The role of the dorsal attention network in attention bias modification for social anxiety disorder

Identifying reliable biomarkers of treatment response is central to advancing personalized psychiatry. While whole-brain functional connectivity models have shown promise in predicting clinical outcomes, especially for broad-spectrum interventions like cognitive-behavioral therapy, targeted treatments may benefit from more specific neuromarkers. In social anxiety disorder (SAD), Gaze-Contingent Music Reward Therapy (GC-MRT) is a novel attention bias modification (ABM) intervention designed to reduce preferential attentional allocation to socially threatening stimuli. Given the dorsal attention network's (DAN) key role in top-down attentional control, we tested whether resting-state intra-network DAN connectivity could serve as a neural predictor of response to GC-MRT. Participants with SAD were randomized to either receive GC-MRT (n = 22) or to a waitlist control condition (n = 24). Resting-state fMRI data were collected before and after the intervention. Intra-DAN connectivity at baseline and post-treatment were associated with post-treatment symptom severity in the GC-MRT group. Post-treatment intra-DAN connectivity significantly differed in the GC-MRT group relative to controls. These findings suggest that intra-network connectivity within the DAN may have the potential to function both as a predictive biomarker and as a neural marker of successful intervention. Our findings highlight the role of the DAN in attention-based clinical interventions and show that network-specific connectivity metrics may offer a more precise understanding of how targeted neuromodulation affects symptom change in SAD.