mathalon D.H.
      • Lim K.O.
      A controlled study of cortical gray matter and ventricular changes in alcoholic men over a 5-year interval.
      ,
      • Pfefferbaum A.
      • Lim K.O.
      • Zipursky R.B.
      • Mathalon D.H.
      • Rosenbloom M.J.
      • Lane B.
      • et al.
      Brain gray and white matter volume loss accelerates with aging in chronic alcoholics: A quantitative MRI study.
      ,
      • Pfefferbaum A.
      • Rosenbloom M.J.
      • Chu W.
      • Sassoon S.A.
      • Rohlfing T.
      • Pohl K.M.
      • et al.
      White matter microstructural recovery with abstinence and decline with relapse in alcohol dependence interacts with normal ageing: A controlled longitudinal DTI study.
      ,
      • Thayer R.E.
      • Hagerty S.L.
      • Sabbineni A.
      • Claus E.D.
      • Hutchison K.E.
      • Weiland B.J.
      Negative and interactive effects of sex, aging, and alcohol abuse on gray matter morphometry.
      ). In the Biobank, because the dataset has increased from ∼10,000 to ∼25,000, while general whole-brain reduction was reported to relate to alcohol consumption with replication (
      • Evangelou E.
      • Suzuki H.
      • Bai W.
      • Pazoki R.
      • Gao H.
      • Matthews P.M.
      • Elliott P.
      Alcohol consumption in the general population is associated with structural changes in multiple organ systems.
      ,
      • Topiwala A.
      • Ebmeier K.P.
      • Maullin-Sapey T.
      • Nichols T.E.
      No safe level of alcohol consumption for brain health: Observational cohort study of 25,378 UK Biobank participants.
      ), the dACC/dmPFC can also be seen with peak reductions. The dACC is closely located to the ventricles, and enlarged ventricles with atrophy in surrounding brain structures are seen in alcohol misuse (
      • Yang X.
      • Tian F.
      • Zhang H.
      • Zeng J.
      • Chen T.
      • Wang S.
      • et al.
      Cortical and subcortical gray matter shrinkage in alcohol-use disorders: A voxel-based meta-analysis.
      ). Because alcohol use has a similar whole-brain atrophy pattern to aging effects (
      • Pardo J.V.
      • Lee J.T.
      • Sheikh S.A.
      • Surerus-Johnson C.
      • Shah H.
      • Munch K.R.
      • et al.
      Where the brain grows old: Decline in anterior cingulate and medial prefrontal function with normal aging.
      ,
      • Madole J.W.
      • Ritchie S.J.
      • Cox S.R.
      • Buchanan C.R.
      • Hernández M.V.
      • Maniega S.M.
      • et al.
      Aging-sensitive networks within the human structural connectome are implicated in late-life cognitive declines.
      ), alcohol adding extra structural atrophy over and above aging effects may be implicated (
      • Thayer R.E.
      • Hagerty S.L.
      • Sabbineni A.
      • Claus E.D.
      • Hutchison K.E.
      • Weiland B.J.
      Negative and interactive effects of sex, aging, and alcohol abuse on gray matter morphometry.
      ,
      • Ning K.
      • Zhao L.
      • Matloff W.
      • Sun F.
      • Toga A.W.
      Association of relative brain age with tobacco smoking, alcohol consumption, and genetic variants.
      ,
      • Sullivan E.V.
      • Pfefferbaum A.
      Brain-behavior relations and effects of aging and common comorbidities in alcohol use disorder: A review.
      ). Critically, we show that alcohol use and aging are independently related to reduction in gray matter, white matter, nodal degree, and FC.
      Although the dACC/dmPFC show convergent effects in alcohol use, smoking, and aging, FC shows distinct patterns. Decreased dACC/dmPFC and striatal FC was observed across both alcohol and smoking, reported previously in rfMRI in nicotine addiction (
      • Hong L.E.
      • Gu H.
      • Yang Y.
      • Ross T.J.
      • Salmeron B.J.
      • Buchholz B.
      • et al.
      Association of nicotine addiction and nicotine’s actions with separate cingulate cortex functional circuits.
      ) and in response inhibition task-based fMRI in alcohol misuse (
      • Courtney K.E.
      • Ghahremani D.G.
      • Ray L.A.
      Fronto-striatal functional connectivity during response inhibition in alcohol dependence.
