Addiction Biology

ORIGINAL ARTICLE OPEN ACCESS

Neural, Motivational, and Psychological Measures of Pain

Avoidance Predict Future Alcohol Use in Adult Drinkers
Thang M. Le1 | F. AnNa Hughes2 | Takeyuki Oba3 | Chiang-­Shan R. Li1,4,5,6

1Department of Psychiatry, Yale University School of Medicine, New Haven, Connecticut, USA | 2Department of Psychology, Villanova
University, Villanova, Pennsylvania, USA | 3Human Informatics and Interaction Research Institute, the National Institute of Advanced Industrial
Science and Technology (AIST), Tsukuba, Japan | 4Department of Neuroscience, Yale University School of Medicine, New Haven, Connecticut,
USA | 5Interdepartmental Neuroscience Program, Yale University School of Medicine, New Haven, Connecticut, USA | 6Wu Tsai Institute, Yale
University, New Haven, Connecticut, USA

Correspondence: Thang M. Le ([email protected])

Received: 10 June 2024 | Revised: 13 January 2025 | Accepted: 15 January 2025

Funding: This work was supported by Brain and Behavior Research Foundation and National Institute on Alcohol Abuse and Alcoholism.

Keywords: alcohol | drinking prediction | fMRI | pain | proactive avoidance

ABSTRACT
Drinking as a coping method in response to pain is a complex behaviour, involving multiple neural, motivational, and psycho-
logical factors. Among these factors, pain sensitivity and pain-­related drinking motive can significantly promote alcohol use. In
contrast, proactive avoidance – a beneficial strategy of initiating overt actions to avoid negative outcomes – reduces harmful con-
sumption. Yet, these factors have not been assessed as potential predictors of future drinking behaviour. Here, in a longitudinal
study we collected fMRI data in 50 drinkers who, at baseline, performed a probabilistic learning go/nogo task that involved pro-
active avoidance of painful electric shocks. Pain-­related psychological measures and the neural correlates of proactive avoidance
were examined in relation to participants' alcohol use and craving in the following 12 months. We found that deficits in proactive
avoidance were associated with future drinking severity. Importantly, diminished activation of the dorsal anterior cingulate
cortex (dACC) during proactive avoidance also predicted subsequent percentage of heavy drinking days. Using Bayesian net-
work modelling, we established a potential pathway in which drinkers' heightened pain sensitivity led to greater pain-­avoidance
drinking motive and alcohol craving. Both craving and weakened dACC activation to proactive avoidance predicted higher levels
of drinking during the follow-­up period. Taken together, our study identified pain sensitivity, pain-­avoidance drinking motive,
and impaired proactive avoidance as predictors of future alcohol use severity. These findings highlight the roles of pain response,
thus potentially informing interventions for individuals at risk of alcohol use disorders.

1   |   Introduction and improve outcomes [2]. Nevertheless, characterizing fac-

tors with potentially causal relationships with drinking has
The prediction of alcohol use over time is an important goal proven challenging. First, alcohol use is a complex behaviour.
for researchers and clinicians alike. Accurately predicting fu- In addition to psychological (e.g., personality traits), motiva-
ture drinking can help identify individuals at risk of develop- tional (e.g., drinking motives), and contextual (e.g., negative
ing alcohol use disorders (AUD) and formulate interventions experiences) [3, 4] influences, the brain processes underlying
to prevent the escalation to dependence [1]. For the treatment reinforcement learning in response to reward and punishment
of AUD, establishing the determinants of chronic drinking may also serve in a predictive role [5, 6]. Thus, it is crucial that
can promote targeting of critical psychological processes relevant variables from different clinical and neural domains

This is an open access article under the terms of the Creative Commons Attribution-­NonCommercial-­NoDerivs License, which permits use and distribution in any medium, provided the original
work is properly cited, the use is non-­commercial and no modifications or adaptations are made.

© 2025 The Author(s). Addiction Biology published by John Wiley & Sons Ltd on behalf of Society for the Study of Addiction.

Addiction Biology, 2025; 30:e70020 1 of 11

https://doi.org/10.1111/adb.70020
 13691600, 2025, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/adb.70020 by Nigeria Hinari NPL, Wiley Online Library on [24/02/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
are examined comprehensively. Second, unique predictors In addition, we acquired imaging data during drinkers' perfor-
may interact to produce a collective effect that is distinct from mance of a probabilistic learning go/nogo task that operational-
the direct influence of individual factors on alcohol use [7]. As ized proactive avoidance. Combining clinical assessments and
such, there is a need for investigations that not only account brain imaging, we identified the predictors of drinking severity
for multiple neurocognitive predictors but also assess their po- in the following 12 months. Finally, we employed Bayesian net-
tential interactions. work to describe the conditional relationships among the neuro-
cognitive predictors and quantify their causal contributions to
Among the motivational factors that facilitate alcohol use, pain future alcohol use. As our work shed light on the determinants
relief as a motive has been found to significantly promote drink- of harmful drinking, it may inform intervention strategies for
ing [8]. Motivational model of alcohol use [9] posits that indi- those at risk of dependency or those vulnerable to relapse follow-
viduals may seek alcohol to avoid or reduce negative physical ing treatment for AUD.
or emotional experiences. These experiences increase craving,
which then leads to drinking as a maladaptive coping strategy
[10]. Paradoxically, chronic alcohol intake heightens sensitivity 2   |   Methods
to pain [11], which further motivates drinking to cope, thus trap-
ping drinkers in a spiralling cycle of drinking and distress. As 2.1   |   Participants
such, pain sensitivity, maladaptive avoidance of pain, and crav-
ing are among the major components of the pathophysiological Fifty adult drinkers (23 females, 35.3 ± 11.7 years in age) were
pathway to harmful alcohol use [8]. recruited from New Haven, CT and surrounding areas via ad-
vertisement. Participants were screened to exclude major med-
To break the cycle of drinking and distress, clinical research ical, neurological, and Axis I psychiatric disorders. Exclusion
has demonstrated the efficacy of proactive avoidance, or the criteria additionally included MR contraindications and history
initiation of overt actions to prevent or address negative situa- of seizures, traumatic brain injury or concussions. No partici-
tions [12–14]. Taking steps to actively solve problems, seek sup- pants were currently on psychotropic medications, and all tested
port, assume control, and set goals has been reported to reduce negative for illicit substances and alcohol use via a drug and
craving [15, 16] and maintain abstinence in AUD patients un- breathalyser tests on the day of the baseline visit. Subjects pro-
dergoing treatment [17]. In contrast, those impaired in proac- vided written informed consent after details of the study were
tive avoidance are prone to harmful drinking and relapse after explained per the institute guidelines and procedures approved
treatment [18]. Despite the potential role of proactive avoidance by the Yale Human Investigation Committee.
in reducing and preventing alcohol misuse, most studies inves-
tigating proactive avoidance have employed cross-­sectional de-
signs. Thus, whether or how measures of proactive avoidance 2.2   |   Drinking Assessment and Follow-­Up
can help predict future drinking represents a promising avenue
for research. During the baseline visit, participants performed a probabi-
listic learning go/nogo task (PLGT, see Behavioural Task) and
Imaging studies investigating the neural correlates of drinking completed detailed assessment of their current alcohol use
behaviour have typically focused on the reward circuit. For in- (Table 1). The assessment included the Alcohol Use Disorders
stance, higher activations of the medial orbitofrontal cortex [19] Identification Test (AUDIT) [28]. AUDIT scores are calculated
and ventral striatum [20, 21] in response to alcohol cues have from the sum of 10 self-­report questions involving quantity of
been associated with greater drinking severity and risky be- alcohol use, alcohol-­ related problems and adverse reactions,
haviours. However, as drinking escalates, there is a fundamen- and drinking behaviour. A higher score indicates greater risk
tal shift of motivation from positive to negative reinforcement. for having or developing an AUD. Consistent with past research
This transition is characterized by the increasingly important [29], 25 participants met the criteria of problem drinkers who
role of the brain circuits involved in pain response and pain were defined as those with [1] AUDIT score ≥ 7 and [2] weekly
avoidance [22, 23]. Therefore, it is plausible that the neural pro- consumption of at least 14 drinks for men and seven drinks for
cesses underlying proactive avoidance of painful outcomes may women. The remaining 25 were considered social drinkers.
have an outsize effect in predicting alcohol misuse. Our recent Participants also completed the 42-­item Inventory of Drinking
work examining the neural correlates of proactive avoidance Situations (IDS-­42) [30]. The IDS-­42 assesses relative frequency
has identified several structures including the dorsal anterior of drinking behaviour across eight categories of drinking situa-
cingulate cortex (dACC), insula, and middle frontal gyrus [24]. tions and determines the drinking motives leading to excessive
Other studies additionally showed that reduced functional and alcohol use. We focused on the Physical Discomfort subscale
structural integrity of these regions were associated with greater as it is directly related to pain sensitivity and pain avoidance.
pain sensitivity [25], craving [26], and drinking severity [27]. Participants also completed the Pain Sensitivity Questionnaire,
However, it remains unclear whether the brain activities during which contains 17 items measuring pain intensity ratings of
proactive avoidance may also be predictive of subsequent alco- painful situations occurring in daily life [31]. The greater the
hol use. score, the higher the individual's sensitivity to pain. Finally, par-
ticipants completed the Beck's depression inventory [32] and the
The present study sought to identify and characterize neural, Generalized anxiety disorder 7-­item anxiety scale (GAD-­7) [33].
motivational, and psychological predictors of alcohol use lon- We included depression and anxiety scores, as well as sex and
gitudinally. We focused on pain-­related factors at baseline in- age, as covariates in all regression analyses to control for their
cluding pain sensitivity and pain-­avoidance drinking motive. effects in drinking behaviour.

