Appetitive action-tendencies and
transcranial direct current stimulation
(tDCS)
Research Report
Sören Mohr
Student number: 5841577 Supervisor: Thomas Gladwin Co-assessor: Reinout Wiers
1. Introduction
Dual-process models of addiction explain addictive behaviors in terms of an imbalance between two semi-independent cognitive systems [23]. An “impulsive” system is characterized by automatic appraisal of stimuli in terms of their emotional and motivational significance, whereas a slower “reflective” system governs controlled processes related to conscious deliberations [23]. Indirect measurement procedures such as the Implicit Association Test (IAT) [10] have become increasingly popular in
psychological research because of their potential to reveal changes in automatic processes, which are mainly outside conscious control and hardly elusive by means of direct self-assessment. This is important since repeated drug use might cause the
impulsive system to become sensitized to drugs and drug-cues [23]. This sensitization is reflected in an attentional bias for drug-related cues [17]. Drug cues are very salient for addicted individuals. They automatically grab attention and might ultimately produce compulsive drug-seeking behavior. To make things worse, this sensitization is
accompanied by an impaired reflective system to inhibit those automatic associations. The attentional bias may foster approach action tendencies towards the drug. Such approach tendencies for drug-related cues have been found with the approach-avoidance task (AAT). There are several variants of the AAT, which all measure the reaction times for approaching and avoiding certain categories of stimuli. One such variant is the Stimulus-Response Compatibility (SRC) task [7]. A stimulus appears in the center of the screen and a manikin appears either above or below the stimulus. Participants are
instructed to move the manikin either toward one category or away from another category by pressing a keyboard button. The difference score between the two response
assignments (approach vs. avoid) indicates the relative approach-avoidance bias for a category of stimuli. This bias has been postulated to reflect a combination of automatic and controlled processes [6]. Wiers, Rinck, Dictus and Van den Wildenberg (2009) developed an AAT in which participants respond to stimuli by pushing or pulling a joystick. Upon a pull movement the stimuli on the screen increases in size and upon a push movement it decreases. They found that heavy drinkers were faster to pull than to push pictures of alcoholic drinks. Intriguingly, this approach bias for alcohol-related cues was successfully modified [22]. In one condition, participants pushed most alcoholic and pulled most non-alcoholic drinks, while being instructed to respond to the picture format (landscape vs. portrait). In other words, they were implicitly trained to avoid alcohol and approach non-alcoholic drinks. Participants trained to avoid alcohol became faster in pushing alcohol pictures away. Further, this training effect generalized to untrained pictures and to a different test assessing approach-avoidance tendencies (i.e., IAT). These generalized effects across stimuli and measures were also demonstrated with a clinical sample [20]. Hence, changing the percentage of alcoholic and non-alcoholic drinks, which are to be pushed or pulled, appears a promising technique to reduce an approach bias for alcohol.
Another potential method for modifying the approach bias for alcohol-related stimuli might be the application of transcranial Direct Current Stimulation (tDCS). TDCS refers to the delivery of a constant low current to a brain region of interest. To this end, two NaCl-soaked electrodes are placed on the scalp. One electrode is generally referred to as the stimulating electrode which is placed above the region of interest, whereas the other one is the reference electrode. Research has shown that tDCS increases or decreases
the excitability of the cortex (i.e. the firing threshold), depending on the polarity of the current (i.e. anodal vs. cathodal current) [12]. It has been shown that the induced enhancement of cortical excitability can transiently modulate behavioral performance. For instance, risk-taking behavior as measured by the Balloon Analog Risk Task (BART) was decreased by tDCS over the dorsolateral prefrontal cortex (DLPFC) [8]. More
relevant to addiction, craving for alcohol has been reduced by tDCS over the DLPFC [3]. The lateral PFC has been implicated in cognitive control in general and in decision making processes underlying action selection in particular [13]. By applying transcranial magnetic stimulation (TMS), research has established a causal role of the left DLPFC in controlling associative biases [4]. It is plausible to assume that the left DLPFC is also involved in overcoming an approach bias for alcohol-related stimuli.
