VALUE-BASED ATTENTION:
THE EFFECTS OF NATURAL
VALUE STIMULI ON
ATTENTIONAL CAPTURE &
ATTENTIONAL NARROWING
Internship
report
EMESE KROON
10556818 – 5 JULI 2017
1
Recent experiments have shown that salience is not the only factor influencing attentional guidance. Learned value of task irrelevant distractors seems to capture our attention in a similar way. If value in general can affect guidance of our attention this effect might be there for stimuli with a natural value too. A clear example of a stimulus type with a natural value is the value of alcohol stimuli to people with an alcohol addiction. This natural value should be reflected in an attentional bias towards alcohol-related stimuli, even when the stimulus is task irrelevant. Our study focused on the effect of different types of natural value stimuli on attentional capture and attentional narrowing. Besides that, we tried to conceptually replicate the effect of learned value on attentional capture, using picture stimuli. Our results show that the presentation of alcohol-related stimuli caused no attentional narrowing and did not affect attentional capture on average. However, we did find a positive correlation between the distraction effect of alcohol stimuli and two alcohol consumption related outcome measures, suggesting that heavy drinking students are more distracted by alcohol stimuli than their less heavy drinking peers. This attentional bias can be considered a risk factor for developing an alcohol addiction. Although improvements should be made in the experimental design, we failed to replicate the earlier found learned value effect when using picture stimuli, indicating that more complex visual stimuli might capture attention in a different way than previously used color stimuli.
Introduction
Imagine yourself cycling in Amsterdam. You are rushing to an important appointment, but the clock is ticking. You see the next traffic light turning green when you hear your phone ringing. You answer the call and at the same time you miss the traffic light turning red when you approach it. The next thing you hear is vehicle horn close by, and you are only just able to stop in time. You failed to notice the red light, because your attention was focused on your phone.
This is just one example demonstrating the importance of attentional processes in daily life decision making. Attention is an important gating mechanism that enables us to focus on certain stimuli while ignoring other stimuli (Le Pelley, Mitchell, Beesly, George, & Wills, 2016). In general, two mechanisms that are important in guiding our attention have been proposed; top-down and bottom-up control (Anderson, Laurent, & Yantis, 2014; Le Pelley et al., 2016). Top-down attention, also referred to as goal-directed attention, is the process of deciding to focus attention to relevant stimuli and to ignore the stimuli that are irrelevant in a given situation. However, most of the time attention is involuntarily driven by stimuli themselves. This is commonly referred to as bottom-up, or stimulus-driven, attention.
2 It is well established that the salience of stimuli affects the chance of attending those stimuli (Beck, 1963; Theeuwes, 1992; Yantis & Jonides, 1984). The moment we hear that vehicle horn ring it is impossible to ignore the stimulus; the salience of the stimulus captures our attention quick and involuntary. However, salience is not the only factor that guides our attention. More recently researchers showed that within-task learned value of stimuli also affects attentional capture (Anderson & Yantis, 2013; Le Pelley, Pearson, Griffiths, & Beesley, 2015). When stimuli were previously related to a high-value outcome, those stimuli captured more attention than stimuli previously related with neutral or low-value outcomes, even when attending to the stimuli was counterproductive for the task at hand.
Counterproductive attentional selection has also been studied in relation to alcohol addiction (Field & Cox, 2008; Sharma, Albery, & Cook, 2001; Wiers et al., 2007). In heavy drinkers, alcohol-related stimuli capture more attention than non-alcohol-related stimuli do. This attentional bias is of great influence on the maintenance of addiction and can hinder rehabilitation processes (Field & Cox, 2008). Alcohol-related stimuli are a good example of stimuli that have a natural value. The stimuli are of high natural value for the addicted person, where they are of neutral or low natural value for non-addicted people. Another attentional process that could be considered counterproductive in relation to alcohol-related stimuli is attentional narrowing. Studies on attentional breadth showed that breadth of attention, and the related ability to detect changes outside of our attentional focus, can be altered by the relevance of the stimuli we focus on (Bosmans, Braet, Koster, & De Raedt, 2009; Grol & De Raedt, 2015). This suggests that salient or high-value stimuli, as is the case for alcohol stimuli in alcohol dependence, can narrow attentional breadth.
Although attentional bias is most prominent in dependent drinkers, studies have shown that non-dependent, social drinkers also show some bias towards alcohol-related stimuli (Duka & Townshend, 2004; Sharma et al., 2001). Combining these results with our current knowledge about the effect of attentional bias on alcohol addiction, the attentional bias towards alcohol-related stimuli seen in non-dependent social drinkers might increase risk of developing alcohol addiction (Field & Cox, 2008; Stacy & Wiers, 2010). Our study focused on students, known to vary in their alcohol consumption. We looked into the effect of natural versus learned value on attentional capture and the effect of natural value on attentional narrowing. Attentional capture was measured using an additional singleton paradigm as was earlier used in relation to learned value stimuli by Le Pelley et al. (2015). While focusing on a fixation cross, participants were told to search for one diamond shaped stimulus that was located on an imaginary circle, accompanied by five grey circles. They needed to respond to the orientation of a line within this diamond shape as quickly as possible and ignore the circle shapes, while a distracting picture was presented in one
3 of the circles in all but the distractor-absent trials. Slow reaction times would indicate a distraction effect of the presented picture. We expected that stimuli with a high learned or natural value would capture more attention and would therefore be more distracting than stimuli with low or neutral reward.
Attentional narrowing was studied using the attentional breadth task, developed by Bosmans et al. (2009). While focusing on a centrally presented picture that participants needed to categorize, they were asked to report the location of a target presented either close to or far from the central picture. Attentional narrowing, related to a specific stimulus type was measured by subtracting accuracy scores on the far trials from the those of close trials for the same picture type. We expected that stimuli with a high natural value would cause more attentional narrowing than stimuli with a low or neutral natural reward.
We expect that high natural value stimuli capture more attention and cause more attentional narrowing than stimuli with a low or neutral value. However, the natural reward of stimuli is dependent on individual experiences and habits. People who consume more alcohol are expected to attribute a higher value to alcohol-related stimuli compared to less heavy drinkers. Although there will be variety in the attributed value of alcohol-related stimuli, nude stimuli have been found to capture the attention of most individuals in a similar way (Jiang, Costello, Fang, Huang, & He, 2006; Most, Smith, Cooter, Levy, & Zald, 2007; Prause, Janssen, & Hetrick, 2008). Therefore, we expect nude stimuli to be of a general high value. This attentional bias towards alcohol stimuli, specifically found in heavy drinkers, is often found to be correlated with two cognitive factors: working memory (WM) capacity and impulsivity. People with a higher attentional bias for alcohol-related stimuli score on average lower on WM capacity tasks and show more impulsive behavior (Dick et al., 2010; Thush et al., 2008; Wiers et al., 2007). We expect to find similar correlations between attentional narrowing and distraction related to alcohol stimuli and these executive function measures (Anderson, Laurent, & Yantis, 2011).