2 of 11 Addiction Biology, 2025

 13691600, 2025, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/adb.70020 by Nigeria Hinari NPL, Wiley Online Library on [24/02/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
All participated in six follow-­up visits over a 12-­month period, period. In addition, at each follow-­up visit, participants reported
with the first visit at four weeks after the baseline session and their alcohol craving during the previous week, using the Penn
all subsequently visits approximately 10 weeks apart. Alcohol Alcohol Craving Scale [36], with greater scores indicating higher
use was assessed using the Form 90 substance use calendar [34] craving. We computed the average scores across the six visits to
and the timeline follow-­back method (Supplementary methods). obtain a single craving measure for the follow-­up period.
From the 12-­month follow-­up period, we calculated for each
subject percentage of heavy drinking days (%HDD), percentage
of drinking days, peak alcohol use, and total drinking quantity 2.3   |   Behavioural Task
(Table 1). As per previous research [35], we used the %HDD (de-
fined as five or more standard drinks per drinking day in men, Participants underwent fMRI while performing the PLGT
and four or more standard drinks per drinking day in women) as (Figure 1) [24]. In each run of this event-­related design, a cue (frac-
the primary measure of drinking severity during the follow-­up tal) image was presented at the beginning of a trial to signal one
of the four contingencies: go to win $1, nogo to win $1, go to avoid
a painful shock, or nogo to avoid shock. There were eight images,
TABLE 1    |    Participants' demographics and drinking assessments.
two per cue category, with cue-­outcome mappings randomized
Demographics across participants. The cue was displayed for 2 s and participants
were instructed to decide whether to press a button (go) or not
Age (years, SD) 35.3 (11.7)
(nogo) before it disappeared. After a randomized interval of 1 to
Sex (% female, male, other) 46, 48, 6 5 s, feedback of reward (win trials), shock (avoid trials), or “null”
(both win and avoid trials) was delivered. The inter-­trial interval
Education (years, SD) 16.2 (2.4)
varied randomly from 2 to 8 s. The randomization and intervening
Current smokers (%) 8.0 time intervals enabled modelling of distinct regional responses to
anticipation and feedback. The outcome was probabilistic, with
Baseline assessments
80%/20% of correct/incorrect responses in the win trials rewarded
AUDIT 8.4 (7.0) and the remaining 20%/80% of correct/incorrect responses leading
Physical Discomfort drinking motive 5.5 (2.1) to a null outcome. In avoid trials, electric shocks were avoided (a
null outcome) on 80%/20% of correct/incorrect responses, with the
Pain sensitivity 70.3 (29.4) remainder leading to electric shocks. In reality, despite the feed-
Depression (BDI) 8.2 (9.5) back display of the shock image, shocks were randomly delivered
only half of the times to minimize head movements. Each shock
Anxiety (GAD7) 4.0 (4.5) was followed by a 20-­s rest window to allow neural and physiolog-
Follow-­up assessments ical responses to return to baseline.

Craving ratings 4.7 (4.1) With four 10-­min runs, there were approximately 50 trials for
Drinking quantity (#drinks/year) 335.6 (459.7) each cue. Participants won ~$71 on average (plus a base pay-
ment of $25) and experienced a total of 16 actual shocks and 16
Heavy drinking days (%) 6.8 (10.4) omitted shocks on average. Prior to MRI, the appropriate pain
Drinking days (%) 21.1 (15.0) intensity level was calibrated for each participant so that the
shocks would be painful but tolerable (Supplemental Methods).
Peak alcohol use (# drinks) 11.8 (13.4)
Participants also practiced on a version of the task with identical

FIGURE 1    |    Probabilistic learning go/nogo task (PLGT): Participants learned to respond to four cue categories to avoid electric shocks (i.e., go-­to-­
avoid, nogo-­to-­avoid) and gain monetary rewards (i.e., go-­to-­w in, nogo-­to-­w in), with two different images per cue category. Correct responses yield-
ed favourable outcomes 80% of the times whereas incorrect responses yielded unfavourable outcomes 80% of the times. Shocks were only delivered
in 50% of the shock feedback instances to reduce head movement.

3 of 11
 13691600, 2025, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/adb.70020 by Nigeria Hinari NPL, Wiley Online Library on [24/02/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
rules but different visual cues. To ensure the subjects' efficient action initiation and the avoidance of an aversive outcome, the
task comprehension, there was one image per cue category in two primary components of proactive avoidance. Serial autocor-
the practice version. relation caused by aliased cardiovascular and respiratory effects
were corrected by the FAST model.

2.4   |   Imaging Protocol and Data Preprocessing To identify the neural processes predicting individual dif-
ferences in drinking severity (%HDD), we used whole-­brain
Conventional T1-­weighted spin echo sagittal anatomical im- multiple regressions with the brain activity to proactive avoid-
ages were acquired for slice localization using a 3 T scanner ance during the cue period as the predictor of %HDD from the
(Siemens Trio). Anatomical images of the functional slice loca- subsequent 12 months. The results of all whole-­brain analyses
tions were next obtained with spin echo imaging in the axial were evaluated with voxel p < 0.001 in combination with clus-
plane parallel to the AC–PC line with (TR) = 1900 ms, echo time ter p < 0.05, corrected for family-­w ise error of multiple com-
(TE) = 2.52 ms, bandwidth = 170 Hz/pixel, FOV = 250 × 250 mm, parisons, according to current reporting standards [37, 38]. All
matrix = 256 × 256, 176 slices with slice thickness = 1 mm and no peaks of activation were reported in MNI coordinates.
gap. Functional blood oxygenation level-­dependent (BOLD) sig-
nals were acquired using multiband imaging (multiband accel-
eration factor = 3) with a single-­shot gradient echo echoplanar 2.6   |   Hierarchical Linear Regression Analysis
imaging sequence. Fifty-­one axial slices parallel to the AC–PC
line covering the whole brain were acquired with TR = 1000 ms, As the study was not pre-­registered and thus exploratory in na-
TE = 30 ms, bandwidth = 2290 Hz/pixel, flip angle = 62°, field ture, we first employed hierarchical linear regression to confirm
of view = 210 × 210 mm, matrix = 84 × 84, with slice thick- that pain sensitivity, Physical Discomfort drinking motive, crav-
ness = 2.5 mm and no gap. ing, and the brain activity during proactive avoidance (i.e., acti-
vation of the dACC; see Results) significantly contributed to the
Imaging data were preprocessed using SPM12 (Wellcome Trust prediction of future %HDD. Previous research suggested these
Centre for Neuroimaging). Subjects with BOLD runs with sig- factors' potential contributions to harmful alcohol use [39–41]
nificant motion (> 3-­mm translation peak-­to-­peak movement but did not address whether they may interact with the neural
and/or 1.5° rotation) were removed. Furthermore, we calculated processes underlying proactive avoidance to predict alcohol
framewise displacement (FD) for each task run and removed consumption. Here, we employed hierarchical regression deter-
subjects with averaged FD of greater than 0.2. This resulted in mine whether the variables of interest (i.e., pain related metrics)
the removal of three drinkers, leaving a sample of 50 subjects explain a statistically significant amount of variance of the de-
as reported above. Images from the first five TRs at the begin- pendent variable (i.e., %HDD from the follow-­up period) after
ning of each run were discarded to ensure only BOLD signals accounting for other variables. This is achieved by building a se-
at steady-­state equilibrium between RF pulsing and relaxation ries of regression models. In each model, a new variable is added
were included in analyses. Physiological signals including res- as an additional predictor while retaining all variables from the
piration and heart rate were regressed out to minimize the in- prior models. Each model is then assessed to determine whether
fluence of these sources of noise. Images of each subject were it accounts for a significant improvement in the variance (R 2)
first realigned (motion corrected) and corrected for slice tim- when compared to the previous one.
ing. A mean functional image volume was constructed for each
subject per run from the realigned image volumes. These mean In step 1, we quantified the variance in %HDD accounted for by
images were co-­registered with the high-­resolution structural the activation of the dACC during proactive avoidance. In step 2
image and then segmented for normalization with affine reg- to 4, we separately and individually added craving score, Physical
istration followed by nonlinear transformation. The normaliza- Discomfort drinking motive, and pain sensitivity as additional
tion parameters determined for the structure volume were then predictors. These models allowed us to determine whether each
applied to the corresponding functional image volumes for each variable accounted for additional variance significantly over and
subject. Images were resampled to 2.5 mm isotropic voxel size. above the previous variables. Finally, in step 5, we constructed a
Finally, the images were smoothed with a Gaussian kernel of model with the brain activity during proactive avoidance, crav-
4-­mm FWHM. ing, and the interaction between pain sensitivity, pain-­avoidance
drinking motive, and craving as the predictors. We included this
interaction term as it was previously reported to be important for
2.5   |   Imaging Data Modelling and Group Analyses the promotion of harmful drinking [8]. Each model was exam-
ined to determine whether the newly added variable produced a
As with previous work [24], we constructed a general linear significant improvement in R 2. The effects of sex, age, depression,
model (GLM) to examine the brain processes underlying the and anxiety were controlled for in all models.
initiation of an action to avoid painful shocks. Cue onsets of
individual trials of the four conditions were convolved with a
canonical HRF and with the temporal derivative of the canon- 2.7   |   Bayesian Network
ical HRF and entered as regressors in the GLM. We used the
contrast Go (Avoid–Win) > Nogo (Avoid–Win) (proactive avoid- Once the predictors of %HDD had been identified (see Results),
ance, hereafter) to identify regional activities during proactive we conducted a Bayesian network analysis to estimate the directed
avoidance for individual subjects. This contrast enabled us to ex- and conditional relationships among the predictors and in relation
amine the neural processes involved in the interaction between to %HDD based on their probabilistic dependencies. Bayesian

