(PS2-4) An Examination of an Integrated Attentional Bias Modification Training for Social Anxiety Disorder: A Focus on the Dynamic Attentional Process

Attentional bias (AB) toward threat is one of the primary cognitive processes that contribute to the development and maintenance of social anxiety disorder (SAD). More recent research has argued that AB has a dynamic and fluctuating nature, which can manifest as attentional fluctuation (AF) (Iacoviello et al., 2014; Zvielli, 2014). Previous studies have developed attentional bias modification (ABM) training to help disengage attention from threatening stimuli, but the results have been mixed. Mogg and Bradley (2018) proposed a cognitive-motivational framework that emphasizes the interplay between bottom-up and top-down cognitive processes. Based on this framework, the current study aimed to develop a mobile-based integrated ABM (I-ABM) training that target both cognitive processes and examined its efficacy.

Sixty-two diagnosed with SAD were randomly assigned to either the I-ABM (n=34) or Placebo Training (PLT; n=28) group. Participants were asked to complete a dot-probe task as well as the self-report questionnaires (e.g., Liebowitz Social Anxiety Scale; LSAS, Mini-Social Phobia Inventory; MSPIN) at the baseline, post-training, and a 2-week follow-up. The I-ABM training was designed to incorporate the threat disengagement and inhibitory control components to potentiate the effect of training. The PLT followed the identical training structure as the I-ABM but lacked active training components. Participants were required to complete 9 training sessions over a three-week period using their mobile. The AB score was calculated by subtracting the mean RT of threat-toward trials from that of threat-away trials. The AF score was calculated using Zvielli's (2014) methodology.

Data analysis included those who completed the post-training assessment (I-ABM: 21, PLT: 20, mean age=26.20, female 85.4%, White 73.2%, Non-Hispanic 87.8%). A repeated measures ANOVA revealed a significant main effect of time on the LSAS at post-training (F(1,35)=6.294, p=.017, η² =0.153) (ABM: 9.40% vs. PLT: 10.74%), with a similar pattern at the follow-up (F(1,23)=9.342, p=.006, η² =0.289). There was a marginally significant interaction of time and group on MSPIN (F(1,35)=3.494, p=.070, η² =0.091) (ABM: 15.91% vs. PLT: -2.42%). This pattern remained at the follow-up (F(1,22)=4.049, p=.057, η² =0.155). There was a main effect of time on the AF score (F(1,37)=9.100, p=.005, η² =0.197), with an overall reduction in AF scores across groups after training. There were no main or interaction effects on the AB score. Correlation analyses revealed that changes in AF scores were negatively associated with changes in LSAS scores in the I-ABM group (r=-.508, p=.044) and MSPIN scores in the PLT group (r=-.467, p=.044).

The current findings suggest that both the I-ABM and PLT groups had symptom reduction after training, with the I-ABM group improving more than the PLT group at the follow-up. Furthermore, the attention training had an effect on reducing AF, but the change in AF was negatively related to the change in social anxiety symptoms. This suggests that, while attention bias training can help mitigate unstable attentional processes, the change of AF may not be associated with a reduction in social anxiety symptoms. Further research with a larger sample size is required.