Przestrzenna antycypacyjna tendencyjność uwagi dotycząca alkoholu. Wstępne doniesienie na temat rzetelności pomiaru i związku z ryzykownym piciem

■ INTRODUCTION

The dot-probe task [1, 2] is often used to measure spatial attentional biases. In trials of this task, first two cues are shown from two different categories, such as alcohol versus soft drink, and subsequently probe stimuli are presented at the cue locations. Differences in responses to the probes that depend on cue category at their location suggest that the cues caused a bias in processing; e.g., if one of the cue types tends to capture attention, then responses to probes at its location should be faster. Alcohol-related stimulus have been found to affect spatial attention [3-6]. In individuals reporting heavy social drinking, attention appears to be drawn towards alcohol cues [4, 5], as would be expected given theory on incentive salience [7, 8]. However, complex, time-dependent patterns have been found in dependence, with a shift from initial orienting to attentional disengagement [6, 9, 10].
Further, a problem with the dot-probe task is that it has been found to have low reliability in a number of studies [11-19]. This would seem to pose a serious threat to at least some forms of attentional bias research using behavioural measures. If the current literature is taken to imply that any behavioural measure of attentional bias is unacceptably likely to be noisy, this could lead to a shift to interesting but expensive and less widely accessible psychophysiological methods. It is essential to ask whether this shift is truly necessary and whether it is justified to consider behavioural measures as inherently problematic [20].
One alternative approach is to explore novel versions of the dot-probe task: perhaps a reliable behavioural measure is possible but not yet known. One such task variant was suggested by the R3 model – the Reprocessing and Reinforcement model of Reflectivity [21, 22]. In this model, automatic attentional biases can occur due to the predicted outcomes of attentional shifts (note that “prediction” here is conceived of as a low-level, underlying process caused by prior reinforcement learning rather than an effortful attempt to make a prediction about the future). For example, if an individual believes that something scary is lurking behind a door, the anticipation of what could happen on opening the door will affect their attention towards it. Or, if there might be something tasty out of sight in a cupboard, the predicted outcome of finding a treat might involuntarily affect attention. It thus seems that attentional biases could occur in response to information on what is likely to happen if attention is directed to one location or another: will something attractive or aversive appear at that location? And could that kind of prediction of outcome automatically affect attentional shifts? A cued version of the dot-probe task, termed the cued Visual Probe Task, cVPT [23], was developed to measure such anticipatory or outcome-related effects. In this task, instead of presenting emotional cues intended to evoke an automatic stimulus-driven response, two visually neutral predictive cues are presented at the start of each trial. On one half of the trials, the predictive cues are replaced by affective stimuli, one cue always being replaced by a stimulus from one category (e.g., alcohol or threat) and the other by a stimulus from another category (e.g., water or safe). On the other half of the trials, probe stimuli are presented requiring a response; on these trials, no affective stimuli are presented at all. Thus, performance on probes is never influenced by the direct presentation of an emotional stimulus, only by the location of visually neutral cues predicting stimulus categories. Further, it has been found that performing a training version of the cVPT induces an attentional bias to stimuli belonging to the trained predicted categories [24]. This supports the interpretation of effects on the cVPT being due to anticipatory processes. It could be expected that performing the cVPT only involves the visually neutral cues acquiring salience, rather than outcome-related processes. However, this would not be expected to lead to the predicted stimulus categories acquiring a bias following the cVPT-training, rather than just the predictive cues.
In the context of threat-related attentional bias, reliability was found to be improved in the cVPT relative to a usual VPT in which emotional cues were presented before probes [25]. The improved reliability was suggested to be due to the removal or mitigation of noisy influences that could play a role when actually presenting emotional stimuli, such as varying responses to particular exemplars, or potentially complex patterns of multiple cognitive responses to actually presented emotional stimuli. While the anticipatory attentional bias was previously used to study alcohol-related attentional bias and bias variability [23, 26], its reliability has not yet been evaluated for alcohol-related stimuli. Further, the validity of the bias as a reflection of processes related to alcohol and risky drinking requires additional support.
The current study therefore aimed first, to determine the reliability of the anticipatory attentional bias for alcohol and, second, to explore correlations between the anticipatory bias and alcohol-related individual differences.

