Eye tracking detected cannabis-cue attention better than the dot-probe task

Attentional bias toward substance-related cues is a key feature of addiction and may contribute to the continued use of substances, such as cannabis. However, the sensitivity of different paradigms for detecting this bias remains underexplored. This study compared three visual attentional bias paradigms: dot probe (DP), eye movement (EM), and MouseView.js (MV) in detecting cannabis-related attentional biases among cannabis users and non-users. Thirty-one participants (i.e., 16 users and 15 non-users) completed all measures: DP (reaction time to probes replacing cannabis or neutral stimuli), EM (measured via eye-tracking), and MV (cursor-based dwell time). Linear mixed-effects models examined the effects of stimulus type, group, craving, and cannabis use severity. The EM and MV tasks revealed significant stimulus effects, with non-users showing greater attention to neutral images, while cannabis users exhibited more evenly distributed attention across stimuli. In the EM task, a significant stimulus × severity interaction indicated reduced attentional differentiation with increased cannabis use severity. MV results followed a similar, but less precise pattern. In contrast, the DP task showed minimal group or interaction effects and lacked sensitivity to detect attentional differences. These findings suggest that EM-based measures are more sensitive for detecting cannabis-related attentional biases, particularly when accounting for cannabis use severity. MV showed some alignment with EM patterns but demonstrated weaker effects. The DP task, despite its widespread use, may be insufficient for capturing cannabis-related attentional differences. Results highlight the importance of task selection in attentional bias research and the potential influence of individual differences in cannabis use severity.