Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Apr 6;50(4):e70293. doi: 10.1111/acer.70293

Impact of Visual Masking on Attentional Bias in Cross‐Sectional Clinical and Healthy Populations: Differences in Conscious and Unconscious Processing of Alcohol‐Related Cues

Isabella Fuchs‐Leitner 1,2,✉, Kurosch Yazdi‐Zorn 1,2, Jan Rosenleitner 1,2
PMCID: PMC13053913  PMID: 41943269

ABSTRACT

Background

Attentional bias (AB) toward substance‐related cues is a core cognitive mechanism in alcohol use disorder (AUD) and closely associated with craving and relapse risk. Yet, the extent to which AB reflects automatic bottom‐up processes remains unclear.

Methods

This study (N = 107) examined alcohol‐related AB in clinical populations (inpatients and outpatients with AUD) and healthy controls (occasional and non‐drinkers). Alcohol cues were presented under masked and unmasked conditions in a dot‐probe task, and AB was measured via reaction times. Patients also reported alcohol craving and dependence levels. The study assessed whether AB patterns differed between clinical and healthy groups, whether cue awareness modulated attentional processing, and whether alcohol‐type preference within the clinical sample influenced AB. Finally, associations between craving, dependence, and AB were explored.

Results

As predicted, unmasked alcohol cues elicited significant AB in both inpatient and outpatient groups, but not in healthy controls. No significant AB emerged in masked trials for any subgroup, suggesting that alcohol‐related AB is not fully automatic but depends on conscious awareness. Patients reporting current craving showed greater AB than those without craving. However, AB magnitude did not correlate with craving intensity, AUD severity scores, or alcohol‐type preference.

Conclusions

These findings indicate that alcohol‐related AB in AUD is shaped by conscious processing rather than purely automatic mechanisms, underscoring the role of top‐down factors in attentional processes relevant to addiction.

Keywords: alcohol use disorder, attentional bias, conscious processing, dot‐probe task, masking

This study examined how individuals with alcohol use disorder and healthy controls respond to alcohol‐related cues, using masked and unmasked presentations to explore attentional biases and the role of cue awareness in cognitive processes related to craving and addiction.

graphic file with name ACER-50-0-g001.jpg

1. Introduction

1.1. Selective Visual Attention

The processes and time course involved in selective visual attention have been intensely studied for decades, describing the selectivity as a competition between bottom‐up versus top‐down control (e.g., Corbetta and Shulman 2002; Desimone and Duncan 1995; Egeth and Yantis 1997). In this view, bottom‐up attention is stimulus‐driven, rapidly capturing attention through salient features, such as abrupt onsets, high contrast, or unique visual characteristics (Theeuwes 2010; Yantis and Jonides 1984). Top‐down attention is goal‐directed, guided by task demands, expectations, and motivational states, biasing perception toward behaviorally relevant information (Folk et al. 1992; Folk and Remington 2006). This traditional concept of two distinct mechanisms has been extended to more integrated frameworks, in which past selection history, current goals and physical salience were shaped into common unified priority maps or unified systems (Awh et al. 2012). Saghravanian (2025) proposes that attentional selection emerges from a unified system in which bottom‐up and top‐down influences reflect the dynamic weighting of priorities shaped by evolution, experience and current goals.

The guidance of selective visual attention is also debated within research on unconscious processing, highlighting its relevance for understanding how attention and awareness interact (Prasad and Mishra 2019). Our previous work (Fuchs et al. 2013; Fuchs and Ansorge 2012) demonstrates that abrupt‐onset visual cues can capture attention rapidly in a bottom‐up manner even without conscious awareness, but this effect is influenced by task relevance, highlighting the interaction between stimulus‐driven and top‐down processes. Masking procedures typically involve very short stimulus presentations, a visual mask, or a combination of both. Masking conditions disrupt recurrent cortical processing and prevent conscious perception (Lamme and Roelfsema 2000), allowing researchers to test attentional guidance independently of awareness. This approach motivated the current study, which examines alcohol‐related attentional bias (AB) in clinical and control populations under masked and unmasked conditions to assess the role of conscious awareness and automaticity.

1.2. Implications for Substance‐Related Attentional Bias

Attentional bias toward substance‐related cues is often described as the preferential allocation of attention to drug‐related compared with neutral stimuli. AB has been proposed as a key cognitive mechanism underlying craving, maintenance of use, and relapse in addictive disorders (Dal Lago et al. 2024; Field and Cox 2008). Within the general framework of visual attention, AB can be conceptualized as the selective orienting toward cues influenced by both stimulus salience (bottom‐up) and higher‐order goals (top‐down). Incentive sensitization theory argues that addiction‐related stimuli become highly salient and automatically attract attention, even when processed unconsciously (Berridge and Robinson 2016; Robinson and Berridge 2008), whereas Tiffany's (1990) automated action schema model similarly proposes that addiction‐related cues can capture attentional resources outside of awareness. Empirically, however, findings remain mixed: preconscious AB has been observed in some contexts, but not consistently across substances (e.g., smoking, Bradley et al. 2004). Although substance‐related AB is often described as an automatic, bottom‐up process, if this were strictly the case, alcohol cues should capture attention independently of masking conditions, even without conscious awareness.

1.3. Methodological Considerations in AB Research

AB has been demonstrated across paradigms and populations, including visual probe (dot‐probe) tasks, Stroop tasks, visual search paradigms, and eye‐tracking measures (Pennington et al. 2019; Roy‐Charland et al. 2017; Simon et al. 2022). In the dot‐probe task, alcohol and neutral stimuli are presented and a probe replaces one of the images; faster responses to probes replacing alcohol stimuli indicate AB toward alcohol cues. Early work showed that heavy social drinkers display greater AB than occasional social drinkers, supporting the notion that alcohol cues capture attention preferentially in higher‐consuming individuals (Townshend and Duka 2001). This suggests that individuals with alcohol use disorder (AUD), who typically consume alcohol at higher levels, may be particularly relevant for studying ABs toward alcohol cues.

A key methodological factor in AB research is stimulus exposure duration. In theory, short cue‐target‐intervals (i.e., stimulus onset asynchronies = SOAs) are thought to reflect early, automatic orienting, whereas longer SOAs demonstrate maintained attention or difficulty disengaging from cues. Therefore, SOA manipulations often reveal different AB patterns: initial orienting to alcohol‐related cues at short SOAs can be followed by avoidance or disengagement at longer SOAs, particularly in clinical populations, indicating an approach–avoidance or ambivalence pattern in treatment‐seeking individuals (Bollen et al. 2021; Vollstadt‐Klein et al. 2009).