4 of 11 Addiction Biology, 2025

 13691600, 2025, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/adb.70020 by Nigeria Hinari NPL, Wiley Online Library on [24/02/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
networks is a statistical graphical model that allows for condi- motivational goal (p = 0.33) or response type (p = 0.08). However,
tional inter-­dependence between the predictors, thus capturing there was a significant interaction effect (F(1, 48) = 5.03, p = 0.026).
their collective effects (i.e., joint probability) [42]. An advantage of Post hoc analyses revealed that performance accuracy was signifi-
the Bayesian network approach is the ability to make inferences on cantly higher for go-­to-­win condition than nogo-­to-­win (p = 0.025),
the direction of relationships. Furthermore, the model's output is go-­to-­avoid (p = 0.002), and nogo-­to-­avoid (p = 0.045) conditions.
explicitly a probability, which can be used as a confidence measure Avoid and win trials or go trials and nogo trials did not signifi-
to evaluate the level of certainty in drinking prediction. cantly differ in performance accuracy (p's > 0.12).

The model's structure consists of nodes (factors of interest) and Next, we examined the relationships between proactive avoid-
directed arcs (or edges) between nodes that indicate probabilis- ance task performance and drinking behaviour (Figure 2). Go-­to-­
tic conditional dependency. A node of interest (i.e., %HDD) is avoid performance accuracy was significantly correlated with the
described by the set of other nodes (i.e., predictive factors) that Physical Discomfort drinking motive (r = −0.43, p = 0.003), the de-
influence the node of interest. Here, the predictors included gree of problem drinking (i.e., AUDIT scores) at baseline (r = −0.47,
pain sensitivity, Physical Discomfort drinking motive, alcohol p < 0.001) as well as with total drinking quantity during the 12-­
craving, and the brain activation of the dACC during proactive month follow-­up period (r = −0.36, p = 0.01). The relationship be-
avoidance. The arcs represent the relationships between nodes tween proactive performance accuracy and %HDD, however, was
and the direction of an arc indicates the direction of dependence. not significant (r = −0.27, p = 0.07). Performance accuracy of other
The model describes the data well when the joint probability of task conditions did not show a significant relationship with drink-
the nodes is maximized. As joint probability varies depending on ing behaviour (p's > 0.33, Supplementary Table S1).
the direction of arcs, causes and consequences may be inferred.

Prior to model construction, we applied nonparanormal transfor- 3.2   |   Imaging Results

mation to all variables to correct for skewness and kurtosis using
the Huge package in R [43]. Next, to identify the optimal network We conducted a whole-­brain multiple regression to identify the
structure, we used the bnlearn R package with hill-­climbing learn- neural correlates of proactive avoidance that may also serve as a
ing algorithm [44]. Specifically, we determined the initial struc- predictor of drinking severity during the follow-­up period. Using
ture by adding arcs, removing them, and reversing their direction the contrast Go (Avoid–Win) > Nogo (Avoid–Win), we found acti-
to achieve the largest joint probability with the smallest number of vation in the dACC (coordinates xyz = 10, 24, 38, t = 4.75) during
arcs. Goodness-­of-­fit was measured by the Bayesian Information proactive avoidance negatively predicted %HDD (Figure 3).
Criterion (BIC). To minimize false positive in the identification
of arcs, we bootstrapped 10 000 samples, computed a network for
each sample, and averaged across the resulting networks to pro- 3.3   |   Predictors of Drinking Severity
duce an additional network structure. Finally, we created a final
structure containing arcs that appeared in at least 70% of bootstrap Hierarchical linear regression showed that the model with the
sampling in the previous step. dACC activation during proactive avoidance as the sole pre-
dictor was significant [F(1, 48) = 16.76, p < 0.001, R 2 = 0.24]. In
the next model, we included craving scores as the additional
3   |   Results predictor. The model was again significant (F(1, 47) = 14.8,
p < 0.001, adjusted R 2 = 0.36), and significantly improving
3.1   |   Behavioural and Clinical Results upon the previous model (ΔR 2 = 0.12, p = 0.001). The sub-
sequent model with the inclusion of drinking motive as an
We conducted a 2 (motivational goal: avoid vs. win) × 2 (response additional predictor was also significant (F(1, 46) = 12.57,
type: go vs. nogo) repeated-­measures ANOVA of performance p < 0.001, adjusted R 2 = 0.41), and significantly improving
accuracy. The results showed no significant main effect of upon the previous model (ΔR 2 = 0.05, p = 0.016). The next

FIGURE 2    |    Proactive avoidance performance accuracy showed a significant and negative relationship with (A) Physical discomfort drinking
motive, (B) problem drinking at baseline, (C) total drinking quantity during the 12-­month follow-­up period, and (D) percentage of heavy drinking
days. Note that scatter plots show regression residuals after the effects of depression, anxiety, sex, and age were removed.

5 of 11
 13691600, 2025, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/adb.70020 by Nigeria Hinari NPL, Wiley Online Library on [24/02/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
FIGURE 3    |    Whole-­brain multiple regression showed that activation in the dorsal anterior cingulate cortex during proactive avoidance negatively
predicted percentage of heavy drinking days during the 12-­month follow-­up period.

model with the addition of pain sensitivity as the predictor did

not significantly improve upon the previous model (p = 0.88).
However, the final model with the brain activity, craving,
and the interaction between pain sensitivity and Physical
Discomfort drinking motive was significant (F(1, 41 = 6.34,
p < 0.001, adjusted R 2 = 0.49), and significantly improving
upon the previous model (ΔR 2 = 0.08, p = 0.03, Supplementary
Table S2). Thus, measures of the brain activation during
proactive avoidance, craving, Physical Discomfort drinking
motive, and pain sensitivity at the baseline visit were found to
be predictors of %HDD during the follow-­up period.

