Received: 21 July 2019 Revised: 31 July 2020 Accepted: 5 October 2020 DOI: 10.1111/adb.12980

ORIGINAL ARTICLE

Chronic alcohol exposure differentially modulates structural and functional properties of amygdala: A cross-sectional study

Csaba Orban1,2,3 | John McGonigle1 | Remy S.A. Flechais1 | Louise M. Paterson1 | Rebecca Elliott4 | David Erritzoe1 | Karen D. Ersche5,6 | Anna Murphy4 | Liam J. Nestor1,6 | Filippo Passetti1,5,6 | Laurence J. Reed1 | Andre S. Ribeiro1 | Dana G. Smith5,7 | John Suckling5,6,8 | Eleanor M. Taylor4 | Adam D. Waldman9 | Victoria C. Wing1 | J.F. William Deakin4 | Trevor W. Robbins5,7 | David J. Nutt1 | Anne R. Lingford-Hughes1 | ICCAM Platform*

1Neuropsychopharmacology Unit, Centre for Psychiatry, Division of Brain Sciences, Imperial College London, London, UK

2Centre for Sleep and Cognition, National University of Singapore, Singapore

3N.1 Institute for Health, ECE & CIRC, National University of Singapore, Singapore

4Neuroscience and Psychiatry Unit, Institute of Brain, Behaviour and Mental Health, The University of Manchester, Manchester, UK

5Behavioural and Clinical Neuroscience Institute, , Cambridge, UK

6Department of Psychiatry, University of Cambridge, Cambridge, UK

7Department of Psychology, University of Cambridge, Cambridge, UK

8Cambridgeshire and Peterborough NHS Foundation Trust, Cambridgeshire, UK

9Centre for Neuroinflammation and Neurodegeneration, Imperial College London, London, UK

Correspondence Anne R. Lingford-Hughes, Abstract Neuropsychopharmacology Unit, Centre for Animal models have shown that chronic alcohol exposure is associated with persistent Psychiatry, Division of Brain Sciences, Imperial College, London, UK. neuroadaptations in amygdala synaptic function, whereas human studies have consis- Email: [email protected] tently reported amygdala grey-matter volume (GMV) reductions in alcohol dependent

Funding information patients (ADP). We hypothesised that chronic alcohol use associated with GlaxoSmithKline; Medical Research Council, neuroadaptations may entail a reconfiguration of the amygdala's functional interactions Grant/Award Number: G1000018; Singapore National Research Foundation (NRF) and that these mechanisms may be affected by structural atrophy. We compared amyg- Fellowship; NUS YIA; Singapore NMRC, Grant/ dala resting state functional connectivity (RSFC) using a whole brain seed-based approach Award Number: CBRG/0088/2015; NUS SOM Aspiration Fund, Grant/Award Number: and amygdala GMV in abstinent ADP (n = 20) and healthy controls (HC; n =39),balanced R185000271720; NUS Strategic Research, for age, gender and levels of head motion. The potential moderating influence of age, Grant/Award Number: DPRT/944/09/14; Singapore MOE Tier 2, Grant/Award Number: cumulative alcohol exposure, abstinence length and head motion was further examined in MOE2014-T2-2-016; Medical Research the two groups separately using correlational analyses. We found increased amygdala Council Doctoral Training Program Studentship RSFC with substantia nigra/ventral tegmental area (SN/VTA) in ADP compared with HC. As expected, amygdala GMV was lower in ADP. Multiple regression analyses of the

* ICCAM Platform collaborators: David Nutt, Anne Lingford-Hughes, Louise Paterson, John McGonigle, Remy Flechais, Csaba Orban, Bill Deakin, Rebecca Elliott, Anna Murphy, Eleanor Taylor, Trevor Robbins, Karen Ersche, John Suckling, Dana Smith, Laurence Reed, Filippo Passetti, Luca Faravelli, David Erritzoe, Inge Mick, Nicola Kalk, Adam W aldman, Liam Nestor, Shankar Kuchibatla, Venkataramana Boyapati, Antonio Metastasio, Yetunde Faluyi, Emilio Fernandez-Egea, Sanja Abbott, , Valerie Voon, Ilan Rabiner.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2020 The Authors. Addiction Biology published by John Wiley & Sons Ltd on behalf of Society for the Study of Addiction

Addiction Biology. 2020;e12980. wileyonlinelibrary.com/journal/adb 1of10 https://doi.org/10.1111/adb.12980 2of10 ORBAN ET AL.

ADP group showed that amygdala-SN/VTA RSFC increases were primarily associated with cumulative alcohol exposure rather than age, whereas amygdala GMV reductions were primarily associated with age rather than cumulative alcohol exposure. The same association between age and amygdala GMV was not observed amongst HC. Importantly, amygdala GMV and amygdala-SN/VTA RSFC were uncorrelated in ADP, and neither measure was correlated with abstinence length. These results suggest that chronic alcohol exposure is associated with persistent elevations in amygdala-SN/VTA RSFC and acceler- ated age-related grey-matter atrophy through potentially distinct mechanisms.

