Neural responses to alcohol-related cues and its associations with attributed
cue-value and cue-valence, and cue-elicited craving: A pilot fMRI study in
heavy drinking students
Annelieke Hagen1 Abstract
Most studies in alcohol cue-reactivity research, to date, have examined the association between blood-oxygenation-level-dependent (BOLD) responses to alcohol-related cues versus alcohol-unrelated cues and self-reported craving. In this pilot study this research is extended by exploring the associations between alcohol-cue-induced BOLD responses and measures of perceived stimulus value (assessed via a "willingness-to-pay" question), and of stimulus valence (assessed via a picture IAT) in twelve heavy drinking students (5 female) aged 18-21. In addition the relationship between severity of drinking (assessed with the AUDIT), self-reported alcohol craving, perceived stimulus value, and stimulus valence are explored. No correlations between severity of drinking and any of the behavioural measures were found. A high positive correlation between the perceived value of the alcohol-related stimuli and self-reported alcohol craving was found. Self-reported craving correlated negatively with increased activation within the superior frontal cortex. Perceived value of the alcohol stimuli correlated positively with increased activations in the superior frontal cortex, inferior temporal cortex and inferior orbital cortex. Neither positive nor negative correlations between alcohol-cue-induced BOLD responses and valence of the alcohol-related stimuli were found. Results, implications and limitations are discussed.
Keywords: fMRI, cue-reactivity, value, valence, craving, alcohol, incentive sensitization, addiction
1. Introduction
According to the incentive sensitization theory of addiction (Robinson & Berridge, 1993), repeated drug use leads to changes in the mesocorticolimbic dopamine reward system, which projects from the ventral tegmental area (VTA) of the midbrain to the nucleus accumbens (NAc), amygdala, and prefrontal cortex. When faced with an appetitive stimulus, there is an immediate increase in phasic dopamine transmission in the VTA, which signals that a reward is expected, and influences the
behaviour, attention, and decision making of an organism. All drugs of abuse increase dopamine levels in the VTA. Consequently, following repeated administration of a drug, the reward circuit becomes
1Department of developmental psychology, University of Amsterdam Student number: 10001503, contact: anneliekehagen@hotmail.com Supervised by: dr. K. Nikolaou, contact: k.nikolaou@uva.nl
hypersensitive in a way that a tremendous, pathological level of incentive salience is attributed to the drug and to stimuli that predict the availability of the drug.
Alcohol-related cues (such as the flavour of beer) can stimulate dopamine release from mesocorticolimbic dopamine neurons (Oberlin et al., 2013). These stimuli, relative to alcohol-unrelated stimuli, can also elicit brain activations [as assessed by functional magnetic resonance imaging (fMRI), positron emission tomography (PET) and single photon emission computed
tomography (SPECT)] in a number of regions within the mesocorticolimbic reward system, including the VTA, the ventral striatum (VS), and ventromedial prefrontal (vmPFC) and orbitofrontal (OFC) cortices (see Schacht, Anton & Myrick, 2012). The pattern of cue-induced brain activation has been suggested to represent a core component of alcohol-induced neuropathology (Heinz, Beck, Grüsser, Grace & Wrase, 2009).
Consequently, more and more attention has been devoted to alcohol cue reactivity (i.e. the reaction to alcohol-related stimuli) in the study of the neural aspects of alcohol use disorders. In fMRI cue reactivity studies, heavy alcohol users or alcohol dependent patients, are typically exposed to alcohol-related cues and to alcohol-unrelated control cues. These cues are often visual but can also be auditory, tactile, haptic, olfactory or gustatory. The within-subject measure of interest is the
comparison between the neural responses to alcohol-related cues, versus the neural responses to alcohol-unrelated cues (e.g. Filbey et al., 2008; Park et al., 2007).
Even though different fMRI studies reveal variance in activation of brain regions in response to alcohol-related stimuli, some main regions have been identified by most of the studies
(Mendelssohn, Kasper & Tauscher 2004; cited in Heinz et al., 2009). Increased blood-oxygenation-level-dependent (BOLD) responses while viewing alcohol-related relative to alcohol-unrelated cues have been found in: (a) ventral [including the nucleus accumbens (NAc)], and dorsal regions of the striatum, which have been respectively associated with connecting motivational aspects of salient stimuli with motor reactions and with reward (Braus et al., 2001; Kühn & Gallinat, 2011; Wrase et al., 2002, 2007; Volkow, Wang, Fowler, Tomasi & Telang, 2010), and behavioural habit formation (Modell & Mountz, 1995; Grüsser et al., 2004; Volkow et al., 2010);(b) limbic and inferior temporal regions such as the amygdala and the hippocampus, reflecting conditioned learning and the processing of learned reward values of conditioned cues and context (Jasinska, Stein, Kaiser, Naumer &
Yalachkov., 2014); (c) the anterior cingulate cortex (ACC) and the adjacent prefrontal cortex (PFC), which are associated with attention and memory processes, encoding the motivational value of stimuli, and executive control (Grüsser et al., 2004; Heinz et al., 2004; Myrick et al., 2004; Tapert et al., 2004; Schacht et al., 2012; Volkow et al., 2010); (d) the OFC, which is involved in the evaluation of the value of goals in decision making (Wrase et al., 2002; Myrick et al., 2004; Volkow et al., 2010); as well as in motor and visual cortices (Schacht et al., 2012).
Robinson and Berridge (1993, 2002, 2008) suggested that repeated drug use only sensitizes neural systems that mediate the motivational process of incentive salience [(wanting/craving) i.e.
dopaminergic reward circuits] but not those that mediate the pleasurable effects of the drug (liking). So in the development of addiction the degree of wanting/craving the drug increases
disproportionately to the degree of liking the drug (Robinson & Berridge, 2008). It is therefore expected that measures of wanting/craving alcohol, should correlate positively with the severity of alcohol use. Behavioural studies that have assessed the degree of craving alcohol, in the presence of alcohol-predicting cues, have shown that heavy drinkers, as assessed by the AUDIT, report more craving for alcohol than light drinkers (Papachristou, Nederkoorn, Havermans, van der Horst & Jansen, 2012).