We will look into the effects of natural value stimuli on two different attentional processes, and its correlations with individual differences in WM capacity, impulsivity and alcohol consumption. This can provide us with new insights into the attentional bias for natural value stimuli in relation to heavy alcohol use in young adults, and two well-established cognitive moderators of alcohol addiction. This information can in turn help us to expand our attentional framework of alcohol addiction and improve current interventions and treatments that focus on attentional biases (Wiers et al., 2015; Cox, Fadardi, Intriligator, & Klinger, 2014).
4
Methods
Subjects
Sixty-one students (40 female), with ages ranging between 18 and 28 years old (M = 22.1, SD = 2.6), participated in the experiment. Our sample size was comparable to earlier studies using similar tasks and was supported by a power analysis (f = .25, α = .05, power = .95, correlation = .2). Participants gave written informed consent before participating in the experiment, that was approved by the ethics committee of the department of Psychology of the University of Amsterdam. They received either course credits or a monetary reward for their participation.
Experimental design
The experiment consisted of three different tasks, all conducted in a single two-hour lab session. Before starting with the tasks, participants indicated their current hunger rate on a 7 point Likert scale and filled in the Barrett Impulsivity Scale 11 questionnaire (BIS-11; Patton & Stanford, 1995; Stanford et al., 2009). Besides that, they performed a quick WM assessment using a backwards digit span task (Conway et al. 2005). The first task used is a variation of the attentional breadth paradigm developed by Bosmans et al. (2009) that measures attentional narrowing. The attentional breadth task was followed by two variations of an additional singleton paradigm, that was already used in a similar way by Le Pelley et al. (2015), to assess attentional capture. After finishing the third task, participants filled in the Alcohol Use Disorders Identification Test (AUDIT), Rutgers Alcohol Problem Index (RAPI), Binge Eating Scale (BES) and restraint eating questionnaires to assess their drinking and eating behaviour (AUDIT: Reinert & Allen, 2002; 18-item RAPI: White & Labouvie, 1989; White & Labouvie, 2000; BES: Duarte, Pinto-Gouveia, Ferreira, 2015; Restraint eating: Herman & Polivy, 1980; van Strien, Herman, Engels, Larsen, & Leeuwe, 2007). Before debriefing, their height, weight, hip and waist were measured. For the sake of this article the analyses will not focus on food related outcome measures.
Task 1
Equipment, stimuli and trial design
The attentional breadth task was programmed using INQUISIT millisecond software (2016). All stimuli were presented on a black background using a 23” ASUS LCD monitor. Participant were seated in front of the screen using a chin rest to keep their eyes at a fixed 27cm from the screen, focusing on the center of the screen. Each trial started with the presentation of 16 grey circles (ø 2.0 cm), located on two imaginary circles; 8 of them located on the inner circle at 4.5cm from the center of the screen (10 degrees of visual angle (dva)), and 8 of them located on the outer circle at 11.2cm from the center of the screen (25 dva;
5 Figure 1). After 42ms, a picture stimulus (2.5x2.5cm) was presented in the center of the screen, accompanied by a black target circle (ø 1.3 cm), that was presented in one of the already present grey circles. The picture stimulus and the target circle were presented for 70ms, after which a response screen was presented asking to categorize the presented picture stimulus as edible or non-edible (Figure 1). This question was followed by a second question asking on which axis the black target circle had appeared. Four different stimulus categories were used, and could be partly selected based on the participants preferences (Table 1).
Procedure
The task started with a short practice phase in which the participant was first presented with an example trial in which the location of the target circle was indicated with a red arrow. The example was followed by 18 practice trials in which the presentation time of the picture and target decreased from 500ms (first 4 trials), over 250ms (trial 5-10), to the experimental presentation time of 70ms (last 8 trials). The picture stimuli in the practice trials were from the same categories, but opposite of the participants preference
Figure 1. Trial design attentional breadth task. Participants performed an attentional breadth task to
measure attentional narrowing. We measured attentional narrowing by presenting 16 circles, from which 8 were located in an inner circle and 8 were located in the outer circle. A centrally located picture and a black target circle (inside one already present grey circle) were presented for 70ms. After disappearance of the stimuli, a first question was asked about the picture category (“To which category did the picture in the center of the screen belong?” 1. Edible, 2. Non-edible). Afterwards a second question was asked about the location of the target circle (“On which axis was the target circle presented?” answers from 1 till 8). The difference in accuracy between the close (inner circle) and far (outer circle) trials was calculated as being the attentional narrowing index (ANI: close - far).
6 (e.g. when the participant choose wine stimuli for the experimental condition the participant would be presented with beer stimuli during the practice phase). After completing the practice phase, participants continued with 5 experimental blocks of 48 trials, resulting in a total of 240 trials. All categories included 6 different pictures (e.g. six different pictures from the category wine), that were all presented twice in each block, in a randomized order. Location of the target circle was also randomized.
Data Analysis
Accuracy on target localization, when stimulus categorization question was answered correctly, was used as the main outcome measure. All trials in which the stimulus categorization question was answered incorrectly were deleted from the analysis. If participants were incorrect on this question too often (3SD above the mean), which resulted in too many deleted trials, participants were excluded from the analysis. Because the chance of answering correctly on the target localization question was 12.5%, we excluded all participants who performed below 15% on either the close (inner circle) or the far trials (outer circle).
We performed a 2 (Distance: close, far) x 3 (Picture Type: alcohol, dressed, nude) repeated measures ANOVA on the accuracy data. To assess the effect of distance on performance per picture categories, we calculated Attentional Narrowing Indices (ANI = accuracy on close trials minus accuracy on far trials), and performed a one-way repeated measures ANOVA to compare the different categories. Higher ANI scores implicate more attentional narrowing. Additionally, we calculated correlations of the ANI with the scores on the WM task, the BIS-11, the RAPI and the AUDIT.
Table 1. Type of picture stimuli and their corresponding categories used in the attentional breadth task Type of stimuli Categories Main category
Alcohol Beer/wine* Edible
Food (Part of another project) Savory/sweet* Edible
General high natural value Male nudes or female nudes* Non-Edible
General low natural value Male dressed or female
dressed**
Non-Edible *Participants indicated their stimulus preference before starting the experiment.
7
Task 2
Equipment, stimuli and trial design
The additional singleton task, measuring attentional capture, was programmed using Presentation software (2017). All stimuli were presented on a black background using a 23” ASUS LCD monitor. Participants were seated in front of the screen focusing on a white fixation cross located in the center of the screen(400ms). Each trial started with the presentation of 5 grey circles (ø 2.3 dva) and 1 grey diamond shaped target (2.3x2.3 dva), placed on an imaginary circle (10.1 dva) at equal intervals (Figure 2). One of the grey circles was filled with a picture stimulus (Table 2), while all other circles were filled with a white oblique (45◦ tilted left or right) line segment (length 0.76 dva). The diamond shape target contained a white line segment (length 0.76 dva) that was either horizontal or vertical. Participants were asked to respond to the orientation of the line inside the target by pressing one of two possible response buttons as fast as possible (counterbalanced over participants). Orientation of the lines within the target and non-target shapes were randomly assigned within each trial. Pictures were assigned randomly, while presenting each picture once in each block. Two trials in each block were used as no distractor control trials, by presenting a line segment in all circles instead of filling one of the circles with one of the available pictures. The location of the target shape and distractor were randomly assigned. Feedback was provided after each trial, stating “Correct”, accompanied by a +1-point bonus, “Incorrect”, or “Too slow!”. Feedback was presented for 700ms on the correct trials, compared to 1000ms on incorrect or too slow trials (>750ms response time). No distractor control trials were not followed by a reward, even when answered correctly. The feedback was followed by a blank screen presented for 1000ms before the new fixation cross appeared.