KEYWORDS alcohol, amygdala, fMRI

1 | INTRODUCTION related to structural atrophy.13,14 Smaller amygdala grey-matter volume (GMV) is widely documented in alcohol dependent populations (ADP) Chronic, excessive alcohol consumption is associated with widespread and15–18 is associated with a greater likelihood of relapse,16 yet its rela- homeostatic neuroadaptations, whose effects become unmasked during tionship to amygdala RSFC has not been specifically investigated. withdrawal as intense autonomic and emotional disturbances, which in We present here data from the ICCAM Experimental Medicine turn negatively reinforce alcohol consumption.1 Relapse is common Platform19 comparing amygdala RSFC and amygdala GMV in absti- within the initial months of abstinence,2 whereas in those who achieve nent ADP with age-matched healthy controls (HC). Because animal longer-term abstinence, symptoms of anxiety and dysphoria often per- models implicate multiple effector regions which may be chronically sist.3 An influential model proposes that cumulative alcohol exposure modulated by local neuroadaptations in the amygdala,7 we applied a elicits an allostatic load, a persistent shift in the baseline function of whole brain approach to our RSFC analyses. We hypothesised that motivational and stress circuits, thereby predisposing individuals to longer cumulative alcohol exposure in ADP would be associated with relapse to their original state of alcohol consumption.4 greater alterations in amygdala RSFC and reduced amygdala GMV. Animal models have highlighted the amygdala as a critical site for allostatic changes which drive escalation of alcohol intake and stress- induced reinstatement.5 The amygdala is an evolutionarily well- 2 | MATERIALS AND METHODS conserved complex of subcortical nuclei involved in fear conditioning, acquisition of stimulus-reward associations and bodily responses to 2.1 | Study description stressors.6 Chronic alcohol exposure has been shown to augment GABA release in the central amygdala7 via upregulation of ‘pro-stress’ All data presented here were collected during the first visit of the neuropeptides.8 The resulting increase in local inhibition in the central ICCAM platform study: a three-center (Imperial College London, Univer- amygdala has been hypothesised to drive the net disinhibition of sity of Cambridge and University of Manchester), multisession neuroim- downstream effectors, for example, hypothalamic, dopaminergic or aging study to test the effectiveness of compounds in modulating noradrenergic brainstem nuclei, thereby providing a possible feed- putative brain pathways in addiction processes.19 During the study visit, forward mechanism for maintaining positive and negative reinforce- 68 HC and 28 ADP underwent a 5-min resting state functional mag- ment of alcohol consumption.5,9,10 netic resonance imaging (fMRI) scan with eyes closed. This visit did not Although animal studies have made great strides in elucidating local include drug administration. Participants provided their written neuroadaptations within circumscribed brain regions, a more compre- informed consent in accordance with the Declaration of Helsinki. Ethical hensive neurobiological model of alcohol dependence requires a approval was obtained from West London and GTAC NRES committee systems-level mapping of interactions between multiple brain regions in (11/H0707/9), and R&D approval was obtained from relevant Research human studies. Resting state functional connectivity (RSFC) has been Governance and PIC (Participant Identification Centre) sites. increasingly applied to characterise network-level differences in sub- stance dependent populations. Studies that have looked at amygdala RSFC in alcohol,11 cocaine12 and opioid-dependent patients13 have 2.2 | Inclusion and exclusion criteria reported lower amygdala RSFC relative to healthy participants. How- ever, to our knowledge, no study has specifically explored chronic alter- Inclusion criteria required ADP to meet DSM-IV criteria for prior alco- ations in amygdala RSFC with reference to cumulative alcohol exposure hol dependence, without current or prior diagnoses of dependence on in abstinent alcohol-dependent patients. Additionally, there is evidence other drugs, except nicotine. ADP were also required to have been that alterations in RSFC in substance-dependent populations may be abstinent for at least 1 month prior to scanning, and none were ORBAN ET AL. 3of10 currently in treatment or receiving medication for alcohol depen- (SSAI22). IQ and handedness were assessed using Weschler's Adult dence. HC were not permitted to any history of drug or alcohol Intelligence Scale23 and The Edinburgh Inventory,24 respectively. dependence, except for nicotine. Resting heart rate (HR), systolic and diastolic blood pressure were The following exclusion criteria were applied for both ADP and HC: measured in a sitting position on the same day prior to the scan. (i) current primary axis I diagnosis; (ii) current or past history of enduring severe mental illness, including psychosis; however, secondary or life- time history of depression or anxiety was permitted due to high comor- 2.4 | MRI acquisition bidity; (iii) history of significant neurological diagnosis; (iv) prescription medication that could interfere with study integrity or safety; (v) any Full details of the data acquisition, preprocessing and statistical MRI contraindication; and (vi) positive breath alcohol or positive urine modelling are provided in the Supporting Information. A brief over- drug test on scan day (amphetamines, barbiturates, cocaine, opiates and view is given here. All sites operated MRI machines with a main mag- benzodiazepines). On study days, smokers were permitted to smoke netic field of three tesla (T)—Siemens Tim Trio systems in London and cigarettes up to 1 h prior to scanning in order to avoid withdrawal. Cambridge and a Philips Achieva in Manchester. High-resolution In the analyses presented here, seven participants (3 HC and structural scans were acquired using magnetisation-prepared rapid 4 ADP) were excluded due to containing less than 5 min of resting gradient echo (MPRAGE) sequences to enable registration to a stan- state data after censoring of volumes with excessive head motion dard space. Functional imaging was performed using a multiecho gra- (>0.5-mm framewise displacement). Four additional ADP were dient echo echoplanar imaging (EPI) sequence (TR = 2,000 ms, excluded: one due to corrupted acquisition and three due to failing to TE = 31 ms) with an in-plane resolution of 3.516 × 3.516 mm and a pass inclusion criterion for abstinence at the time of baseline scan. slice thickness of 3.000 mm. For the resting state scan, 183 volumes From the remaining pool of 65 HC, 39 were selected to maximise were acquired over 6 min (the first three volumes were subse- matching to the ADP group on age, gender, smoking status and head quently discarded to allow for T1 saturation effects). For each volume, motion in scanner (mean framewise displacement). 36 oblique axial slices (34 in Manchester) at an angle of around 30 to the anterior (AC) and posterior commissure (PC) line were acquired. Oblique acquisition enabled greater coverage of inferior regions at the 2.3 | Materials expense of signal dropout affecting the most superior 9 mm aspect of the brain. There was minimal variation in the time of day of scans A detailed account of lifetime drug and alcohol use of each participant across all three sites (earliest: noon, latest: 3:30 pm), thus minimising was obtained through a combination of clinical interviews, the potential effects of diurnal variation in resting state activity.25 For Alcohol, Smoking and Substance Involvement Screening Test20 and more detail on the acquisition parameters, see McGonigle et al.26 detailed timeline followback which codes drug and alcohol use per year of life. The presence of Axis I psychiatric diagnoses was screened using a summarised version of the Structured Clinical Interview for 2.5 | Preprocessing DSM-IV (MINI21), and a study physician undertook a further psychiat- ric history, including family history. Preprocessing of the MRI data was performed using a combination of A cumulative alcohol exposure measure was calculated for all AFNI,27 ADP taking into account the quantity, frequency and pattern of their FreeSurfer,28 ANTs,29 FSL30 and in-house Python code. drinking during their lifetime. For each year to be counted as ‘expo- Structural T1 images were intensity normalised, brain extracted sure’, male participants had to exceed 400 g of alcohol per week, or and nonlinearly registered to an MNI152 2-mm brain template. GMV have more than three drinking episodes of at least 70 g per week for maps were computed by deriving partial volume estimates of grey- at least 6 months of that year, and female participants had to exceed matter from T1 images, registering these to MNI152 space (using the 280 g per week or more than three drinking episodes of at least 56 g previously computed transformation matrices), applying Jacobian per week for at least 6 months of that year. For each ADP, a cumula- modulation and spatial smoothing at 6-mm FWHM. tive alcohol exposure score was calculated that was equivalent to the EPI images initially underwent despiking, slicing timing correction, number of years exceeding these criteria. rigid-body motion correction, brain extraction and boundary-based All psychiatric and substance dependence histories were registration to their corresponding native T1 image. EPI images were reviewed by two addiction psychiatrists (RSAF and FP) to ensure uni- spatially smoothed at 6-mm FWHM. After registration of EPI images, formity of diagnostic thresholds across sites and any discrepancies data underwent motion scrubbing (framewise displacement > 0.5 mm), arbitrated by a third addiction psychiatrist (ARLH). Abstinence length bandpassing (0.01–0.1 Hz) and detrending. Noise parameters was measured from last date of alcohol consumption. Cigarette pack- regressed out of the data included six motion parameters and mean age years (pack years) were defined as number of packs of cigarettes signal from the lateral ventricles (LV), draining veins (DV) and local (i.e., 20 cigarettes) smoked per day multiplied by the number of years white matter (WMlocal31). smoked. Symptoms of anxiety were characterised using Spielberger A bilateral amygdala region of interest was defined by placing two Trait Anxiety Scale (STAI18) and Spielberger State Anxiety Scale 5-mm radius spheres centred within activation clusters (MNI: 22, −4, 4of10 ORBAN ET AL.