      ). In heavy alcohol users, our findings that negative relationships in FC between the dACC and dorsolateral and ventrolateral prefrontal regions are consistent with impairments observed in executive functioning in addiction, such as working memory, planning, attentional set shifting, and inhibitory processes (
      • Thoma R.J.
      • Monnig M.A.
      • Lysne P.A.
      • Ruhl D.A.
      • Pommy J.A.
      • Bogenschutz M.
      • et al.
      Adolescent substance abuse: The effects of alcohol and marijuana on neuropsychological performance.
      ,
      • O’Reilly R.C.
      The what and how of prefrontal cortical organization.
      ,
      • Corbetta M.
      • Shulman G.L.
      Control of goal-directed and stimulus-driven attention in the brain.
      ). The unique FC with the amygdala is consistent with the role of negative emotional theories in alcohol addictions (
      • Marinkovic K.
      • Oscar-Berman M.
      • Urban T.
      • O’Reilly C.E.
      • Howard J.A.
      • Sawyer K.
      • Harris G.J.
      Alcoholism and dampened temporal limbic activation to emotional faces.
      ,
      • Volkow N.D.
      The reality of comorbidity: Depression and drug abuse.
      ,
      • Wiers C.E.
      • Shumay E.
      • Volkow N.D.
      • Frieling H.
      • Kotsiari A.
      • Lindenmeyer J.
      • et al.
      Effects of depressive symptoms and peripheral DAT methylation on neural reactivity to alcohol cues in alcoholism.
      ). In contrast, dACC FC in smoking participants shows lower connectivity with the superior prefrontal cortex, a region implicated in the development of alcohol misuse in adolescents (
      • Whelan R.
      • Watts R.
      • Orr C.A.
      • Althoff R.R.
      • Artiges E.
      • Banaschewski T.
      • et al.
      Neuropsychosocial profiles of current and future adolescent alcohol misusers.
      ).
      The peak of white matter FA reduction was seen in the genu and body of the corpus callosum and cingulum for alcohol use, smoking, and aging, which converge with the T1 and resting-state findings. In a longitudinal study in relapsed alcohol users, white matter in the callosal genu and body showed accelerated decline (
      • Pfefferbaum A.
      • Rosenbloom M.J.
      • Chu W.
      • Sassoon S.A.
      • Rohlfing T.
      • Pohl K.M.
      • et al.
      White matter microstructural recovery with abstinence and decline with relapse in alcohol dependence interacts with normal ageing: A controlled longitudinal DTI study.
      ), whereas abstinence shows the capacity to improve white matter integrity (
      • Pfefferbaum A.
      • Rosenbloom M.J.
      • Chu W.
      • Sassoon S.A.
      • Rohlfing T.
      • Pohl K.M.
      • et al.
      White matter microstructural recovery with abstinence and decline with relapse in alcohol dependence interacts with normal ageing: A controlled longitudinal DTI study.
      ). Decreased FA in white matter tracts such as the internal capsule may represent the structural basis of the FC relationships observed in frontostriatal substrates.

      Prospective Alcohol Resilience: vmPFC and Its FC

      Greater vmPFC, ventral striatal, and amygdala structural and/or functional integrity related to alcohol use resilience at 2 years. The opposing interpretation similarly may be applied of lower structural and functional integrity relating to greater subsequent alcohol use. The vmPFC finding converges with observations of impaired vmPFC state-prediction error activity in goal-directed control predicting alcohol relapse behaviors (
      • Sjoerds Z.
      • de Wit S.
      • van den Brink W.
      • Robbins T.W.
      • Beekman A.T.F.
      • Penninx B.W.J.H.
      • Veltman D.J.
      Behavioral and neuroimaging evidence for overreliance on habit learning in alcohol-dependent patients.
      ). In an addiction and obesity study, the vmPFC activity was anticorrelated with the return of appetitive conditioned responding (
      • Ebrahimi C.
      • Koch S.P.
      • Pietrock C.
      • Fydrich T.
      • Heinz A.
      • Schlagenhauf F.
      Opposing roles for amygdala and vmPFC in the return of appetitive conditioned responses in humans.
      ). Our finding of lower FC between the vmPFC and the frontoparietal network supports this relationship with impairments in representation of goal-directed control and executive impairments (
      • Sebold M.
      • Nebe S.
      • Garbusow M.
      • Guggenmos M.
      • Schad D.J.
      • Beck A.
      • et al.