3.4   |   Bayesian Network

Using the brain activation during proactive avoidance (i.e.,

dACC), Physical Discomfort drinking motive, and pain sensitiv-
ity, and alcohol craving, we described their probabilistic condi-
tional relationships in a prediction model of %HDD during the
follow-­up period. The model structure was consistent across the
three models we tested, including the initial model, averaged
model across bootstrap sampling models, and model with arcs
present in at least 70% of the bootstrap samplings. This structure
(Figure 4A) showed that heightened pain sensitivity at baseline
gave rise to the motive of drinking to avoid physical discomfort
which, in turn, increased the degree of craving. Both craving
and deficient activation of the dACC during proactive avoidance FIGURE 4    |    Bayesian network results showed the conditional proba-
predicted greater subsequent %HDD. bilistic relationships between the predictors at baseline and percentage
of heavy drinking days during the 12-­month follow-­up period. Arrow
thickness indicates arc strength, as measured by regression coefficients.
4   |   Discussion Direction of arrows indicates direction of dependence. The sign of the
arc strength indicates whether the relationship was positive or negative.
Using a probabilistic learning go/nogo task, we found that the
performance of proactive avoidance was negatively associated elevated craving ultimately led to higher %HDD over the follow-
with the degree of pain-­avoidance drinking motive, and prob- ing 12 months. Overall, by demonstrating a potentially depen-
lem drinking at baseline. This behavioural evidence indicated a dent relationship between deficits in proactive avoidance and
facilitating effect of proactive avoidance deficits on alcohol use. harmful drinking, we identified a pathway linking maladaptive
Our imaging results were converging, showing that activation in pain avoidance to alcohol misuse.
the dACC during proactive avoidance was also inversely related
with future %HDD. These regional activations, along with pain
sensitivity, pain-­avoidance drinking motive, and craving sig- 5   |   Proactive Avoidance Deficits and Alcohol use
nificantly predicted subsequent alcohol use severity. Employing
Bayesian network analysis, we identified a potential pathway of In response to negative physical and emotional states, drink-
harmful drinking in which drinkers' heightened pain sensitivity ers may engage in reactive avoidance and excessive use of al-
promoted motivation to drink as a strategy to avoid physical dis- cohol to alleviate pain. This maladaptive coping method is
comfort. This drinking motive then directly increased alcohol associated with the development and maintenance of AUD [45].
craving. Reduced brain activation to proactive avoidance and Alternatively, one can adopt proactive avoidance, a beneficial

6 of 11 Addiction Biology, 2025

 13691600, 2025, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/adb.70020 by Nigeria Hinari NPL, Wiley Online Library on [24/02/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
strategy that involves actively solving problems, regulating emo- Specifically, diminished activation of the dACC during pro-
tions, gaining sense of control, and seeking support in response active avoidance was predictive of %HDD during follow-­up.
to aversive events. As a component in the treatment of alcohol Previous research has suggested the critical involvement of
misuse, proactive avoidance has been shown to decrease con- the neural processes underlying avoidance learning in the
sumption in problem drinkers and the risk of relapse in those re- formation and maintenance of harmful drinking behaviour
ceiving treatment for AUD [46, 47]. Despite the protective role of [22]. However, little is known about their relationship with
proactive avoidance in the prevention and treatment of alcohol subsequent alcohol misuse. Here, we present evidence that
misuse, past studies have largely relied on questionnaires and the dACC may be part of a circuit that modulates proactive
self-­reports of drinkers' coping methods to measure proactive avoidance. Importantly, the dysfunctions of this region likely
avoidance. A study employing the Reward Incentive Delay with have a direct impact on future drinking.
Shock task recently examined the link between active avoidance
and impulsivity in heavy drinkers [21] but not in a longitudi- The role of the dACC in pain avoidance learning is complex.
nal setting. While the relationship of proactive avoidance with The dACC shows robust response to both pain anticipation
drinking behaviour represents an opportunity for its application [51] and physical pain induction [52] in neurotypical individu-
as a predictor of prospective alcohol use, this avenue was not als. Imaging meta-­analyses further reported increased dACC
previously explored. To our knowledge, the current work with activation during fear conditioning that involved responses to
the PLGT is the first to operationalize proactive avoidance and conditioned fear-­inducing vs. nonconditioned neutral stimuli
provide empirical evidence for the impact of its impairment on [53, 54]. These findings are consistent with electrophysiological
both current and future drinking behaviour. recordings in both humans and monkeys, which showed pain-­
responsive neuronal activities during pain anticipation [55], es-
In the PLGT, proactive avoidance referred to the ability to cape from pain [56], and learning of pain avoidance [57]. Such
learn via the action-­outcome association to initiate a response evidence supports dACC's role not only in pain processing but
to avoid painful electric shocks. Our results showed that go-­ also in pain avoidance and pain avoidance learning. It is worth
to-­avoid accuracy was negatively associated with drinking nothing that the dACC has significant projections to the motor
severity. This evidence suggests those with more severe prob- system [58] and to the midline thalamic nuclei, which them-
lem drinking at baseline may have difficulty learning to avoid selves contain nociceptive neurons [59]. There is also rich lit-
shocks via the proactive initiation of an action. Such difficulty erature of reinforcement learning documenting dACC response
then likely exerted a long-­term effect on subsequent drinking to prediction error [60, 61]. Thus, the dACC is likely critical in
behaviour. Thus, our finding of impaired action initiation adds integrating nociceptive information to influence motor response
to the literature which has focused on deficient inhibitory con- selection and learning associated with avoidance of noxious
trol in those with AUD and substance use disorders. Our work outcomes [62]. Inter-­connected with the amygdala, insula, and
is also in line with a previous study showing learning dysfunc- spinothalamic system, the dACC may channel the pain inputs
tions in drinkers who exhibited preference for behaviours with to subcortical and cortical motor centres to bias goal-­directed
risky consequences [21]. Thus, cognitive dysfunctions involved behaviours [63].
in alcohol misuse are likely complex, including failures to both
initiate beneficial and inhibit unfavourable responses. Which Animal lesion studies additionally implicated dACC in avoid-
impairment manifests as an avoidance learning dysfunction ance learning, specifically the acquisition and maintenance
likely depends on the specific motivational goals. of avoidance responses. For instance, dACC lesions in rabbits
severely disrupted the acquisition of discriminative avoidance
It is worth noting that there was not a significant relationship be- learning involving footshocks [64]. Similarly, lesions in rats
tween the drinking measures and go-­to-­w in accuracy. Therefore, reduced the aversive quality of repeated noxious tactile stim-
the poor learning observed in the current study was not due to ulation [65] and significantly decreased escape/avoidance be-
a failure to initiate an action in general but specific to actions haviour [66]. In contrast, glutamatergic activation of the dACC
aimed at avoiding pain. In the context of alcohol use, the impair- in rodents during conditioning produced avoidance learning in
ment may manifest as drinking as a reactive coping method. Such the absence of a peripheral noxious stimulus, leading to subse-
findings are consistent with earlier evidence of negative urgency quent avoidance behaviour [67].
[48] as well as drinking to alleviate physical pain and negative
emotions [49] in individuals with AUD. In contrast, those who In those with problem drinking, proactive avoidance dysfunc-
employ proactive problem-­focused strategies to manage negative tions may be related to structural abnormalities of the dACC.
situations are found to rely less on alcohol to cope, thus protect- Metabolic, molecular, and structural alterations of the dACC
ing them from engaging in detrimental drinking behaviour [50]. have indeed been implicated in alcohol misuse. In studies
Taken together, the current work provides support for a critical with MR spectroscopy, the levels of choline were found to be
role of proactive avoidance in mitigating alcohol use, offering po- elevated in the ACC in patients with neuropathic pain [68]
tential implications for the diagnosis and treatment of AUD. and predictive of alcohol craving and use severity in those
with AUD [69]. In a study of positron emission tomography
imaging, striatal dopamine D2/3 receptor availability was
6   |   Brain Correlates of Proactive Avoidance and negatively correlated with the levels of gamma-­a minobutyric
Alcohol use acid in the dACC in healthy controls but not in dependent
drinkers [70]. The latter finding suggested disruptions of fron-
We found that the neural correlates of proactive avoidance tal inhibitory GABAergic mechanisms, as also documented
played a significant role in determining future alcohol use. in individuals misusing other substances [71]. Further,