Table 2. Type of picture stimuli used in the natural-value attentional capture task Type of stimuli Categories
Food (Part of another project) Unhealthy savory or sweet food*
Alcohol Beer or wine*
Low-value food and drink Healthy food and drinks
General high natural value Male nude or female nude*
General low natural value control Dressed male or dressed female**
*Participants indicated their stimulus preference before starting the experiment. **Other sex that than chosen in the general high natural value category.
8
Procedure
The task started with a practice block consisting of 10 trials, from which 7 had a yellow circle as distractor and 3 had no distractor. Participants continued with the practice until they performed with 70% accuracy. After practice, participants performed 10 blocks of 32 trials (6 pictures for all 5 stimulus types (see Table 2) and 2 distractor-absent trials), resulting in a total of 320 experimental trials. Participants could take a break after block 2, 5, 7 and 10. A non-variable total earning of 1.03 euros and 2.36 euros was shown after block 5 and 10 respectively. The earnings in task 2 were standardized to prevent influence of variability in earnings on task 3.
Data Analysis
Reaction time was used as the main outcome measure. Just as Le Pelley et al. (2015) did, we deleted the first two trials after the breaks and deleted all trials with a reaction time bellow 150ms. Besides that, all participants with an accuracy score over 3 SDs lower than average were discarded.
We performed two 2 (Expected Reward: high, low) x 2 (Picture Type: drinks, people) repeated measures ANOVAs to compare the alcohol stimuli with the healthy stimuli and compare the nude stimuli
Figure 2. Trial design attentional capture tsak. An additional singleton paradigm was used to assess
attentional capture. Each trial started with presentation of a target circle, participants were instructed to focus on. After 400ms, 5 circles and one diamond shaped target were presented. All but one circles were filled with an oblique line (45◦ tilted left or right). The remaining circle was filled with a distractor picture. Participants were instructed to search for the diamond shaped target and report whether the line inside this target was horizontal or vertical as quickly as possible. When answered correctly, participants saw a feedback screen indicating their performance. Presentation time of the feedback was performance dependent. Feedback was followed by a blank screen presented for 1000ms.
9 with the dressed stimuli. After this, a one way repeated measures ANOVA was used to compare alcohol stimuli trials, with healthy stimuli trials and absent trials and the same analysis was used to compare nude stimuli trials, with dressed stimuli trials and absent trials. Additionally, we did a correlational analysis on the relevant difference scores and the scores on the WM capacity test, the BIS-11, RAPI and AUDIT.
Task 3
Equipment, stimuli and trial design
Task 3 was an additional singleton task, very similar to the one used in task 2 (Figure 2), that was used to try to conceptually replicate the findings of Le Pelley et al. (2015) with regards to the learned value of task irrelevant distractors. Instead of using red and blue distractor circles that were linked to either a high (+10 cents) or a low (+1 cents) reward, we used red and blue pieces of furniture linked to either a high or low learned reward. All stimuli were presented on a black background using a 23” ASUS LCD monitor. Tasks 3 only differed from task 2 in the fact that the different natural value stimuli used in task 2 were replaced by pictures of red and blue furniture (Seven different pictures per category, all presented twice in each block, accompanied by four distractor-absent trials in each block; Table 3) and the fact that correct responses on the high reward distractor trials earned 10 cents, compared to only 1 cent when correct on the low reward distractor trials. Monetary gains were provided based on performance, in contrast to the standardized gain in task 2.
Procedure
The task contained 6 blocks of 32 trials, which is similar to the amount of trials needed to show a stable performance in Le Pelley (2015). Participants could take a break after each two blocks and would be provided with truthful feedback about the money they had gained so far. After finishing the last block, contingency questions were asked to check whether participants had learned the association between the furniture color and the monetary gain. The total monetary gain of task 2 and task 3 was added to the participation reward or course credits.
Table 3. Type of picture stimuli used in the learned-value attentional capture task
Type of stimuli Categories
High learned value (+10 cents) Furniture (blue or red)*
Low learned value (+1 cent) Furniture (blue or red)*
10
Data Analysis
Reaction time was used as the main outcome measure. Just as Le Pelley et al. (2015), we deleted the first two trials after the breaks and deleted all trials with a reaction time bellow 150ms. Besides that, all participants with an accuracy score over 3 SDs lower than average were discarded. We performed a 6 (Block: 1-6) x 3 (Reward: absent, low, high) repeated measures ANOVAs to look at the effect of the different distractor types over blocks.
Results
Task 1
Preliminary analysis
All trials in which the picture categorization question was answered incorrectly were deleted. When the amount of deleted trials was over three standard deviations from the mean, the participant was discarded from analysis. This criterion resulted in discarding one participant. From the remaining participants, an average of 0.97% of the trials were deleted. All individual pictures were identified with at least 93% accuracy. Our second exclusion criterion; scoring at least 15% correct on the target location question in both the close and far trials (12,5% being chance performance), resulted in discarding another 10 participants. Leaving us with the remaining 49 participants (31 female).
Assumption checks
Before performing a repeated measures ANOVA, we checked the corresponding assumptions of sphericity and normality. Shapiro Wilk’s test of normality showed that all but the far trials from the dressed category violated the assumption of normality (p < .05). However, when combining both close and far trials into an average score, only the alcohol trials violated the assumption (W(49) = .95, p = .03) and when calculating the ANI, none of the difference scores violated the assumption of normality. Additionally, Mauchly’s test of sphericity showed that none of the stimulus categories did violate the assumption of sphericity (p > .05).
Accuracy analysis
A 2 (Distance: close, far) x 3 (Picture Type: alcohol, dressed, nude) repeated measures ANOVA revealed a main effect of both Picture Type (F(2,96) = 4.01, p = .021, η² = .08; Figure 3b) and Distance (F(1,48) = 427.08, p < .001, η² = .90). No Distance by Picture Type interaction was found (F < 1; Figure 3a). Subsequent post-hoc t-tests (Bonferroni corrected) showed that accuracy on the alcohol trials (M = .62, SD = .03) was significantly higher than the accuracy on the nude trials (M = .59, SD = .02; t(48) = 2.85, p = .02; Figure 3b; Table 4). There was no difference between the alcohol trials (M = .62, SD = .03) and the dressed trials (M
11 = .60, SD = .02; t(48) = 2.01, p = .14), and the nude trials (M = .59, SD = .02) and the dressed trials (M = .60,
SD = .02; t(48) = .60, p = 1.00).
Figure 3. Results of the attentional breadth task. (A) Accuracy on all types of picture stimuli for both far and close trials separately.