      When habits are dangerous: Alcohol expectancies and habitual decision making predict relapse in alcohol dependence.
      ,
      • Voon V.
      • Reiter A.
      • Sebold M.
      • Groman S.
      Model-based control in dimensional psychiatry.
      ,
      • Gan G.
      • Guevara A.
      • Marxen M.
      • Neumann M.
      • Jünger E.
      • Kobiella A.
      • et al.
      Alcohol-induced impairment of inhibitory control is linked to attenuated brain responses in right fronto-temporal cortex.
      ). Previous studies have found that reduced gray matter concentration/volume in the frontal executive control network are risk factors for alcohol misuse (
      • Baranger D.A.A.
      • Demers C.H.
      • Elsayed N.M.
      • Knodt A.R.
      • Radtke S.R.
      • Desmarais A.
      • et al.
      Convergent evidence for predispositional effects of brain gray matter volume on alcohol consumption.
      ,
      • Whelan R.
      • Watts R.
      • Orr C.A.
      • Althoff R.R.
      • Artiges E.
      • Banaschewski T.
      • et al.
      Neuropsychosocial profiles of current and future adolescent alcohol misusers.
      ,
      • Squeglia L.M.
      • Ball T.M.
      • Jacobus J.
      • Brumback T.
      • McKenna B.S.
      • Nguyen-Louie T.T.
      • et al.
      Neural predictors of initiating alcohol use during adolescence [published correction appears in Am J Psychiatry 2017; 174:80].
      ), and our study supports the prospective resilience role of the control network in alcohol use from the connectivity-based analysis. A recent co-twin control study suggested that alcohol use can result in reduced cortical thickness in the frontal cognitive control and salience networks by comparing alcohol use twins with co-twins who drank less (
      • Harper J.
      • Malone S.M.
      • Wilson S.
      • Hunt R.H.
      • Thomas K.M.
      • Iacono W.G.
      The effects of alcohol and cannabis use on the cortical thickness of cognitive control and salience brain networks in emerging adulthood: A co-twin control study.
      ). Longitudinal studies in adolescents and young adults have reported extensive evidence that lower frontal volume is related to predisposition toward alcohol misuse (
      • Lees B.
      • Garcia A.M.
      • Debenham J.
      • Kirkland A.E.
      • Bryant B.E.
      • Mewton L.
      • Squeglia L.M.
      Promising vulnerability markers of substance use and misuse: A review of human neurobehavioral studies.
      ). Using an alcohol-related Pavlovian-to-instrumental transfer paradigm, the activation of medial frontal related to the relapse of use in young adults (
      • Sekutowicz M.
      • Guggenmos M.
      • Kuitunen-Paul S.
      • Garbusow M.
      • Sebold M.
      • Pelz P.
      • et al.
      Neural response patterns during Pavlovian-to-instrumental transfer predict alcohol relapse and young adult drinking.
      ). Our study suggested that the reverse direction, that higher connectivity between the vmPFC and the executive control network associates with reduction in alcohol use in older adults, is also plausible, indicating potentially modifiable brain networks for clinical intervention.
      The limbic substrates’ relationship with subsequent behaviors highlights the critical role of incentive motivation and negative emotionality theories and extinction processes in alcohol misuse (
      • Voon V.
      • Grodin E.
      • Mandali A.
      • Morris L.
      • Doñamayor N.
      • Weidacker K.
      • et al.
      Addictions NeuroImaging Assessment (ANIA): Towards an integrative framework for alcohol use disorder.
      ). Reward-related ventral striatal activity and threat-related amygdala activity are correlated with drinking initiation in the first stage of alcohol addiction (
      • Koob G.F.
      • Volkow N.D.
      Neurobiology of addiction: A neurocircuitry analysis.
      ). Increased dopamine and opioid transmission in basal ganglia and extended amygdala are a well-documented finding in driving the rewarding properties in alcohol consumption in humans and animals (
      • Koob G.F.
      • Volkow N.D.
      Neurobiology of addiction: A neurocircuitry analysis.
      ). Stress cues and negative images decrease limbic and prefrontal activity and intrinsic connectivity in abstinent subjects with AUD (
      • Yang H.
      • Devous M.D.
      • Briggs R.W.
      • Spence J.S.
      • Xiao H.
      • Kreyling N.
      • Adinoff B.
      Altered neural processing of threat in alcohol-dependent men.
      ,
      • O’Daly O.G.
      • Trick L.
      • Scaife J.