In addition, alcohol-cue-induced BOLD responses in regions associated with the reward network, should also correlate positively with the degree of craving alcohol in the presence of alcohol-related stimuli. Indeed, fMRI studies in which participants are asked to indicate how much they crave alcohol, either during presentation of alcohol-related vs. alcohol-unrelated cues in the scanner (e.g. Park et al., 2007), or post-scanning while seeing these cues for the second time (e.g. Fryer et al., 2013), have found that heavy social drinkers and individuals with alcohol use disorder (AUD) showed positive correlations between self-reported alcohol craving and alcohol-cue-induced activity in the VS (including the NAc; see Kühn and Gallinat, 2011), medial prefrontal cortex (mPFC; Fryer et al., 2013), and the OFC (Filbey et al., 2008; Myrick et al., 2004). Positive correlations with self-reported craving have also been found with activations within the DS and motor regions (e.g. pre- and post-central gyri; Kühn and Gallinat, 2011; Park et sl., 2007; Tapert et al., 2003); in temporal and parietal areas (e.g. fusiform gyrus; Park et al., 2007; Tapert et al., 2003; parahippocampal gyrus; Park et al., 2007), in visual cortical regions (e.g. lingual gyrus; Kühn and Gallinat, 2011), as well as in regions of the prefrontal cortex (e.g. ACC; Myrick et al., 2004; and dorsolateral prefrontal cortex (DLPFC; Park et al., 2007) and the cerebellum (Fryer et al., 2013). Negative correlations were found with the right precuneus / cuneus and right inferior frontal cortex (Tapert et al., 2003).
The valence alcohol implicit association test (valence-IAT) has been used to measure the implicit positive/negative valence attributed to alcohol-related cues. It is therefore not a direct measure of alcoholliking, but it is plausible that attributions towards an alcoholrelated (i.e. alcohol
-predicting) cue, correlate with positive/negative attributions toward alcohol itself, since the alcohol-related cues have become conditioned stimuli alcohol-related to the drug (Carter & Tiffany, 1999). In the valence-IAT, stimuli words or pictures that belong to one of two target- (e.g. alcohol / soda), or one of two attribution (e.g. positive / negative) categories, appear one by one on the computer screen.
Participants categorize the presented stimuli by pressing one of two keys. The logic behind the IAT is that performance should be better (i.e. faster and more accurate) when associated categories are assigned to the same response, than when associated categories are assigned to different responses. Using a word valence-IAT, some behavioural studies have reported associations between the ‘alcohol’ target and the ‘negative’ attribution category in heavy and light drinkers, but this association was not correlated with drinking behaviour (Wiers, van Woerden, Smulders & de Jong, 2002; Wiers, van de
Luitgaarden, van den Wildenberg & Smulders, 2005; Houben & Wiers, 2007). Another study however, showed that implicit positive associations with alcohol, correlated with a higher AUDIT score, as assessed by a unipolar (i.e. with one target category) word valence IAT (Houben & Wiers, 2008).
In this pilot study, heavy alcohol drinking students were presented with pictures of alcohol-containing drinks, and with pictures of drinks that did not contain alcohol in a cue-reactivity task in the MRI scanner. Post-scanning, participants saw the same images again, and were asked to rate how much they would like to drink alcohol (i.e. self-reported craving for alcohol). Participants also completed a picture alcohol valence-IAT that used the same images. Based on previous studies, we predicted that the severity of drinking, as assessed with the AUDIT, would correlate positively with both self-reported craving for alcohol in the presence of alcohol-related cues, and implicit positive attributions to alcohol-associated cues (i.e. valence-IAT scores). With respect to the fMRI data, we predicted that viewing alcohol-related cues relative to alcohol-unrelated cues would be related with increased BOLD responses in regions within the brain’s reward system (i.e. regions within the striatum; amygdala; OFC; and vmPFC), as well as in the hippocampus, visual, and motor cortices, consistent with previous studies. In addition, we predicted that increased activations in the VS, DS, OFC, motor, temporal, visual and prefrontal regions, would correlate positively with self-reported cue-induced craving for alcohol, consistent with previous studies.
The present study extended previous studies in two ways. The first was by examining the association between alcohol-cue-induced BOLD responses and IAT scores. To our knowledge no study to date has examined the relationship between implicit positive/negative valence attributed to alcohol-associated cues and alcohol cue-induced BOLD-responses. This extension is therefore exploratory. However a study in which increased BOLD responses were correlated with valence ratings of emotional pictures, showed that there were negative correlations between watching pictures with positive valence and increased BOLD responses in the dorsomedial prefrontal cortex (dmPFC; bilaterally), the ACC (bilaterally), the insula (bilaterally), and midbrain tegmentum (Viinikainen, Jääskeläinen, Alexandrov, Balk, Autti & Sams, 2010). Viewing pictures with negative valence correlated positively with increased BOLD responses in the right lateral sulcus, insula (bilaterally) right dorsolateral prefrontal cortex (dlPFC) and right amygdala (Viinikainen et al., 2010). Based on this previous study, and the previous finding that both heavy and light drinkers have an association between alcohol and the negative attribution category (Wiers, van Woerden, Smulders & de Jong, 2002; Wiers, van de Luitgaarden, van den Wildenberg & Smulders, 2005; Houben & Wiers, 2007), we predicted that the valence IAT scores would correlate positively with increased activation in the right lateral sulcus, insula (bilaterally), right dlPFC, and the right amygdala.
The second extension involved the inclusion of a behavioural measure of the perceived value attributed to alcohol-related cues. In the post scanning presentation of the cue-reactivity task’s images, participants did not only rate how much they would like to drink alcohol (craving), but also how much
they would be willing to pay (WTP) for a glass of the depicted drink. WTP ratings are seen as a measure of the subjective value attributed to a stimulus, as a result of repeated pairings with the drug reward, and should therefore correlate with craving and the severity of alcohol use. Since people who crave alcohol more will be more inclined to act on this and do more (pay more) to get the drink they crave (e.g. Plassman, O’Doherty & Rangel, 2007; Schonberg et al., 2014). However, no study to date has explored these associations. We predicted that the perceived stimulus value, as assessed by WTP, would correlate positively with both self-reported alcohol craving and AUDIT scores. Finally, no study to date has explored the relationship between alcohol-cue induced BOLD responses and the perceived stimulus value. Therefore a regression analysis was included to examine this association. It was hypothesized that there would be increased activation in some similar, but not the exact same regions, as for self-reported craving. In addition we hypothesized, based on an earlier study in which normal-weight subjects provided WTP for different types of junk food, that WTP would correlate positively with increased activity in the OFC (Plassman et al., 2007). In this previous study evidence was found that the OFC encoded the subjects WTP for the food-items, which supports the hypothesis that the OFC encodes the value of goals in decision-making (Plassman et al., 2007). Even though this study investigated different (food instead of alcohol) cues, a correlation between perceived stimulus value and activation within the OFC should be found, if the OFC encodes the value of goals
independent of what those goals are. 2.Methods
2.1 Participants
Heavy drinking students were recruited via an advertisement on the University of Amsterdam online participant recruitment website, and with flyers in the faculty buildings. The flyers and
advertisement stated that participants were needed for an fMRI experiment in which they would receive computer-based cognitive training that could help them reduce their drinking behaviour. Inclusion criteria were: (a) age between 18 and 21 years; (b) drinking at least 10 alcoholic beverages a week as measured with the Alcohol Use Questionnaire (AUQ; Mehrabian et al, 1978), a score between 8 and 25 on the Alcohol Use Disorder Identification Test (AUDIT; Saunders et al., 1993), indicating a high likelihood of hazardous drinking; and (c) normal or corrected-to-normal visual acuity. Exclusion criteria consisted of: (a) A history of mental illness or neurological problems; (b) a history of drug or alcohol abuse; (c) smoking more than 10 cigarettes per day;(d) regular weekly use of cannabis; (e) being on any medication for any psychological or physical condition (including paracetamol and antibiotics, but excluding the contraceptive pill); and (f) MRI contraindications (e.g. pregnancy, metal implants, tattoos, cardiac pacemaker).