7 of 11
 13691600, 2025, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/adb.70020 by Nigeria Hinari NPL, Wiley Online Library on [24/02/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
meta-­a nalyses showed significantly reduced dACC grey mat- relates maladaptive regulation of negative states with alcohol as
ter volumes in those with AUD relative to healthy controls a pain-­relieving method [78]. Further, in our pathway, craving
[72, 73]. Importantly, the reduction was linked to changes in increased the probability of subsequent alcohol use while the
cognitive functions, including action execution, fear, and pain brain processes underlying proactive avoidance exerted the op-
perception [73]. As these domains are components of proac- posite effect. It should be noted that our Bayesian network re-
tive avoidance, it is plausible that volumetric reduction of the sults showed that the impacts of these pain-­related factors were
dACC contributes to deficits in proactive avoidance learning indirect as they interacted to predict future drinking via alcohol
and escalating drinking, as demonstrated in a longitudinal craving. Thus, our results highlight the importance of consid-
study of adolescents [74]. ering the interactions among different predictors as such inter-
actions may produce unique effects unaccounted for by each
Taken together, past research has widely implicated dACC dys- factor's separate contribution.
function in problem drinking. The current work extends this
literature by characterizing dACC's weakened role in proactive
avoidance in a potential pathway to dependent alcohol use. As 8   |   Limitations and Conclusions
proactive avoidance becomes impaired due to the dACC's disen-
gagement, individuals may resort to alcohol use as maladaptive The current study has several limitations. First, we did not have a
coping strategy to avoid painful physical and affective states. direct measure of pain avoidance during the follow-­up period. The
We showed that dACC dysfunction during proactive avoidance participants provided a self-­report of physical pain experienced on
learning may serve as a biomarker for future drinking sever- the scale of 0 (no pain whatsoever) to 10 (worst pain possible) in
ity, thus making the region a potential target for the treatment the preceding 30 days at each of the follow-­up interviews. While
of AUD. its relationship with drinking severity was significant, it did not
survive after controlling for age, sex, depression, and anxiety (data
not shown). Thus, it is important that future longitudinal stud-
7   |   The Relationship of Pain and Future Drinking ies directly assess pain avoidance and its effect on alcohol use.
Behaviour Second, our Bayesian network models did not show connections
for the neural correlates of proactive avoidance with pain sensitiv-
The current study provided convergent neural and behavioural ity and physical discomfort drinking motive. It is plausible that the
evidence for a robust link between pain-­related measures, crav- dACC's activity during proactive avoidance learning was primar-
ing, and future drinking severity. Heightened pain sensitivity ily involved in the integration of pain anticipation and avoidance
likely promoted the motive of alcohol seeking to avoid physical behaviour. Previous research has found that other subdivisions of
aversive states, subsequently eliciting alcohol craving and drink- the ACC (e.g., pregenual) or the somatosensory cortices as well
ing. Another significant pain-­related predictor of alcohol use is the as midcingulate cortex may be more selective to pain sensitivity
brain processes involved in proactive avoidance. Hypoactivity of [79]. Third, when we controlled for baseline drinking severity (i.e.,
the dACC during action initiation to avoid painful outcomes was AUDIT scores), in addition to age, sex, depression, and anxiety, the
associated with greater drinking severity during follow-­up. As final model in the hierarchical analysis did not significantly im-
such, our findings highlight responses to pain avoidance in the prove upon previous models. It is plausible that drinking severity
dACC as potential biomarkers for future drinking behaviour. It at baseline played a role in predicting future alcohol use. There
should be noted that we did not find a significant path between was a significant relationship between AUDIT scores and phys-
dACC and pain sensitivity or physical discomfort as a drinking ical discomfort as a drinking motive (data not shown), which in
motive. Thus, the model did not conclusively support our primary turn was positively associated with %HDD, suggesting that pain
hypothesis that the neural correlates of proactive avoidance defi- avoidance contributed to both current and future drinking se-
cits would modulate pain-­related factors in their prediction of fu- verity. Finally, we did not assess participants' smoking behaviour
ture drinking. during the follow-­up period. As smoking could represent another
pain avoidance strategy and potentially interact with alcohol use, it
Individuals with AUD and problem drinking have been found would be beneficial to consider smoking in the prediction of future
to be more sensitive to pain than healthy participants [25]. This drinking.
sensitivity motivates drinking [75] and is likely mediated by the
analgesic effects of ethanol [76] as well as reactive avoidance The current work demonstrates that drinking as a pain avoid-
as a maladaptive coping strategy [77]. Those with AUD prefer ance strategy is a complex behaviour, and the prediction of
drinking to alleviate pain to proactive avoidance which, while which requires assessment of multiple neural, motivational,
beneficial in the long run, is more effortful [18]. Despite such and psychological factors. Our findings specifically shed light
evidence, past literature had not characterized the link between on how pain sensitivity, pain-­avoidance drinking motive, crav-
pain sensitivity, avoidance drinking motive, and future alcohol ing, and weakened brain activation during proactive avoidance
use. Here, employing a longitudinal design and Bayesian net- may serve to predict subsequent alcohol use. Additionally, their
work analysis, we established a pathway to provide additional inter-­relationships represent a possible trajectory of escalation
information regarding the direction and dependency among to dependent drinking in those at risk of developing AUD or
the predictors. Specifically, we found that the pain sensitivity relapse in those seeking treatment for their harmful drinking
may have promoted future drinking indirectly via the motive behaviour. Thus, our findings offer further understanding of
of drinking to avoid physical discomfort which then promoted the neural, motivational, and cognitive processes perpetuat-
alcohol craving. Craving is a critical component of avoidance ing alcohol misuse and critical insights into the importance of
coping in those with problem drinking and AUD as it directly managing pain and negative emotions in the treatment of AUD.