Closely related is the distinction between conscious and unconscious processing, using clearly visible unmasked and masked cue presentations, respectively. Masked presentations probe automatic processing and assess whether AB emerges without conscious awareness. Findings from masked paradigms are mixed: some studies report early orienting to alcohol cues in heavy drinkers or detoxified patients, whereas others find attentional avoidance or null effects under brief exposures (Baker et al. 2014; Monem and Fillmore 2019). Experimental manipulations of implicit motivational states (e.g., subliminal primes) have not consistently altered AB, suggesting a complex relationship between preconscious cue processing and overt attentional indices (Baker et al. 2014; Monem and Fillmore 2019; Roberts and Fillmore 2015). Furthermore, cognitive bias measures derived from different tasks and awareness levels (masked/unmasked Stroop and visual probe tasks) were found to be uncorrelated, indicating different underlying processes (Mogg and Bradley 2002). Theoretical and empirical inconsistencies highlight the need to clarify whether substance‐related ABs emerge below and/or above the threshold of awareness.

1.4. Substance‐Related AB in Clinical Populations

Literature dedicated to AB in clinical samples with (severe) AUD is heterogeneous. Investigations of detoxified inpatients or outpatients sometimes reveal heightened AB relative to controls, whereas other studies show no difference or even avoidance of alcohol cues. These inconsistencies may depend on abstinence duration, chronicity of dependence, therapeutic context, and SOA employed (Bollen et al. 2021, 2022; Loeber et al. 2009). For example, the above‐mentioned approach–avoidance pattern might be moderated by treatment stage, current craving or relapse risk (Bollen et al. 2022; Vollstadt‐Klein et al. 2009). Additionally, heterogeneity may arise from the use of different experimental paradigms (e.g., Stroop, dot‐probe, free viewing) and AB indices, which assess distinct components of attention. Indirect measures based on reaction times, such as the visual probe or Stroop task, have shown limited reliability in AUD samples (Bollen et al. 2022), whereas more direct measures, such as eye‐tracking, provide continuous indices of gaze allocation and dwell time and tend to yield higher reliability (Christiansen et al. 2015). Such variability highlights the need for studies that systematically examine masked and unmasked presentations across clinical subgroups. Importantly, both state and trait components of AB should be further considered, along with related factors, such as craving.

1.5. Alcohol‐Type Preference (Personalized Stimuli)

Including alcohol‐type preference (beer, wine, and liquor) can increase ecological and motivational validity in AB research. Personalized cues are thought to be more salient and meaningful, potentially eliciting stronger AB. Evidence is mixed: Fridrici et al. (2013) found no greater Stroop interference for preferred versus general alcohol words, Christiansen et al. (2015) reported higher internal consistency for visual probe tasks using preferred‐drink images, but Jones et al. (2018) found that even with such procedures, reliability remained weak. Overall, it remains unclear how strongly alcohol‐type preference influences AB.

Including alcohol‐type preference (beer, wine, and liquor) as a step toward individualized stimuli reflects an important effort to increase ecological and motivational validity in AB research. In theory, personalized cues (i.e., stimuli matching a participant's preferred drink) should be more salient and thus more likely to elicit AB, because they are more meaningful to the individual. However, the literature on the effectiveness of individualized alcohol cues is mixed. Fridrici et al. (2013) found that patients suffering from AUD did not show greater Stroop interference for personalized (i.e., preferred) alcohol‐related words compared with general alcohol words, suggesting limited benefit of highly individualized stimuli in that population. On the contrary, some studies suggest that the type of alcoholic beverage can impact attention. For example, Christiansen et al. (2015) demonstrated that the internal consistency of AB was higher when the visual probe task used alcohol images tailored to participants' preferred drink rather than generic alcohol cues. Recording eye‐tracking measures (e.g., gaze dwell time) in addition to reaction times (RTs), increased Cronbach's alpha values for personalized stimuli (α = 0.73 for RT and α = 0.76 for gaze dwell‐time measures). However, it is worth noting that later work by Jones et al. (2018) found that, even when applying recommended procedures, the internal consistency and test–retest reliability of the alcohol visual probe task remained weak. This underlines how challenging it is to obtain reliable AB measures and suggests that whereas personalization improves reliability relative to standard tasks, there is still room for methodological refinement. However, including multiple types of alcohol in a stimulus set might better capture attentional dynamics.

1.6. Synthesis and Implications for the Present Study

The present study investigated alcohol‐related AB in clinical and healthy populations using a visual dot‐probe paradigm. The clinical sample included inpatients and outpatients with AUD, whereas the control sample consisted of non‐drinkers and occasional drinkers. Alcohol cues were presented under masked (32 ms followed by a visual mask) and unmasked (200 ms) conditions, with total cue‐plus‐mask duration equated across conditions. Reaction times indexed AB, and participants also completed measures of craving and alcohol dependence to examine relationships with attentional processing.

The study aimed to determine whether AB differs between clinical and control groups, how cue awareness modulates AB, and whether RT measures are sensitive to these effects. If alcohol cues exert purely automatic, bottom‐up capture, AB should emerge even under masked conditions. Conversely, if AB depends on conscious processing, incentive salience, craving, or other top‐down factors, it should be reduced or absent under masking. This design allows a direct test of automaticity as a bottom‐up mechanism in alcohol‐related AB.

Based on previous research and theoretical frameworks, we expected unmasked alcohol cues to elicit greater AB in both inpatient and outpatient groups relative to healthy controls, reflecting heightened conscious incentive salience. We also anticipated higher AB in patients experiencing current craving compared to those without craving, along with positive correlations with AUD severity scores, underscoring its clinical relevance. Finally, we explored whether AB differed depending on participants' preferred alcohol type, contributing to hypotheses about the role of individualized cues in capturing attention. These predictions shaped a hierarchical analysis plan, including assessments of masking visibility, group‐wise differences on congruency and masking effects, and the influence of craving, alcohol type, and dependence measures within the clinical sample.

2. Materials and Methods

2.1. Participants and Setting

A total of 107 participants took part in the study. The design included two populations: clinical and healthy controls, each divided into two subgroups (see Table 1). Among the clinical sample with AUD (N = 60), 30 inpatients and 30 outpatients undergoing treatment at Kepler University Hospital were included. The healthy control (HC) group (N = 47) was split into non‐drinkers (N = 27) and occasional drinkers (N = 20), based on AUDIT scores (cut‐off of 4) and self‐reported alcohol consumption within the past week (cut‐off: 7 days). Non‐drinkers were recruited through a call targeting individuals who never or only rarely consumed alcohol due to health, religious, or lifestyle reasons. Sample size was determined based on adequate power (80%) to detect medium effects (f = 0.25) at α = 0.05 within each group, consistent with effect sizes typically reported in dot‐probe AB research. Separate within‐group ANOVAs were planned a priori to maximize sensitivity to condition effects by reducing between‐group heterogeneity. Written informed consent was obtained from all participants. The study was approved by the local ethics committee and conducted in accordance with the principles of the Declaration of Helsinki.

TABLE 1.

Demographic data and indices of alcohol dependence for the clinical sample and healthy controls, including all four subgroups.