Accuracy was on average higher for close than for far trials. Error bars reflect the standard error of the mean. (B) Accuracy per picture category, reflecting the average score of both close and far trials for the same stimulus category taken together. Accuracy of the alcohol trials was higher than accuracy on the nude trials. The overlapping error bars reflect the general average standard error from the mean, not reflecting the repeated measures nature of the data. See Table 4 for the difference scores. (C) Attentional narrowing indices (ANI) per stimulus category were calculated by subtracting the accuracy score on the far trials from the accuracy score on the close trials. ANI did not differ between different stimulus categories. Error bars reflect the standard error from the mean.
0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
Alcohol Dressed Nude
Acc
u
ra
cy
Accuracy for close and far trials per stimulus type
Close Far A 0,52 0,54 0,56 0,58 0,6 0,62 0,64 0,66 A cc u racy
Accuracy per picture category
Alcohol Dressed Nude
B 0,42 0,44 0,46 0,48 0,5 0,52 ANI
Attentional narrowing index (ANI)
per stimulus category
Alcohol Dressed Nude
C
Table 4. Post-hoc results on the difference between different stimuli types regardless of distance
Stimulus 1 Stimulus 2 Mean difference SE t p-value (Bonferroni corrected)
Alcohol Dressed 0.02 0.01 1.81 0.23
Alcohol Nude 0.03 0.01 3.05 0.01
12
Attentional narrowing
We calculated Attentional Narrowing Indices (ANI) for all picture types, by subtracting the accuracy score on the close trials from the accuracy score of the far trials. A one-way repeated measures ANOVA, comparing the ANI of the different picture types, showed no effect of picture type (F < 1; Figure 3c).
Correlational analysis
One additional participant was discarded from the correlational analysis due to incomplete questionnaire assessment (N = 48, 30 female). Results showed that there was a positive correlation between RAPI and AUDIT score (r(46) = .74, p < .001), implicating a consistent measure of drinking behavior. Besides that, both the AUDIT and RAPI scores were positively correlated with the total BES scores (AUDIT: r(46) = .32, p = .03; RAPI: r(46) = .30, p = .04). No other significant correlations were found.
Task 2
Preliminary analysis
One participant was discarded from the analysis, leaving us with 60 participants that completed the entire task (39 female). A picture reliability check showed that there was no reliability problem regarding picture stimuli that represented the same stimulus category (Cronbach’s alpha > .85).
Assumption checks
We checked the available data for the assumptions of normality and sphericity, before running our analyses. Shapiro-Wilk’s test of normality showed that the assumption of normality was not violated (p > .05). Besides that, Mauchly’s test for sphericity showed that combined dataset of dressed, nude and distractor-absent trials and the interaction effect of this data with gender violated the assumption of sphericity (χ2(60) = .87, p = .02). Therefore, we corrected the degrees of freedom by using the Greenhouse-Geisser estimate of sphericity (ε = .89).
Reaction time analysis
To check whether the expected effect of natural reward influenced reaction time, we performed a 2 (Expected Reward: high, low) x 2 (Picture Type: drinks, people) repeated measures ANOVA. The results show that there is no effect of Expected Reward (F < 1), Picture Type (F(1,58) = 5.12, p = .88, η² = .00) nor an interaction effect (F < 1; Figure 4a).
A one way repeated measures ANOVA comparing alcohol trials (M = 614.3, SD = 47.3) with healthy trials (M = 612.2, SD = 44.3) and distractor-absent trials (M = 608.5, SD = 50.5) revealed no effect of Picture Type (F(2,116) = 2.79, p = .07, η² = .04), indicating no increased distraction caused by alcohol stimuli compared to healthy stimuli and distractor-absent trials. However, an exploratory analysis, adding Gender
13 as a between subjects factor, showed that there was a significant interaction between Picture Type and Gender (F(2,116) = 4.82, p = .01, η² = .07; Figure 4b).
Another one way repeated measures ANOVA comparing nude trials (M = 612.2, SD = 48.5) with dressed trials (M = 612.0, SD = 49.9) and distractor-absent trials (M = 608.5, SD = 50.5) revealed no effect of Picture Type either (F(1.74,116) = 2.52, p = .09, η² = .04). This indicates that nude stimuli were no more distracting than the dressed stimuli and the trials without a distractor. However, similar as in the previous exploratory comparison, adding Gender as a between subject factor showed a similar significant interaction between Picture Type and Gender (F(1.74,116) = 8.00, p = .001, η² = .12; Figure 4c).
A paired t-test showed that male participants were over all significantly faster on the distractor-absent trials, then women were (t(20) = 2.25, p = .04; Figure 4b and 4c).
Correlational analysis
Difference scores between several picture types were calculated to look into the effect of one distractor, compared to the other distractors and the distractor-absent trials. One additional participants was discarded from this analysis due to incomplete questionnaire assessment (N = 59, 38 female). A correlational analysis was done to check whether these difference scores, indicating a distraction effect, correlated with either WM capacity or RAPI, AUDIT or BIS-11 scores. Analysis showed that there is a positive correlation between RAPI and AUDIT score (r(57)= .78, p < .001), again implicating a consistent measure of drinking behavior. Besides that, there was a positive correlation between AUDIT (r(57) = .35,
p = 0.01) and RAPI (r(57) = .30, p = .02) score and the distractor effect of alcohol stimuli (alcohol minus
absent, Figure 5a and 5b).
Task 3
Preliminary analysis
Just as in the analysis of task two, we adopted the following two trial exclusion criteria; the first two trials after each break and all trials with a reaction time below 150ms were discarded. No participants were discarded based on the additional criterion that accuracy should be within three standard deviations from the mean. Reaction time data of all 61 participants (40 female; M = 22.1, SD = 2.6) was analyzed as being the dependent variable. A picture reliability check showed that there was no reliability problem regarding picture stimuli that represented the same stimulus category (Cronbach’s alpha > .94).
14
Figure 4. Results of the natural-value attentional capture task. The effect of the distractor is reflected in the reaction times.
Stimuli that capture more attention will show a longer reaction time. (A) Effect of expected reward of specific stimuli on reaction time. The expected high-value drinks (alcohol) and people (nude), did not differ from the expected low-value drinks (healthy) and people (dressed). Error bars reflect the standard error of the mean per group. (B) Effect of healthy and alcohol stimuli (compared to absent) per gender. Error bars reflect the standard error of the mean per group. (C) Effect of dressed and nude stimuli (compared to absent) per gender. Error bars reflect the standard error of the mean per group.
600 605 610 615 620 625 High Low Re act ion t im e (m s)
Distraction effect of expected reward
Drinks People A 570 590 610 630
Absent Healthy Alcohol
R eac ti o n ti me ( ms )
Distraction effect of healthy and
alcohol stimuli per gender
Male Female B 570 590 610 630
Absent Dressed Nude
Re acti o n ti me (ms )
Distraction effect of dressed and
nude stimuli per gender
Male Female
C
Figure 5. correlational results. A difference score between reaction time on alcohol stimuli trials and distractor-absent trials
was calculated to reflect the distraction effect of alcohol stimuli. Correlational analysis showed a positive correlation between the distraction by alcohol stimuli and AUDIT score (A) and between the distraction by alcohol stimuli and RAPI score (B).