      • Marshall J.
      • Ball D.
      • Phillips M.L.
      • et al.
      Withdrawal-associated increases and decreases in functional neural connectivity associated with altered emotional regulation in alcoholism.
      ). Abstinence in AUD is also associated with lower striatal D2/3 (
      • Martinez D.
      • Gil R.
      • Slifstein M.
      • Hwang D.R.
      • Huang Y.
      • Perez A.
      • et al.
      Alcohol dependence is associated with blunted dopamine transmission in the ventral striatum.
      ), increased μ-opioid receptor availability (
      • Weerts E.M.
      • Wand G.S.
      • Kuwabara H.
      • Munro C.A.
      • Dannals R.F.
      • Hilton J.
      • et al.
      Positron emission tomography imaging of mu- and delta-opioid receptor binding in alcohol-dependent and healthy control subjects.
      ), and blunted amphetamine-induced dopamine and dexamphetamine-induced opioid striatal release (
      • Turton S.
      • Myers J.F.
      • Mick I.
      • Colasanti A.
      • Venkataraman A.
      • Durant C.
      • et al.
      Blunted endogenous opioid release following an oral dexamphetamine challenge in abstinent alcohol-dependent individuals.
      ).
      Our effect size is from small (cross-sectional and longitudinal analyses of all participants) to medium (longitudinal analysis of the frequent drinking group). The relationship between the sample size and effect size in large datasets has been examined (
      • Marek S.
      • Tervo-Clemmens B.
      • Calabro F.J.
      • Montez D.F.
      • Kay B.P.
      • Hatoum A.S.
      • et al.
      Towards reproducible brain-wide association studies.
      ), showing that as sample size increased, the effect size stabilized at around r = 0.01 across all brainwide associations with behaviors. The strongest correlation between a brain metric and a behavioral measure was |r| = 0.16, using Adolescent Brain Cognitive Development Study 3 (n = 11,878) (
      • Marek S.
      • Tervo-Clemmens B.
      • Calabro F.J.
      • Montez D.F.
      • Kay B.P.
      • Hatoum A.S.
      • et al.
      Towards reproducible brain-wide association studies.
      ). Our effect size matches their findings. In addition, there is a critical distinction between population risk and individual risk. A statistically significant relationship might occur at the population level but might reveal very little about the likely impact on an individual member of the population (
      • Bellinger D.C.
      Interpretation of small effect sizes in occupational and environmental neurotoxicology: Individual versus population risk.
      ). In a population-level study, the impact of a factor depends not only on the magnitude (effect size) but also on the distribution of the factor. Given the common use of alcohol, the small effect size may have a considerable impact at the population level (
      • Bellinger D.C.
      Interpretation of small effect sizes in occupational and environmental neurotoxicology: Individual versus population risk.
      ).

      Limitations and Conclusions

      The Biobank data represents a normative population from 45 to 81 years old with a range of alcohol use and may not be generalizable to a younger population or more severe alcohol misuse. Restricted by the available behavioral tests provided by the Biobank, we used alcohol intake frequency instead of AUDIT scores in the longitudinal analysis. However, we show a strong correlation between AUDIT and the first time point alcohol intake frequency scores, and the same brain correlates are found using either measure. We have two time points of behavioral scores as the longitudinal measure because the Biobank tested twice prospectively, which ideally could include more time points. For the longitudinal analysis, GRF-corrected voxel p < .01 is on the liberal side (
      • Woo C.W.
      • Krishnan A.
      • Wager T.D.
      Cluster-extent based thresholding in fMRI analyses: Pitfalls and recommendations.
      ). However, we would like to highlight that it is based on a large sample with a medium effect size, suggesting that the effect is meaningful and not trivial.
      Put together, we show a dissociation between dorsal and ventral mesial prefrontal limbic substrates and their core cognitive processes in adult alcohol misuse cross-sectionally or longitudinally. These findings have important implications for anatomical targeting for novel neuromodulatory approaches.

      Acknowledgments and Disclosures

      This study is supported by a Medical Research Council Senior Clinical Fellowship (Grant No. MR/P008747/1 [to VV]). NS is supported by an Academic Clinical Fellowship from University of Cambridge/CPFT. Data were curated and analysed using a computational facility funded by a Medical Research Council research infrastructure award (MR/M009041/1) to the School of Clinical Medicine, University of Cambridge, and supported by the mental health theme of the NIHR Cambridge Biomedical Research Centre. The views expressed are those of the authors and not necessarily those of the NIH, NHS, NIHR, or Department of Health and Social Care. This research has been conducted using the UK Biobank Resource under Application Number 64044. The authors thank the anonymous reviewers for helping improve the quality of the article.