Population Subgroup N Sex (m/f) Age AUDIT SDS Last alcohol consumption
Clinical sample Inpatients 30 20/10 45.7 27.6 [17–39] 9.6 [3–14] 17.5 [6–50]
Outpatients 30 21/9 45.7 23.8 [10–37] 7.3 [2–13] 85.9 [1–210]
Healthy controls Occasional drinkers 20 13/7 29.9 5.5 [0–10] 0.9 [0–4] 4.7 [1–31]
Non‐drinkers 27 7/20 36.3 1.3 [0–4] 0.1 [0–1] 532.1 [7+]
Open in a new tab

Note: Mean age in years, Sex: number of males (m) and females (f) per subgroup. AUDIT and SDS: Mean total scores for Alcohol Use Disorder Identification Test (AUDIT) and Severity of Dependence Scale (SDS) with range in brackets. Last alcohol consumption in days.

Testing took place individually, and each session lasted approximately 30 min. Participants were seated at a workstation with an adjustable chair, using a chin and forehead rest positioned 65 cm from a 22‐in. monitor (1680 × 1050 px, 60 Hz) and a standard keyboard. Stimulus presentation was controlled using Python with PsychoPy, and responses were registered using two clearly marked keys (“X” for the left index finger and “Comma” for the right). Each session consisted of two parts: first, participants completed questionnaires in an interview format, including sociodemographic information; second, they completed a computerized Visual Probe Task.

2.2. Stimuli

To elicit robust attentional effects while maintaining ecological validity, we controlled for visual saliency and presentation order. Crucially, each stimulus was presented only once to avoid confounding memory effects that could influence attentional capture. Stimuli consisted of custom photographs of individual bottles, cans, and cartons representing familiar Austrian brands, with backgrounds removed and sizes standardized. The final set included 96 image pairs, each showing one alcoholic and one non‐alcoholic beverage on a neutral white background with a 7‐cm horizontal gap and six additional pairs for practice trials. Pairs were evenly divided across three categories (beer, wine, spirits; 32 pairs each), with alcoholic beverages paired with varying non‐alcoholic counterparts. Each image measured 308 × 461 pixels (8.66 × 13 cm), and the presentation order was randomized for each participant, with left–right positions fully counterbalanced.

Visual saliency was controlled using the computational model of Itti et al. (1998) implemented in MATLAB (Saliency Toolbox by Walther and Koch 2006). Stimulus pairs were recombined until the most salient point, as predicted by the model, occurred equally often in alcoholic and non‐alcoholic images and was balanced across screen positions. For the masked condition, a visual mask was created to be presented immediately after the subliminal presentation of the stimulus pair, with the aim of disrupting conscious processing. The mask was generated by combining and scrambling the two images of each pair (see Figure 1), while preserving the original color palette of the beverages.

FIGURE 1.

FIGURE 1

Open in a new tab

Stimuli and visual mask. Example of a stimulus pair (left panel), showing one alcoholic and one non‐alcoholic beverage, alongside the corresponding mask (right panel). Alcoholic beverages belonged to the three alcohol types: beer, wine, or liquors. All photographs were taken by the authors. Any visible brand logos or names were included solely for scientific purposes in the study and in the figures.

2.3. Procedure

In the first part, participants completed several questionnaires to assess alcohol use with the Alcohol Use Disorders Identification Test (AUDIT, German Version: Dybek et al. 2006; Saunders et al. 1993), and alcohol dependence, including the Severity of Dependence Scale (SDS, Gossop et al. 1995). Craving was assessed by a Visual Analogue Scale (VAS), on which participants rated their currently perceived craving intensity from 0 (“not at all”) to 100 (“very strong”). These measures provided complementary information on dependence severity, alcohol consumption, and current craving to contextualize behavioral results.

In the second part, a visual probe paradigm was employed to assess AB, with trials presented in either a masked or unmasked condition. In the unmasked condition, stimulus pairs were displayed briefly (200 ms) but clearly, whereas in the masked condition, stimuli were presented subliminally and immediately followed by a visual mask to disrupt conscious perception (see Figure 2). Each trial began with a fixation cross presented for 500 ms. In the unmasked condition, the stimulus pair of beverages was displayed for 200 ms. In the masked condition, the same stimulus pair was presented subliminally for 32 ms and immediately followed by a visual mask for 168 ms, keeping the total stimulus duration equal to 200 ms. Following stimulus presentation, a target probe (dot) appeared on the screen, either on the side of the alcoholic beverage (congruent) or on the non‐alcoholic side (incongruent), and remained visible until the participant responded via button press, and reaction times (RTs) were recorded. Each trial concluded with a blank screen presented for a jittered duration to reduce temporal predictability and prevent anticipatory responses.

FIGURE 2.

FIGURE 2

Open in a new tab

Trial sequence for unmasked (top) and masked (bottom) conditions. Each trial began with a fixation cross (500 ms), followed by a beverage pair (200 ms). In the masked condition, stimuli were presented subliminally (32 ms) and immediately followed by a visual mask (168 ms). A target probe (dot) then appeared either congruently or incongruently with the alcoholic beverage and remained until response. Trials ended with a blank screen of jittered duration.

2.4. Data Preparation

Reaction times (RTs) were first aggregated per participant for each combination of the within‐subject factors Congruency (congruent: target on the same side as the alcohol image; incongruent: target on the opposite side) and Masking (unmasked vs. masked). These aggregated RTs provided the basis for the repeated‐measures design, with Subgroup (inpatients, outpatients, healthy controls and social drinkers) included as the between‐subject factor. Prior to aggregation, trials with incorrect responses were removed (i.e., 0.73% of all trials). RTs that deviated by more than two standard deviations from each participant's mean within each factor combination (i.e., both faster and slower responses) were excluded, resulting in the removal of 4.5% of all correct trials.

Furthermore, AB scores were computed as the difference between incongruent and congruent conditions (AB = RT_incongruent − RT_congruent), separately for masked and unmasked conditions, resulting in one AB value per masking condition per participant. Effect sizes were calculated for t‐tests (Cohen's d) and ANOVA results (η 2 = eta squared), where Bonferroni‐corrected t‐tests were applied for post hoc comparisons.

2.5. Statistical Analysis

A stepwise procedure was conducted, including an assessment of masking visibility, a hierarchical analysis comparing populations and examining AB patterns across subgroups and conditions, and an evaluation in the clinical sample of the effects of alcohol type, craving, and correlations with dependence scores.