-100 -50 0 50 100 0 5 10 15 20 25 30 D ist rac ti o n ef fe ct ( m s) Audit score
Correlation between alcohol distraction effect and AUDIT score
A -100 -50 0 50 100 0 5 10 15 20 25 30 D ist rac ti o n ef fe ct ( m s) RAPI score
Correlation between alcohol distraction effect and RAPI score
15
Assumption checks
As for the previous tasks we ran an analysis on the available data to check for the assumptions of normality and sphericity. Shapiro-Wilk’s test of normality showed that data for none of the reward types violated the assumption of normality (p > .05). However, Mauchly’s sphericity test showed that both reward type (χ2(2) = .83, p = .004) and the reward type interaction with block (χ2(10) = .06, p < .001), did violate the assumption of sphericity. The degrees of freedom were corrected using a greenhouse-Geisser correction (ε = .85, and ε = .60). The same applied to the interaction between reward type and contingency awareness and the interaction between reward type and gender.
Reaction time analysis
A 6 (Block: 1-6) x 3 (Reward: absent, low, high) repeated measures ANOVA showed no effect of Block (F(5,300) = 1.36, p = .24, η² = .02), but did reveal an effect of Reward (F(1.71,120) = 12.10, p < .001, η² = .17; Figure 6a). The interaction between Block and Reward was not significant (F < 1).
Subsequent post-hoc t-tests (Bonferroni corrected) revealed that the reward effect was driven by the lower reaction times on the distractor-absent trials. There was a significant difference between high reward (M = 609.8, SD = 64.0) and distractor-absent trials (M = 598.2, SD = 71.1; t(59) = 3.37, p = .002), and between low reward (M = 609.2, SD = 65.8) and distractor-absent trials (t(59) = 3.51, p = .001). However, there was no difference between the high reward and low reward trials (t < 1).
Because there was no Block effect and the Reward effect could be driven by contingency awareness, we ran an exploratory one-way repeated measures ANOVA on Reward, with Contingency Awareness as a between-subjects factor. As there were four questions on contingency awareness, people were only considered aware of the association when scoring at least three out of four correct. The ANOVA results showed the Reward effect that was also present in the block wise analysis (F(1.63,118) = 6.01, p = .01, η² = .09). However, there was no interaction between Reward and Contingency Awareness (F < 1) and the between subjects effect of Contingency Awareness was not significant either (F(1,59) = 3.74, p = .06,
η² = .06; Figure 6b).
As another exploratory analysis, based on the interaction with Gender found for task 2, we ran a one-way repeated measures ANOVA with Gender as a between subject variable. The results showed, additionally to the Reward effect mentioned above, that there was a significant Gender by Reward interaction (F(1.68,118) = 3.74, p = .01, η² = .08; Figure 6c), showing a similar pattern as can be seen in Figure 4c. However, the between subjects effect of Gender was not significant (F(1,59) = 1.86, p = .18, η² = .03).
16 To check whether the difference between male and female on the absent trials in task 2 was consistent over tasks, we checked whether male and female differed on the distractor-absent trials in task 3. A paired t-test (Bonferroni corrected) showed that male participants (M = 574.1, SD = 67.6) were indeed significantly faster on the distractor-absent trials than female participants were (M = 610.8, SD = 70.5;
t(20) = 2.22, p = .04).
Figure 6. Results of the learned-value attentional capture task. The effect of the distractor is reflected in the reaction time.
Stimuli that capture more attention will show a longer reaction time. (A) Effect of the three different learned value stimuli on reaction time per block. Error bars display the mean standard error of the group mean. (B) The effect of the three different learned value stimuli on reaction time for the group of participants that was aware of the stimulus reward association, compared to the participants that were unware of the association. Error bars display the mean standard error of the group mean. (C) The effect of the three different learned value stimuli on reaction time for both male and female participants separately. Error bars display the mean standard error from the group mean.
560 570 580 590 600 610 620 1 2 3 4 5 6 Re act ion t im e (m s)
Effect of distractor type per block
Absent Low High
AA
550 600 650
Absent Low High
R eac ti o n ti me ( ms )
Effect of distractor: variable
contingency awareness
Aware Unaware B 550 600 650Absent Low High
R eac ti o n ti me ( ms )
Effect of distractor: gender
Male Female
17
Discussion
Although it is well established that salience of stimuli affects the chance we attend those stimuli, less is known about the more recently found effect of stimulus value on attentional processes. The aim of the current study was to examine whether the natural value of picture stimuli affects attentional narrowing (task 1) and attentional capture (task 2). Besides that, we tried to conceptually replicate the learned value effect, as was found in the study of Le Pelley et al. (2015), by using colored picture stimuli linked to a monetary reward (task 3).
Unlike expected, the results on the attentional breadth task showed that there was no difference in attentional narrowing between trials with either alcohol, nude or dressed stimuli. However, average accuracy was higher on the alcohol trials than it was on the nude trials, while no difference was found between the dressed trials and the alcohol nor nude trials. These results are inconsistent with the expectation that nude stimuli are of a high natural value to most people and therefore elicit more attentional narrowing than the dressed stimuli (Jiang et al., 2006; Most et al., 2007; Prause et al., 2008). Furthermore, no correlation was found between attentional narrowing for alcohol stimuli and scores on the AUDIT and RAPI questionnaires, indicating that the amount of attentional narrowing by alcohol stimuli could not be explained by individual differences in drinking behavior. It might be that the complexity of the alcohol stimuli was not comparable to the complexity of the human stimuli. This could have made them easier to identify, enabling broadening of attention and causing higher average accuracy on the alcohol trials. The results of the attentional breath task were consistent with earlier experiments in that participants were better at identifying the location of a target close to the central stimuli than they were at identifying targets presented further to the periphery. Participants seemed to perform slightly better than seen in previous experiments using a similar task, which was probably caused by the at least 15% correct on both close and far trials exclusion criteria that we adopted. Although the additional criterion might have reduced comparability with previous studies, it seemed reasonable to use a slightly more conservative criterion than the 12,5% (chance) criterion, to make sure that participants made a reasoned decision. Besides that, AUDIT and RAPI scores were positively correlated with impulsivity scores, which is consistent with previous studies on the relation between executive control and alcohol consumption. We did not find the often-found negative correlation between either of these scores and WM capacity (Dick et al., 2010; Thush et al., 2008; Wiers et al., 2007).