      The authors report no biomedical financial interests or potential conflicts of interest.

      Supplementary Material

      References

        • Hasin D.S.
        • Stinson F.S.
        • Ogburn E.
        • Grant B.F.
        Prevalence, correlates, disability, and comorbidity of DSM-IV alcohol abuse and dependence in the United States: Results from the National Epidemiologic Survey on Alcohol and Related Conditions.
        Arch Gen Psychiatry. 2007; 64: 830-842
        • World Health Organization
        Global Status Report on Alcohol and Health 2018.
        World Health Organization, Geneva2019
        • Burton R.
        • Henn C.
        • Lavoie D.
        • Wolff A.
        • Marsden J.
        • Sheron N.
        • et al.
        The Public Health Burden of Alcohol and the Effectiveness and Cost-Effectiveness of Alcohol Control Policies: An Evidence Review.
        Public Health England, London2016
        • Kwako L.E.
        • Momenan R.
        • Litten R.Z.
        • Koob G.F.
        • Goldman D.
        Addictions Neuroclinical Assessment: A neuroscience-based framework for addictive disorders.
        Biol Psychiatry. 2016; 80: 179-189
        • Voon V.
        • Grodin E.
        • Mandali A.
        • Morris L.
        • Doñamayor N.
        • Weidacker K.
        • et al.
        Addictions NeuroImaging Assessment (ANIA): Towards an integrative framework for alcohol use disorder.
        Neurosci Biobehav Rev. 2020; 113: 492-506
        • Baranger D.A.A.
        • Demers C.H.
        • Elsayed N.M.
        • Knodt A.R.
        • Radtke S.R.
        • Desmarais A.
        • et al.
        Convergent evidence for predispositional effects of brain gray matter volume on alcohol consumption.
        Biol Psychiatry. 2020; 87: 645-655
        • Whelan R.
        • Watts R.
        • Orr C.A.
        • Althoff R.R.
        • Artiges E.
        • Banaschewski T.
        • et al.
        Neuropsychosocial profiles of current and future adolescent alcohol misusers.
        Nature. 2014; 512: 185-189
        • Squeglia L.M.
        • Ball T.M.
        • Jacobus J.
        • Brumback T.
        • McKenna B.S.
        • Nguyen-Louie T.T.
        • et al.
        Neural predictors of initiating alcohol use during adolescence [published correction appears in Am J Psychiatry 2017; 174:80].
        Am J Psychiatry. 2017; 174: 172-185
        • Norman A.L.
        • Pulido C.
        • Squeglia L.M.
        • Spadoni A.D.
        • Paulus M.P.
        • Tapert S.F.
        Neural activation during inhibition predicts initiation of substance use in adolescence.
        Drug Alcohol Depend. 2011; 119: 216-223
        • Wetherill R.R.
        • Squeglia L.M.
        • Yang T.T.
        • Tapert S.F.
        A longitudinal examination of adolescent response inhibition: Neural differences before and after the initiation of heavy drinking.
        Psychopharmacology (Berl). 2013; 230: 663-671
        • Heitzeg M.M.
        • Nigg J.T.
        • Hardee J.E.
        • Soules M.
        • Steinberg D.
        • Zubieta J.K.
        • Zucker R.A.
        Left middle frontal gyrus response to inhibitory errors in children prospectively predicts early problem substance use.
        Drug Alcohol Depend. 2014; 141: 51-57
        • Nguyen-Louie T.T.
        • Courtney K.E.
        • Squeglia L.M.
        • Bagot K.
        • Eberson S.
        • Migliorini R.
        • et al.
        Prospective changes in neural alcohol cue reactivity in at-risk adolescents.
        Brain Imaging Behav. 2018; 12: 931-941
        • Sudlow C.
        • Gallacher J.
        • Allen N.
        • Beral V.
        • Burton P.
        • Danesh J.
        • et al.
        UK Biobank: An open access resource for identifying the causes of a wide range of complex diseases of middle and old age.
        PLoS Med. 2015; 12e1001779
        • Saunders J.B.
        • Aasland O.G.
        • Babor T.F.
        • de la Fuente J.R.