−12) reported in a previously published activation likelihood estimation estimates, again, after controlling for effects of site and ICV (for GMV analysis of amygdala responsivity32 (Figure S1). Amygdala timeseries only). Linear relationships between clinical variables and brain measures were extracted from the fully preprocessed resting state EPI of each were examined using Pearson correlations and multiple regression. participant and modelled within the framework of a GLM. The resulting statistical maps provided a voxel-wise estimate of amygdala RSFC for each subject. This same amygdala region of interest was used both for 3 | RESULTS the resting state fMRI and amygdala GMV analyses (Figure S1). As an additional control, we repeated the GMV analysis with a This study examined amygdala RSFC and GMV in 20 ADP and 39 HC. structurally defined segmentation of the amygdala that more closely Demographic and clinical characteristics of the two groups are pres- estimated total amygdala volume than our spherical region of interest. ented in Table 1. No significant group differences were observed for Here, amygdala GMV (total amygdala GMV) was computed in native age, body weight, handedness, IQ score, resting HR, systolic and space by adding up voxel-wise grey-matter partial volume estimate diastolic blood pressure or head motion. ADP reported significantly values within individual-level bilateral masks of the amygdala derived fewer years in full time education (t(37) = −2.4, p = 0.02), more frequent using FreeSurfer. Again, whole brain volume and site effects were family history of alcohol dependence (χ2 =5.2,p = 0.02) and scored sig- regressed out before further analyses. nificantly higher on measures of trait (t(37) = 3.9, p < 0.001) and state anxiety (t(37) = 3.5, p < 0.001), compared with HC. Proportion of smokers were higher among ADP than HC (χ2 =4.1,p = 0.04), whereas 2.6 | Statistical modelling lifetime cigarette package years were not significantly different between smoker ADP and smoker HC (t(30) = 1.9, p =0.06). Group comparison of amygdala RSFC in ADP versus HC was implemented in a mixed-effects model, controlling for site (FSL's FLAME1). Results were thresholded using clusters determined by 3.1 | Amygdala-substantia nigra/ventral tegmental Z > 2.3 and a corrected cluster significance threshold of p < 0.05. Group area RSFC is positively correlated with cumulative comparison of amygdala GMV in ADP versus HC was carried out using alcohol exposure atwo-samplet test, after controlling for site and intracranial volume (ICV) via regression. Further analyses of amygdala RSFC or GMV in Comparison of voxel-wise amygdala RSFC between ADP and HC rev- regions of interest were undertaken using subject-level parameter ealed significantly elevated RSFC in patients in a single cluster located

TABLE 1 Demographics and clinical measures

ADP (n = 20) HC (n = 39)

Measures Mean ± SD Range Mean ± SD Range tp Age 46.0 ± 8.1 30 to 60 44.3 ± 7.4 31 to 56 0.8 0.43 Sex 15 male, 5 female 32 male, 7 female χ = 0.1 0.77 Body weight, kg 77.6 ± 17.9 49.6 to 117.0 77.6 ± 13.4 51.4 to 103.6 −0.01 0.99 Handedness score 50.0 ± 75.3 −100 to (+100) 51.9 ± 62.9 −100 to (+100) −0.1 0.92 Subject motion, FD in mm 0.20 ± 0.08 0.081 to 0.32 0.17 ± 0.06 0.089 to 0.31 1.5 0.15 WTAR IQ score 105.0 ± 7.7 91 to 118 108.0 ± 9.1 78 to 120 −1.2 0.22 Full time education, y 12.0 ± 2.4 9 to 19 13.6 ± 2.5 10 to 18 −2.4 0.02 Family history of alcohol dependence 7 positive, 13 negative 3 positive, 36 negative χ = 5.2 0.02 Cumulative alcohol exposure, y 21.7 ± 8.4 10 to 41 –– Abstinence, mo 11.2 ± 10.8 1 to 38 –– Smoking status 16 smoker, 4 nonsmoker 19 smoker, 20 nonsmoker χ = 4.1 0.04 Cigarette pack years among smokersa 30.0 ± 17.9 6.7 to 77.9 19.9 ± 11.1 0.1 to 41.6 1.9 0.06 Spielberger trait anxiety score 39.6 ± 12.5 21 to 72 29.4 ± 7.5 20 to 48 3.9 <0.001 Spielberger state anxiety score 33.4 ± 11.2 21 to 61 25.7 ± 6 20 to 48 3.5 <0.001 Resting heart rate, bpmb 70.1 ± 8.8 56 to 86 67.1 ± 10.4 42 to 97 1.1 0.3 Systolic blood pressure 127.4 ± 13.5 99 to 150 125 ± 14.5 93 to 164 0.6 0.6 Diastolic blood pressure 78.2 ± 9.0 62 to 95 78.6 ± 9.9 55 to 107 −0.2 0.9

Abbreviations: ADP, alcohol dependent patients; HC, healthy controls. aMissing data: 1 ADP, 2 HC. bMissing data: 1 HC. ORBAN ET AL. 5of10

FIGURE 1 Cumulative alcohol exposure is associated with elevated amygdala-SN/VTA RSFC resting state functional connectivity (RSFC) in alcohol dependent patients (ADP). (A) Whole brain voxel-wise group comparison shows significantly elevated amygdala RSFC in the midbrain for the ADP > healthy controls (HC) contrast. Significant voxels were determined by a threshold of Z > 2.3 and a corrected cluster significance threshold of p < 0.05. Slice coordinates are presented in MNI space. (B) Amygdala-SN/VTA RSFC is significantly increased in ADP relative to HC (t(57) = 3.3, p = 0.002). Group difference remains significant when excluding the ADP participant with the highest amygdala-SN/VTA RSFC (t(46) = 2.9, p = 0.005) or when excluding all participants with family history of alcohol dependence (t(47) = 2.5, p = 0.02). Bars denote group means. Legend denotes first-degree family history of alcohol dependence. (C) Amygdala-SN/VTA RSFC is positively correlated with cumulative alcohol exposure (r = 0.68, p = 0.001). PE, parameter estimate