Two versions of the additional singleton paradigm were used to assess attentional capture by either natural or learned value stimuli. Results on the natural-value attentional capture task did not match our expectations with regards to the natural value of stimuli. Even stimuli that were expected to be of a
18 natural high value (nudes) affected attentional capture no more than the dressed stimuli or even the distractor-absent trials. For the alcohol stimuli, we found no distraction effect on average, but the distraction effect of alcohol was positively correlated with both AUDIT and RAPI scores. This indicates that those who consume more alcohol and show more worrisome drinking behavior, which is reflected in higher scores on both questionnaires, were more distracted by alcohol-related stimuli of their preference. These results are consistent with previous studies that found that dependent and non-dependent heavy drinkers show an increased attentional bias towards alcohol stimuli (Duka & Townshend, 2004; Fadardi & Cox., 2009; Field & Cox., 2008; Sharma et al., 2001). This attentional bias is one important factor in the development and maintenance of alcohol addiction and the fact that this attentional bias is clearly visible in our student population points out the importance of developing interventions that focus on reducing attentional bias. Different attentional bias modification treatments have proven to be effective in reducing attentional bias and improving treatment progression in patients with an alcohol addiction (Schoenmakers et al., 2010; Wiers, Eberl, Rinck, Becker,& Lindenmeyer, 2011). Attentional bias modification is a way of retraining attentional bias away from the addiction related cues. This disengagement procedure changes automatic tendencies towards the stimuli and could therefore be beneficial in the treatment of addiction. Although studies show that the results are still somewhat inconsistent, this method of treatment focusing on attentional bias is a promising approach for future treatment of alcohol addiction (Clerkin, Magee, Wells, Beard, & Barnett, 2016; Schoenmakers, Wiers, Jones, Bruce, & Jansen, 2007). The treatment, can easily be used in combination with other, well-established, treatment protocols and development of an online protocol seems to be a reasonable future step.
In the learned-value attentional capture task we found an effect of distracter type. Participants are generally faster on the distractor-absent trials then they are on the high and low value distractor trials, but unlike expected there was no difference in reaction time between the high and low value distractor trials. This effect could have been caused by the presence of any distractor compared to no distractor at all, what could be explained as a salience effect. However, such an explanation would not fit the results on the natural value version of the task, in which we found no difference between distractor-absent and distractor-present trials. Further analyses showed that there was a marginally significant difference in reaction time between participants who were aware of the furniture color and reward association and the participants who were not aware of this association. Participants who were aware of the association were faster in performing the task in general. We would expect that people who are aware of the high value of a certain color stimulus would be more distracted by that stimulus than people who watch the same stimulus as a stimulus with no value. The current results do not match this expectation. It could have been
19 the case that people who responded faster in general pick up the association quicker. However, our male participants respond faster on average, while only 38% of the male participants were aware of the association compared to 57% of the female participants (Hodgkins, 1963; Shelton & Kumar, 2010). Therefore, this explanation is rather unlikely. On average, only 51% of the participants were aware of the association between color and reward. This relatively low number of aware subjects suggests that further studies using a similar task should consider being explicit about the association or use more blocks to increase the learning period. When we compare our results on the learned-value attentional capture task to those of Le Pelley et al. (2015), the absence of a learning effect over blocks stands out. It is a good possibility that we did not replicate this learning effect because of the amount of similar trials our participants already performed during the natural value task.
An interesting exploratory result, seen in both versions of the additional singleton task, is the interaction between gender and stimulus type. Where males were generally just slightly faster than women on the distractor-present trials, there was a big difference on the distractor-absent trials. Like expected, males were faster on the distractor-absent trials, while females were equally distracted by the distractor-present and distractor-absent trials. The combined scores of males and females could have averaged out a possible finding in the natural value version of the task, where the interaction effect was slightly clearer than it was in the learned value version. Low power when analyzing the groups separately made it impossible to draw correlational conclusions based on single gender data. Further experiments, using the same type of task, should check for gender as a possible confounding factor.
Although some results clearly met our expectations, there were unexpected results that highlighted some limitations of the experimental and task designs. First, we should mention some general limitations of the picture stimuli we used, that could have affected the results of all three tasks in a different way. Like already mentioned, the different picture categories used in attentional breadth and the natural-value attentional capture task did differ in picture complexity. For the attentional breadth task, it could be the case that the less complex pictures were easier to identify and caused less attentional narrowing than the more complex pictures. It might be worth trying to present only one picture type per block to ensure that participants are unable to compare the stimuli with previous stimuli and react based on the difference between these stimuli instead of the stimuli itself. Another change that would be beneficial for stimulus identification in all tasks, is increasing the size of the pictures. Picture stimuli were relatively small and it could be the case that we did not find an effect of picture type in the natural-value attentional capture task because of this small picture size in combination with the complexity of the stimuli. However, the fact that the only the distraction effect of alcohol stimuli was positively correlated
20 with the AUDIT and RAPI scores indicates that it is unlikely that participants were unable to identify the stimuli as belonging to a specific stimulus type. Lastly, the stimuli used are not as brightly colored as the stimuli generally used in the attentional capture task (Le Pelley et al. 2015). It might be that our stimuli were less salient and therefore easier to ignore than the stimuli originally used.
A second general limitation worth mentioning is the fixed task order we used. The same picture stimuli were used in both task 1 and task 2 without counterbalancing the order of the tasks. Although this fixed order removed possible effects of the additional monetary reward in task 2 on performance in task 1, it could be that picture habituation reduced the chances of finding a distraction effect in task 2. Besides that, the high amount of trials participants performed in task 2, is most likely the reason that we did not find a learning effect in task 3.
Third, the consistently found variety of gender differences in response to nude and dressed stimuli, from both the same and opposite sex, could have influenced the results on both the nude and dressed trials (Chivers, Seto, & Blanchard, 2007; Costa, Braun, & Birbaumer, 2003; Nummenmaa, Hietanen, Santtila, & Hyönä, 2012; Wright & Adams, 1999). Future experiments using these type of stimuli, while testing both male and female participants, should be aware of this possible confounding factor and check for gender differences in response if power is sufficient. To reduce the risk of facing low power in separate gender analyses, like we experienced in the current experiment, it would be reasonable to recruit more participants with a balanced amount of male and female participants, or to recruit male or female participants only.
Fourth, we made some changes in the attentional breadth paradigm that reduced comparability with previous studies and might have influenced the effects we measured. Unlike previous experiments, we used stimulus types that were very dissimilar from each other. Generally, attentional breadth tasks use more comparable pictures, such as different face expression, in which a slight nuance would cause a difference in accuracy scores. The difference in attentional narrowing we would measure by presenting our picture stimuli would certainly not be based on such a nuance. Our stimuli are easier to differentiate and could also induce an attentional narrowing effect that is affected by differences in stimulus complexity. Another difference is that our stimuli were all relatively appetitive, while most often stimuli from a broad valence spectrum are used (e.g. sad faces, neutral faces, and happy faces). It would be better to add non-appetitive control stimuli that are expected to cause less attentional narrowing than the more appetitive stimuli. Besides that, it might be that the picture identification question was not sufficient to make a good comparison between picture categories. The question asking whether the stimulus could be categorized as edible or non-edible, makes a categorization in which all human pictures are categorized as
21 non-edible. This might have caused a broad categorization, instead of the specific stimulus identification we expected to see; all human pictures are directly filtered to be non-edible. However, previous experiments focusing on subliminal effects of nude and dressed stimuli suggest that the effect of both types of stimuli are present in reaction time data, even without participants being conscious of the stimulus itself (Jiang et al., 2006). The current task was not developed for using reaction time as an outcome measure, but we cannot exclude that the processing of nude and dressed stimuli, even when belonging in the same category, did alter reaction time instead of the accuracy we measured. Another interesting observation that could have affected out results is that participants consistently reported to get bored during the attentional breadth task. This was mainly caused by the fact that most participants were not aware of the fact that they saw the target circle on the far trials and did not receive any feedback about their performance. Although it did not cause lower accuracy scores over time we cannot exclude that participants would not have improved over time if they were persistently motivated during the task. An increase in average performance could lead to discarding less participants. Besides that, it could be that the reduced motivation has consistently affected the results of the tasks that followed. Future experiments using this task should consider presenting feedback in between trials to keep participants motivated during the task. Last, a more exploratory question, where future research could focus on, is whether the close and far trials measure the same construct. We found that participants consistently report to be unaware of the target location in the far trials, while still responding well above chance. This differs from the response given in the close trials where participants are well aware of the target location and respond with far higher accuracy than on the far trials. Adding a confidence rating might be good to experimentally confirm this trend. We can conclude that methodological improvements need to be made to be able to use the attentional breadth task for assessment of attentional narrowing caused by more complex picture stimuli.