  1. Visibility of masking conditions: In a first step, visibility of the masked and unmasked stimuli was evaluated.

  2. Group‐wise comparison based on RT analysis: Given the cross‐sectional design, the sample was divided into clinical participants with AUD (inpatients and outpatients) and healthy controls (occasional and non‐drinkers). Recruiting non‐drinking controls was challenging, and these participants could not be matched to the clinical sample on age or gender. Furthermore, as optimal matching criteria for AB research remain unclear anyway, no matching was attempted, and the populations may differ fundamentally in clinical characteristics, alcohol exposure, and baseline ABs. Pooling all subgroups could inflate Type 1 error and obscure subgroup‐specific effects. To address this, we conducted two separate repeated‐measures ANOVAs with congruency and masking as within‐subject factors to assess conscious (unmasked) and unconscious (masked) alcohol‐related SBs, whereas subgroup served as a between‐subject factor for direct comparisons. This approach maximized sensitivity to detect subgroup‐specific attentional effects while maintaining interpretability and statistical rigor. Significant effects were followed up with Bonferroni‐adjusted pairwise comparisons. Additionally, planned comparisons of congruency effects across all four subgroups and masking conditions were conducted, which are valid even without testing the full interaction structure and supported in the methodological literature (e.g., Ruxton and Beauchamp 2008).

  3. Impact of craving on AB and alcohol dependence correlates: Differences in AB as a function of craving were assessed. Additionally, Pearson correlations were performed to examine associations between AB‐masked and AB‐unmasked scores and alcohol dependence measures, including AUDIT and SDS scores.

  4. Impact of alcohol‐type preference on AB within the clinical sample: For each alcohol type (Beer, Wine, Liquor), patients were categorized according to whether they preferred that drink or not. Exploratory analyses with independent‐samples t‐tests were then conducted to compare AB‐masked and AB‐unmasked scores between preference and non‐preference groups for each alcohol type.

3. Results

3.1. Visibility of Masking Conditions

We aimed for limitation of conscious perception and processing of the visual stimuli in the masked condition, applying very short stimulus presentation times (32 ms) followed by a visual mask. As manipulation check, participants were instructed to determine the position of the alcoholic beverage (left or right) in a separate block following the dot‐probe task. Mean percentage of correct responses were used to determine stimulus visibility. In case of perfect masking and elimination of conscious visual perception, performance at chance level (i.e., 50% correct trials) would be expected.

Overall, correctness scores in both masking conditions were significantly higher than chance level (M masked = 65.4%, M unmasked = 94.6, both ps < 0.001). Importantly, mean correctness scores for masked trials were significantly lower than for unmasked conditions, t(106) = 22.1, p = 0.013, Cohen's d = 2.1. This aligns with subjective reports, in which participants indicated low to no visibility in the masked conditions and described a feeling of “guessing” the correct position. Moreover, because visibility was assessed in a separate block, focusing on this particular task might even have led to an overestimation of cue visibility compared with the dot‐probe task in the first block. In sum, the masked condition appeared to effectively reduce conscious perception of the cues, both objectively and subjectively.

3.2. Population‐Wise Comparisons Based on RT Analysis

3.2.1. Clinical AUD Sample

A 2 (Congruency: congruent vs. incongruent) × 2 (Masking: unmasked vs. masked) repeated‐measures ANOVA with subgroup (inpatients vs. outpatients) as a between‐subject factor on mean RTs revealed the following effects. A significant main effect of Congruency, F(1, 58) = 10.44, p = 0.002, η 2 = 0.153, with generally faster responses in congruent (429 ms) compared with incongruent trials (435 ms). A significant main effect for masking was found, F(1,58) = 7.70, p = 0.007, η 2 = 0.117, reflecting generally faster responses in the masked (429 ms) compared with the unmasked condition (435 ms). We found no significant main effect for Subgroup (p = 0.395).

The interaction between Congruency × Subgroup was marginally significant, F(1, 58) = 3.78, p = 0.057, η 2 = 0.061, indicating that the congruency effect might differ systematically across the two masking conditions. Post hoc Bonferroni‐adjusted t‐tests revealed a significant AB (RT‐incongruent minus RT‐congruent) in the unmasked (AB = 8.9 ms, p = 0.003), but not in the masked condition (AB = 2.2 ms, p = 0.259). All other effects were nonsignificant (all ps > 0.398).

3.2.2. Healthy Controls

The same 2 Congruency × 2 Masking design for a repeated‐measures ANOVA on RTs with subgroup (occasional vs. non‐drinkers) as a between‐subject factor was evaluated. Again, a significant main effect of Masking, F(1, 45) = 17.11, p < 0.001, η 2 = 0.275, with generally faster responses in the masked (361 ms) compared with the unmasked condition (386 ms). A significant main effect for group, F(1, 45) = 4.94, p = 0.031, η 2 = 0.099, reflected generally faster responses by the occasional (351 ms) compared with non‐drinkers (378 ms). All other effects were not significant (for Congruency × Masking p = 0.631, all other ps > 0.153).

Planned exploratory analyses based on our a priori hypotheses regarding congruency effects across masking conditions and all four subgroups was conducted (see Table 2). In the unmasked condition, both of the HC subgroups showed no meaningful congruency effect (occasional drinkers: AB = 2.0 ms, non‐drinkers: AB = −1.1 ms), whereas both clinical subgroups showed an effect: inpatients (AB = 7.6 ms, p = 0.030) and outpatient (AB = 10.2 ms, p = 0.004), indicating slower responses in the incongruent compared with the congruent condition. In the masked condition, no subgroup exhibited any significant congruency effect (all ps > 0.131, see Table 2 and Figure 3). Overall, these results suggest that AB effects were present in both subgroups of the clinical AUD sample with clearly visible alcohol cues in the unmasked condition, whereas masking generally reduced or eliminated congruency effects across all subgroups.

TABLE 2.

Reaction times (RTs) and attentional bias (AB) for the four subgroups, across masking conditions.

Masking Subgroup Mean RT‐incongruent Mean RT‐congruent AB p Cohen's d
Unmasked Inpatient 445.7 438.1 7.6 0.030 0.40
Outpatient 433.2 423.0 10.2 0.004 0.54
Occasional drinkers 354.7 352.7 2.0 0.642 0.47
Non‐drinkers 381.6 384.0 −2.4 0.510 0.13
Masked Inpatient 437.4 436.8 0.6 0.797 0.05
Outpatient 423.1 419.3 3.8 0.131 0.28
Occasional drinkers 348.4 347.7 0.6 0.839 0.05
Non‐drinkers 371.4 374.8 −3.4 0.202 0.25
Open in a new tab

Note: AB (attentional bias) = Mean RT‐incongruent minus Mean RT‐congruent. Two‐tailed p values indicate Bonferroni‐adjusted pairwise comparisons for congruency effects (congruent vs. incongruent).

FIGURE 3.

FIGURE 3

Open in a new tab

Mean attentional bias across masking conditions and subgroups. Depicted are the mean ABs (mean RTs for incongruent minus congruent trials in ms), separately for unmasked (left) and masked conditions (right) for all four subgroups. Clinical populations are depicted with striped bars (inpatients and outpatients) and healthy controls using solid bars (occasional and non‐drinkers). Error bars indicate standard errors of the mean.