■ MATERIAL AND METHODS

Participants

The experiment was completed online by an analytical sample of 83 healthy adult participants (75 female, 8 male; age 19.7, SD = 2.95). A further 15 participants performed the experiment but were excluded due to inadequate performance (mean accuracy below 0.90) indicating, given the simplicity of the tasks and usual accuracy levels, that these participants were not performing the task as required. Participants gave informed consent, and the study was approved by the institutional ethics committee.

Questionnaires

The following questionnaires were used to assess individual differences related to alcohol use. The 3-item Alcohol Use Disorders Identification Test – Consumption, AUDIT-C was used to measure hazardous drinking [27-30]. The Drinking Motives Questionnaire Revised, DMQ-R [31] provided measures of Enhancement, Social, Coping, and Conformity motives. The Reasons for Abstaining or Limiting Drinking questionnaire, RALD [32, 33] was used to measure motives to refrain from drinking: Loss of Control, Adverse Consequences, and Convictions. Finally, aspects of craving for alcohol were measured with the Alcohol Craving Questionnaire – Short Form, ACQ [34, 35]: Compulsivity, Expectancy, Purposefulness and Emotionality.

Tasks

The cued Visual Probe Task consisted of 9 blocks of 40 trials, preceded by a training phase of 8 blocks of 40 trials. Trials were divided into Picture and Probe trials.
Probe trials started with the presentation of a central fixation cross for 300, 400 or 500 ms, followed by the presentation of two visually neutral cues: the symbols O O O O O and | | | | |, presented in yellow versus blue. The cues were onscreen for a Cue-Stimulus Interval of 400 ms. The cues were located on alternating diagonals per trial: either on the top-left and bottom-right of the screen, or on the bottom-left and top-right of the screen. Following the cue period, a probe stimulus, >><<, was presented at one of the cue locations, and a distractor stimulus, // or //, at the other location. The probe stimulus was presented for 1000 ms, or until a response was given if faster than 1000 ms. The task was to quickly and accurately press a key (R, F, I or J) corresponding to the probe location. Errors were followed by a red –1 for incorrect responses, and a red “Too late!” if no response was given, for 200 ms.
On Picture trials, the cues were replaced by pictures presented at the cue locations. One of the cues was always replaced by an alcoholic stimulus (a colour picture of a glass or bottle of an alcoholic beverage) centred on the cue location. The other cue was always replaced by a non-alcoholic stimulus (a colour picture of a non-alcoholic beverage). The mapping of cues to stimulus category was randomized over subjects. The pictures remained onscreen for 1000 ms, followed by 200 ms of empty screen.

Procedure

Participants performed the experiment fully online. They first filled in the questionnaires and subsequently completed the training and assessment phase of the task. For the sake of transparency, we note that participants performed additional tasks and sessions unrelated to the current study.

Preprocessing and statistical analyses

Trials that were likely not to reflect normal task performance were removed and included the first four task trials, trials following an error and the first trial of each block. Of the remaining probe trials, the median reaction time was calculated for the Probe-on-Alcohol and Probe-on-Non-alcohol predictive cue locations. The reliability of the bias, i.e., the median RT for the Alcohol minus Non-alcohol locations, was tested using Spearman-Brown formula for the split-half reliability of the task, which was divided in sets of even and odd numbered blocks. The effect of probe location was tested using paired-sample t-tests, and correlations were tested between questionnaire scores and the individual bias scores. Based on previous results involving the same Cue-Stimulus Interval of 400 ms [26], we could hypothesize an increasing bias towards alcohol (i.e., more negative bias scores) with increasing scores on the AUDIT, ACQ-Compulsivity, ACQ-Expectancy, and RALD-Convictions and therefore used one-sided tests (p < 0.1 criterion) for these scales.
Data and software will be made available upon request.