Before further use of the attentional capture task with picture stimuli, some improvements need to be made on the stimuli themselves and the way they are presented. Besides that, we were not able to replicate the learned value effect Le Pelley et al. (2015) found by using picture stimuli. Nonetheless, there are some interesting ways in which this task could be used to further investigate the effect of natural value on attentional capture. One possibility is to use a task design similar to the one used by Le Pelley et al. (2015) in which they use colored circle, linked to a specific value, as distractors. The first blocks would present the original task in which the color value association is learned. After that, we could present pictures from a specific picture category inside the circles from a specific learned value (color) category in half of the trials and present only the colored circles in the other half of the trials. In this way, you can
22 combine both the learned and natural value version of the task and assess the additional effect of natural value on the learned value affect, which will be reflected in reaction times. Another adjustment that might improve the current design is a stimulus related reward. In the original learned value version of the task, the amount of points gained is translated into a monetary reward. Therefore, specific colors are linked to a higher amount of money. In our current natural value task, there is no such a stimulus related reward. It is a possibility to design a task in which the amount of points earned for specific pictures will be translated in a picture related reward such as a certain amount of a free alcoholic consumption, or in case of a food related study a certain amount of sweets. Adding this reward to the task increases comparability with the learned value version of the task and may induce an increased attentional bias towards stimuli that are of high value to the participant, because he or she would actually receive a reward related to those stimuli. In sum, the current study showed that students with higher scores related to alcohol consumption and worrisome drinking behavior are more distracted by alcohol-related cues. The fact that we found this attentional bias towards alcohol-related stimuli is, that is a well-known risk factor for the development and maintenance of alcohol addiction, in a young adult student sample illustrates once more how important it is to acknowledge drinking problems in students and work on interventions that use attentional bias modification to reduce this attentional bias. In this way studying different attentional processes in relation to natural value stimuli can help us to reveal specific and individual relationships between attention for addiction specific stimuli and its cognitive and behavioral moderators. This will help us to explore the utility of attentional bias modification, in combination with other cognitive or behavioral treatment, for intervention and treatment of addiction.
23
References
Anderson, B. A., Laurent, P. A., & Yantis, S. (2011). Value-driven attentional capture. Proceedings of the
National Academy of Sciences, 108(25), 10367-10371. https://doi.org/10.1073/pnas.1104047108
Anderson, B. A., Laurent, P. A., & Yantis, S. (2014). Value-driven attentional priority signals in human basal ganglia and visual cortex. Brain Research, 1587, 88-96. https://doi.org/10.1016/j.brainres.2014 .08.062
Anderson, B. A., & Yantis, S. (2013). Persistence of value-driven attentional capture. Journal of
Experimental Psychology: Human Perception and Performance, 39(1), 6-9. https://doi.org/10.1037/a0030860
Beck, S. B. (1963). Eyelid conditioning as a function of CS intensity, UCS intensity, and Manifest Anxiety Scale score. Journal of Experimental Psychology, 66(5), 429-438. https://doi.org/10.1037/h0042820
Bosmans, G., Braet, C., Koster, E., & Raedt, R. D. (2009). Attachment security and attentional breadth toward the attachment figure in middle childhood. Journal of Clinical Child & Adolescent
Psychology, 38(6), 872-882. https://doi.org/10.1080/15374410903258926
Chivers, M. L., Seto, M. C., & Blanchard, R. (2007). Gender and sexual orientation differences in sexual response to sexual activities versus gender of actors in sexual films. Journal of Personality and
Social Psychology, 93(6), 1108. https://doi.org/10.1037/0022-3514.93.6.1108
Clerkin, E. M., Magee, J. C., Wells, T. T., Beard, C., & Barnett, N. P. (2016). Randomized controlled trial of attention bias modification in a racially diverse, socially anxious, alcohol dependent sample.
Behaviour Research and Therapy, 87, 58-69. https://doi.org/10.1016/j.brat.2016.08.010
Conway, A. R., Kane, M. J., Bunting, M. F., Hambrick, D. Z., Wilhelm, O., & Engle, R. W. (2005). Working memory span tasks: A methodological review and user’s guide. Psychonomic Bulletin & Review,
12(5), 769-786. https://doi.org/10.3758/BF03196772
Costa, M., Braun, C., & Birbaumer, N. (2003). Gender differences in response to pictures of nudes: a magnetoencephalographic study. Biological Psychology, 63(2), 129-147. https://doi.org/10.1016/S0301-0511(03)00054-1
Cox, W. M., Fadardi, J. S., Intriligator, J. M., & Klinger, E. (2014). Attentional bias modification for addictive behaviors: clinical implications. CNS spectrums, 19(03), 215-224. https://doi.org/10.1017/S1092852914000091
24 Dick, D. M., Smith, G., Olausson, P., Mitchell, S. H., Leeman, R. F., O'Malley, S. S., & Sher, K. (2010). Review: understanding the construct of impulsivity and its relationship to alcohol use disorders. Addiction
Biology, 15(2), 217-226. https://doi.org/10.1111/j.1369-1600.2009.00190.x
Duarte, C., Pinto-Gouveia, J., & Ferreira, C. (2015). Expanding binge eating assessment: Validity and screening value of the Binge Eating Scale in women from the general population. Eating Behaviors,
18, 41-47. https://doi.org/10.1016/j.eatbeh.2015.03.007
Duka, T., & Townshend, J. M. (2004). The priming effect of alcohol pre-load on attentional bias to alcohol-related stimuli. Psychopharmacology, 176(3-4), 353-361. https://doi.org/10.1007/s00213-004-1906-7
Fadardi, J. S., & Cox, W. M. (2009). Reversing the sequence: reducing alcohol consumption by overcoming alcohol attentional bias. Drug and Alcohol Dependence, 101(3), 137-145. https://doi.org/10.1016/j.drugalcdep.2008.11.015
Field, M., & Cox, W. M. (2008). Attentional bias in addictive behaviors: a review of its development, causes, and consequences. Drug and Alcohol Dependence, 97(1), 1-20. https://doi.org/10.1016/j.drugalcdep.2008.03.030
Herman, C. P., & Polivy, J. (1980). Restrained eating. In A. J. Stunkard (Ed.), Obesity (pp. 208-225). Philadelphia: W.B. Saunders.