To summarize, the clinical and healthy populations exhibited distinct RT patterns. Notably, a Masking × Congruency interaction approached significance in the clinical AUD group, but not in controls. Post hoc and planned comparisons revealed a pattern consistent with theoretical predictions: significant AB emerged in both inpatients and outpatients for clearly visible, unmasked alcohol cues, whereas healthy controls showed no such effects. Under masked conditions, no significant congruency effects were observed in any of the four subgroups.

3.3. Impact of Craving on AB, and Alcohol Dependence Correlates

Only around a third of patients reported current cravings with relatively low intensity ratings (19.2 out of 100 on the VAS). Comparisons of patients based on whether or not they indicated craving (18 craving, 42 non‐craving) revealed a significant difference in the strength of AB in unmasked conditions (M craving = 14.2 ms, M non‐craving = 6.6 ms), t(58) = 1.24, p = 0.042, Cohen's d = 0.35. In masked conditions, no such difference was found (M = 2.1 ms and 2.3 ms for craving and non‐craving respectively, p = 0.343). Overall AB did not differ depending on craving groups (M = 8.0 ms and 4.4 ms for craving and non‐craving, p = 0.644). No significant correlations for the overall AB, AB‐masked, or AB‐unmasked and AUDIT and SDS scores were found for the clinical population (all ps > 0.092).

3.4. Impact of Alcohol Preference on AB Within the Patient Sample

Data were restricted to patients only, and individual ABs were calculated based on mean RTs for incongruent minus congruent trials. Patients indicated whether or not they preferred the three types of alcoholic beverages used as stimuli in this study (beer, wine, and liquor). Independent‐samples t‐tests were conducted to compare AB scores between participants who preferred the beverage and those who did not under unmasked and masked conditions. Results are displayed in Table 3.

TABLE 3.

Attentional Bias (AB) for preference of alcohol type and masking conditions. Results of comparisons between patients with and without preference for the respective alcohol type.

Alcohol type Condition Preference N Mean t p Cohen's d
Beer AB_unmasked No 27 6.5 −0.5 0.65 0.12
Yes 33 10.0
AB_masked No 27 3.4 −0.3 0.77 0.08
Yes 33 5.2
AB_total No 27 5.2 −0.5 0.30 0.13
Yes 33 7.3
Wine AB_unmasked No 38 9.9 −0.9 0.39 0.24
Yes 22 16.2
AB_masked No 38 −0.1 −0.5 0.59 0.15
Yes 22 3.6
AB_total No 38 4.7 −0.8 0.41 0.23
Yes 22 9.6
Liquor AB_unmasked No 44 5.2 −0.3 0.80 0.08
Yes 16 8.2
AB_masked No 44 1.2 0.2 0.86 0.05
Yes 16 −0.1
AB_total No 44 3.4 −0.1 0.91 0.03
Yes 16 4.2
Open in a new tab

Note: AB_unmasked = mean attentional bias in unmasked conditions, AB_masked = mean attentional bias in masked conditions, AB_total = AB mean over both masking conditions.

We found no significant differences between preference and non‐preference groups for any of the alcohol types or masking conditions (all ps > 0.38). After Bonferroni correction for six comparisons (adjusted α = 0.008), none of the effects approached significance.

4. Discussion

The present study provides novel insight into alcohol‐related SB AB in clinical and healthy populations by examining both conscious and unconscious processing using masked and unmasked dot‐probe paradigms. The main finding is that AB emerged strongly in the clinical AUD sample only under clearly visible (unmasked) conditions and was largely absent under masked conditions. This indicates that conscious cue perception contributes substantially to attentional capture in individuals with AUD, challenging the view that AB operates purely automatically. RT measures were sensitive to these effects, and AB was dependent on the presence of current craving, whereas trait‐level alcohol dependence scores (AUDIT, SDS) showed no significant associations.

4.1. Alcohol‐Related AB in the Clinical Populations

Within the clinical sample, no significant differences were observed between inpatient and outpatient subgroups, indicating that the AB is robust across different treatment settings and levels of clinical severity. Outpatients displayed numerically higher ABs in both masking conditions, but the lack of statistical significance warrants caution. However, the trend is in line with the finding that alcohol‐related ABs are reliably observed in subclinical populations, whereas studies involving patients with severe AUD frequently fail to detect such biases or consistent approach–avoidance patterns (Bollen et al. 2022). Comparing the two populations, patients with AUD in general exhibited heightened AB toward alcohol cues relative to healthy controls, but this effect was largely restricted to unmasked trials.

4.2. Conscious Versus Unconscious Processing and Automaticity

A central theoretical implication of our findings is that alcohol‐related AB appears to require conscious processing. The absence of AB in masked trials argues against a purely automatic, bottom‐up account, suggesting instead that attentional capture is contingent on awareness of the substance‐related cues. This is consistent with recent research indicating that AB in AUD is driven more by controlled, later stages of attentional processing than early automatic capture (Bollen et al. 2024; Topbaş et al. 2025). In our study, we found generally faster responses in masked compared with unmasked conditions. One possible explanation for this observation is that the mask onset acted as a brief attentional boost, which accelerates response activation. Research has shown that alerting signals can speed up responses to subliminal primes (Fischer et al. 2013).

4.3. Role of Craving and Clinical Relevance

In their recent review, Bollen et al. (2024) investigated whether AB in severe AUD might reflect either a trait‐like characteristic, being relatively stable and linked to overall disorder severity, or a state‐dependent phenomenon, fluctuating with momentary motivational and affective conditions, such as craving or stress. Evidence suggests that AB is more strongly influenced by transient states than by stable traits. In the current study, we observed that craving indicated AB in the clinical group: patients reporting craving showed stronger bias effects than those without craving in unmasked conditions. However, craving was overall low in our clinical population, and we found no significant correlations between reported strength of craving and AB. More general trait‐level measures of alcohol use severity (AUDIT) and dependence (SDS) did not significantly correlate with AB. Our current findings are therefore fully consistent with and can be incorporated into the proposed state‐versus‐trait framework. (Bollen et al. 2024).

Incorporating this idea into general addiction models, substance‐related AB aligns with dynamic frameworks, which propose that attentional and cognitive biases fluctuate with motivational and affective states. In Incentive Sensitization Theory (IST, Berridge and Robinson 2016), repeated drug use sensitizes the brain's “wanting” system, making drug‐related cues highly motivating even if the actual pleasure (“liking”) from the drug does not increase. This distinction helps explain why attentional and cognitive biases toward these cues can vary depending on current states, such as craving, stress, or withdrawal: the expression of bias is dynamic, whereas the underlying sensitized “wanting” system represents a more stable vulnerability. Furthermore, in clinical populations, cognitive bias modification (CBM) interventions often find that reductions in craving occur alongside reductions in bias, suggesting a close link between these processes (Luehring‐Jones et al. 2017). On a broader level of attentional guidance, our findings also suggest that top‐down factors, such as craving or motivational relevance, can dynamically shape attentional and cognitive biases.