■ RESULTS

The mean sample scores on the questionnaires are given in Table I. Using cut-offs for risky drinking in a student population [36], the percentage of female participants with an AUDIT-C score of at least 5 was 35%, and the percentage of male participants with an AUDIT-C score of at least 7 was 25%.
The reliability of bias scores was 0.84. Removal of extreme points (z-score > 3 on either the “even” or “odd” bias, n = 2) resulted in a reliability of 0.74.
There was no overall probe location effect, with reaction times of 486 ms (SD = 48 ms) for probes on the non-alcohol location and 489 ms (SD = 44 ms) for probes on the alcohol location. Of the a priori expected one-sided relationships with questionnaires, only higher AUDIT scores were associated with a bias towards alcohol, r = –0.22, p = 0.050. There were no other significant correlations.

■ DISCUSSION

The current results agree with previous studies in two main ways. First, the anticipatory attentional bias for alcohol revealed good split-half reliability. This is of interest in relation to the concern with reliability for the usual dot-probe task [18]. The current task’s reliance on only anticipatory attentional processes may aid reliability. Any interfering processes evoked by actual stimulus presentation do not occur on probe trials, and there is no variability due to the immediate presentation of different specific exemplars (although there could of course still be more complex history effects related to the particular sequence of presented stimuli). Any bias must be due to the learned relationship between the simple, non-varying cues and the overall stimulus categories. The diagonalized form of the task may also contribute as neither responses nor stimulus locations were ever immediately repeated, reducing potential trial-to-trial carryover effects.
The availability of a form of spatial attentional bias with good reliability would be of interest when studying correlations between attentional bias and other individual differences, or when the bias is used as a dependent variable in a training study. Further, this finding indicates that the anticipatory bias is a consistent individual difference as individuals systematically differ in the degree to which their attention is affected by predicted outcomes of attentional shifts towards or away from cued locations. The current findings may thus be useful for research aimed at a better understanding of the nature of attentional biases and automaticity, and of the influences of task variations on psychometric properties.
Second, supporting the interpretation of individual differences in the bias in terms of attention to predicted outcome categories and hence the validity of the bias, hazardous drinking was indeed associated with an anticipatory bias towards predicted alcohol location, as in a previous study [26]. However, as other predicted effects were not found and the effect size of the current result was not large and the test would not survive correction for multiple testing, future research should focus on replicating the specific correlation between risky drinking and anticipatory attentional bias. The relatively weak correlation between bias and risky drinking may have been due to the limitation of a low overall level of risky drinking in the current sample. Further research is now necessary in larger samples with a wider range of risky drinking, and in clinical groups compared to control groups. Another important limitation of the current opportunistic sample is the unequal distribution of male and female participants. Future work should consider moderating effects of gender on the relationships between anticipatory bias and alcohol-related individual differences.
Another limitation is the use of only a single Cue-Stimulus Interval of 400 ms. While this was selected based on previous work, it must be acknowledged that the current results are specific to this interval, and longer or shorter durations could well produce different reliabilities and effects due to the role of temporal dynamics in attentional biases [6, 9, 10].

■ CONCLUSIONS

The current results provide further support for the concept and measurability of an anticipatory attentional bias for alcohol. Risky drinking may be related to enhanced salience of predicted alcohol stimuli. More generally, biases can be due to the selection of cognitive responses, such as attentional shifts, based on their predicted outcome; e.g., whether they are likely to focus attention on upcoming salient stimuli. The results also add to the evidence suggesting that behavioural measures of attentional bias can, in principle, be both reliable and valid, but it may be necessary to use novel variations of tasks and concepts.

Conflict of interest/Konflikt interesów

None declared./Nie występuje.

Financial support/Finansowanie

None declared./Nie zadeklarowano.