        • Grant M.
        Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO Collaborative Project on Early Detection of Persons With Harmful Alcohol Consumption–II.
        Addiction. 1993; 88: 791-804
        • Allen E.A.
        • Pasley B.N.
        • Duong T.
        • Freeman R.D.
        Transcranial magnetic stimulation elicits coupled neural and hemodynamic consequences.
        Science. 2007; 317: 1918-1921
        • Yan C.G.
        • Wang X.D.
        • Zuo X.N.
        • Zang Y.F.
        DPABI: Data processing & analysis for (resting-state) brain imaging.
        Neuroinformatics. 2016; 14: 339-351
        • Xia M.
        • Wang J.
        • He Y.
        BrainNet Viewer: A network visualization tool for human brain connectomics.
        PLoS One. 2013; 8e68910
        • Wang J.
        • Wang X.
        • Xia M.
        • Liao X.
        • Evans A.
        • He Y.
        GRETNA: A graph theoretical network analysis toolbox for imaging connectomics [published correction appears in Front Hum Neurosci 2015; 9:458].
        Front Hum Neurosci. 2015; 9: 386
        • Menon V.
        Large-scale brain networks and psychopathology: A unifying triple network model.
        Trends Cogn Sci. 2011; 15: 483-506
        • Schacht J.P.
        • Anton R.F.
        • Myrick H.
        Functional neuroimaging studies of alcohol cue reactivity: A quantitative meta-analysis and systematic review.
        Addict Biol. 2013; 18: 121-133
        • Yücel M.
        • Lubman D.I.
        • Harrison B.J.
        • Fornito A.
        • Allen N.B.
        • Wellard R.M.
        • et al.
        A combined spectroscopic and functional MRI investigation of the dorsal anterior cingulate region in opiate addiction.
        Mol Psychiatry. 2007; 12 (691-702): 611
        • Kaufman J.N.
        • Ross T.J.
        • Stein E.A.
        • Garavan H.
        Cingulate hypoactivity in cocaine users during a GO-NOGO task as revealed by event-related functional magnetic resonance imaging.
        J Neurosci. 2003; 23: 7839-7843
        • Grecco G.G.
        • Chumin E.J.
        • Dzemidzic M.
        • Cheng H.
        • Finn P.
        • Newman S.
        • et al.
        Anterior cingulate cortex metabolites and white matter microstructure: A multimodal study of emergent alcohol use disorder.
        Brain Imaging Behav. 2021; 15: 2436-2444
        • Moffett J.R.
        • Ross B.
        • Arun P.
        • Madhavarao C.N.
        • Namboodiri A.M.A.
        N-acetylaspartate in the CNS: From neurodiagnostics to neurobiology.
        Prog Neurobiol. 2007; 81: 89-131
        • Pfefferbaum A.
        • Sullivan E.V.
        • Rosenbloom M.J.
        • Mathalon D.H.
        • Lim K.O.
        A controlled study of cortical gray matter and ventricular changes in alcoholic men over a 5-year interval.
        Arch Gen Psychiatry. 1998; 55: 905-912
        • Pfefferbaum A.
        • Lim K.O.
        • Zipursky R.B.
        • Mathalon D.H.
        • Rosenbloom M.J.
        • Lane B.
        • et al.
        Brain gray and white matter volume loss accelerates with aging in chronic alcoholics: A quantitative MRI study.
        Alcohol Clin Exp Res. 1992; 16: 1078-1089
        • Pfefferbaum A.
        • Rosenbloom M.J.
        • Chu W.
        • Sassoon S.A.
        • Rohlfing T.
        • Pohl K.M.
        • et al.
        White matter microstructural recovery with abstinence and decline with relapse in alcohol dependence interacts with normal ageing: A controlled longitudinal DTI study.
        Lancet Psychiatry. 2014; 1: 202-212
        • Thayer R.E.
        • Hagerty S.L.
        • Sabbineni A.
        • Claus E.D.
        • Hutchison K.E.
        • Weiland B.J.
        Negative and interactive effects of sex, aging, and alcohol abuse on gray matter morphometry.
        Hum Brain Mapp. 2016; 37: 2276-2292
        • Evangelou E.
        • Suzuki H.
        • Bai W.
        • Pazoki R.
        • Gao H.
        • Matthews P.M.
        • Elliott P.
        Alcohol consumption in the general population is associated with structural changes in multiple organ systems.