The goal of the present study was to apply tDCS as a means of modulating approach and avoidance behavior towards alcohol-related stimuli. We used a newly developed SRC task in which the manikin was replaced by a simulated hand. The bias for alcohol stimuli was hypothesized to be reduced with anodal stimulation of the left
DLPFC.
2. Experiment 1
2.1 Materials and Methods
Participants
Twenty-eight students (24 female) from the University of Amsterdam participated in the first experiment. The mean age was 22.6 years (SD=3.3). Participants gave informed
written consent before taking part in the study, which was approved by the Ethics Committee of the Psychology Faculty of the University of Amsterdam. Participants received one research credit or €7 for participation.
Preference task
A preference task was used to select individualized stimuli. Pairs of stimuli were taken from a set of 24 pictures of beverages (12 alcoholic, 12 soft drink). Participants had to indicate which of the two stimuli they most preferred. After that, they saw one novel stimulus, which was successively shown with the most to least preferred stimuli. The four most preferred alcohol pictures and the four most preferred soft drink pictures were selected for the SRC task.
SRC Tasks
In the first study, two versions of the SRC task were compared. The version that yielded the most pronounced approach bias was selected for the main study. All participants completed both tasks in one session. In one version, participants moved a manikin on the screen. In the other version this manikin was replaced by a simulated hand. Within each version, there were feature-relevant trials and feature-irrelevant trials. On feature-relevant trials, participants were instructed to approach alcohol and to avoid soft drinks, or vice versa. On feature-irrelevant trials, either the letter ‘X’ or the letter ‘O’ appeared in the middle of the screen and participants were instructed to approach ‘X’ and to avoid ‘O’, or vice versa. On each trial, a beverage appeared in the middle of the screen and a
interval between the appearance of the beverage and the manikin/hand, interstimulus interval (ISI), was randomized to be either 0ms, 300ms, 600ms or 900ms.
Two categories were used: alcoholic drinks and soft drinks. Each category consisted of 11 pictures. The pictures depicted bottles and cans that were positioned on a table and photographed from above. For the hand variant, two pictures of a hand were used, one depicting a male right hand, another depicting the corresponding left hand. The hand was in a grasping position. For the manikin variant, three pictures of a stick-figure were used. One picture depicted a manikin in a standing position, whereas the other two pictures depicted a manikin with its left or right leg lifted. Upon pressing a button, all three pictures appeared one after another, which created the impression of a moving manikin. The hand/manikin could be moved across the screen using four different keyboard buttons: D and F for the left side and J and K for the right side. With the outer buttons (D and K) participants could move the hand away from the beverage. The inner buttons (F and J) were used to move the hand towards the beverage. One key press was sufficient to move the hand/manikin across the screen. There were two different response assignments. At the beginning of a block, participants were instructed to approach
alcohol and to avoid soft drinks, or vice versa. On each trial, a picture of a beverage was presented in the middle of the screen and was immediately followed by the hand/manikin which randomly appeared either on the right or the left side of the beverage. Participants were instructed to respond as fast and as accurately as possible. The time between the onset of the hand and the first key press was measured as dependent variable.
AUDIT
Alcohol use and problems were assessed with the Alcohol Use Disorders Identification Test (AUDIT). The AUDIT has been shown to be a valid screening instrument for alcohol-related problems in the general population and in students [18].
Procedure
Upon arrival, participants signed informed consent and were seated in front of a
computer. The task was explained orally and participants performed two training blocks. Depending on conditions, participants first completed the manikin task and then the hand task or vice versa. Participants completed 16 experimental blocks per version. Each block consisted of 36 trials. This resulted in a total of 32 blocks and 72 trials. After completing the experimental task, participants filled in the AUDIT questionnaire and received one research credit or were paid.
Data Analysis Strategy
Analysis of reaction times requires correction of outliers. However, there is no well-established cut-off point. Therefore, we compared three upper-bound cut-off points (1000ms, 1500ms, 2500ms). All reaction times below 150 ms were discarded. For each participant, mean reaction times (RTs) were calculated for approaching and avoiding alcohol and soft drink pictures. This was done for the two version of the task (hand/manikin), the trialtype (feature-relevant trials/feature-irrelevant trials), and the four ISIs (0ms,300ms,600ms,900ms), which yielded 64 RTs per subject. The RTs over
The results from the ANOVA mentioned below are based on the 2500ms cut-off point which yielded the strongest effects. In a preliminary analysis, it was investigated whether the results were affected by the order in which the version of the task (hand vs. manikin) was presented. Since there was no significant main effect of this factor, it was not included in the main analysis. The four different combinations of version and trialtype were analyzed separately in an analysis of variance (ANOVA). All four combinations were subjected to a 2x2x2 repeated measure ANOVA with responsetype (approach vs. avoid) and drinktype (alcohol vs. softdrink) as within-subjects factors, and AUDIT score as between subjects factor.