3 in the midbrain (peak Z: 4.1, pcorrected < 0.05, 3,936 mm , MNI: 8, −20, nonsmoker ADP than the mean amygdala-SN/VTA RSFC of the −14), overlapping substantia nigra (SN) and ventral tegmental area nonsmoking HC group (Figure S2). Amygdala-SN/VTA RSFC did not (VTA) bilaterally (Figure 1A). No additional clusters showed significant significantly differ between smoker and nonsmoker HC subgroups increases or decreases in amygdala RSFC in ADP compared with nor between smoker and nonsmoker ADP. Furthermore, amygdala- HC. To refine the anatomical specificity of this midbrain cluster, an SN/VTA RSFC was not significantly correlated with abstinence SN/VTA region of interest was constructed (Figure S1) using a proba- length, head motion, state anxiety, trait anxiety, amygdala GMV nor bilistic atlas.33 Mean RSFC between amygdala and SN/VTA was sig- SN/VTA GMV across ADP. nificantly elevated in ADP compared with HC (t(57) = 3.3, p = 0.002; Figure 1B). This effect remained significant when excluding the ADP participant with the highest amygdala-SN/VTA RSFC (t(46) = 2.9, 3.2 | Accelerated age-related atrophy of amygdala in p = 0.005) or when comparing subgroups of ADP and HC with no ADP, but not in HC family history of alcohol dependence (t(47) = 2.5, p = 0.02). In ADP, amygdala-SN/VTA RSFC was positively correlated with ADP exhibited significantly lower amygdala GMV, in the same region both cumulative alcohol exposure (r = 0.68, p = 0.001; Figure 1C) of interest where RSFC with SN/VTA was found to be elevated and age (r = 0.49, p = 0.03; Figure S2), when examined separately. (Figure 1), relative to HC (t = −3.0, p = 0.004; Figure 2A). This effect However, when the two variables were entered into a multiple remained significant when examining subgroups without a family his- regression model (by themselves), only cumulative alcohol exposure tory of alcohol dependence (t(47) = −3.6, p = 0.0007). Age (Figure 2B; was significantly associated with amygdala-SN/VTA RSFC (t r = −0.63, p = 0.003) and cumulative alcohol exposure (r = −0.48, (17) = 2.7, p = 0.02), but not age (t(17) = −0.36, p = 0.7). To further p = 0.03; Figure S3) were both negatively correlated with amygdala rule out effects of GMV atrophy on RSFC, the voxel-wise amygdala GMV in ADP. In order to separate their relative contributions, the two RSFC group comparison was repeated while accounting for scanning measures were entered (by themselves) into a multiple regression. In site, amygdala GMV, total ICV and voxel-wise GMV in a single model. contrast to our findings with amygdala-SN/VTA RSFC, amygdala This more conservative model produced almost identical results GMV remained significantly associated with age (t(17) = −2.3, (Figure S2) to our original findings (Figure 1A). We also examined the p = 0.04), but not with cumulative alcohol exposure (t(17) = 0.2, potential impact of smoking status by comparing ADP (n = 20) with p = 0.85) in the multiple regression. The observed age-related decline smoker (n = 20) and nonsmoker (n = 19) subgroups of HC. ADP of amygdala GMV was specific to ADP, as HC did not show the same exhibited significantly greater amygdala-SN/VTA RSFC relative to pattern (Figure 2B). An analysis of variance (ANOVA) revealed a sig- both smoker and nonsmoker HC (Figure S2). As most ADP partici- nificant group-by-age interaction (F1,55 = 5.8, p = 0.02) for amygdala pants were smokers (16 out of 20), we were unable to split ADP into GMV. Amygdala GMV was significantly smaller in ADP relative to subgroups by smoking status for statistical analyses. Nonetheless, we both HC smokers (t(37) = −2.5, p = 0.02) and HC nonsmokers observed greater amygdala-SN/VTA RSFC in three out of the four (t(37) = −2.5, p = 0.02; Figure S3). Amygdala GMV was not 6of10 ORBAN ET AL.

FIGURE 2 Amygdala gray matter volume (GMV) exhibits accelerated age-related atrophy in alcohol dependent patients (ADP). (A) ADP showed significantly reduced amygdala GMV relative to HC (t(47) = −3.0, p = 0.004). Group difference remained significant when excluding those with family history of alcohol dependence (t = −3.6, p = 0.0007). Bars denote group means. Legend denotes first-degree family history of alcohol dependence. (B) Amygdala GMV was inversely correlated with age in ADP (r = −0.63, p = 0.003), but not in HC (r = −0.02, 0.87), exhibiting a significant group-by-age interaction (F1, 55 = 5.8, p = 0.02). PE, parameter estimate significantly different between smoker and nonsmoker HC subgroups correlated (r=0.79, p<0.001), that is, older ADP were more likely to nor between smoker and nonsmokers ADP. have accumulated lifetime years with heavy levels of alcohol con- Similar to our findings with amygdala-SN/VTA RSFC, amygdala sumption. However, cumulative alcohol exposure showed a stronger GMV was not significantly correlated with abstinence length in ADP. association with amygdala-SN/VTA, whereas age showed a stronger SN/VTA GMV was significantly smaller in ADP relative to HC association with amygdala GMV, than vice versa. Notably, amygdala (t(37) = −2.6, p = 0.01). However in ADP, SN/VTA GMV was not GMV and amygdala-SN/VTA RSFC were not significantly correlated significantly correlated with age, cumulative alcohol exposure or with each other (r=−0.15, p = 0.5). abstinence length.

3.4 | Control analyses 3.3 | No association between amygdala-SN/VTA RSFC and amygdala GMV Levene's tests were nonsignificant for both amygdala GMV and for amygdala-SN/VTA RSFC; thus, the observed group differences are The relationships between amygdala-SN/VTA RSFC, amygdala GMV, unlikely to have been driven by unequal group variances. We also age and cumulative alcohol exposure are summarised in Figure 3. undertook permutation analyses to see whether having unequal Cumulative alcohol exposure and age were strongly positively sample sizes (n = 20 vs. n = 39) was affecting our results. We cre- ated a null distribution from 100,000 permutations, where on each permutation, we (i) randomly selected 20 out of 39 HC (to match the 20 ADP in numbers), (ii) shuffled the group labels across these 40 participants (20 HC + 20 ADP) (iii) and compared the two shuf- fled groups using an independent t test. The overall group differ- ences computed via this permutation test remained significant for both amygdala SN/VTA-RSFC (p = 0.002) and for amygdala GMV (p = 0.004). Our spherical amygdala region of interest was motivated by our RSFC analyses and kept the same for the amygdala GMV analyses to enable direct comparison between the modalities. To control for potential biases arising from our choice of amygdala definition, we repeated our amygdala GMV analyses using structurally defined total amygdala volumes (see Section 2). Broadly consistent with our original FIGURE 3 Summary diagram showing linear relationships analysis, we found a significant negative correlation between age and between amygdala-SN/VTA RSFC, amygdala GM volume, cumulative total amygdala GMV (r = −0.56, p = 0.01) in ADP, whereas correla- alcohol exposure and age in ADP (n = 20). Red and blue lines indicate tions between total amygdala GMV and cumulative alcohol exposure significant positive and negative correlations, respectively, whereas or abstinence length were not significant. The group difference black line indicates the absence of a significant correlation. All statistics refer to Pearson correlations. Dotted lines indicate weak but between ADP and HC for total amygdala GMV also remained signifi- significant correlations cant (p = 0.008). ORBAN ET AL. 7of10