Twelve Dutch-speaking students were included in this pilot (5 female; mean age in years = 20.08, SD = 1.17; mean age in months = 248.42, SD = 12.85). On average participants consumed 20.78 standard Dutch units of alcohol per week (SD = 10.69, range = 8.50 - 41). Participants had an
average AUDIT score of 13.50 (SD = 6.68, range = 6 - 25). All participants fulfilled the inclusion criteria during screening. All participants provided written informed consent. The Ethics Committee of the University of Amsterdam approved the study.
2.2 Design & procedure
This pilot study was part of a larger investigation in which participants received cognitive training to reduce their drinking behaviour. The larger study consists of 6 sessions: a baseline assessment, three cognitive training sessions, a post-training assessment, and a final follow-up questionnaire session. For this pilot study, we only analysed the functional imaging data, as well as data from certain behavioural measures, collected during the baseline assessment session.
After giving informed consent participants underwent a dummy scan to familiarize themselves with the scanning environment. They then filled out a pen and paper version of the Nuffield Hospitals Medical History Questionnaire. Next they were placed in the MRI scanner where they completed the alcohol cue-reactivity task. Immediately after scanning participants rated the pictures they had seen during scanning based on: (a) How much money they would be willing to pay for each picture cue (“Willingness-to-pay” ratings); and (b) The degree of craving elicited by each picture cue. This was followed by the completion of an alcohol Implicit Association Test, the provision of demographic information, and the completion of the AUDIT (Saunders et al., 1993), amongst other assessments that will not be discussed here.
2.3 Materials and measures
All tasks, materials and questionnaires were presented in Dutch.
2.3.1 Severity of drinking
The AUDIT (Saunders et al., 1993) was included as a measure of the severity of alcohol use.
2.3.2 Cue reactivity task in the MRI scanner
During scanning participants completed a cue reactivity task. It consisted of three image conditions that included: (a) 18 alcohol-related pictures (images of bottles of alcohol-containing drinks displayed next to glasses full of the drink. Each picture was presented twice, creating 36 alcohol-related trials); (b) 18 alcohol-unalcohol-related pictures (images of bottles of non-alcohol-containing drinks, displayed next to glasses full of the drink. Each picture was presented twice, creating 36 control trials); and (c) 15 images of animals.
Each trial began with the presentation of a fixation cross for a jittered duration (range 1-7 seconds). This was followed by the presentation of one of the images for 3 seconds. The sequence of the fixation cross jitter-time for each trial was optimized using Optseq
additionally used to optimise which image-condition would be presented on each trial. While the order of image-condition presentation was identical for each participant, the image presented was chosen randomly for each participant.
During the task, participants were instructed to pay careful attention to each picture and to press the response button with the index finger of their right hand whenever they saw a picture of an animal. These images were included to make sure that participants paid attention to the stimuli.
2.3.3 Value of the alcohol stimuli and self-reported craving
Immediately after scanning, participants were presented with the 18 alcohol-related and the 18 control images, that they had seen during the cue-reactivity task, in a random order and were asked to rate: (a) “How much money they would be willing to pay for a glass of the drink depicted in the picture” [“Willingness-to-pay” (WTP) ratings]; and (b) “How much they would like to drink alcohol?” [“Degree of craving alcohol” (DC)]. Participants responded by pressing a key from one to ten, where one represented one euro or not wanting alcohol, and ten represented ten euro’s or really wanting alcohol a lot. The rating procedure was self-paced, and the next image was presented only after the participants had placed their rating.
For the WTP and DC ratings, an average score was calculated for the responses given in the presence of the alcohol-related images (WTP-alcohol, DC-alcohol), and for the responses given in the presence of the non-alcohol-related pictures (WTP-control, DC-control). Two difference scores were computed by subtracting the average rating score given to the control pictures from the average scores given to the alcohol-associated pictures (WTP-diff, DC-diff).
2.3.4 Valence of the alcohol stimuli
Implicit positive/negative associations with alcohol were assessed using a valence alcohol Implicit Associations Test (IAT; based on Greenwald, Nosek & Banaji, 2003; Houben et al., 2010, 2013). In the IAT the same images were used as presented during the cue-reactivity task.
The IAT comprised seven blocks with a total of 216 trials. The first three blocks were practise blocks. In the first block the target categories alcohol and soda, were shown in the upper left and upper right corner of the screen respectively. Images of alcohol- and non-alcohol-containing drinks were shown in the middle of the screen and participants had to allocate them to the correct category (‘e’ key for left category, ‘i’ key for right category). In the second block the two categories were now the attribute categories “pleasant” and “unpleasant” respectively. In the middle of the screen, words appeared that were either pleasant (e.g. ‘kiss’) or unpleasant (e.g. ‘harm’). Again participants had to allocate them to the correct category with the ‘e’ and ‘i’ key. In the third and fourth block a target category was paired with an attribute category (e.g. alcohol and pleasant in the left corner, soda and unpleasant in the right corner), and participants allocated the images and the words. In the fifth block the participants were shown only the target categories again to which they had to allocate alcohol and
soda images. However whether alcohol was in the left and soda in the right corner or vice versa was now opposite as what it was in the earlier blocks. The sixth and seventh blocks were identical to block three and four only the pairing of target and attribute category had now changed (e.g. if first alcohol and pleasant were paired together, now alcohol was paired with unpleasant). The position where the target category first appeared, and the order of target and attribute pairing, was counterbalanced across participants.