        Elife. 2021; 10e65325
        • Topiwala A.
        • Ebmeier K.P.
        • Maullin-Sapey T.
        • Nichols T.E.
        No safe level of alcohol consumption for brain health: Observational cohort study of 25,378 UK Biobank participants.
        medRxiv. 2021; https://doi.org/10.1101/2021.05.10.21256931
        • Yang X.
        • Tian F.
        • Zhang H.
        • Zeng J.
        • Chen T.
        • Wang S.
        • et al.
        Cortical and subcortical gray matter shrinkage in alcohol-use disorders: A voxel-based meta-analysis.
        Neurosci Biobehav Rev. 2016; 66: 92-103
        • Pardo J.V.
        • Lee J.T.
        • Sheikh S.A.
        • Surerus-Johnson C.
        • Shah H.
        • Munch K.R.
        • et al.
        Where the brain grows old: Decline in anterior cingulate and medial prefrontal function with normal aging.
        Neuroimage. 2007; 35: 1231-1237
        • Madole J.W.
        • Ritchie S.J.
        • Cox S.R.
        • Buchanan C.R.
        • Hernández M.V.
        • Maniega S.M.
        • et al.
        Aging-sensitive networks within the human structural connectome are implicated in late-life cognitive declines.
        Biol Psychiatry. 2021; 89: 795-806
        • Ning K.
        • Zhao L.
        • Matloff W.
        • Sun F.
        • Toga A.W.
        Association of relative brain age with tobacco smoking, alcohol consumption, and genetic variants.
        Sci Rep. 2020; 10: 10
        • Sullivan E.V.
        • Pfefferbaum A.
        Brain-behavior relations and effects of aging and common comorbidities in alcohol use disorder: A review.
        Neuropsychology. 2019; 33: 760-780
        • Hong L.E.
        • Gu H.
        • Yang Y.
        • Ross T.J.
        • Salmeron B.J.
        • Buchholz B.
        • et al.
        Association of nicotine addiction and nicotine’s actions with separate cingulate cortex functional circuits.
        Arch Gen Psychiatry. 2009; 66: 431-441
        • Courtney K.E.
        • Ghahremani D.G.
        • Ray L.A.
        Fronto-striatal functional connectivity during response inhibition in alcohol dependence.
        Addict Biol. 2013; 18: 593-604
        • Thoma R.J.
        • Monnig M.A.
        • Lysne P.A.
        • Ruhl D.A.
        • Pommy J.A.
        • Bogenschutz M.
        • et al.
        Adolescent substance abuse: The effects of alcohol and marijuana on neuropsychological performance.
        Alcohol Clin Exp Res. 2011; 35: 39-46
        • O’Reilly R.C.
        The what and how of prefrontal cortical organization.
        Trends Neurosci. 2010; 33: 355-361
        • Corbetta M.
        • Shulman G.L.
        Control of goal-directed and stimulus-driven attention in the brain.
        Nat Rev Neurosci. 2002; 3: 201-215
        • Marinkovic K.
        • Oscar-Berman M.
        • Urban T.
        • O’Reilly C.E.
        • Howard J.A.
        • Sawyer K.
        • Harris G.J.
        Alcoholism and dampened temporal limbic activation to emotional faces.
        Alcohol Clin Exp Res. 2009; 33: 1880-1892
        • Volkow N.D.
        The reality of comorbidity: Depression and drug abuse.
        Biol Psychiatry. 2004; 56: 714-717
        • Wiers C.E.
        • Shumay E.
        • Volkow N.D.
        • Frieling H.
        • Kotsiari A.
        • Lindenmeyer J.
        • et al.
        Effects of depressive symptoms and peripheral DAT methylation on neural reactivity to alcohol cues in alcoholism.
        Transl Psychiatry. 2015; 5: e648
        • Sjoerds Z.
        • de Wit S.
        • van den Brink W.
        • Robbins T.W.
        • Beekman A.T.F.
        • Penninx B.W.J.H.
        • Veltman D.J.
        Behavioral and neuroimaging evidence for overreliance on habit learning in alcohol-dependent patients.
        Transl Psychiatry. 2013; 3: e337
        • Ebrahimi C.
        • Koch S.P.
        • Pietrock C.
        • Fydrich T.
        • Heinz A.
        • Schlagenhauf F.
        Opposing roles for amygdala and vmPFC in the return of appetitive conditioned responses in humans.
