Alcohol-related attentional biases in recently detoxified inpatients with severe alcohol use disorder: an eye-tracking approach.

Background: Dominant theoretical models consider that attentional biases (AB) towards alcohol-related stimuli play a key role in the development and maintenance of alcohol use disorder (AUD). Their assessment has however showed high inconsistencies and has been mostly based on unreliable behavioral measures. This study evaluated the presence and extent of alcohol-related AB in recently detoxified inpatients with severe AUD by combining the visual probe task (VPT) paradigm with eye-tracking measures, known to improve the VPT reliability in subclinical populations. Methods: We recruited 24 patients and 27 matched healthy controls. They performed the VPT (measuring reaction time when processing visual targets preceded by alcoholic and matched non-alcoholic pictures) combined with eye-tracking measures (dwell time, first fixation direction/duration, second fixation direction) during two sessions. Estimates of internal consistency, split-half reliability, and test-retest reliability were measured. Results: Patients showed shorter dwell time for alcohol cues (p = .004, d=.853) and reduced number of fixations towards alcohol after a first fixation on non-alcohol cues (p = .012, d=.758) compared to controls. These findings suggest the presence of alcohol-related avoidance AB in detoxified patients with severe AUD. The VPT achieved excellent reliability for these eye-tracking measures. Reaction times and first fixation measures did not indicate any AB pattern and showed poor reliability. Conclusions: The VPT, when combined with dwell time and second fixation direction, constitutes a reliable method for assessing AB in detoxified patients. It showed the presence of an alcohol-related avoidance bias in this clinical population, in contradiction with the approach bias predicted by theoretical models