The IAT d-score was calculated, from the 144 non-practise trials, with a slightly adapted version of the algorithm postulated by Greenwald et al. (2003). A higher d-score means more positive implicit associations towards alcohol related stimuli.
2.5 Functional magnetic resonance imaging procedures.
BOLD data were obtained using a 3 Tesla scanner (Philips Achieva TX MRI-scanner; 32-channel head coil).
For the functional scans a single shot fast field echo planar (EPI; MS-FFE) sequence was obtained covering the entire brain (sequential acquisition; flip angle = 76.1°, repetition time = 2s, echo time = 27.63ms, 37 slices, slice thickness 2mm, slice gap 0.3mm, voxel dimensions 3x3x3 mm, field of view 240 x 240 mm, matrix size 80x80 pixels). The functional data were acquired in one
continuous session (252 volumes per subject; the initial 4 volumes were discarded to ensure steady state B0 magnetization).
In addition, for the structural scans, a fast Magnetization Prepared Rapid Acquisition Gradient Echo (MPRAGE) sequence was obtained (flip angle 8°, repetition time = 8.2s, echo time = 3.8 ms, 220 slices, voxel dimensions 1x1x1mm, field of view 240 x 188, matrix size 240).
2.6 Data analysis
2.6.1 Behavioural data analysis
The behavioural data were analyzed using IBM SPSS statistics 22.0. Paired samples t-tests were performed to check whether participants craved alcohol more when watching alcoholic drinks than when watching non-alcoholic drinks (DC-alcohol vs. DC-control), and to check whether they were willing to pay more for alcoholic drinks than for non-alcoholic drinks (alcohol vs. WTP-control).
In the rest of the analyses the DC-diff and WTP-diff scores were used as measures of the degree of alcohol craving and the attributed value to alcohol cues when controlling for the degree of alcohol craving and the attributed value associated with the control cues. Correlations were computed to examine the relationship between: (a) the severity of alcohol use and the WTP-diff, DC-diff, and IAT scores (bonferroni corrected, p-values below .05/3 = .017 are regarded statistically significant); (b) the WTP-diff and DC-diff scores. The assumption of normality was violated for WTP-diff. For this measure Spearman’s correlations were performed instead of Pearson’s correlations.
2.6.2 fMRI data analysis
Pre-processing and the statistical analyses of the fMRI data were done using SPM8 (Wellcome Trust Centre for Neuroimaging, London, UK). The scans were spatially realigned to correct for head movement during the session and after that corrected for slice timing. Using co-registration T2 and T1 images were aligned. Next the scans were normalized to the Montreal Neurological Institute’s (MNI ; Quebec, Canada) brain template. Finally the scans were smoothed using an isotropic Gaussian kernel of 8 mm at FWHM. One participant had to be excluded due to excessive head movement; therefore the fMRI analyses were performed with eleven participants.
We modelled all conditions based on the onset of each stimulus type. The six movement regressors from realignment were included as regressors of no interest. For every participant a first level analysis was performed in which a contrast of alcohol-related versus control cues was computed.
These contrast images were used in a second level one-sample t-test to investigate significant activations associated with the passive observation of alcohol-related vs. control cues. This t-test used a whole brain significance threshold of .005 and a voxel threshold of 10. This is quite a strict
significance level for a whole brain analysis on only 11 people [for comparison, in a recent
neuroimaging study with a sample of 326 participants a significance level of p < .05 was used (Claus, Ewing, Filbey, Sabbineni & Hutchison, 2011)]. The resulting thresholded statistical parametric map (SPM) was saved and was used as an inclusive mask for all subsequent regression analyses, since we wanted to investigate correlations between regions that were more active when watching alcohol-related images than when watching alcohol-unalcohol-related images and the behavioural measures.
Three regression models were computed to examine correlations between regions that were more active for watching alcohol-related than for watching alcohol-unrelated stimuli and WTP-diff, DC-diff and IAT-Dscores respectively. For each regression model both positive and negative correlations were examined.
Anatomical localization of significant activations was assessed by superimposition of the SPM maps on the single-subject T1-weighted MNI standard brain supplied by SPM8 and MRIcron
(http://www.mccauslandcenter.sc.edu/mricro/index.html). Anatomical localization of subcortical regions was assessed using Duvernoy’s anatomical atlas (Duvernoy, 1999).
3. Results
3.1 Behavioural results
In Table 1 the means and standard deviations for all behavioural measures are depicted. Participants craved alcohol more when watching images of alcoholic drinks than when watching images of non-alcoholic drinks [t(11) = 3.81, p = .003]. For the WTP-alcohol the assumption of normality was violated, however since the paired samples t-test is fairly robust to violations of this assumption it was decided to do the paired samples t-test. Participants were willing to pay more for
alcoholic drinks than for non-alcoholic drinks, however this difference was not statistically significant [t(11) = 1.11, p = .292].
There was a high positive correlation between WTP-diff and DC-diff, [r (10) = .726, p = .007], suggesting that the more participants were willing to pay for the depicted alcoholic drink compared to non-alcoholic drinks, the more they craved alcohol when watching alcohol-related images compared to when watching alcohol-unrelated images. However, the AUDIT did not correlate with any of the behavioural measures [r < .284, non-significant].
Table 1
Means and standard deviations (SD) of the behavioural measures
Measure Mean SD AUDIT 13.50 6.68 WTP-alcohol 2.25 .52 WTP-control 1.98 .82 WTP-diff .27 .83 DC-alcohol 3.66 1.84 DC-control 2.01 1.11 DC-diff 1.66 1.51 IAT-Dscore .04 .43 N = 12 3.2 fMRI results
Increased activations in response to alcohol-associated relative to control cues were found in voxels within the inferior occipital cortex, the superior frontal cortex, the inferior temporal cortex, the inferior orbital cortex/insula and the hippocampus (see Table 2). Activation within the superior frontal cortex correlated negatively with DC-diff and positively with WTP-diff (see Figure 1). Activation within the inferior temporal cortex correlated positively with WTP-diff (see Figure 2) and activation within the inferior orbital cortex/insula also correlated positively with WTP-diff (see Figure 3). Neither positive nor negative correlations between the IAT-Dscores and the regions with increased activations in response to alcohol-associated relative to control cues were found.