4 | DISCUSSION studies have also reported correlations between duration of substance use and RSFC in other circuits, in recently-abstinent We have shown that in addition to ADP exhibiting lower amygdala ADP,39 currently using cocaine12 and prescription opioid dependent GMV, they also show greater amygdala-SN/VTA RSFC compared with populations.13 In the current study, abstinence length was not associ- age-matched HC. Multiple regression analysis revealed that ated with recovery of amygdala-SN/VTA RSFC, perhaps due to the amygdala-SN/VTA RSFC was positively associated with cumulative limited range of abstinence in our ADP sample (1 to 38 months). alcohol exposure, whereas amygdala GMV was negatively associated Another important question is the extent to which alterations of with age. Amygdala GMV did not have the same inverse relationship this amygdala-SN/VTA circuit might characterise other forms of sub- with age in HC, suggesting accelerated age-related atrophy in ADP. stance dependence. In the current study, nicotine dependence alone Neither amygdala-SN/VTA RSFC nor amygdala GMV showed an was not found to be associated with elevated amygdala-SN/VTA association with length of abstinence, indicating persistent effects. RSFC in HC based on a group comparison of HC smokers and non- These findings suggest long-term, enduring changes to structural and smokers. Because in the current study there were only four ADP that functional properties of the amygdala in ADP and hint at distinct were nonsmokers, it remains to be established whether elevated underlying processes. However, given the small sample size and amygdala-SN/VTA RSFC generalises to nonsmoking ADP. cross-sectional design of this study, these findings would benefit from In theory, it is possible that increased amygdala-SN/TA RSFC could replication in large-scale longitudinal studies. predate alcohol dependence. However, in our study, younger partici- Our findings of increased amygdala-SN/VTA RSFC in abstinent pants showed similar levels of amygdala-SN/VTA RSFC in ADP and HC ADP are consistent with animal studies showing synaptic groups, whereas the group differences appeared to be driven by older neuroadaptations in the amygdala6,27 and VTA34,35 following chronic participants with a longer cumulative alcohol exposure. Furthermore, alcohol exposure. In theory, synaptic-level changes in response to amygdala-SN/VTA RSFC's association with cumulative alcohol exposure chronic alcohol exposure could lead to persistent increases in large- and its apparent lack of association with family history of alcohol depen- scale interactions between spatially distributed regions. The amygdala dence in our study also suggest neuroadaptive effects. and VTA are increasingly viewed as hubs that integrate circuits Consistent with previous literature, ADP showed GMV reduc- processing both aversive and appetitive motivational signals.36 tions in the amygdala and SN/VTA, relative to HC.17 Alterations of Although the biological significance of RSFC remains elusive, one amygdala GMV, like amygdala-SN/VTA RSFC, were more severe for model proposes that RSFC patterns might reflect, in part, statistical ADP that were older and in those with longer cumulative alcohol histories of large-scale circuit interactions encoded as relative synaptic exposure. Surprisingly, in contrast to our hypothesis, based on previ- weights.37 Based on this model, chronic coactivation of the amygdala ous findings in prescription opioid dependent populations,13 amygdala and SN/VTA during the cycles of anticipation, intoxication and with- GMV reductions were not correlated with amygdala-SN/VTA RSFC in drawal that characterise alcohol dependence38 could produce persis- ADP. Thus, variation in amygdala GMV did not seem to underlie varia- tent elevations of RSFC, as observed in our study. tion in amygdala-SN/VTA RSFC in ADP, potentially hinting at distinct An elevation of amygdala-SN/VTA RSFC in ADP from ‘normal’ underlying processes. levels in HC could be interpreted as evidence of a cumulative pathologi- When our analyses accounted for, both age and cumulative alco- cal shift in brain function. Alternatively, increased amygdala-SN/VTA hol exposure, amygdala GMV reductions in ADP were only signifi- RSFC could also reflect a mechanism for achieving and maintaining cantly associated with age, and not with cumulative alcohol exposure. abstinence. Of note, Beck et al.18 showed increased SN/VTA-amygdala One possible interpretation is that some features of alcohol depen- connectivity during cue-reactivity in early abstinent ADP patients who dence, other than duration of exposure, might interact with age, 3 months later managed to maintain abstinence versus those who thereby producing accelerated age-related amygdala atrophy. relapsed. They proposed that the hyperconnectivity of this circuit might According to this account, the same duration of alcohol dependence facilitate maintenance of abstinence through eliciting aversive affective might be associated with different degrees of amygdala GMV reduc- states in response to alcohol-cues. Trait and state anxiety, two mea- tion depending on an individual's age. This interpretation is consistent sures associated with aversive emotional states in the current dataset, with a recent longitudinal study which found accelerated atrophy of were not significantly correlated with amygdala-SN/VTA RSFC in ADP. the frontal cortex in recently abstinent ADP that was unrelated to the However, the functional significance of amygdala-SN/VTA connectivity number of years of heavy drinking.40 Other studies have also reported could be context dependent, showing distinct effects at rest, as in the accelerated cortical atrophy in recently abstinent ADP41 and cocaine- current study, and during exposure to alcohol cues, as in the study of dependent population42 as well as age-related atrophy of the hippo- Beck et al.18 campus in heavy nondependent drinkers.43 Our findings build on this To our knowledge, this is the first study to show a linear relation- work by presenting evidence suggestive of accelerated age-related ship between cumulative alcohol exposure and an amygdala resting atrophy specifically in the amygdala and in longer-term abstinent state circuit in abstinent ADP. Previously, we have observed a reduc- ADP. Nonetheless, the exact set of conditions that are necessary to tion of amygdala-insula RSFC in another abstinent ADP cohort, an observe age-associated atrophy in ADP remains unknown, for exam- effect that was not found to be associated with cumulative alcohol ple, minimum duration of diagnosis, severity of addiction or minimum exposure, and which was not replicated in the current cohort.11 Other amount of alcohol exposure that is necessary. 8of10 ORBAN ET AL.

