Participants will receive six sessions (three sessions per week for two weeks; session duration = 30 mins) of approach-avoidance training in the virtual reality. Alcoholic beverages are pushed away with a controller and soft-drinks will be pulled towards oneself. Training will begin approximately three weeks before discharge from the inpatient clinics to measure the add-on effect and to ensure that the add-on treatment does not extend the treatment period.

Participants will receive six sessions (three sessions per week for two weeks; session duration = 30 mins) of approach-avoidance training on the computer. Alcoholic beverages are pushed away with a joystick and soft-drinks will be pulled towards oneself. Training will begin approximately three weeks before discharge from the inpatient clinics to measure the add-on effect and to ensure that the add-on treatment does not extend the treatment period.

Participants will receive treatment as usual on the wards. For ethical reasons, participants in this condition will get the offer to undertake the already scientifically validated computer-based approach avoidance training after their completion of the study.

In the VR-based AAT, patients are situated in a bar environment where drinks appear on a bar table in front of them. Patients are instructed to react to 100% of the alcoholic beverages with a push response with a controller, and to non-alcoholic drinks in 100% of the trials with a pull response. The stimuli set consists of 50 different beverages, which are displayed three times in total.

Alcohol consumption measures will be derived from the Timeline Followback (TLFB). The TLFB method involves using a calendar to identify alcohol consumption patterns in terms of number of drinks per day during the last 30 days. One drink is defined as 12 grams of ethanol. The Visual Analogue Scale (VAS) comprises single items used to measure the degree of alcohol cravings on scales ranging from 0-100, with 0 representing no cravings and 100 an extreme degree of cravings. The scale is presented visually on a ruler, and the individual will be requested to report the frequency, mean level and the peak level of cravings experienced during the last 30 days.

The 14-item obsessive-compulsive drinking scale (OCDS) measures the degree of cravings experienced during the last seven days. Items are rated on scales ranging from 0 to 4. The higher the score, the more pronounced the cravings. The scale has two subscales: obsessive cravings score and compulsive cravings score. A total score is calculated based on these subscales, and it is possible to calculate a resistance/impairment and interference subscale score.

The diagnostic alcohol-approach-avoidance task (alcohol-AAT) measures approach biases. During the alcohol- AAT, AUD individuals are requested to react to pictures of alcoholic and non-alcoholic drinks using approach and avoidance responses (by pulling or pushing a joystick (comuter-based AAT) or a controller (VR-AAT)). 50% of the alcoholic/non-alcoholic beverages are pushed away and 50% pulled towards oneself. An approach bias is indicated when reaction times are faster for approaching alcohol cues than for avoiding them, whereas the opposite indicates avoidance bias.