Table 2
Activations resulting from the alcohol vs. control cue one sample t-Test
Region Lateralization Number of voxels T-value MNI coordinates (x y z) Inferior occipital cortex L 296 6.73 -42 -70 -8 5.57 -39 -88 -5 4.56 -30 -91 -11 Superior frontal cortexa L 46 5.65 -21 50 34 4.43 -12 59 31 Inferior temporal cortexb R 24 4.9 36 -58 -5 3.7 45 -64 -8 Inferior orbital cortex / Insulac L 26 4.46 -33 35 -2 Hippocampus L 15 4.03 -36 -7 -26 3.93 -27 -10 -23 P<.005, voxel threshold = 10
a Correlated negatively with DC-diff and positively with WTP-diff b Correlated positively with WTP-diff
c Correlated positively with WTP-diff
Figure 1
The correlations (resulting from the regression analyses) between the superior frontal cortex and WTP-diff and DC-diff respectively.
Figure 2
The correlation (resulting from the regression analysis) between the Inferior temporal cortex and WTP-diff.
Figure 3
The correlation (resulting from the regression analysis) between the inferior orbital cortex/ insula and WTP-diff.
4. Discussion
4.1 Summary of predictions and results
This study aimed to investigate the relationship between self-reported alcohol craving and perceived stimulus value, and between the severity of drinking and self-reported craving, perceived stimulus value and stimulus valence. Participants craved alcohol more when watching images of alcoholic drinks, than when watching images of non-alcoholic drinks, and participants were willing to pay more for alcoholic than for non-alcoholic drinks (although this difference was not significant).
This indicated that the difference scores used in the rest of the analyses indeed reflected craving of alcohol in the presence of alcohol related stimuli, and perceived value of alcohol-related stimuli when controlling for the perceived value of alcohol-unrelated stimuli. It was predicted that there would be a positive correlation between self-reported craving and perceived stimulus value, and positive
correlations between severity of drinking and self-reported craving, perceived stimulus value and stimulus valence.
In line with these predictions a high positive correlation between self-reported craving and perceived stimulus value was found. Contrary to these predictions no correlation between severity of drinking and self-reported craving, perceived stimulus value or stimulus valence was found.
In addition this study aimed to investigate the BOLD responses to alcohol-related cues versus alcohol-unrelated cues. We predicted that viewing alcohol-related cues relative to alcohol-unrelated cues would be related with increased BOLD responses in regions within the brain’s reward system (i.e. regions within the striatum; amygdala; OFC; and vmPFC), as well as in the hippocampus, visual, and motor cortices. Increased activations in response to alcohol-related relative to control cues were found in: the occipital cortex, which is part of the visual cortex, and has been found in earlier studies (Schacht et al., 2012); the PFC, which is associated with attention and memory processes, encoding the motivational value of stimuli, and executive control, which has also been found in earlier studies (Grüsser et al., 2004; Heinz et al., 2004; Myrick et al., 2004; Tapert et al., 2004; Schacht et al., 2012; Volkow et al., 2010); the OFC, which is involved in the evaluation of the value of goals in decision-making, and has also been found in earlier studies (Wrase et al., 2002; Myrick et al., 2004; Volkow et al., 2010); and inferior temporal regions, including the hippocampus, reflecting conditioned learning and the processing of learned reward values of conditioned cues and context, which have also been found in earlier studies (Jasinska et al., 2014). In this regard our findings were in line with our
predictions and consistent with previous research, however there were also some areas like the ventral striatum that we did not find, contrary to our expectations based on results from earlier studies (Braus et al., 2001; Kühn & Gallinat, 2011; Wrase et al., 2002, 2007; Volkow et al., 2010).
Furthermore this study aimed to investigate the relationship between the BOLD responses to alcohol-related cues versus alcohol-unrelated cues and reported craving. We predicted that self-reported craving would correlate positively with increased activations within the VS, DS, OFC, motor, temporal, visual and prefrontal regions. Contrary to our predictions we found no positive correlations with self-reported craving. We did find a negative correlation between activation in the left superior frontal cortex and self-reported craving.
In addition to the relationship with self-reported craving this study also investigated the relationship between the BOLD responses to alcohol-related cues versus alcohol-unrelated cues and stimulus valence. We predicted that stimulus valence would correlate positively with BOLD responses in regions within the brain’s reward system (i.e. regions within the striatum; amygdala; OFC; and vmPFC), and with the right lateral sulcus, insula (bilaterally) right dlPFC and right amygdala.
Contrary to our predictions neither positive nor negative correlations between BOLD responses to alcohol-related cues versus alcohol-unrelated cues and stimulus valence were found.
Finally this study aimed to investigate the relationship between the BOLD responses to alcohol-related cues versus alcohol-unalcohol-related cues and perceived stimulus value. We predicted positive correlations between the BOLD responses to alcohol-related cues versus alcohol-unrelated cues and some similar regions as that have been associated with self-reported craving (i.e. VS, OFC, motor, temporal, visual and prefrontal regions). In addition to this we predicted positive correlations between perceived stimulus value and activations within the OFC. In line with our predictions the perceived value of the alcohol stimuli correlated positively with activations within the orbital cortex. In addition the perceived value of the alcohol stimuli correlated positively with activations in temporal regions and activations in parts of the prefrontal cortex. Temporal and prefrontal regions have, in earlier studies, been correlated with self-reported craving (Park et al., 2007; Fryer, 2013). Therefore these findings were in line with our predictions.
4.2. Discussion behavioural findings
Participants were willing to pay more for alcoholic than for non-alcoholic drinks (although this difference was not significant). This could either have been because participants based their answers on the actual cost of a drink or because they based their answers on the perceived value of the drink. In the first case it might not be so much a measure of the value they put on the stimulus but more of the knowledge of the price of a drink. However, this was probably not the case since the difference score of the WTP was correlated with the difference score of self-reported craving, suggesting it has more to do with the perceived value of the drink than the actual cost of the drink. However, this is still an open question, and for the overarching research project we will additionally include a measure of the actual price of the drinks in the correlational and regression analyses.
Contrary to our expectations the AUDIT did not correlate with any of the behavioural measures. This could have been due to the small sample size of this pilot study. The fact that we did not find a correlation with the IAT might also be because of the bipolar (alcoholic relative to non-alcoholic drinks) nature of the IAT we used. To see if implicit positive associations with alcohol correlate with severity of drinking a unipolar IAT (i.e an IAT with one target category), as used by Houben and Wiers (2008), might be needed. Furthermore, previous research has shown that a picture IAT, compared to a word IAT, shows faster reaction times and smaller IAT effects (Foroni & Bel-Bahar (2010). In this study it was suggested that different types of stimuli have a different level of
representation and therefore induce processing differences. Pictures may engage visual and perceptual systems, while (more abstract) words engage mostly semantic and long-term memory systems. Since stronger effects have been found for a word IAT it is possible that, if there are actually (small) effects, these could be detected with a word IAT.