2.2 Results
The mean AUDIT score was 5.7 (SD=3.4).
All effects significant at p<0.05 are reported. First, the feature-relevant hand version was analyzed. On reaction times, there was a significant effect of responsetype, F(1,16)= 20.4, p=0.0004, due to faster approach than avoid responses. An interaction was found between AUDIT scores, responsetype and drinktype, F(1,16)=4.9, p=0.042. This was due to AUDIT scores being associated with faster approach-alcohol responses relative to all three other trial types (p<0.05 for all correlations between AUDIT and relevant contrast scores). On accuracy, responsetype and drinktype interacted, F(1,16)=6, p=0.03, due to avoid-soft drink trials being significantly less accurate than both avoid-alcohol and approach-soft drink trials (all p<0.05).
Second, the feature-irrelevant hand version was analyzed. On RT, approach trials were faster than avoid trials, F(1,16)=12.5, p=0.003. On accuracy, responsetype and
drinktype interacted, F(1,16)=5.2, p=0.04: for approach trials only, alcohol trials were more accurate than soft drink trials.
Third, the feature-relevant manikin version was analyzed. Approach trials were faster than avoid trials, F(1,16)=5.5, p=0.03, and AUDIT interacted with responsetype and drinktype, F(1,16)=7.2, p=0.016, due to faster approach-alcohol responses than either avoid-alcohol or approach-soft drink (all p<0.05). No effects on accuracy were found. Last, the feature-irrelevant manikin version was analyzed. No effects were significant.
2.3 Discussion of Experiment 1
The primary result was that only the feature-relevant tasks showed an interaction with AUDIT scores. Of the two feature-relevant tasks, only the hand version showed an overall (that is, not in a three-way interaction with AUDIT scores) interaction of responsetype by drinktype, which involved what appeared to be a problem inhibiting approach responses in the alcohol-approach, soft-drink-avoid block. We therefore selected the relevant-feature hand task for Experiment 2.
3. Experiment 2
3.1 Materials and Methods
ParticipantsSeventeen students (13 female) from the University of Amsterdam participated in the second experiment. The mean age was 21.7 years (SD=2.4). Participants gave informed written consent before taking part in the study, which was approved by the Ethics
Committee of the Psychology Faculty of the University of Amsterdam. The experiment consisted of three sessions. Participants received one research credit or €7 per session.
SRC Tasks
As follow-up analyses showed no interactions with intervals between the beverage picture and hand, we used an ISI of 0ms, which means that the appearance of the beverage was immediately followed by the hand. The rest of the task was exactly the same as in experiment 1.
TDCS
Direct current was induced by two NaCl-soaked sponge electrodes and delivered by a battery-driven stimulator. The device is shown in Figure 1. The electrodes were attached to the scalp and stabilized with elastic bands. The stimulating electrode (anode) was placed above the left dorsolateral prefrontal cortex (LDLPFC), which corresponds to F3 of the 10-20 system. The reference electrode (cathode) was placed above the right orbit. This electrode arrangement has been shown to reduce craving in alcohol-dependent patients [3]. The same arrangement was used in both experimental conditions. However, in the active stimulation condition, a current of 1 mA was turned on for a period of 10 min, whereas in the sham stimulation condition, a current of 1 mA was turned on for only 60 sec.
Figure 1 Direct Current Stimulator with sponge electrodes
AUDIT
Alcohol use and problems were assessed with the Alcohol Use Disorders Identification Test (AUDIT), which was used in the first experiment.