4.4. Influence of Alcohol‐Type Preference

Exploratory analyses of the impact of alcohol type (beer, wine, or liquor) on AB revealed no effect in our study, suggesting a generalized tendency toward alcohol cues rather than category‐specific effects, which aligns with evidence of limited benefits of highly individualized cues in clinical populations (Fridrici et al. 2013). Our design included relatively few trials per alcohol type (32 trials), likely limiting power to detect subtle differences, and AB is generally difficult to measure reliably (Jones et al. 2018). Thus, small beverage‐specific effects may have gone undetected. Nevertheless, including multiple alcohol types remains a valuable exploratory approach, increasing ecological validity and potentially enhancing task consistency, which is important for developing personalized interventions, such as CBM training (Batschelet et al. 2020).

4.5. Clinical Implications and Future Directions

Our results carry clinical and therapeutical implications. Although prior ABM studies in AUD have shown mixed results, a recent quasi‐randomized trial combining ABM with cognitive behavioral therapy (CBT) did not significantly reduce relapse risk or craving compared with CBT alone (Amano et al. 2025), highlighting the need to refine ABM approaches. In the current study, we observed AB under unmasked conditions, indicating the necessity of conscious exposure. For ABM, ensuring active engagement during the tasks might improve therapeutic efficacy. Furthermore, the presence of craving affected the strength of AB. This suggests that current craving states should be considered during interventions.

4.6. Limitations

Overall, considerable effort was invested into the study design, including the creation of our own stimuli to control for salience and memory effects. We carefully planned the analyses, incorporating separate examinations of clinical and control populations to preserve subgroup‐specific effects, and used a cross‐sectional sample to provide a clear snapshot of AB across groups. Nevertheless, several limitations must be noted. First, although masking was used to reduce conscious processing, we did not include subjective or objective measures of awareness per masked trial (e.g., forced‐choice detection or visibility ratings). Without these, we cannot be entirely certain that participants were truly unaware of the stimuli in the masked condition. Future research should incorporate awareness checks to better delineate the boundary between unconscious and conscious processing. Second, whereas we found associations with the presence of craving, our design is cross‐sectional, limiting causal inferences. Longitudinal studies could further shed light on how fluctuations in craving and motivational states dynamically influence AB over time. Third, although alcohol‐type preference did not moderate AB in our sample, our measure of preference may have been too broad. Future studies might examine beverage‐specific bias with more specific assessments (e.g., frequency, craving intensity for specific alcohol types) or neurobiological measures. Fourth, the current dot‐probe design does not allow a formal separation of attentional orienting and disengagement processes, as all trials included alcohol‐neutral pairs and no neutral–neutral baseline. Consequently, the standard ABA measure reflects a combination of facilitated orienting and delayed disengagement rather than independent components. Future studies could include a neutral baseline or alternative paradigms to more precisely assess the temporal dynamics of AB toward alcohol cues. Fifth, outlier exclusion procedures differ across RT studies, which can make comparisons between studies challenging. Simulation work by Berger and Kiefer (2021) directly compared several common outlier exclusion methods and found that no single approach is clearly superior, indicating that there is no gold standard. These considerations highlight the importance of carefully choosing and reporting the method used for trimming extreme responses in reaction‐time analyses. Finally, participants were not matched for age or sex, which could potentially influence AB. Future studies with larger samples may benefit from considering these factors, either through careful matching or targeted designs, to better understand their impact on attentional processes.

5. Conclusion

In conclusion, our study suggests that in clinical AUD populations, AB for alcohol is largely a phenomenon of conscious, rather than unconscious, processing. Masked presentations failed to elicit AB, challenging strict automaticity accounts and emphasizing the role of top‐down, goal‐directed processes in attentional capture. Strength of AB is influenced by the current state of craving, whereas not correlation to traits like dependence severity or beverage preference was found. These findings emphasize the role of motivational states in the development and manifestation of AB, and have practical clinical implications for interventions aimed at reducing AB and associated relapse risk.

Author Contributions

I.F.‐L.: conceptualization, data curation, formal analysis, methodology, project administration, writing – original draft. J.R.: conceptualization, investigation, validation, writing – review and editing. K.Y.‐Z.: resources, writing – review and editing.

Funding

The authors have nothing to report.

Ethics Statement

The study was conducted ethically in accordance with the World Medical Association Declaration of Helsinki. The study protocol was approved by the local ethics committee on human research: JKU MED Ethics Committee, at the Kepler University Hospital Linz (study no.: 1342/2023). Written informed consent was obtained from all individuals to participate in the study.

Conflicts of Interest

The authors declare no conflicts of interest.

Acknowledgments

Many thanks to Nikolas Gerstgrasser and Silvia Lindorfer for their valuable help with recruiting our inpatient sample. We also thank Maximilian Loegler and Silvan Hartl for their support in data acquisition and recruitment of participants in our healthy control group. Some minor language editing and refinement suggestions were provided by ChatGPT, an AI language model (OpenAI 2025). The content, ideas, and analysis presented in this work are solely those of the authors. Open Access funding provided by Johannes Kepler Universitat Linz/KEMÖ.

Data Availability Statement

Data and protocols are not publicly available to ensure special caution in handling patient information, but they can be provided upon reasonable request.