In line with our expectations there was a high positive correlation between perceived value of the alcohol stimuli and self-reported alcohol craving, the more participants were willing to pay for an alcoholic drink the more they craved alcohol. This confirms that WTP is indeed related to self-reported craving and that individuals who crave alcohol more are more inclined to act upon this craving, and do what is needed to acquire the craved drink. In researching possibilities to help people reduce their drinking, or to determine people who are at risk of drinking too much/relapse, WTP is possibly an interesting measure since it might help to determine a persons readiness to actually buy (and therefore consume) a drink.
4.2. Discussion fMRI findings
Our findings regarding increased activations in response to related versus alcohol-unrelated cues were in line with our predictions. However, based on earlier studies we had additionally predicted to find correlations with activations within ventral [including the nucleus accumbens
(NAc)], and dorsal regions of the striatum, which have been respectively associated with connecting motivational aspects of salient stimuli with motor reactions and with reward (Braus et al., 2001; Kühn & Gallinat, 2011; Wrase et al., 2002, 2007; Volkow, Wang, Fowler, Tomasi & Telang, 2010), and behavioural habit formation (Modell & Mountz, 1995; Grüsser et al., 2004; Volkow et al., 2010); and the anterior cingulate cortex (ACC), which is associated with attention and memory processes, encoding the motivational value of stimuli, and executive control (Grüsser et al., 2004; Heinz et al., 2004; Myrick et al., 2004; Tapert et al., 2004; Schacht et al., 2012; Volkow et al., 2010). That we did not find all regions that have been found earlier could have to do with the small number of participants in our study. In addition our sample consisted of heavy drinking students who were not (yet) alcohol dependent but were drinking hazardously. In many of the earlier studies alcohol dependent individuals and healthy (irregular drinkers) controls participated (e.g. Fryer, 2013; Park, 2007). Therefore it is possible that our sample has some actual different neural responses towards alcohol-related versus alcohol-unrelated cues, compared to samples from earlier studies. They could be in the state where they have had many positive experiences with alcohol, making them like the alcohol, but where they have not yet developed a disproportionate sense of craving for alcohol (Robinson and Berridge, 1993). Therefore, especially areas associated with craving [including VS and DS (e.g. Kühn and Gallinat, 2011) and ACC (Myrick et al., 2004)] that have been found in alcohol dependent subjects might not show (as much) increased activations in subjects who are not yet alcohol dependent. The dorsal striatum for example is associated with behavioural habit formation (Modell & Mountz, 1995; Grüsser et al., 2004; Volkow et al., 2010), which might not be as much developed in heavy social drinkers as in alcohol dependent individuals. Although participants did show significantly more craving for alcohol when watching alcohol-related images versus alcohol-unrelated images, this still might be less craving for alcohol than alcohol dependent subjects would have shown.
We did not manage to replicate any of the positive correlations found with self-reported craving in earlier studies. This was a bit surprising since correlations with self-reported craving have been found in many studies. A possible explanation however might be, as stated above, that the subjects in our sample craved alcohol less than alcohol dependent subjects that are often used in studies. In addition we used an inclusive mask of the increased bold activations for alcohol-related versus alcohol-unrelated cues in the regressions, which could partly explain the lack of correlations found with self-reported craving. Since areas in which no increased activation of alcohol-related versus alcohol-unrelated cues were found, were excluded from the regressions. Finally the sample size might again have been too small to find effects that in reality are there. The only correlation that we found with self-reported craving was a negative correlation with the left superior frontal cortex. The left superior frontal cortex has been associated with self-awareness and self-judgement (Goldberg, Harel & Malach, 2006). One way to explain our finding might be that subjects who were more self-aware and more occupied with self-judgement were more inclined to make themselves look better and report lower craving scores. However this evidence is to scarce to draw firm conclusions from it and this relationship should be further investigated in future studies with a larger sample size.
In line with our predictions based on the study of Plassman et al. (2007) the perceived value of the alcohol stimuli correlated positively with increased activation within the OFC. This provides further support for the hypothesis that the OFC plays a role in the encoding of the value of goals in decision-making. Also the correlation between perceived stimulus value and activation within prefrontal and temporal regions was in line with our expectations, based on the relationship between perceived stimulus value and self-reported craving. Temporal regions reflect conditioned learning and the processing of learned reward values of conditioned cues and context (Jasinska et al., 2014), which suggests that the perceived value of alcohol-related stimuli has to do with the learned/conditioned reward an individual puts on this cue. It is however surprising that we did not also find these
correlations between activation within temporal and prefrontal regions and self-reported craving since this has been found in previous studies (Park et al., 2007; Fryer, 2013). It is possible (as stated above) that our subjects did not crave alcohol as much as participants in previous studies.
Contrary to our expectations neither positive nor negative associations were found between the valence of the alcohol-related stimuli and any of the brain regions with increased activation in response to alcohol-associated relative to control cues. The small number of participants in this study could be a reason for this. Secondly, in the study of Viinikainen et al. (2010) negative and positive emotional pictures were used that were not drug-related. Valence processes for general
positive/negative pictures might be different from valence of a stimulus related to a person’s drug of abuse. Another possibility is that the valence alcohol-IAT used in this study was not appropriate. Research into the IAT suggest that in some cases it may be better to personalize (I like / I dislike) the IAT or use a unipolar IAT to be able to look specifically at positive or negative associations with alcohol-related stimuli (Houben & Wiers, 2007; Houben & Wiers, 2008). Interesting results may also
be found when using an arousal/relaxation IAT or an approach/avoidance IAT since heavy but not light drinkers have implicit arousal associations with alcohol-related stimuli (Wiers et al., 2002), and hazardous drinkers show implicit approach associations between approach and alcohol-related stimuli (Palfai & Ostafin, 2003). Furthermore it has been stated that picture IAT’s will give smaller effects than word IAT’s (Foroni & Bel-Bahar, 2010). Thus it is possible that if there actually was an effect it might have been detected with a word IAT. However the advantage of using the picture IAT was that we could use the same images for the craving, perceived stimulus value and valence measures, controlling for differences due to differences in stimuli.
4.3. Limitations and future research
The most serious limitation of this study is that we had a sample of only twelve participants (and eleven in the fMRI analyses). With a sample this small, power is very low and it is possible that there were effects that we could not identify due to the small sample size. This however was a pilot study and in the overarching project we will test about 80 participants, so we will be able to check our preliminary results with more power.
