Transferring the Approach Avoidance Task into Virtual Reality: A Study in Patients with Alcohol Use Disorder vs. Healthy Controls

Abstract

Study aims were to (I) transfer the measurement of the approach bias (Apb) related to alcoholic stimuli via the Approach Avoidance Task (AAT) into Virtual Reality (VR), (II) check whether measuring Apb in VR lead to similar or different results compared to the classical PC-based version, (III) check the validity of VR vs. PC-based bias scores in terms of relatedness to clinical variables. Different ‘grasping-conditions’ were tested and contrasted in VR concerning (Ia) feasibility (performance): (1) never grasp, (2) always grasp, (3) grasp when PULLing stimuli towards oneself. (Ib) Differences in the bias scores between patients with alcohol use disorder (AUD) and healthy controls (HC) were examined for each grasping-condition. (II) PC-based bias scores were computed and contrasted for AUD vs. HC. (III) correlations of the different VR- vs. PC-based bias scores with AUD symptom severity and impulsivity were checked to evaluate validity. (Ia) Grasping-condition 1, followed by 3, showed acceptable (>50%) and good (>80%) rates of correct performances allowing for robust median estimation. (Ib) Significant differences in the resulting bias scores emerged between AUD and HC only for grasping-condition 1 (p = .034) and 3 at trend-level (p = .093). (II) The PC-based bias scores did not discriminate between AUD vs. HC groups. (III) Grasping-condition 1 and 3 VR-based bias scores correlated significantly with impulsivity. In sum, transferring the AAT into VR is feasible, valid, and best implemented without an additional grasping-component when using the VR-controller. This way of Apb assessment represents a viable, perhaps even superior, alternative to PC-based assessments.