Ethics/Etyka

The work described in this article has been carried out in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki) on medical research involving human subjects, EU Directive (210/63/EU) on protection of animals used for scientific purposes, Uniform Requirements for manuscripts submitted to biomedical journals and the ethical principles defined in the Farmington Consensus of 1997.
Treści przedstawione w pracy są zgodne z zasadami Deklaracji Helsińskiej odnoszącymi się do badań z udziałem ludzi, dyrektywami UE dotyczącymi ochrony zwierząt używanych do celów naukowych, ujednoliconymi wymaganiami dla czasopism biomedycznych oraz z zasadami etycznymi określonymi w Porozumieniu z Farmington w 1997 roku.

References/Piśmiennictwo

1. Mogg K, Bradley BP. Orienting of Attention to Threatening Facial Expressions Presented under Conditions of Restricted Awareness. Cogn Emot 1999; 13(6): 713-40.
2. MacLeod CM, Mathews A, Tata P. Attentional bias in emotional disorders. J Abnorm Psychol 1986; 95(1): 15-20.
3. Field M, Cox WM. Attentional bias in addictive behaviors: a review of its development, causes, and consequences. Drug Alcohol Depend 2008; 97(1-2): 1-20.
4. Field M, Mogg K, Zetteler J, Bradley BP. Attentional biases for alcohol cues in heavy and light social drinkers: the roles of initial orienting and maintained attention. Psychopharmacology (Berl) 2004; 176(1): 88-93.
5. Townshend JM, Duka T. Attentional bias associated with alcohol cues: differences between heavy and occasional social drinkers. Psychopharmacology (Berl) 2001; 157(1): 67-74.
6. Townshend JM, Duka T. Avoidance of alcohol-related stimuli in alcohol-dependent inpatients. Alcohol Clin Exp Res 2007; 31(8): 1349-57.
7. Volkow ND, Koob GF, McLellan AT. Neurobiologic Advances from the Brain Disease Model of Addiction. N Engl J Med 2016; 374(4): 363-71.
8. Robinson TE, Berridge KC. Addiction. Annu Rev Psychol 2003; 54: 25-53.
9. Noël X, Colmant M, Van Der Linden M, Bechara A, Bullens Q, Hanak C, et al. Time course of attention for alcohol cues in abstinent alcoholic patients: the role of initial orienting. Alcohol Clin Exp Res 2006; 30(11): 1871-7.
10. Vollstädt-Klein S, Loeber S, von der Goltz C, Mann K, Kiefer F. Avoidance of alcohol-related stimuli increases during the early stage of abstinence in alcohol-dependent patients. Alcohol Alcohol 2009; 44(5): 458-63.
11. Brown HM, Eley TC, Broeren S, MacLeod CM, Rinck M, Hadwin JA, et al. Psychometric properties of reaction time based experimental paradigms measuring anxiety-related information-processing biases in children. J Anxiety Disord 2014; 28(1): 97-107.
12. Dear BF, Sharpe L, Nicholas MK, Refshauge K. The psychometric properties of the dot--probe paradigm when used in pain-related attentional bias research. J Pain 2011; 12(12): 1247-54.
13. Puls S, Rothermund K. Attending to emotional expressions: no evidence for automatic capture in the dot-probe task. Cogn Emot 2018; 32: 450-63.
14. Schmukle SC. Unreliability of the dot probe task. Eur J Pers 2005; 19(7): 595-605.
15. Waechter S, Nelson AL, Wright C, Hyatt A, Oakman J. Measuring attentional bias to threat: Reliability of dot probe and eye movement indices. Cognit Ther Res 2014; 38(3): 313-33.
16. Kappenman ES, Farrens JL, Luck SJ, Proudfit GH. Behavioral and ERP measures of attentional bias to threat in the dot-probe task: Poor reliability and lack of correlation with anxiety. Front Psychol 2014; 5: 1368.
