Sleep-Related Attentional Bias in Insomnia: A Drift Diffusion Model Approach

Cognitive models propose that insomnia is maintained in part by selective attention to sleep-related information, yet reaction-time indices alone offer limited mechanistic specificity. We investigated sleep-related attentional bias in adolescents and young adults with insomnia disorder (n = 201; aged 15-24 years; DSM-5) using a sleep-related dot-probe task with sleep-related and neutral Cantonese Chinese word pairs. Trial-level responses were analysed with Hierarchical Drift Diffusion Modelling (HDDM) to estimate drift rate (v), the speed of evidence accumulation for probe response choices and to examine moderation by anxiety symptoms. Drift rates were higher on congruent than incongruent trials (q = 0.036), indicating faster evidence accumulation when the probe appeared in the location of sleep-related words, consistent with sleep-related attentional bias indexed indirectly via probe responses. Higher anxiety was associated with faster drift rates across both trial types (q = 0.023 and q = 0.024), consistent with generalised hyperarousal rather than selective enhancement of sleep-related bias. The congruency × anxiety interaction was not significant (95% HDI [-0.10, 0.31]). These findings provide computational evidence consistent with sleep-related attentional bias in young people with insomnia and suggest that comorbid anxiety is associated with broadly increased evidence accumulation rather than cue-specific amplification.