        Transl Psychiatry. 2019; 9: 148
        • Sebold M.
        • Nebe S.
        • Garbusow M.
        • Guggenmos M.
        • Schad D.J.
        • Beck A.
        • et al.
        When habits are dangerous: Alcohol expectancies and habitual decision making predict relapse in alcohol dependence.
        Biol Psychiatry. 2017; 82: 847-856
        • Voon V.
        • Reiter A.
        • Sebold M.
        • Groman S.
        Model-based control in dimensional psychiatry.
        Biol Psychiatry. 2017; 82: 391-400
        • Gan G.
        • Guevara A.
        • Marxen M.
        • Neumann M.
        • Jünger E.
        • Kobiella A.
        • et al.
        Alcohol-induced impairment of inhibitory control is linked to attenuated brain responses in right fronto-temporal cortex.
        Biol Psychiatry. 2014; 76: 698-707
        • Harper J.
        • Malone S.M.
        • Wilson S.
        • Hunt R.H.
        • Thomas K.M.
        • Iacono W.G.
        The effects of alcohol and cannabis use on the cortical thickness of cognitive control and salience brain networks in emerging adulthood: A co-twin control study.
        Biol Psychiatry. 2021; 89: 1012-1022
        • Lees B.
        • Garcia A.M.
        • Debenham J.
        • Kirkland A.E.
        • Bryant B.E.
        • Mewton L.
        • Squeglia L.M.
        Promising vulnerability markers of substance use and misuse: A review of human neurobehavioral studies.
        Neuropharmacology. 2021; 187: 108500
        • Sekutowicz M.
        • Guggenmos M.
        • Kuitunen-Paul S.
        • Garbusow M.
        • Sebold M.
        • Pelz P.
        • et al.
        Neural response patterns during Pavlovian-to-instrumental transfer predict alcohol relapse and young adult drinking.
        Biol Psychiatry. 2019; 86: 857-863
        • Koob G.F.
        • Volkow N.D.
        Neurobiology of addiction: A neurocircuitry analysis.
        Lancet Psychiatry. 2016; 3: 760-773
        • Yang H.
        • Devous M.D.
        • Briggs R.W.
        • Spence J.S.
        • Xiao H.
        • Kreyling N.
        • Adinoff B.
        Altered neural processing of threat in alcohol-dependent men.
        Alcohol Clin Exp Res. 2013; 37: 2029-2038
        • O’Daly O.G.
        • Trick L.
        • Scaife J.
        • Marshall J.
        • Ball D.
        • Phillips M.L.
        • et al.
        Withdrawal-associated increases and decreases in functional neural connectivity associated with altered emotional regulation in alcoholism.
        Neuropsychopharmacology. 2012; 37: 2267-2276
        • Martinez D.
        • Gil R.
        • Slifstein M.
        • Hwang D.R.
        • Huang Y.
        • Perez A.
        • et al.
        Alcohol dependence is associated with blunted dopamine transmission in the ventral striatum.
        Biol Psychiatry. 2005; 58: 779-786
        • Weerts E.M.
        • Wand G.S.
        • Kuwabara H.
        • Munro C.A.
        • Dannals R.F.
        • Hilton J.
        • et al.
        Positron emission tomography imaging of mu- and delta-opioid receptor binding in alcohol-dependent and healthy control subjects.
        Alcohol Clin Exp Res. 2011; 35: 2162-2173
        • Turton S.
        • Myers J.F.
        • Mick I.
        • Colasanti A.
        • Venkataraman A.
        • Durant C.
        • et al.
        Blunted endogenous opioid release following an oral dexamphetamine challenge in abstinent alcohol-dependent individuals.
        Mol Psychiatry. 2020; 25: 1749-1758
        • Marek S.
        • Tervo-Clemmens B.
        • Calabro F.J.
        • Montez D.F.
        • Kay B.P.
        • Hatoum A.S.
        • et al.
        Towards reproducible brain-wide association studies.
        bioRxiv. 2020; https://doi.org/10.1101/2020.08.21.257758
        • Bellinger D.C.
        Interpretation of small effect sizes in occupational and environmental neurotoxicology: Individual versus population risk.
        Neurotoxicology. 2007; 28: 245-251
        • Woo C.W.
        • Krishnan A.
        • Wager T.D.
        Cluster-extent based thresholding in fMRI analyses: Pitfalls and recommendations.
        Neuroimage. 2014; 91: 412-419