8 of 11 Addiction Biology, 2025

 13691600, 2025, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/adb.70020 by Nigeria Hinari NPL, Wiley Online Library on [24/02/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Alcohol use,” Journal of Personality and Social Psychology 69, no. 5
(1995): 990–1005, https://​doi.​org/​10.​1037/​0 022-­​3514.​69.5.​990.
Author Contributions
10. M. D. Skinner and H. J. Aubin, “Craving's Place in Addiction The-
Study concept and design: Thang Le and Chiang-shan Li. Data collec- ory: Contributions of the Major Models,” Neuroscience and Biobehav-
tion: Thang Le and F. AnNa Hughes. Analysis and interpretation of ioral Reviews 34, no. 4 (2010): 606–623, https://​doi.​org/​10.​1016/j.​neubi​
data: Thang Le, Chiang-shan Li, and Takeyuki Oba. Drafting of the orev.​2 009.​11.​024.
manuscript: Thang Le. Critical revision of the manuscript for important
intellectual content: Chiang-shan Li. Statistical analysis: Thang Le and 11. R. Fu, D. Gregor, Z. Peng, J. Li, A. Bekker, and J. H. Ye, “Chronic
Takeyuki Oba. Obtained funding: Thang Le. Study supervision: Thang Intermittent Voluntary Alcohol Drinking Induces Hyperalgesia in
Le and Chiang-shan Li. All authors have made substantial contribu- Sprague-­Dawley Rats,” International Journal of Physiology, Pathophysiol-
tions to the design of the work, the analysis and interpretation of the ogy and Pharmacology 7, no. 3 (2015): 136–144.
data, or drafting of the work or its critical revision. 12. J. M. Townshend and T. Duka, “Avoidance of Alcohol-­Related Stim-
uli in Alcohol-­Dependent Inpatients,” Alcoholism, Clinical and Experi-
Acknowledgements mental Research 31, no. 8 (2007): 1349–1357, https://​doi.​org/​10.​1111/j.​
1530-­​0277.​2 007.​0 0429.​x.
The study was supported by NIH grants K99AA029716 (Le) and BBRF-­
Young Investigator Grant (Le). The authors are grateful for the support and 13. K. M. Stormark, N. P. Field, K. Hugdahl, and M. Horowitz, “Selec-
resources generously provided by the Connecticut Mental Health Center. tive Processing of Visual Alcohol Cues in Abstinent Alcoholics: An
Approach-­Avoidance Conflict?” Addictive Behaviors 22, no. 4 (1997):
509–519, https://​doi.​org/​10.​1016/​S 0306​- ­​4 603(96)​0 0051​-­​2 .
Conflicts of Interest
14. T. Chung, J. Langenbucher, E. Labouvie, R. J. Pandina, and R. H.
The authors declare no conflicts of interest. Moos, “Changes in Alcoholic patients' Coping Responses Predict 12-­
Month Treatment Outcomes,” Journal of Consulting and Clinical Psychol-
ogy 69, no. 1 (2001): 92–100, https://​doi.​org/​10.​1037/​0 022-­​0 06X.​69.1.​92.
Data Availability Statement
15. D. Brevers, A. Bechara, C. D. Kilts, et al., “Competing Motivations:
The data that support the findings of this study are available on request
Proactive Response Inhibition Toward Addiction-­Related Stimuli in
from the corresponding author. The data are not publicly available due
Quitting-­Motivated Individuals,” Journal of Gambling Studies 34, no. 3
to privacy or ethical restrictions.
(2018): 785–806, https://​doi.​org/​10.​1007/​s1089​9 -­​017-­​9722-­​2 .
16. D. Y. Kim and J. H. Lee, “The Effects of Training to Reduce Au-
References tomatic Action Tendencies Toward Alcohol Using the Virtual Alcohol
Approach-­Avoidance Task in Heavy Social Drinkers,” Cyberpsychology,
1. I. Rossow and E. Kuntsche, “Early Onset of Drinking and Risk of Behavior and Social Networking 22, no. 12 (2019): 794–798, https://​doi.​
Heavy Drinking in Young Adulthood—A 13-­Year Prospective Study,” org/​10.​1089/​cyber.​2 019.​0121.
Alcoholism, Clinical and Experimental Research 37 (2013): 37(s1),
https://​doi.​org/​10.​1111/j.​1530-­​0277.​2 012.​01924.​x. 17. A. E. Moser and H. M. Annis, “The Role of Coping in Relapse Crisis
Outcome: A Prospective Study of Treated Alcoholics,” Addiction 91, no. 8
2. K. Witkiewitz, R. Z. Litten, and L. Leggio, “Advances in the Science (1996): 1101–1114, https://​doi.​org/​10.​1046/j.​1360-­​0 443.​1996.​91811​013.​x.
and Treatment of Alcohol use Disorder,” Science Advances 5, no. 9
(2019): eaax4043, https://​doi.​org/​10.​1126/​sciadv.​aax4043. 18. C. Levin, M. Ilgen, and R. Moos, “Avoidance Coping Strategies Mod-
erate the Relationship Between Self-­Efficacy and 5-­Year Alcohol Treat-
3. S. J. Adamson, J. D. Sellman, and C. M. A. Frampton, “Patient Pre- ment Outcomes,” Psychology of Addictive Behaviors 21, no. 1 (2007):
dictors of Alcohol Treatment Outcome: A Systematic Review,” Journal 108–113, https://​doi.​org/​10.​1037/​0893-­​164X.​21.1.​108.
of Substance Abuse Treatment 36, no. 1 (2009): 75–86, https://​doi.​org/​10.​
1016/j.​jsat.​2 008.​05.​0 07. 19. A. Beck, T. Wüstenberg, A. Genauck, et al., “Effect of Brain Struc-
ture, Brain Function, and Brain Connectivity on Relapse in Alcohol-­
4. K. Witkiewitz and N. A. Villarroel, “Dynamic Association Between Dependent Patients,” Archives of General Psychiatry 69, no. 8 (2012):
Negative Affect and Alcohol Lapses Following Alcohol Treatment,” 842–852, https://​doi.​org/​10.​1001/​archg​enpsy​chiat​r y.​2 011.​2 026.
Journal of Consulting and Clinical Psychology 77, no. 4 (2009): 633–644,
https://​doi.​org/​10.​1037/​a0015647. 20. A. C. Lim, R. J. Green, E. N. Grodin, et al., “Alcohol Cue-­
Induced Ventral Striatum Activity Predicts Subsequent Alcohol Self-­
5. A. Heinz, A. Beck, S. M. Grüsser, A. A. Grace, and J. Wrase, “Identify- Administration,” Alcoholism, Clinical and Experimental Research 44,
ing the Neural Circuitry of Alcohol Craving and Relapse Vulnerability,” no. 6 (2020): 1224–1233, https://​doi.​org/​10.​1111/​acer.​14342​.
Addiction Biology 14, no. 1 (2009): 108–118.
21. E. N. Grodin, L. Sussman, K. Sundby, et al., “Neural Correlates of
6. S. Edwards, L. F. Vendruscolo, N. W. Gilpin, M. Wojnar, and Compulsive Alcohol Seeking in Heavy Drinkers,” Biological Psychiatry:
K. Witkiewitz, “Alcohol and Pain: A Translational Review of Preclinical Cognitive Neuroscience and Neuroimaging 3, no. 12 (2018): 1022–1031,
and Clinical Findings to Inform Future Treatment Strategies,” Alco- https://​doi.​org/​10.​1016/j.​bpsc.​2 018.​0 6.​0 09.
holism, Clinical and Experimental Research 44, no. 2 (2020): 368–383,
https://​doi.​org/​10.​1111/​acer.​14260​. 22. A. V. Apkarian, V. Neugebauer, G. Koob, et al., “Neural Mechanisms
of Pain and Alcohol Dependence,” Pharmacology, Biochemistry, and Be-
7. K. G. Hill, J. D. Hawkins, J. A. Bailey, R. F. Catalano, R. D. Abbott, havior 112 (2013): 34–41, https://​doi.​org/​10.​1016/j.​pbb.​2 013.​0 9.​0 08.
and V. B. Shapiro, “Person-­Environment Interaction in the Prediction
of Alcohol Abuse and Alcohol Dependence in Adulthood,” Drug and 23. M. Egli, G. F. Koob, and S. Edwards, “Alcohol Dependence as a
Alcohol Dependence 110, no. 1–2 (2010): 62–69, https://​doi.​org/​10.​1016/j.​ Chronic Pain Disorder,” Neuroscience and Biobehavioral Reviews 36,
druga​lcdep.​2 010.​02.​0 05. no. 10 (2012): 2179–2192, https://​doi.​org/​10.​1016/j.​neubi​orev.​2 012.​
07.​010.
8. E. L. Zale, S. A. Maisto, and J. W. Ditre, “Interrelations Between Pain
and Alcohol: An Integrative Review,” Clinical Psychology Review 37 24. T. M. Le, T. Oba, L. Couch, L. McInerney, and C. R. Li, “The Neu-
(2015): 57–71, https://​doi.​org/​10.​1016/j.​cpr.​2 015.​02.​0 05. ral Correlates of Individual Differences in Reinforcement Learning
During Pain Avoidance and Reward Seeking,” eNeuro 11(2):ENEU-
9. M. L. Cooper, M. R. Frone, M. Russell, and P. Mudar, “Drinking to RO.0437-­23.2024 (2024): 1–16, https://​doi.​org/​10.​1523/​ENEURO.​0 437-­​
Regulate Positive and Negative Emotions: A Motivational Model of 23.​2 024.