Procedure
The experiment consisted of three sessions. In the first session, participants were given written information about the nature of the experiment and the application of tDCS. An experimenter was available for any questions and ensured that the participants had no contraindications to brain stimulation. Participants signed informed consent and an appointment was made for the next session. In the first experimental session, the
computer task was explained to the participant and two practice blocks were performed. The task was explained orally and participants performed two training blocks. Then the two electrodes were attached to the scalp. The current was turned on and it was ensured that the participant felt comfortable. Participants completed 16 experimental blocks and each block consisted of 16 trials. After completing the task, an appointment was made for the second experimental session, which proceeded in the same manner as the first one. Each participant took part in all three sessions, which were separated by an interval of at least 24 hr. One session lasted approximately one hour. By the end of the last session, participants filled in the AUDIT questionnaire. They were asked whether they could
distinguish between the active and sham stimulation and then received three credits or €21.
Data Analysis Strategy
We used an upper-bound cut-off point of 2500ms. All reaction times below 150 ms were discarded. For each participant, mean RTs were calculated for approaching and avoiding alcohol and soft drink pictures. This was done for both experimental conditions. The RTs over subjects were then subjected to an analysis of variance (ANOVA) for repeated measures. Mean RTs were subjected to a 2x2x2x2 analysis of variance (ANCOVA) with responsetype (approach vs. avoid), drinktype (alcohol vs. softdrink), and condition (stimulation vs. sham) as within-subjects factor, and order (stimulation first vs. sham first) as between-subjects factor.
3.2 Results
The mean AUDIT score was 7.4 (SD=4.1).
All effects are reported as significant at p<0.05. The analysis revealed a significant main effect of responsetype, F(1,12)=12.434, p=0.004. Participants were faster to approach the stimuli (582.467ms) than to avoid them (595.567ms). There was also a significant main effect of drinktype, F(1,12)=15.813, p=0.002. Participants were faster to respond to alcohol stimuli (579.094ms) than to soft drinks (598.94ms). There were no significant interactions in this analysis.
In another analysis the reduction in the bias was correlated to the AUDIT score. The correlation was not significant.
4. Discussion
The first experiment revealed that the feature-relevant task versions revealed an approach bias in RTs for alcohol stimuli that was dependent on risky drinking. The hand task further revealed an overall interaction between approach-avoid responses and alcohol versus soft drink stimuli, that appeared to reflect an “overgeneralized” approach response in the context of approaching alcohol. Participants presumably to some extent identify themselves with the manikin (as a kind of avatar) or the hand. The appearance of a grasping hand might further facilitate this identification and thereby lead to the additional effect on accuracy. In the feature-irrelevant version, attention is directed towards a
feature unrelated to the contents of the presented picture (in this case the letters X and O). Some studies imply that stimulus valence affects approach-avoidance reactions even without an intention to evaluate [11]. At first sight, the present study suggests otherwise. However, for the feature-irrelevant hand version the interaction of responsetype and drinktype was almost significant and there was a significant interaction of responsetype and AUDIT for the feature-irrelevant manikin version. This suggests that if only riskier drinkers were included, the feature-irrelevant versions would probably affect approach-avoidance reactions to a greater extent. Furthermore, studies using positive and negative words as stimuli cannot be directly compared to the present study in which pictures of drinks were used, since the difference in evaluations of alcohol and soft drinks is less obvious than that of positive and negative stimuli. However, the approach bias we found in the feature-relevant hand version is in line with the assumption that the outcome of intentional evaluations is stronger than the outcome of unintentional evaluations [7]. An
approach or avoidance bias which is independent of the intention to evaluate the pictures' content might only be detectable in individuals with strong evaluative associations (e.g., heavy drinkers).
The feature-relevant hand version was selected for the second experiment. In this experiment, the approach bias for alcohol was attempted to be reduced by applying anodal transcranial Direct Current Stimulation (tDCS) to the left dorsolateral prefrontal cortex (DLPFC). According to the results, the stimulation had no effect on approach-avoidance behavior. What might be the reason for the lack of an effect? One reason might be the low amplitude of 1 mA. Boggio et al. (2008), for instance, used 2 mA to modify levels of craving for alcohol using tDCS. Even then, the effects were rather small. Of course, it could be the case that stimulation reduces an approach bias only for participants who show a strong bias in the first place. Hence, the inclusion of additional riskier
drinkers could possibly yield more positive results. However, the absence of significant results concerning any alcohol-related bias may also mean that tDCS actually strongly influenced the bias, but that this was not detectable in statistical analyses due to crossover effects. That is, perhaps subjects who performed the task with tDCS in their first session had their bias reduced, and this effect persisted in their second session. This would make effects of tDCS difficult to detect.