Secondly our sample is not completely comparable to many of the studies done previously in this field. Most previous studies have used alcohol dependent subject and healthy controls that drink irregularly (e.g. Fryer, 2013; Park, 2007). This makes it a bit harder to interpret our findings in light of previous research. However this is a group of drinkers who are at risk of developing alcohol
dependency and already display hazardous drinking behaviour, but are not yet alcohol dependent. Therefore it is an important group to study in order to understand the development of alcohol use disorders and find suitable treatments to prevent these drinking patterns from escalating into alcohol dependency.
Furthermore the IAT that was used might not have been the best choice of IAT. To find an effect of stimulus valence a word alcohol valence IAT (personalized and/or unipolar) might be warranted in future research. In addition it might be interesting to further explore the relationship between an approach/avoidance and arousal/relaxation IAT and increased BOLD responses in response to alcohol-related stimuli.
The perceived value of alcohol-related stimuli seems to be an interesting measure, which is positively correlated with self-reported craving, activation within the superior frontal cortex, and activation within the orbital and temporal cortices. The perceived value of a stimulus is related with self-reported craving and the conditioned reward that an individual assigns to a cue. How much value an individual attributes to a stimulus (and the actual alcohol that it depicts), might also show how ready an individual is to actually act on his or her alcohol craving. It is therefore plausible that the perceived value of a stimulus plays a role in actual drinking behaviour, although we did not find a correlation between the severity of drinking and perceived value of the stimuli. Since this was the first study to look into the relationship between perceived value of alcohol-related stimuli and self-reported
craving, severity of drinking, and increased activations in response to alcohol-associated relative to control cues, future studies should (with a larger sample size) further examine these relationships. Acknowledgements
I thank dr. K. Nikolaou for her supervision, editing and comments on earlier versions of this report, drs. W. J. Boendermaker for programming the IAT, and T. de Jong for his technical assistance in the use of a program to administer all tasks and measurements. Costs of scanning were covered by a NWO brain & cognition PhD grant to drs. W. J. Boendermaker.
References
Aron, A. R., Robbins, T. W., & Poldrack, R. A. (2004). Inhibition and the right inferior frontal cortex.
Trends in Cognitive Sciences, 8, 170-177.
Bouton, M. E., Westbrook, R. F., Cocoran, K. A., & Maren, S. (2006). Contextual and temporal modulation of extinction: behavioral and biological mechanisms. Biological
Psychiatry, 60, 352-360.
Braus, D. F., Wrase, J., Grüsser, S., Hermann, D., Ruf, M., Flor, H., Mann, K., & Heinz, A. (2001). Alcohol-associated stimuli activate the ventral striatum in abstinent alcoholics.
Journal of Neural Transmission, 108, 887-894.
Carter, B. L., & Tiffany, S. T. (1999). Meta-analysis of cue-reactivity in addiction research. Addiction,
94, 327-340.
Claus, E. D., Ewing, S. W., Filbey, F. M., Sabbineni, A., & Hutchison, K. E. (2011). Identifying neurobiological phenotypes associated with alcohol use disorder severity.
Neuropsychopharmacology, 36, 2086-2096.
Drummond, D. C., Cooper, T., & Glautier, S. P. (1990). Conditional learning in alcohol dependence: implications for cue exposure treatment. British Journal of Addiction, 85, 725-743.
Duvernoy, H. M. (1999). The Human Brain: Surface, Blood Supply and Three-Dimensional Anatomy. 2nd edn completely revised ednSpringer: Wien, Germany; New York, NY, USA.
Filbey, F. M., Claus, E., Audette, A. R., Niculescu, M., Banich, M. T., Tanabe, J., Du, Y. P., & Hutchison, K. E. (2008). Exposure to the taste of alcohol elicits activation of the
mesocorticolimbic neurocircuitry. Neuropsychopharmacology, 33, 1391-1401.
Foroni, F., & Bel-Bahar, T. (2010). Picture-IAT versus Word-IAT: level of stimulus representatation influences on the IAT. European Journal of Social Psychology, 40, 321-337.
Fryer, S. L., Jorgensen, K. W., Yetter, E. J., Daurignac, E. C., Watson, T. D., Shanbhag, H., Krystal, J. H., & Mathalon, D. H. (2013). Differential brain response to alcohol cue distractors across stages of alcohol dependence. Biological Psychology, 92, 282-291.
during sesorimotor processing. Neuron, 50, 329-339.
Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the
implicit association test: I. An improved scoring algorithm. Journal of personality and
social psychology, 85, 197.
Grüsser, S. M., Wrase, J., Klein, S., Hermann, D., Smolka, M. N., Ruf, M., Weber-Fahr, W., Flor, H., Mann, K., Braus, D. F., & Heinz, A. (2004). Cue-induced activation of the striatum and medial prefrontal cortex is associated with subsequent relapse in abstinent alcoholics. Psychopharmacology 175, 296–302.
Heinz, A., Beck, A., Grüsser, S. M., Grace, & A. A., Wrase, J. (2009). Identifying the neural circuitry of alcohol craving and relapse vulnerability. Addiction Biology, 14, 108-118.
Heinz, A., Siessmeier, T., Wrase, J., Hermann, D., Klein, S., Grüsser-Sinopoli, S. M., … Bartenstein, P. (2004). Correlation between dopamine D2 receptors in the ventral striatum and central processing of alcohol cues and craving. American Journal of
Psychiatry, 161, 1783-1789.
Houben, K., Havermans, C., Nederkoorn, C., & Jansen, A. (2013). Beer à no- go: learning to stop responding to alcohol cues reduces alcohol intake via reduced affective associations rather than increased response inhibition. Addiction, 107, 1280-1287.
Houben, K., Schoenmakers, T. M., & Wiers, R. W. (2010). I didn’t feel like drinking but I don’t know why: The effects of evaluative conditioning on alcohol-relates attitudes, craving and behavior. Addictive Behaviors, 35, 1161-1163.
Houben, K., & Wiers, R. W. (2007). Personalizing the alcohol-IAT with individualized stimuli: relationship with drinking behavior and drinking-related problems. Addictive Behaviors, 32, 2852-2864.
Houben, K., & Wiers, R. W. (2008). Implicitly positive about alcohol? Implicit positive associations predict drinking behavior. Addictive Behaviors, 33, 979-986.
Jasinska, A. J., Stein, E. A., Kaiser, J., Naumer, M. J., & Yalachkov, Y. (2014). Factors modulating neural reactivity to drug cues in addiction: a survey of human neuroimaging studies.