References

  1. Amano, Y. , Koizumi K., Takizawa H., Tasaka S., and Hamaguchi T.. 2025. “Effectiveness of Attentional Bias Modification Combined With Cognitive Behavioral Therapy in Reducing Relapse Risk and Cravings in Male Patients With Alcohol Use Disorder: A Quasi‐Randomized Controlled Trial.” Neuropsychopharmacology Reports 45, no. 1: e70002. 10.1002/npr2.70002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Awh, E. , Belopolsky A. V., and Theeuwes J.. 2012. “Top‐Down Versus Bottom‐Up Attentional Control: A Failed Theoretical Dichotomy.” Trends in Cognitive Sciences 16, no. 8: 437–443. 10.1016/j.tics.2012.06.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Baker, S. , Dickson J. M., and Field M.. 2014. “Implicit Priming of Conflicting Motivational Orientations in Heavy Drinkers.” BMC Psychology 2, no. 1: 28. 10.1186/s40359-014-0028-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Batschelet, H. M. , Stein M., Tschuemperlin R. M., Soravia L. M., and Moggi F.. 2020. “Alcohol‐Specific Computerized Interventions to Alter Cognitive Biases: A Systematic Review of Effects on Experimental Tasks, Drinking Behavior, and Neuronal Activation.” Frontiers in Psychiatry 10: 871. 10.3389/fpsyt.2019.00871. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Berger, A. , and Kiefer M.. 2021. “Comparison of Different Response Time Outlier Exclusion Methods: A Simulation Study.” Frontiers in Psychology 12: 675558. 10.3389/fpsyg.2021.675558. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Berridge, K. C. , and Robinson T. E.. 2016. “Liking, Wanting, and the Incentive‐Sensitization Theory of Addiction.” American Psychologist 71, no. 8: 670–679. 10.1037/amp0000059. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bollen, Z. , Field M., Billaux P., and Maurage P.. 2022. “Attentional Bias in Alcohol Drinkers: A Systematic Review of Its Link With Consumption Variables.” Neuroscience and Biobehavioral Reviews 139: 104,703. 10.1016/j.neubiorev.2022.104703. [DOI] [PubMed] [Google Scholar]
  8. Bollen, Z. , Pabst A., Masson N., et al. 2021. “Alcohol‐Related Attentional Biases in Recently Detoxified Inpatients With Severe Alcohol Use Disorder: An Eye‐Tracking Approach.” Drug and Alcohol Dependence 225: 108,803. 10.1016/j.drugalcdep.2021.108803. [DOI] [PubMed] [Google Scholar]
  9. Bollen, Z. , Pabst A., Masson N., Wiers R. W., Field M., and Maurage P.. 2024. “Craving Modulates Attentional Bias Towards Alcohol in Severe Alcohol Use Disorder: An Eye‐Tracking Study.” Addiction 119, no. 1: 102–112. 10.1111/add.16333. [DOI] [PubMed] [Google Scholar]
  10. Bradley, B. , Field M., Mogg K., and De Houwer J.. 2004. “Attentional and Evaluative Biases for Smoking Cues in Nicotine Dependence: Component Processes of Biases in Visual Orienting.” Behavioural Pharmacology 15, no. 1: 29–36. 10.1097/00008877-200,402,000-00004. [DOI] [PubMed] [Google Scholar]
  11. Christiansen, P. , Mansfield R., Duckworth J., Field M., and Jones A.. 2015. “Internal Reliability of the Alcohol‐Related Visual Probe Task Is Increased by Utilising Personalised Stimuli and Eye‐Tracking.” Drug and Alcohol Dependence 155: 170–174. 10.1016/j.drugalcdep.2015.07.672. [DOI] [PubMed] [Google Scholar]
  12. Corbetta, M. , and Shulman G. L.. 2002. “Control of Goal‐Directed and Stimulus‐Driven Attention in the Brain.” Nature Reviews Neuroscience 3, no. 3: 201–215. 10.1038/nrn755. [DOI] [PubMed] [Google Scholar]
  13. Dal Lago, D. , Burns E., Jackson R. C., and Wilcockson T. D. W.. 2024. “Are Alcohol‐Related Attentional Biases and Holistic Perception Independent Processes?” Experimental and Clinical Psychopharmacology 32, no. 5: 579–587. 10.1037/pha0000727. [DOI] [PubMed] [Google Scholar]
  14. Desimone, R. , and Duncan J.. 1995. “Neural Mechanisms of Selective Visual Attention.” Annual Review of Neuroscience 18, no. 1: 193–222. 10.1146/annurev.ne.18.030195.001205. [DOI] [PubMed] [Google Scholar]
  15. Dybek, I. , Bischof G., Grothues J., et al. 2006. “The Reliability and Validity of the Alcohol Use Disorders Identification Test (AUDIT) in a German General Practice Population Sample.” Journal of Studies on Alcohol 67, no. 3: 473–481. 10.15288/jsa.2006.67.473. [DOI] [PubMed] [Google Scholar]
  16. Egeth, H. E. , and Yantis S.. 1997. “Visual Attention: Control, Representation, and Time Course.” Annual Review of Psychology 48, no. 1: 269–297. 10.1146/annurev.psych.48.1.269. [DOI] [PubMed] [Google Scholar]
  17. Field, M. , and Cox W.. 2008. “Attentional Bias in Addictive Behaviors: A Review of Its Development, Causes, and Consequences.” Drug and Alcohol Dependence 97, no. 1–2: 1–20. 10.1016/j.drugalcdep.2008.03.030. [DOI] [PubMed] [Google Scholar]
  18. Fischer, R. , Plessow F., and Kiesel A.. 2013. “The Effects of Alerting Signals in Masked Priming.” Frontiers in Psychology 4: 448. 10.3389/fpsyg.2013.00448. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Folk, C. L. , and Remington R.. 2006. “Top‐Down Modulation of Preattentive Processing: Testing the Recovery Account of Contingent Capture.” Visual Cognition 14, no. 4–8: 445–465. 10.1080/13506280500193545. [DOI] [Google Scholar]
  20. Folk, C. L. , Remington R. W., and Johnston J. C.. 1992. “Involuntary Covert Orienting Is Contingent on Attentional Control Settings.” Journal of Experimental Psychology. Human Perception and Performance 18, no. 4: 1030–1044. 10.1037/0096-1523.18.4.1030. [DOI] [PubMed] [Google Scholar]
  21. Fridrici, C. , Leichsenring‐Driessen C., Driessen M., Wingenfeld K., Kremer G., and Beblo T.. 2013. “The Individualized Alcohol Stroop Task: No Attentional Bias Toward Personalized Stimuli in Alcohol‐Dependents.” Psychology of Addictive Behaviors 27, no. 1: 62–70. 10.1037/a0029139. [DOI] [PubMed] [Google Scholar]
  22. Fuchs, I. , and Ansorge U.. 2012. “Unconscious Cueing via the Superior Colliculi: Evidence From Searching for Onset and Color Targets.” Brain Sciences 2, no. 1: 33–60. 10.3390/brainsci2010033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Fuchs, I. , Theeuwes J., and Ansorge U.. 2013. “Exogenous Attentional Capture by Subliminal Abrupt‐Onset Cues: Evidence From Contrast‐Polarity Independent Cueing Effects.” Journal of Experimental Psychology. Human Perception and Performance 39, no. 4: 974–988. 10.1037/a0030419. [DOI] [PubMed] [Google Scholar]
  24. Gossop, M. , Darke S., Griffiths P., et al. 1995. “The Severity of Dependence Scale (SDS): Psychometric Properties of the SDS in English and Australian Samples of Heroin, Cocaine and Amphetamine Users.” Addiction 90, no. 5: 607–614. 10.1046/j.1360-0443.1995.9056072.x. [DOI] [PubMed] [Google Scholar]