Overall, relatively little is known about the mechanism of tDCS and its effects on cognitive performance. For instance, there are mixed results concerning the effects of unilateral versus bilateral stimulation. When tDCS is used bilateraly, excitatory stimulation (anodal current) is applied to one side (e.g. left DLPFC), coupled with
frontal asymmetry for approach and avoidance behavior to appetitive stimuli. And recently, a study has found increased left (vs. right) DLPFC activation during approach (vs. avoidance) actions [2]. This effect was irrespective of the valence of the stimuli. Bilateral DLPFC stimulation could thus be hypothesized to actually enhance approach behavior and slow down avoidance behavior. However, this could also be an effect of unilateral stimulation, presuming that it leads to an activation asymmetry in the DLPFC. On the other hand, DLPFC activation has been implicated in exercising self-control [16]. Alcohol impairs executive functioning and decreased activity in the DLPFC has been found in alcohol-dependent subjects [14], which is the main reason that it was used as the site of stimulation in the present study. The differential involvement of the DLPFC in approach-avoidance behavior and executive functions might explain the lack of effects. It could be the case that stimulating the DLPFC is successful at improving working
memory [9]. However, it remains unclear what aspect of working memory is affected by stimulation of the DLPFC. In the SRC task, associative responses most probably reflect preexisting associations between stimuli and responses, as well as stimulus-response mappings established during the experiment. In the current experiment, stimulation of the DLPFC might have enhanced responses to both approach alcohol blocks and avoid alcohol blocks.
It is also possible that another electrode arrangement is necessary to selectively modify approach-avoidance behavior. A potential candidate for stimulation could be the right inferior frontal gyrus, which is situated in the lateral prefrontal cortex. This structure has been implicated in go/no-go tasks [1] in which participants are to inhibit prepotent responses. Approach-avoidance tasks and go/no-go tasks are conceptually similar in that
both deal with inhibition of responses and task-sets. The right inferior frontal gyrus might thus also be involved when it comes to overcoming an approach bias for alcohol by exerting its inhibitory function on more automatic processes in subcortical regions. In conclusion, our results show that the Stimulus-Response Compatibility (SRC) task in which a hand is used to approach or avoid a stimulus outperforms the classic version in which a manikin is used. The hand version thus seems a promising new approach and avoidance task for investigating automatic action tendencies. The application of tDCS above the DLPFC did not modify approach and avoidance behavior in the present study. Different stimulation parameters and/or alternative sites of stimulation, specifically the right inferior frontal gyrus, are warranted to explore tDCS as a means of modulating automatic action tendencies towards alcohol.
References
[1] Aron, A. R., Robbins, T. W., & Poldrack, R. A. (2004). Inhibition and the right inferior frontal cortex. Trends in Cognitive Sciences, 8, 170-7.
[2] Berkman, E. T., & Lieberman, M. D. (2010). Approaching the bad and avoiding the good: lateral prefrontal cortical asymmetry distinguishes between action and valence. Journal of Cognitive Neuroscience, 22, 1970-9.
[3] Boggio, P. S., Sultani, N., Fecteau, S., Merabet, L., Mecca, T., Pascual-Leone, A., Basaglia, A., et al. (2008). Prefrontal cortex modulation using transcranial DC stimulation reduces alcohol craving: a double-blind, sham-controlled study. Drug and Alcohol Dependence, 92, 55-60.
[4] Cattaneo, Z., Mattavelli, G., Platania, E., & Papagno, C. (2011). The role of the prefrontal cortex in controlling gender-stereotypical associations: a TMS investigation. NeuroImage, 56, 1839–1846.
[5] Chikazoe, J., Konishi, S., Asari, T., Jimura, K., & Miyashita, Y. (2007). Activation of right inferior frontal gyrus during response inhibition across response modalities. Journal of Cognitive Neuroscience, 19, 69-80.