Another possibility is that smaller amygdala GMV might be a pre- TWR has research grants with Eli Lilly and Company and existing vulnerability to alcohol dependence rather than a conse- Lundbeck; has received royalties from Cambridge Cognition; has quence of chronic alcohol exposure. Indeed, reduced amygdala received editorial honoraria from Springer Verlag, Elsevier, Society for volumes have been observed in offspring or unaffected first-degree Neuroscience; has performed educational lectures for Merck Sharp & relatives of ADP,44–46 and they have also been associated with a Dohme; and performs consultancy work for Cambridge Cognition, Eli greater likelihood of relapse during early abstinence.16 Although this Lilly and Company, Lundbeck, Teva Pharmaceutical Industries and does not provide an explanation for the negative correlation between Shire Pharmaceuticals. age and amygdala GMV, the accounts of preexisting differences and JFWD currently advises or carries out research funded by accelerated-atrophy are by no means mutually exclusive. Autifony Therapeutics, Sunovion Pharmaceuticals, Lundbeck, Amygdala GMV did not significantly vary with abstinence length AstraZeneca and Servier. All payment is to The University of across ADP. This is consistent with previous studies finding amygdala Manchester. GMV atrophy in both shorter (mean: 17 days and longer-term absti- DJN is an advisor to British National Formulary, Medical Research nent ADP populations; 7.1 years; 5.9 years). Nevertheless, a recent Council, General Medical Council, and Department of Health (UK), is cross-sectional study found a positive correlation between amygdala President of the European Brain Council, past President of the British GMV and a wide range of abstinence duration (0.8 to 20.1 years) in Neuroscience Association and European College of former alcohol users, leaving open the possibility that recovery of Neuropsychopharmacology, chair of the Independent Scientific Com- GMV might be possible given a long enough duration of abstinence.47 mittee on Drugs (UK), is a member of the International Centre for Sci- Our study has several limitations. First, given the cross-sectional ence in Drug Policy, advisor to Swedish government on drug, alcohol design, we cannot establish causality between chronic alcohol con- and tobacco research, editor of the Journal of , sumption and variation in amygdala-SN/VTA RSFC or amygdala sits on advisory Boards at Lundbeck, Merck Sharp & Dohme, GMV. In other words, we cannot infer from our data alone how Nalpharm, Orexigen Therapeutics, Shire Pharmaceuticals, has received much of the observed variation in amygdala GMV and RSFC likely to speaking honoraria (in addition to above) from Bristol-Myers Squibb/ have arisen prior to, during or in the aftermath of chronic alcohol Otsuka, GlaxoSmithKline, Eli Lilly and Company, Janssen, Servier, is a exposure (i.e., in abstinence). Second, the timeline followback member of the Lundbeck International Neuroscience Foundation, has approach that we used for estimating cumulative alcohol exposure received grants or clinical trial payments from P1vital, Medical relies heavily on patients' recall. Older participants who would have Research Council, National Health Service, Lundbeck, has share had to recall their drinking patterns further back in time would have options with P1vital, has been expert witness in a number of legal been less likely to accurately recall their alcohol consumption,48 thus cases relating to psychotropic drugs and has edited/written potentially limiting our ability to separate effects of estimated alcohol 27 books—some purchased by pharmaceutical companies. exposure from age. Third, our sample size is relatively modest due to ARL-H has received honoraria from Lundbeck and research our study's strong exclusion criteria, thus limiting our overall statisti- support from GlaxoSmithKline for a PhD studentship. cal power. Fourth, although we tried to balance multiple characteris- All other authors declared no conflict of interest. tics between the two groups, some inherent differences remained, for example, ADP had significantly lower educational levels than FUNDING INFORMATION HC. Fifth, the length of the resting state fMRI scan was relatively This article presents independent research funded by the Medical short at 5 min, which likely affected the reliability of our estimates of Research Council as part of their addiction initiative (grant number amygdala RSFC. A longer scan duration would have also opened up G1000018). GSK kindly funded the functional and structural MRI the possibility of estimating amygdala subdivisions in each individ- scans that took place at Imperial College. Csaba Orban was funded by ual49 rather than relying on a group-level amygdala region of interest. a Medical Research Council Doctoral Training Program Studentship at Last, we did not measure respiratory activity, a confound whose Imperial College and by Singapore MOE Tier 2 (MOE2014-T2-2-016), importance is becoming increasingly recognised in resting-state fMRI National University of Singapore Strategic Research (DPRT/944/ studies.50 09/14), National University of Singapore SOM Aspiration Fund In summary, we showed that ADP exhibit differences in amygdala (R185000271720), Singapore NMRC (CBRG/0088/2015), National structure and function that persist in long-term abstinence and which University of Singapore YIA and the Singapore National Research seem to differentially interact with age and cumulative alcohol expo- Foundation (NRF) Fellowship (Class of 2017). sure, respectively. These findings will require replication in a larger dataset, whereas longitudinal studies will be necessary to characterise AUTHORS CONTRIBUTION the timecourse of these effects and to delineate their potential roles ARLH, DJN, TWR, JFWD, ADW, EMT, JS, DGS, LJR, LJN, AM, KDE, in mechanisms of abstinence and relapse. RE, LMP, RSAF, CO and JM designed research; LMP, AM, EMT, DGS, FP, RSAF, JM, KDE, CO, DE, LJN and LJR collected data; CO and JG DISCLOSURE/CONFLICT OF INTEREST analyzed fMRI data; RSAF, FP and ARLH analyzed clinical data; ASR LJN was a Senior Research Scientist employed by GlaxoSmithKline contributed analytic tools; LMP, RSAF, FP, JS, LJN, DE and VCW during this work. assisted with interpretation of data; CO wrote the first draft of the ORBAN ET AL. 9of10 manuscript; CO, JM and ARLH wrote the manuscript. All authors 6. Janak PH, Tye KM. From circuits to behaviour in the amygdala. contributed to and have approved the final manuscript. . 