Hodgkins, J. (1963). Reaction time and speed of movement in males and females of various ages. Research
Quarterly. American Association for Health, Physical Education and Recreation, 34(3), 335-343.
Inquisit 4.0 [Computer software]. (2016). Seattle. WA: Millisecond Software
Le Pelley, M. E., Mitchell, C. J., Beesley, T., George, D. N., & Wills, A. J. (2016). Attention and associative learning in humans: An integrative review. Psychological Bulletin, 142(10), 1111-1140. https://doi.org/10.1037/bul0000064
Le Pelley, M. E., Pearson, D., Griffiths, O., & Beesley, T. (2015). When goals conflict with values: Counterproductive attentional and oculomotor capture by reward-related stimuli. Journal of
Experimental Psychology: General, 144(1), 158-171. https://doi.org/10.1037/xge0000037
Grol, M., & De Raedt, R. (2015). The influence of psychological resilience on the relation between
automatic stimulus evaluation and attentional breadth for surprised faces. Cognition and Emotion,
29(1), 146-157. https://doi.org/10.1080/02699931.2014.895299
Jiang, Y., Costello, P., Fang, F., Huang, M., & He, S. (2006). A gender-and sexual orientation-dependent spatial attentional effect of invisible images. Proceedings of the National Academy of Sciences,
25 Most, S. B., Smith, S. D., Cooter, A. B., Levy, B. N., & Zald, D. H. (2007). The naked truth: Positive, arousing distractors impair rapid target perception. Cognition and Emotion, 21(5), 964-981. https://doi.org/10.1080/02699930600959340
Nummenmaa, L., Hietanen, J. K., Santtila, P., & Hyönä, J. (2012). Gender and visibility of sexual cues influence eye movements while viewing faces and bodies. Archives of Sexual Behavior, 41(6), 1439-1451. https://doi.org/10.1007/s10508-012-9911-0
Patton, J. H., & Stanford, M. S. (1995). Factor structure of the Barratt impulsiveness scale. Journal of Clinical
Psychology, 51(6), 768-774. https://doi.org/10.1002/1097-4679(199511)51:6<768::AID-JCLP2270510607>3.0.CO;2-1
Presentation 19.0 [computer software]. (2017). Berkeley, CA. Neurobehavioral Systems, Inc.
Prause, N., Janssen, E., & Hetrick, W. P. (2008). Attention and emotional responses to sexual stimuli and their relationship to sexual desire. Archives of Sexual Behavior, 37(6), 934-949. https://doi.org/10.1007/s10508-007-9236-6
Reinert, D. F., & Allen, J. P. (2002). The alcohol use disorders identification test (AUDIT): a review of recent research. Alcoholism: Clinical and Experimental Research, 26(2), 272-279. https://doi.org/10.1111/j.1530- 0277.2002.tb02534.x
Schoenmakers, T. M., de Bruin, M., Lux, I. F., Goertz, A. G., Van Kerkhof, D. H., & Wiers, R. W. (2010). Clinical effectiveness of attentional bias modification training in abstinent alcoholic patients.
Drug and Alcohol Dependence, 109(1), 30-36. https://doi.org/10.1016/j.drugalcdep.2009.11.022
Schoenmakers, T., Wiers, R. W., Jones, B. T., Bruce, G., & Jansen, A. (2007). Attentional re‐training decreases attentional bias in heavy drinkers without generalization. Addiction, 102(3), 399-405. https://doi.org/10.1111/j.1360-0443.2006.01718.x
Sharma, D., Albery, I. P., & Cook, C. (2001). Selective attentional bias to alcohol related stimuli in problem drinkers and non‐problem drinkers. Addiction, 96(2), 285-295. https://doi.org/10.1046/j.1360- 0443.2001.96228512.x
Shelton, J., & Kumar, G. P. (2010). Comparison between auditory and visual simple reaction times.
Neuroscience and Medicine, 1(1), 30. https://doi.org/10.4236/nm.2010.11004
Stacy, A. W., & Wiers, R. W. (2010). Implicit cognition and addiction: a tool for explaining paradoxical behavior. Annual Review of Clinical Psychology, 6, 551-575. https://doi.org/10.1146/annurev.clinpsy.121208.131444
26 Stanford, M. S., Mathias, C. W., Dougherty, D. M., Lake, S. L., Anderson, N. E., & Patton, J. H. (2009). Fifty years of the Barratt Impulsiveness Scale: An update and review. Personality and Individual
Differences, 47(5), 385-395. https://doi.org/10.1016/j.paid.2009.04.008
Theeuwes, J. (1992). Perceptual selectivity for color and form. Attention, Perception, & Psychophysics,
51(6), 599-606. https://doi.org/10.3758/BF03211656
Thush, C., Wiers, R. W., Ames, S. L., Grenard, J. L., Sussman, S., & Stacy, A. W. (2008). Interaction between implicit and explicit cognition and working memory capacity in the prediction of alcohol use in at-risk adolescents. Drug and Alcohol Dependence, 94(1), 116-124. https://doi.org/10.1016/j.drugalcdep.2007.10.019
van Strien, T., Herman, C. P., Engels, R. C., Larsen, J. K., & van Leeuwe, J. F. (2007). Construct validation of the Restraint Scale in normal-weight and overweight females. Appetite, 49(1), 109- 121. https://doi.org/10.1016/j.appet.2007.01.003
White, H. R., & Labouvie, E. W. (1989). Towards the assessment of adolescent problem drinking. Journal of
Studies on Alcohol, 50(1), 30-37. https://doi.org/10.15288/jsa.1989.50.30
White, H. R., & Labouvie, E. W. (2000). Longitudinal trends in problem drinking as measured by the Rutgers Alcohol Problem Index. Alcoholism: Clinical and Experimental Research, 24(5 Suppl 1).
Wiers, R. W., Bartholow, B. D., van den Wildenberg, E., Thush, C., Engels, R. C., Sher, K. J., & Stacy, A. W. (2007). Automatic and controlled processes and the development of addictive behaviors in adolescents: a review and a model. Pharmacology Biochemistry and Behavior, 86(2), 263-283. https://doi.org/10.1016/j.pbb.2006.09.021
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(4), 490-497. https://doi.org/10.1177/0956797611400615 Wiers, R. W., Houben, K., Fadardi, J. S., Van Beek, P., Rhemtulla, M., & Cox, W. M. (2015). Alcohol cognitive
bias modification training for problem drinkers over the web. Addictive Behaviors, 40, 21-26. https://doi.org/10.1016/j.addbeh.2014.08.010
Wright Jr, L. W., & Adams, H. E. (1999). The effects of stimuli that vary in erotic content on cognitive processes. Journal of Sex Research, 36(2), 145-151. https://doi.org/10.1080/00224499909551979 Yantis, S., & Jonides, J. (1984). Abrupt visual onsets and selective attention: evidence from visual search.
Journal of Experimental Psychology: Human Perception and Performance, 10(5), 601-621.