[6] Conrey, F.R., Sherman, J.W., Gawronski, B., Hugenberg, K., Groom, C.J. (2005) Separating multiple processes in implicit social cognition: the quad model of implicit task performance. Journal of Personality and Social Psychology, 89, 469–487.
[7] De Houwer, J., Crombez, G., Baeyens, F., & Hermans, D. (2001). On the generality of the affective Simon effect. Cognition and Emotion, 15, 189-206.
[8] Fecteau, S., Pascual-Leone, A., Zald, D. H., Liguori, P., Théoret, H., Boggio, P. S., & Fregni, F. (2007). Activation of prefrontal cortex by transcranial direct current stimulation reduces appetite for risk during ambiguous decision making. The Journal of Neuroscience, 27, 6212-8.
[9] Fregni, F., Boggio, P. S., Nitsche, M., Bermpohl, F., Antal, A., Feredoes, E., Marcolin, M. a, et al. (2005). Anodal transcranial direct current stimulation of prefrontal cortex enhances working memory. Experimental Brain Research,166, 23-30.
[10] Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. (1998). Measuring individual differences in implicit cognition: the implicit association test. Journal of
Personality and Social Psychology, 74, 1464-80.
[11] Krieglmeyer, R., & Deutsch, R. (2010). Comparing measures of approach avoidance behaviour: The manikin task vs. two versions of the joystick task. Cognition & Emotion, 24, 810-828.
[12] Liebetanz D., Nitsche, M. A., Tergau, F., & Paulus, W. (2002). Pharmacological approach to the mechanisms of transcranial DC-stimulation-induced after-effects of human motor cortex excitability. Brain, 125, 2238-2247.
[13] Lim, S. L., O’Doherty, J. P., & Rangel, A. (2011). The decision value computations in the vmPFC and striatum use a relative value code that is guided by visual attention. Journal of Neuroscience, 31, 13214-13223.
[14] Magalhaes, A. C. (2005). Functional magnetic resonance and spectroscopy in drug and substance abuse. Topics in Magnetic Resonance Imaging, 16, 247-51. [15] Mogg, K., Bradley, B. P., Field, M., & De Houwer, J. (2003). Eye movements to
smoking-related pictures in smokers: relationship between attentional biases and implicit and explicit measures of stimulus valence. Addiction, 98, 825-36. [16] Ridderinkhof, K. R., Van den Wildenberg, W. P. M., Segalowitz, S. J., & Carter, C.
S. (2004). Neurocognitive mechanisms of cognitive control: the role of prefrontal cortex in action selection, response inhibition, performance monitoring, and reward-based learning. Brain and Cognition, 56, 129-40.
[17] Robinson, T. E. & Berridge, K. E. (1993). The neutral basis of drug
craving: an incentive-sensitisation theory of addiction. Brain Research Review, 18, 247-291.
[18] Schoenmakers, T., Wiers, R. W, & Field, M. (2008). Effects of a low dose of alcohol on cognitive biases and craving in heavy drinkers. Psychopharmacology, 197, 169-178.
[19] Spielberg, J. M., Stewart, J. L., Levin, R. L., Miller, G. A, & Heller, W. (2008). Prefrontal Cortex, Emotion, and Approach/Withdrawal Motivation. Social and Personality Psychology Compass, 2, 135-153.
[20] Wiers, R. W., Eberl, C., Rinck, M., Becker, E. S., & Lindenmeyer, J. (2011). Retraining automatic action tendencies changes alcoholic patients’ approach bias for alcohol and improves treatment outcome. Psychological Science, 22, 490– 497.
[21] Wiers, R. W., Rinck, M., Dictus, M., & Van den Wildenberg, E. (2009). Relatively strong automatic appetitive action-tendencies in male carriers of the OPRM1 G-allele. Genes, Brain, and Behavior, 8, 101-6.
[22] Wiers, R. W, Rinck, M., Kordts, R., Houben, K., & Strack, F. (2010). Retraining automatic action-tendencies to approach alcohol in hazardous drinkers. Addiction, 105, 279-87.
[23] Wiers, R. W., & Stacy, A. W. (2006). Implicit cognition and addiction. Current Directions in Psychological Science, 15, 292-296.