2015;517(7534):284-292. https://doi.org/10.1038/ nature14188 7. Roberto M. Increased GABA release in the central amygdala of ACKNOWLEDGEMENTS ethanol-dependent rats. J Neurosci. 2004;24(45):10159-10166. Some of this work has been presented at the Organization from https://doi.org/10.1523/JNEUROSCI.3004-04.2004 Human Brain Mapping, Honolulu, HI, USA, 2015. 8. Roberto M, Cruz MT, Gilpin NW, et al. Corticotropin releasing factor– We wish to thank the Medical Research Council (MRC) for induced amygdala gamma-aminobutyric acid release plays a key role in alcohol dependence. Biol Psychiatry. 2010;67(9):831-839. https:// funding the study and G for supplying the study drug and funding doi.org/10.1016/j.biopsych.2009.11.007 scans at London. 9. Gilpin NW, Roberto M. Neuropeptide modulation of central amygdala The research was carried out at the NIHR/ neuroplasticity is a key mediator of alcohol dependence. Neurosci Bio- Imperial Clinical Research Facility, the NIHR/Wellcome Trust behav Rev. 2012;36(2):873-888. https://doi.org/10.1016/j.neubiorev. 2011.11.002 Cambridge Research Facility and Clinical Trials Unit at Salford Royal 10. Gilpin NW, Herman MA, Roberto M. The central amygdala as an NHS Foundation Trust and is supported by the North West London, integrative hub for anxiety and alcohol use disorders. Biol Psychiatry. Eastern and Greater Manchester NIHR Clinical Research Networks. 2015;77(10):859-869. https://doi.org/10.1016/j.biopsych.2014. The views expressed are those of the author(s) and not necessarily 09.008 those of the Medical Research Council, the NHS, the NIHR or the 11. Orban C, McGonigle J, Kalk NJ, et al. Resting state synchrony in anxiety-related circuits of abstinent alcohol-dependent patients. Department of Health. Am J Drug Alcohol Abuse. 2013;39(6):433-440. https://doi.org/10. We wish to thank Sanja Abbott for help with programming the 3109/00952990.2013.846348 fMRI tasks used at Cambridge and research assistants Claire Whitelock, 12. Gu H, Salmeron BJ, Ross TJ, et al. Mesocorticolimbic circuits are Heather Agyepong, Rania Christoforou and Natalie Cuzen for their help impaired in chronic cocaine users as demonstrated by resting-state functional connectivity. Neuroimage. 2010;53(2):593-601. https://doi. with data collection, MR physicist Rex Newbould and MR technician, org/10.1016/j.neuroimage.2010.06.066 Jonathan Howard for their assistance with MR acquisition and task set- 13. Upadhyay J, Maleki N, Potter J, et al. Alterations in brain structure up, Shane McKie for help with statistical analysis and Martyn and functional connectivity in prescription opioid-dependent patients. McFarquhar for help with task programming and statistical analysis. Brain. 2010;133(7):2098-2114. https://doi.org/10.1093/brain/ Recruitment partners—We wish to thank our recruitment awq138 14. Filbey FM, Aslan S, Calhoun VD, et al. Long-term effects of marijuana partners; Imperial College Healthcare NHS Trust, Central and North use on the brain. Proc Natl Acad Sci. 2014;111(47):16913-16918. West London NHS trust, Camden and Islington NHS trust, Cambridge https://doi.org/10.1073/pnas.1415297111 University Hospitals NHS Foundation Trust, Norfolk and Suffolk NHS 15. Fein G, Landman B, Tran H, et al. Brain atrophy in long-term abstinent Foundation Trust, Cambridge and Peterborough NHS Foundation alcoholics who demonstrate impairment on a simulated gambling Trust, South Staffordshire and Shropshire NHS Foundation Trust, task. Neuroimage. 2006;32(3):1465-1471. https://doi.org/10.1016/j. neuroimage.2006.06.013 Manchester Mental Health NHS and Social Care Trust, Greater 16. Wrase J, Makris N, Braus DF, et al. Amygdala volume associated with Manchester West NHS Foundation Trust, Pennine Care NHS alcohol abuse relapse and craving. Am J Psychiatry. 2008;165(9): Foundation Trust, Salford Royal NHS Foundation Trust, Addaction, 1179-1184. https://doi.org/10.1176/appi.ajp.2008.07121877 Foundation 66 and CRI (Crime Reduction Initiative). 17. Makris N, Oscar-Berman M, Jaffin SK, et al. Decreased volume of the brain reward system in alcoholism. Biol Psychiatry. 2008;64(3): 192-202. https://doi.org/10.1016/j.biopsych.2008.01.018 ORCID 18. Beck A, Wüstenberg T, Genauck A, et al. Effect of brain structure, Csaba Orban https://orcid.org/0000-0001-9133-3561 brain function, and brain connectivity on relapse in alcohol- Anna Murphy https://orcid.org/0000-0002-3529-2766 dependent patients. Arch Gen Psychiatry. 2012;69(8):842-852. https://doi.org/10.1001/archgenpsychiatry.2011.2026 Liam J. Nestor https://orcid.org/0000-0002-8854-9908 19. Paterson LM, Flechais RS, Murphy A, et al. The Imperial College Cambridge Manchester (ICCAM) platform study: an experimental REFERENCES medicine platform for evaluating new drugs for relapse prevention in 1. Koob GF. A role for brain stress systems in addiction. Neuron. 2008; addiction. Part A: study description. J Psychopharmacol. 2015;29(9): 59(1):11-34. https://doi.org/10.1016/j.neuron.2008.06.012 943-960. https://doi.org/10.1177/0269881115596155 2. Charney DA, Zikos E, Gill KJ. Early recovery from alcohol depen- 20. Group WAW. The alcohol, smoking and substance involvement dence: factors that promote or impede abstinence. J Subst Abuse screening test (ASSIST): development, reliability and feasibility. Addic- Treat. 2010;38(1):42-50. https://doi.org/10.1016/j.jsat.2009.06.002 tion. 2002;97(9):1183-1194. https://doi.org/10.1046/j.1360-0443. 3. Heilig M, Egli M, Crabbe JC, Becker HC. Acute withdrawal, protracted 2002.00185.x abstinence and negative affect in alcoholism: are they linked? Addict 21. Sheehan DV, Lecrubier Y, Sheehan KH, et al. The Mini-International Biol. 2010;15(2):169-184. https://doi.org/10.1111/j.1369-1600. Neuropsychiatric Interview (MINI): the development and validation of 2009.00194.x a structured diagnostic psychiatric interview for DSM-IV and ICD-10. 4. Koob GF, Le Moal M. Addiction and the brain antireward system. J Clin Psychiatry. 1998;59:22-33. Annu Rev Psychol. 2008;59(1):29-53. https://doi.org/10.1146/ 22. Spielberger CD, Gorsuch R. State-Trait Anxiety Inventory for Adults: annurev.psych.59.103006.093548 Manual and Sample: Manual, Instrument and Scoring Guide. Consulting 5. Koob GF. Brain stress systems in the amygdala and addiction. Brain Res. Psychologists Press; 1983. 2009;1293:61-75. https://doi.org/10.1016/j.brainres.2009.03.038 23. Wechsler D. Wechsler Adult Intelligence Scale-Fourth. Pearson; 2008. 10 of 10 ORBAN ET AL.