Neuroscience and Biobehavioral Reviews, 38, 1-16.
Kühn, S., & Gallinat, J. (2011). Common biology of craving across legal and illegal drugs – a
quantitative meta-analysis of cue-reactivity brain response. European Journal of Neuroscience,
33, 1318-1326.
Modell, J. G., & Mountz, J. M. (1995). Focal cerebral blood flow change during craving for alcohol measured by SPECT. Journal of Neuropsychiatry and Clinical Neuroscience
7, 15–22.
Psychological Reports, 43, 803-806.
Monti, P. M., Binkoff, J. A., Abrams, D. B., Zwick, W. R., Nirenberg, T. D., & Liepman, M. R. (1987). Reactivity of alcoholics and nonalcoholics to drinking cues. Journal of
Abnormal Psychology, 96, 122-126.
Myrick, H., Anton, R. F., Li, X. B., Henderson, S., Drobes, D., Voronin, K., & George, M. S. (2004). Differential brain activity in alcoholics and social drinkers to alcohol cues: relationship to craving. Neuropsychopharmacology 29, 393–402.
Oberlin, B. G., Dzemidzic, M., Tran, S. M., Soeurt, C. M., Albrecht, D. S., Yoder, K. K., & Kareken, D. A. (2013). Beer flavor provokes stratial dopamine release in male drinkers: mediation by family history of alcoholism. Neuropsychopharmacology, 38, 1617-1624.
Palfai, T. P., & Ostafin, B. D. (2003). Alcohol-related motivational tendencies in hazardous drinkers: assessing implicit response tendencies using the modified-IAT. Behaviour Research and
Therapy, 41, 1149-1162.
Papachristou, H., Nederkoorn, C., Havermans, R., van der Horst, M., & Jansen, A. (2012). Can’t stop the craving: the effect of impulsivity on cue-elicited craving for alcohol in heavy and light social drinkers. Psychopharmacology, 219,511-518.
Park, M. S., Sohn, J. H., Suk, J. A., Kim, S. H., Sohn, S., & Sparacio, R. (2007). Brain substrates of craving to alcohol cues in subjects with alcohol use disorder. Alcohol and Alcoholism, 42, 417-422.
Plassman, H., O’Doherty, J., & Rangel, A. (2007). Orbitofrontal cortex encodes willingness to pay in everyday economic transactions. Journal of Neuroscience, 27, 9984-9988.
Robinson, T. E., & Berridge, K. C. (1993). The neural basis of drug craving: an incentive-alience theory of addiction. Brain Research Reviews, 18, 247-291.
Robinson, T. E., & Berridge, K. C. (2001). Mechanisms of action of addictive stimuli: Incentive-sensitization and addiction. Addiction, 96, 103-114.
Robinson, T. E., & Berridge, K. C. (2008). The incentive sensitization theory of addiction: some current isues. Philosophical Transactions of the Royal Society of Biological Sciences, 363, 3137-3146.
Saunders, J. B., Aasland, O. G., Babor, T. F., De La Fuente, J. R., & Grant, M. (1993). Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption—II. Addiction, 88, 791-804.
Schacht, J. P., Anton, R. F., & Myrick, H. (2012). Functional neuroimaging studies of alcohol cue reactivity: a quantitative meta-analysis and systematic review. Addiction Biology,
18. 121-133.
Schneider, F., Habel, U., Wagner, M., Franke, P., Salloum, J. B., Shah, N. J., Toni, I., Sulzbach, C., Hönig, K., Maier, W., Gaebel, W., Zilles, K. (2001). Subcortical
correlates of craving in recently abstinent alcoholic patients. American Journal of
Psychiatry 158, 1075–1083.
Schonberg, T., Bakkour, A., Hover, A. M., Mumford, J. A., Nagar, L., Perez, J., & Poldrack, R. A. (2014). Changing value through cued approach: an automatic mechanism of behavior change.
Nature Neuroscience, 17, 625-630.
Schultz, W., Dayan, P., & Montague, P. R. (1997). A neural substrate of prediction and reward. Science, 275, 1593-1599.
Tapert, S. F., Brown, G. G., Baratta, M. V., & Brown, S. A. (2004). FMRI BOLD response to alcohol stimuli in alcohol dependent young women. Addictive Behavior 29, 33–50.
Tapert, S. F., Cheung, E. H., Brown, G. G., Frank, L. R., Paulus, M. P., Schweinsburg, A. D., Meloy, M. J., & Brown, S. A. (2003). Neural response to alcohol stimuli in adolescents with alcohol use disorder. Archives of General Psychiatry, 60, 727-735
Viinikainen, M., Jääskeläinen, L. P., Alexandrov, Y., Balk, M. H., Autti, T., & Sams, M. (2010). Nonlinear relationship between emotional valence and brain activity: evidence of separate negative and positive valence dimensions. Human Brain Mapping, 31, 1030-1040.
Volkow, N. D., Wang, G. J., Fowler, J. S., Tomasi, D., & Telang, F. (2010). Addiction: beyond dopamine reward circuitry. Proceedings of the National Acadamy of Sciences, 108, 15037- 15042.
Vollstädt-klein, S., Loeber, S., Richter, A., Kirsch, M., Bach, P., von der Goltz, C., Hermann, D., Mann, K., & Kiefer, F. (2011). Validating incentive salience with functional magnetic resonance imaging: association between mesolimbic cue reactivity and attentional bias in alcohol-dependent patients. Addiction Biology, 17, 807-816.
Wiers, R. W., van de Luitgaarden, J., van den Wildenberg, E., Smulders, F. T. Y. (2005). Challenging implicit and explicit alcohol-related cognitions in young heavy drinkers. Addiction, 100, 806- 819.
Wiers, R. W., van Woerden, N., Smulders, F. T. Y., & de Jong, P. J. (2002). Implicit and explicit alcohol-related cognitions in heavyand light drinkers. Journal of Abnormal Psychology, 111, 648-658.
Wrase, J., Grüsser, S. M., Klein, S., Diener, C., Hermann, D., Flor, H., Mann, K., Braus, D. F., & Heinz, A. (2002). Development of alcohol-associated cues and cue-induced brain activation in alcoholics. European Psychiatry 17, 287–291.
Wrase, J., Schlagenhauf, F., Kienast, T., Wustenberg, T., Bermpohl, F., Kahnt, T., Beck, A., Strohle, A., Juckel, G., Knutson, B., & Heinz, A. (2007). Dysfunction of reward processing correlates with alcohol craving in detoxified alcoholics. Neuroimage 35, 787–794.