  25. Itti, L. , Koch C., and Niebur E.. 1998. “A Model of Saliency‐Based Visual Attention for Rapid Scene Analysis.” IEEE Transactions on Pattern Analysis and Machine Intelligence 20, no. 11: 1254–1259. 10.1109/34.730558. [DOI] [Google Scholar]
  26. Jones, A. , Christiansen P., and Field M.. 2018. “Failed Attempts to Improve the Reliability of the Alcohol Visual Probe Task Following Empirical Recommendations.” Psychology of Addictive Behaviors 32, no. 8: 922–932. 10.1037/adb0000414. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Lamme, V. A. F. , and Roelfsema P. R.. 2000. “The Distinct Modes of Vision Offered by Feedforward and Recurrent Processing.” Trends in Neurosciences 23, no. 11: 571–579. 10.1016/S0166-2236(00)01657-X. [DOI] [PubMed] [Google Scholar]
  28. Loeber, S. , Vollstädt‐Klein S., Von Der Goltz C., Flor H., Mann K., and Kiefer F.. 2009. “Attentional Bias in Alcohol‐Dependent Patients: The Role of Chronicity and Executive Functioning.” Addiction Biology 14, no. 2: 194–203. 10.1111/j.1369-1600.2009.00146.x. [DOI] [PubMed] [Google Scholar]
  29. Luehring‐Jones, P. , Louis C., Dennis‐Tiwary T. A., and Erblich J.. 2017. “A Single Session of Attentional Bias Modification Reduces Alcohol Craving and Implicit Measures of Alcohol Bias in Young Adult Drinkers.” Alcoholism: Clinical and Experimental Research 41, no. 12: 2207–2216. 10.1111/acer.13520. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Mogg, K. , and Bradley B. P.. 2002. “Selective Processing of Smoking‐Related Cues in Smokers: Manipulation of Deprivation Level and Comparison of Three Measures of Processing Bias.” Journal of Psychopharmacology 16, no. 4: 385–392. 10.1177/026988110201600416. [DOI] [PubMed] [Google Scholar]
  31. Monem, R. , and Fillmore M. T.. 2019. “Alcohol Administration Reduces Attentional Bias to Alcohol‐Related but Not Food‐Related Cues: Evidence for a Satiety Hypothesis.” Psychology of Addictive Behaviors 33, no. 8: 677–684. 10.1037/adb0000522. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. OpenAI . 2025. “ChatGPT [Large Language Model].” https://chat.openai.com/.
  33. Pennington, C. R. , Qureshi A. W., Monk R. L., Greenwood K., and Heim D.. 2019. “Beer? Over Here! Examining Attentional Bias Towards Alcoholic and Appetitive Stimuli in a Visual Search Eye‐Tracking Task.” Psychopharmacology 236, no. 12: 3465–3476. 10.1007/s00213-019-05313-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Prasad, S. , and Mishra R. K.. 2019. “The Nature of Unconscious Attention to Subliminal Cues.” Vision 3, no. 3: 38. 10.3390/vision3030038. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Roberts, W. , and Fillmore M. T.. 2015. “Attentional Bias to Alcohol‐Related Stimuli as an Indicator of Changes in Motivation to Drink.” Psychology of Addictive Behaviors 29, no. 1: 63–70. 10.1037/adb0000005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Robinson, T. E. , and Berridge K. C.. 2008. “The Incentive Sensitization Theory of Addiction: Some Current Issues.” Philosophical Transactions of the Royal Society B: Biological Sciences 363, no. 1507: 3137–3146. 10.1098/rstb.2008.0093. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Roy‐Charland, A. , Plamondon A., Homeniuk A. S., Flesch C. A., Klein R. M., and Stewart S. H.. 2017. “Attentional Bias Toward Alcohol‐Related Stimuli in Heavy Drinkers: Evidence From Dynamic Eye Movement Recording.” American Journal of Drug and Alcohol Abuse 43, no. 3: 332–340. 10.1080/00952990.2016.1209511. [DOI] [PubMed] [Google Scholar]
  38. Ruxton, G. D. , and Beauchamp G.. 2008. “Time for Some a Priori Thinking About Post Hoc Testing.” Behavioral Ecology 19, no. 3: 690–693. 10.1093/beheco/arn020. [DOI] [Google Scholar]
  39. Saghravanian, S. J. 2025. “Rethinking Attention: A Unified Perspective on Top‐Down and Bottom‐Up Processes.” Acta Psychologica 258: 105244. 10.1016/j.actpsy.2025.105244. [DOI] [PubMed] [Google Scholar]
  40. Saunders, J. B. , Aasland O. G., Babor T. F., de la Puente J. R., and Grant M.. 1993. “Development of the Alcohol Use Disorders Screening Test (AUDIT). WHO Collaborative Project on Early Detection of Persons With Harmful Alcohol Consumption.” Addiction (Abingdon, England) 88: 791–804. [DOI] [PubMed] [Google Scholar]
  41. Simon, J. , Grogna D., Rivard M.‐C., Heck M., Bouchard S., and Quertemont E.. 2022. “Assessing Attentional Bias for Alcohol‐Related Cues Using Eye Tracking in a Virtual Reality Environment.” Frontiers in Virtual Reality 3: 849840. 10.3389/frvir.2022.849840. [DOI] [Google Scholar]
  42. Theeuwes, J. 2010. “Top–Down and Bottom–Up Control of Visual Selection.” Acta Psychologica 135, no. 2: 77–99. 10.1016/j.actpsy.2010.02.006. [DOI] [PubMed] [Google Scholar]
  43. Tiffany, S. T. 1990. “A Cognitive Model of Drug Urges and Drug‐Use Behavior: Role of Automatic and Nonautomatic Processes.” Psychological Review 97, no. 2: 147–168. 10.1037/0033-295X.97.2.147. [DOI] [PubMed] [Google Scholar]
  44. Topbaş, Z. S. , Günday E. A., Şimşek N., and Usta E.. 2025. “Does Attentional Bias Predict Relapse in Addiction? A Systematic Review of Longitudinal Studies.” Brain and Behavior 15, no. 3: e70300. 10.1002/brb3.70300. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Townshend, J. , and Duka T.. 2001. “Attentional Bias Associated With Alcohol Cues: Differences Between Heavy and Occasional Social Drinkers.” Psychopharmacology 157, no. 1: 67–74. 10.1007/s002130100764. [DOI] [PubMed] [Google Scholar]
  46. Vollstadt‐Klein, S. , Loeber S., Von Der Goltz C., Mann K., and Kiefer F.. 2009. “Avoidance of Alcohol‐Related Stimuli Increases During the Early Stage of Abstinence in Alcohol‐Dependent Patients.” Alcohol and Alcoholism 44, no. 5: 458–463. 10.1093/alcalc/agp056. [DOI] [PubMed] [Google Scholar]
  47. Walther, D. , and Koch C.. 2006. “Modeling Attention to Salient Proto‐Objects.” Neural Networks 19, no. 9: 1395–1407. 10.1016/j.neunet.2006.10.001. [DOI] [PubMed] [Google Scholar]
  48. Yantis, S. , and Jonides J.. 1984. “Abrupt Visual Onsets and Selective Attention: Evidence From Visual Search.” Journal of Experimental Psychology. Human Perception and Performance 10, no. 5: 601–621. 10.1037/0096-1523.10.5.601. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Data and protocols are not publicly available to ensure special caution in handling patient information, but they can be provided upon reasonable request.

Articles from Alcohol, Clinical & Experimental Research are provided here courtesy of Wiley

RESOURCES