24. Oldfield RC. The assessment and analysis of handedness: the compromise. JAMA Psychiat. 2018;75(5):474-483. https://doi.org/10. Edinburgh inventory. Neuropsychologia. 1971;9(1):97-113. https:// 1001/jamapsychiatry.2018.0021 doi.org/10.1016/0028-3932(71)90067-4 41. Pfefferbaum A, Lim KO, Zipursky RB, et al. Brain gray and white mat- 25. Orban C, Kong R, Li J, Chee MWL, Yeo BTT. Time of day is associated ter volume loss accelerates with aging in chronic alcoholics: a quanti- with paradoxical reductions in global signal fluctuation and functional tative MRI study. Alcohol Clin Exp Res. 1992;16(6):1078-1089. connectivity. PLoS Biol. 2020;18(2):e3000602. https://doi.org/10. https://doi.org/10.1111/j.1530-0277.1992.tb00702.x 1371/journal.pbio.3000602 42. Ersche KD, Jones PS, Williams GB, Robbins TW, Bullmore ET. 26. McGonigle J, Murphy A, Paterson LM, et al. The ICCAM platform study: Cocaine dependence: a fast-track for brain ageing? Mol Psychiatry. an experimental medicine platform for evaluating new drugs for relapse 2013;18(2):134-135. https://doi.org/10.1038/mp.2012.31 prevention in addiction. Part B: fMRI description. JPsychopharmacol. 43. Topiwala A, Allan CL, Valkanova V, et al. Moderate alcohol consump- 2017;31(1):3-16. https://doi.org/10.1177/0269881116668592 tion as risk factor for adverse brain outcomes and cognitive decline: 27. Cox RW. AFNI: software for analysis and visualization of functional longitudinal cohort study. BMJ. 2017;357:j2353. https://doi.org/10. magnetic resonance neuroimages. Comput Biomed Res. 1996;29(3): 1136/bmj.j2353 162-173. https://doi.org/10.1006/cbmr.1996.0014 44. Benegal V, Antony G, Venkatasubramanian G, Jayakumar PN. Gray 28. Fischl B. FreeSurfer. Neuroimage. 2012;62(2):774-781. matter volume abnormalities and externalizing symptoms in subjects 29. Avants BB, Tustison NJ, Song G, Cook PA, Klein A, Gee JC. at high risk for alcohol dependence. Addict Biol. 2007;12(1):122-132. A reproducible evaluation of ANTs similarity metric performance https://doi.org/10.1111/j.1369-1600.2006.00043.x in brain image registration. Neuroimage. 2011;54(3):2033-2044. 45. Dager AD, McKay DR, Kent JW, et al. Shared genetic factors influence https://doi.org/10.1016/j.neuroimage.2010.09.025 amygdala volumes and risk for alcoholism. Neuropsychopharmacology. 30. Jenkinson M, Beckmann CF, Behrens TE, Woolrich MW, Smith SM. 2015;40(2):412-420. https://doi.org/10.1038/npp.2014.187 FSL. Neuroimage. 2012;62(2):782-790. 46. Hill SY, De Bellis MD, Keshavan MS, et al. Right amygdala volume in 31. Jo HJ, Gotts SJ, Reynolds RC, et al. Effective preprocessing proce- adolescent and young adult offspring from families at high risk for dures virtually eliminate distance-dependent motion artifacts in rest- developing alcoholism. Biol Psychiatry. 2001;49(11):894-905. https:// ing state FMRI. J Appl Math. 2013;2013:1-9. https://doi.org/10. doi.org/10.1016/S0006-3223(01)01088-5 1155/2013/935154 47. Korponay C, Kosson DS, Decety J, Kiehl KA, Koenigs M. Brain volume 32. Costafreda SG, Brammer MJ, David AS, Fu CHY. Predictors of correlates with duration of abstinence from substance abuse in a amygdala activation during the processing of emotional stimuli: a region-specific and substance-specific manner. Biol Psychiatry Cogn meta-analysis of 385 PET and fMRI studies. Brain Res Rev. 2008; Neurosci . 2017;2(7):626-635. https://doi.org/10.1016/ 58(1):57-70. https://doi.org/10.1016/j.brainresrev.2007.10.012 j.bpsc.2017.03.011 33. Murty VP, Shermohammed M, Smith DV, Carter RM, Huettel SA, 48. Vakili S, Sobell LC, Sobell MB, Simco ER, Agrawal S. Using the Adcock RA. Resting state networks distinguish human ventral teg- timeline followback to determine time windows representative of mental area from substantia nigra. Neuroimage. 2014;100:580-589. annual alcohol consumption with problem drinkers. Addict Behav. https://doi.org/10.1016/j.neuroimage.2014.06.047 2008;33(9):1123-1130. https://doi.org/10.1016/j.addbeh.2008.03.009 34. Diana M, Pistis M, Muntoni A, Gessa G. Mesolimbic dopaminergic 49. Sylvester CM, Yu Q, Srivastava AB, et al. Individual-specific functional reduction outlasts ethanol withdrawal syndrome: evidence of pro- connectivity of the amygdala: a substrate for precision psychiatry. tracted abstinence. Neuroscience. 1996;71(2):411-415. https://doi. Proc Natl Acad Sci. 2020;117(7):3808-3818. https://doi.org/10.1073/ org/10.1016/0306-4522(95)00482-3 pnas.1910842117 35. Shen R-Y, Choong K-C, Thompson AC. Long-term reduction in ventral 50. Power JD, Lynch CJ, Dubin MJ, Silver BM, Martin A, Jones RM. tegmental area dopamine neuron population activity following Characteristics of respiratory measures in young adults scanned at rest, repeated stimulant or ethanol treatment. Biol Psychiatry. 2007;61(1): including systematic changes and “missed” deep breaths. Neuroimage. 93-100. https://doi.org/10.1016/j.biopsych.2006.03.018 2020;204:116234. https://doi.org/10.1016/j.neuroimage.2019.116234 36. Lüthi A, Lüscher C. Pathological circuit function underlying addiction and anxiety disorders. Nat Neurosci. 2014;17(12):1635-1643. https:// doi.org/10.1038/nn.3849 SUPPORTING INFORMATION 37. Fair DA, Cohen AL, Power JD, Dosenbach NU, Church JA, Miezin FM, Additional supporting information may be found online in the Schlaggar BL, Petersen SE Functional brain networks develop from a Supporting Information section at the end of this article. “local to distributed” organization. Sporns O, ed PLoS Comput Biol 2009;5(5):e1000381. https://doi.org/10.1371/journal.pcbi.1000381 38. Koob GF, Volkow ND. Neurocircuitry of addiction. Neuropsychopharmacology. 2010;35(1):217-238. https://doi.org/10. How to cite this article: Orban C, McGonigle J, Flechais RSA, 1038/npp.2009.110 et al. Chronic alcohol exposure differentially modulates 39. Camchong J, Stenger A, Fein G. Resting-state synchrony during early structural and functional properties of amygdala: A cross- alcohol abstinence can predict subsequent relapse. Cereb Cortex. sectional study. Addiction Biology. 2020;e12980. https://doi. 2013;23(9):2086-2099. https://doi.org/10.1093/cercor/bhs190 40. Sullivan EV, Zahr NM, Sassoon SA, et al. The role of aging, drug org/10.1111/adb.12980 dependence, and hepatitis C comorbidity in alcoholism cortical