9 of 11
 13691600, 2025, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/adb.70020 by Nigeria Hinari NPL, Wiley Online Library on [24/02/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
25. M. Adrienne McGinn, K. N. Edwards, and S. Edwards, “Chronic 40. H. Y. Hung, W. C. Chien, C. H. Chung, et al., “Patients With Al-
Inflammatory Pain Alters Alcohol-­Regulated Frontocortical Signaling cohol use Disorder Increase Pain and Analgesics use: A Nationwide
and Associations Between Alcohol Drinking and Thermal Sensitivity,” Population-­Based Cohort Study,” Drug and Alcohol Dependence 229, no.
Neurobiology of Pain 8, no. July (2020): 100052, https://​doi.​org/​10.​1016/j.​ 161 (2021): 109102, https://​doi.​org/​10.​1016/j.​druga​lcdep.​2 021.​109102.
ynpai.​2 020.​100052.
41. K. B. Carey, “Situational Determinants of Heavy Drinking Among
26. J. Bauer, A. Pedersen, N. Scherbaum, et al., “Craving in Alcohol-­ College Students,” Journal of Counseling Psychology 40, no. 2 (1993):
Dependent Patients After Detoxification Is Related to Glutamatergic 217–220.
Dysfunction in the Nucleus Accumbens and the Anterior Cingulate
42. R. R. Bouckaert, Bayesian Belief Networks: From Construction to In-
Cortex,” Neuropsychopharmacology 38, no. 8 (2013): 1401–1408, https://​
ference (Utrecht, Netherlands: University of Utrecht Press, 1995).
doi.​org/​10.​1038/​npp.​2 013.​45.
43. T. Zhao, H. Liu, K. Roeder, J. Lafferty, and L. Wasserman, “The
27. Y. Mashhoon, C. Czerkawski, D. J. Crowley, J. E. Cohen-­Gilbert, J.
Huge Package for High-­Dimensional Undirected Graph Estimation in
T. Sneider, and M. M. Silveri, “Binge Alcohol Consumption in Emerging
R,” Journal of Machine Learning Research 13 (2012): 1059–1062.
Adults: Anterior Cingulate Cortical “Thinness” Is Associated With Al-
cohol use Patterns,” Alcoholism, Clinical and Experimental Research 38, 44. M. Scutari, “Learning Bayesian Networks With the Bnlearn R Pack-
no. 7 (2014): 1955–1964, https://​doi.​org/​10.​1111/​acer.​12475​. age,” Journal of Statistical Software 35, no. 3 (2010): 1–22, https://​doi.​
org/​10.​18637/​​jss.​v 035.​i03.
28. J. B. Saunders, O. G. Aasland, T. F. Babor, J. R. De La Fuente, and M.
Grant, “Development of the Alcohol use Disorders Identification Test 45. P. Hasking, M. Lyvers, and C. Carlopio, “The Relationship Between
(AUDIT): WHO Collaborative Project on Early Detection of Persons Coping Strategies, Alcohol Expectancies, Drinking Motives and Drink-
With Harmful Alcohol Consumption-­I I,” Addiction Published Online ing Behaviour,” Addictive Behaviors 36, no. 5 (2011): 479–487, https://​
88 (1993): 791–804, https://​doi.​org/​10.​1111/j.​1360-­​0 443.​1993.​tb020​93.​x. doi.​org/​10.​1016/j.​addbeh.​2 011.​01.​014.
29. A. Nadkarni, A. Garber, S. Costa, et al., “Auditing the AUDIT: A 46. J. Feil and P. Hasking, “The Relationship Between Personality, Cop-
Systematic Review of cut-­Off Scores for the Alcohol use Disorders Iden- ing Strategies and Alcohol use,” Addiction Research and Theory 16, no. 5
tification Test (AUDIT) in low-­and Middle-­Income Countries,” Drug (2008): 526–537, https://​doi.​org/​10.​1080/​16066​35080​2 025714.
and Alcohol Dependence 202, no. July (2019): 123–133, https://​doi.​org/​
47. C. T. McCabe, S. C. Roesch, and A. A. Aldridge-­G erry, ““Have a
10.​1016/j.​druga​lcdep.​2 019.​0 4.​031.
drink, you'll feel better.” Predictors of daily alcohol consumption
30. H. M. Annis, J. M. Graham, and C. S. Davis, “Inventory of Drinking among extraverts: The mediational role of coping,” Anxiety, Stress
Situations (IDS) Users Guide; 1987; Toronto,” Addict Res Found Ontario Coping 26, no. 2 (2013): 121–135, https://​doi.​org/​10.​1080/​10615​8 06.​
Published online (1987): 1–116. 2012.​657182.
31. R. Ruscheweyh, M. Marziniak, F. Stumpenhorst, J. Reinholz, and S. 48. D. S. Chester, D. R. Lynam, R. Milich, D. K. Powell, A. H. Andersen,
Knecht, “Pain Sensitivity can Be Assessed by Self-­R ating: Development and C. N. DeWall, “How Do Negative Emotions Impair Self-­Control?
and Validation of the Pain Sensitivity Questionnaire,” Pain 146, no. 1–2 A Neural Model of Negative Urgency,” NeuroImage 132 (2016): 43–50,
(2009): 65–74, https://​doi.​org/​10.​1016/j.​pain.​2 009.​06.​020. https://​doi.​org/​10.​1016/j.​neuro​image.​2 016.​02.​024.
32. A. T. Beck, “An Inventory for Measuring Depression,” Archives of 49. N. Maleki, K. Tahaney, B. L. Thompson, and M. Oscar-­Berman, “At
General Psychiatry 4, no. 6 (1961): 561, https://​doi.​org/​10.​1001/​archp​syc.​ the Intersection of Alcohol use Disorder and Chronic Pain,” Neuropsy-
1961.​01710​12003​1004. chology 33, no. 6 (2019): 795–807, https://​doi.​org/​10.​1037/​neu00​0 0558​.
33. R. L. Spitzer, K. Kroenke, J. B. W. Williams, and B. Löwe, “A Brief 50. R. Walker and R. S. Stephens, “Protective Behavioral Strategies Me-
Measure for Assessing Generalized Anxiety Disorder,” Archives of In- diate Problem-­Focused Coping and Alcohol use in College Students,”
ternal Medicine 166, no. 10 (2006): 1092, https://​doi.​org/​10.​1001/​archi​ Addictive Behaviors 39, no. 6 (2014): 1033–1037, https://​doi.​org/​10.​
nte.​166.​10.​1092. 1016/j.​addbeh.​2 014.​02.​0 06.
34. W. R. Miller and F. K. Del Boca, “Measurement of Drinking Behav- 51. S. Palermo, F. Benedetti, T. Costa, and M. Amanzio, “Pain Antic-
ior Using the Form 90 Family of Instruments,” Journal of Studies on ipation: An Activation Likelihood Estimation meta-­A nalysis of Brain
Alcohol Supplement 12, no. 12 (1994): 112–118, https://​doi.​org/​10.​15288/​​ Imaging Studies,” Human Brain Mapping 36, no. 5 (2015): 1648–1661,
jsas.​1994.​s12.​112. https://​doi.​org/​10.​1002/​hbm.​22727​.
35. K. Witkiewitz, A. D. Wilson, M. R. Pearson, et al., “Temporal Sta- 52. A. Xu, B. Larsen, E. B. Baller, et al., “Convergent Neural Represen-
bility of Heavy Drinking Days and Drinking Reductions Among Heavy tations of Experimentally-­Induced Acute Pain in Healthy Volunteers:
Drinkers in the COMBINE Study,” Alcoholism, Clinical and Experimental A Large-­Scale fMRI meta-­A nalysis,” Neuroscience and Biobehavioral
Research 41, no. 5 (2017): 1054–1062, https://​doi.​org/​10.​1111/​acer.​13371​. Reviews 2020, no. 112 (2019): 300–323, https://​doi.​org/​10.​1016/j.​neubi​
orev.​2 020.​01.​0 04.
36. B. A. Flannery, J. R. Volpicelli, and H. M. Pettinati, “Psychometric
Properties of the Penn Alcohol Craving Scale,” Alcoholism, Clinical and 53. A. Etkin and T. D. Wager, “Functional Neuroimaging of Anxiety:
Experimental Research 23, no. 8 (1999): 1289–1295, https://​doi.​org/​10.​ A meta-­A na Lysis of Emotional Processing in PTSD, Social Anxiety
1111/j.​1530-­​0277.​1999.​tb043​49.​x. Disorder, and Specific Phobia,” American Journal of Psychiatry 164,
no. 10 (2007): 1476–1488, https://​doi.​org/​10.​1176/​appi.​ajp. ​2 007.​
37. C. W. Woo, A. Krishnan, and T. D. Wager, “Cluster-­E xtent Based
07030504.
Thresholding in fMRI Analyses: Pitfalls and Recommendations,” Neu-
roImage 91 (2014): 412–419, https://​doi.​org/​10.​1016/j.​neuro​image.​2 013.​ 54. M. L. Mechias, A. Etkin, and R. Kalisch, “A meta-­A nalysis of In-
12.​058. structed Fear Studies: Implications for Conscious Appraisal of Threat,”
NeuroImage 49, no. 2 (2010): 1760–1768, https://​doi.​org/​10.​1016/j.​neuro​
38. A. Eklund, T. E. Nichols, and H. Knutsson, “Cluster Failure: Why
image.​2 009.​0 9.​0 40.
fMRI Inferences for Spatial Extent Have Inflated False-­Positive Rates,”
Proceedings of the National Academy of Sciences 113, no. 28 (2016): 55. W. D. Hutchison, K. D. Davis, A. M. Lozano, R. R. Tasker, and J. O.
7900–7905, https://​doi.​org/​10.​1073/​pnas.​16024​13113​. Dostrovsky, “Pain-­Related Neurons in the Human Cingulate Cortex,”
Nature Neuroscience 2, no. 5 (1999): 403–405, https://​doi.​org/​10.​1038/​8065.
39. J. A. Cucinello-­Ragland and S. Edwards, Neurobiological Aspects of
Pain in the Context of Alcohol Use Disorder, Vol. 157, 1st ed., (Minneapolis, 56. K. Iwata, H. Kamo, A. Ogawa, et al., “Anterior Cingulate Corti-
MN: Elsevier Inc. 2021), 1–29, https://​doi.​org/​10.​1016/​bs.​irn.​2020.​09.​0 01. cal Neuronal Activity During Perception of Noxious Thermal Stimuli