17. McNally RJ. Attentional bias for threat: Crisis or opportunity? Clin Psychol Rev 2019; 69: 4-13.
18. Christiansen P, Schoenmakers TM, Field M. Less than meets the eye: Reappraising the clinical relevance of attentional bias in addiction. Addict Behav 2015; 44: 43-50.
19. Ataya AF, Adams S, Mullings E, Cooper RM, Attwood AS, Munafò MR. Internal reliability of measures of substance-related cognitive bias. Drug Alcohol Depend 2012; 121(1-2): 148-51.
20. MacLeod CM, Grafton B, Notebaert L. Anxiety-Linked Attentional Bias: Is It Reliable? Annu Rev Clin Psychol 2019. DOI: 10.1146/annurev-clinpsy-050718-095505 [In press].
21. Gladwin TE, Figner B, Crone EA, Wiers RW. Addiction, adolescence, and the integration of control and motivation. Dev Cogn Neurosci 2011; 1(4): 364-76.
22. Gladwin TE, Figner B. “Hot” cognition and dual systems: Introduction, criticisms, and ways forward. In: Wilhelms E, Reyna VF (eds.). Frontiers of Cognitive Psychology. Series: Neuroeconomics, Judgment and Decision Making. New York: Psychology Press; 2014, p. 157-80.
23. Gladwin TE. Attentional bias variability and cued attentional bias for alcohol stimuli. Addict Res Theory 2016; 25(1): 32-8.
24. Gladwin TE, Möbius M, Becker ES. Predictive Attentional Bias Modification induces stimulus-evoked attentional bias for threat. Eur J Psychol 2019 [In press].
25. Gladwin TE, Möbius M, Mcloughlin S, Tyndall I. Anticipatory versus reactive spatial attentional bias to threat. Br J Psychol 2019; 110(1): 3-14.
26. Gladwin TE, Vink M. Alcohol-related attentional bias variability and conflicting automatic associations. J Exp Psychopathol 2018; 9(2): 27.
27. Bush K, Kivlahan DR, McDonell MB, Fihn SD, Bradley KA. The AUDIT alcohol consumption questions (AUDIT-C): an effective brief screening test for problem drinking. Ambulatory Care Quality Improvement Project (ACQUIP). Alcohol Use Disorders Identification Test. Arch Intern Med 1998; 158(16): 1789-95.
28. Bradley KA, DeBenedetti AF, Volk RJ, Williams EC, Frank D, Kivlahan DR. AUDIT-C as a brief screen for alcohol misuse in primary care. Alcohol Clin Exp Res 2007; 31(7): 1208-17.
29. Gordon AJ, Maisto SA, McNeil M, Kraemer KL, Conigliaro RL, Kelley ME, et al. Three questions can detect hazardous drinkers. J Fam Pract 2001; 50(4): 313-20.
30. Gual A, Segura L, Contel M, Heather N, Colom J. Audit-3 and audit-4: effectiveness of two short forms of the alcohol use disorders identification test. Alcohol Alcohol 2002; 37(6): 591-6.
31. Cooper ML. Motivations for alcohol use among adolescents: Development and validation of a four-factor model. Psychol Assess 1994; 6(2): 117-28.
32. Epler AJ, Sher KJ, Piasecki TM. Reasons for abstaining or limiting drinking: a developmental perspective. Psychol Addict Behav 2009; 23(3): 428-42.
33. Anderson KG, Grunwald I, Bekman N, Brown SA, Grant A. To drink or not to drink: Motives and expectancies for use and nonuse in adolescence. Addict Behav 2011; 36(10): 972-9.
34. Singleton E, Henningfield J, Heishman T, Douglas E, Tiffany S. Multidimensional aspects of craving for alcohol. In: Proceedings of the 57th Annual Meeting of the College on Problems of Drug Dependence, 1995.
35. Connolly KM, Coffey SF, Baschnagel JS, Drobes DJ, Saladin ME. Evaluation of the Alcohol Craving Questionnaire-Now factor structures: application of a cue reactivity paradigm. Drug Alcohol Depend 2009; 103(1-2): 84-91.
36. DeMartini KS, Carey KB. Optimizing the use of the AUDIT for alcohol screening in college students. Psychol Assess 2012; 24(4): 954-63.