10 of 11 Addiction Biology, 2025

 13691600, 2025, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/adb.70020 by Nigeria Hinari NPL, Wiley Online Library on [24/02/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
in Monkeys,” Journal of Neurophysiology 94, no. 3 (2005): 1980–1991, meta-­A nalysis,” JAMA Psychiatry 74, no. 5 (2017): 511–519, https://​doi.​
https://​doi.​org/​10.​1152/​jn.​0 0190.​2 005. org/​10.​1001/​jamap​sychi​atry.​2 017.​0135.
57. T. Koyama, K. Kato, Y. Z. Tanaka, and A. Mikami, “Anterior Cingu- 72. L. Li, H. Yu, Y. Liu, et al., “Lower Regional Grey Matter in Alcohol
late Activity During Pain-­Avoidance and Reward Tasks in Monkeys,” use Disorders: Evidence From a Voxel-­Based meta-­Analysis,” BMC Psy-
Neuroscience Research 39, no. 4 (2001): 421–430, https://​doi.​org/​10.​1016/​ chiatry 21, no. 1 (2021): 1–11, https://​doi.​org/​10.​1186/​s1288​8-­​021-­​03244​-­​9.
S0168​- ­​0102(01)​0 0197​-­​3.
73. C. Spindler, S. Trautmann, N. Alexander, et al., “Meta-­A nalysis of
58. O. Devinsky, M. J. Morrell, and B. A. Vogt, “Contributions of Ante- Grey Matter Changes and Their Behavioral Characterization in Patients
rior Cingulate Cortex to Behaviour,” Brain 118 (1995): 279–306, https://​ With Alcohol use Disorder,” Scientific Reports 11, no. 1 (2021): 1–15,
doi.​org/​10.​1093/​brain/​​118.1.​279. https://​doi.​org/​10.​1038/​s 4159​8-­​021-­​8 4804​-­​7.
59. B. Vogt, D. Rosene, and D. Pandya, “Thalamic and cortical afferents 74. A. Cheetham, N. B. Allen, S. Whittle, J. Simmons, M. Yücel, and
differentiate anterior from posterior cingulate cortex in the monkey,” D. I. Lubman, “Volumetric Differences in the Anterior Cingulate Cortex
Science (80-­) 204, no. 4389 (1979): 205–207, https://​doi.​org/​10.​1126/​scien​ Prospectively Predict Alcohol-­Related Problems in Adolescence,” Psy-
ce.​107587. chopharmacology 231, no. 8 (2014): 1731–1742, https://​doi.​org/​10.​1007/​
s0021​3 -­​014-­​3 483-­​8.
60. J. W. Brown and T. S. Braver, “Learned Predictions of Error Like-
lihood in the Anterior Cingulate Cortex,” Science (80-­). 2011;1118 307 75. T. M. Le, S. Zhornitsky, W. Wang, J. Ide, S. Zhang, and C.-­S. R. Li,
(2005): 1118–1122, https://​doi.​org/​10.​1126/​scien​ce.​1105783. “Posterior Cingulate Cortical Response to Active Avoidance Mediates
the Relationship Between Punishment Sensitivity and Problem Drink-
61. W. H. Alexander and J. W. Brown, “The Role of the Anterior Cingu- ing,” Journal of Neuroscience 39, no. 32 (2019): 6354–6364, https://​doi.​
late Cortex in Prediction Error and Signaling Surprise,” Topics in Cogni- org/​10.​1523/​jneur​osci.​0508-­​19.​2 019.
tive Science 11, no. 1 (2019): 119–135, https://​doi.​org/​10.​1111/​tops.​12307​.
76. T. Thompson, C. Oram, C. U. Correll, S. Tsermentseli, and B. Stubbs,
62. M. F. S. Rushworth, T. E. J. Behrens, P. H. Rudebeck, and M. E. “Analgesic Effects of Alcohol: A Systematic Review and Meta-­A nalysis
Walton, “Contrasting Roles for Cingulate and Orbitofrontal Cortex in of Controlled Experimental Studies in Healthy Participants,” Journal
Decisions and Social Behaviour,” Trends in Cognitive Sciences 11, no. 4 of Pain 18, no. 5 (2017): 499–510, https://​doi.​org/​10.​1016/j.​jpain.​2 016.​
(2007): 168–176, https://​doi.​org/​10.​1016/j.​tics.​2 007.​01.​0 04. 11.​0 09.
63. A. J. Shackman, T. V. Salomons, H. A. Slagter, A. S. Fox, J. J. Winter, 77. P. L. Brennan, K. K. Schutte, and R. H. Moos, “Pain and use of
and R. J. Davidson, “The Integration of Negative Affect, Pain and Cog- Alcohol to Manage Pain: Prevalence and 3-­ Year Outcomes Among
nitive Control in the Cingulate Cortex,” Nature Reviews. Neuroscience Older Problem and non-­problem Drinkers,” Addiction 100, no. 6 (2005):
12, no. 3 (2011): 154–167, https://​doi.​org/​10.​1038/​nrn2994. 777–786, https://​doi.​org/​10.​1111/j.​1360-­​0 443.​2 005.​01074.​x.
64. M. Gabriel, Y. Kubota, S. Sparenborg, K. Straube, and B. A. Vogt, 78. J. T. Waddell, K. J. Sher, and T. M. Piasecki, “Coping Motives and
“Effects of Cingulate Cortical Lesions on Avoidance Learning and Negative Affect: An Ecological Study of the Antecedents of Alcohol
Training-­Induced Unit Activity in Rabbits,” Experimental Brain Re- Craving and Alcohol use,” Psychology of Addictive Behaviors Published
search 86 (1991): 600. Online 35 (2021): 565–576, https://​doi.​org/​10.​1037/​adb00​0 0696​.
65. C. J. LaBuda and P. N. Fuchs, “Attenuation of Negative Pain Affect 79. A. Rogachov, J. C. Cheng, N. Erpelding, K. S. Hemington, A. P.
Produced by Unilateral Spinal Nerve Injury in the rat Following Ante- Crawley, and K. D. Davis, “Regional Brain Signal Variability: A Novel
rior Cingulate Cortex Activation,” Neuroscience 136, no. 1 (2005): 311– indicator of Pain Sensitivity and Coping,” Pain 157, no. 11 (2016): 2483–
322, https://​doi.​org/​10.​1016/j.​neuro​scien​ce.​2 005.​07.​010. 2492, https://​doi.​org/​10.​1097/j.​pain.​0 0000​0 0000​0 00665.
66. S. C. LaGraize, C. J. Labuda, M. A. Rutledge, R. L. Jackson,
and P. N. Fuchs, “Differential Effect of Anterior Cingulate Cortex Le-
sion on Mechanical Hypersensitivity and Escape/Avoidance Behavior Supporting Information
in an Animal Model of Neuropathic Pain,” Experimental Neurology
Additional supporting information can be found online in the
188, no. 1 (2004): 139–148, https://​doi.​org/​10.​1016/j.​expne​urol.​2 004.​
Supporting Information section.
04.​0 03.
67. J. P. Johansen and H. L. Fields, “Glutamatergic Activation of Ante-
rior Cingulate Cortex Produces an Aversive Teaching Signal,” Nature
Neuroscience 7, no. 4 (2004): 398–403, https://​doi.​org/​10.​1038/​nn1207.
68. E. Widerström-­Noga, P. M. Pattany, Y. Cruz-­A lmeida, et al., “Me-
tabolite Concentrations in the Anterior Cingulate Cortex Predict High
Neuropathic Pain Impact After Spinal Cord Injury,” Pain 154, no. 2
(2013): 204–212, https://​doi.​org/​10.​1016/j.​pain.​2 012.​07.​022.
69. E. N. Grodin, E. M. Burnette, J. O'Neill, J. Alger, and L. A. Ray,
“Alcohol Craving and Severity Are Associated With Dorsal Anterior
Cingulate Choline Levels in Individuals With an Alcohol use Disorder,”
Alcohol and Alcoholism 58, no. 3 (2023): 289–297, https://​doi.​org/​10.​
1093/​alcalc/​agad014.
70. G. Spitta, T. Gleich, A. Rosenthal, et al., “Correlation of Striatal Do-
pamine D2/3 Receptor Availability With GABA Level in the Anterior
Cingulate Cortex in Healthy Controls but Not in Alcohol-­Dependent
Subjects and Individuals at High Risk: A Multimodal Magnetic Reso-
nance Spectroscopy and Positron,” Addiction Biology 29, no. 6 (2024):
1–10, https://​doi.​org/​10.​1111/​adb.​13424​.
71. A. H. Ashok, Y. Mizuno, N. D. Volkow, and O. D. Howes, “Associ-
ation of Stimulant use With Dopaminergic Alterations in Users of Co-
caine, Amphetamine, or Methamphetamine a Systematic Review and

11 of 11