0073 Sleep-Related Attentional Bias in Youth Insomnia – Eye-Tracking Experiment with Hierarchical Drift–Diffusion Modelling (HDDM)

Abstract Introduction Sleep-related attentional bias has long been proposed as a maintaining factor in insomnia, yet reaction-time studies in youth have yielded decidedly mixed results. We therefore combined eye-tracking with hierarchical drift–diffusion modelling (HDDM) to examine whether eye-movement is associated with altered attention and decision making process in youth with insomnia. Methods Seventy-five youth (mean age 17.43 years, 65.33% female) with DSM-5 insomnia and sixty-nine healthy sleepers (mean age: 16.74 years, 62.32% female) completed a dot-probe task presenting sleep-related and neutral image pairs during eye-tracking. First fixation onset latency (FFOL), first fixation duration (FFD) and total fixation duration (TFD) were derived for each image type. Conventional attentional-bias indices and 2×2 ANOVAs (Group × Stimulus type) were calculated. HDDM first estimated baseline drift rate (v) and response threshold (a), and then fitted regression models in which trial-level eye-movement metrics and probe location (sleep-prompting vs non–sleep-prompting) predicted v and a, with starting point (z) and non-decision time (t) treated as random intercepts. Models were compared by deviance information criterion (DIC). Results Groups did not differ in reaction-time–based attentional-bias scores, nor were there Group × Stimulus type interactions for any eye-movement metric. Baseline HDDM demonstrated a lower global drift rate in adolescents with insomnia relative to healthy sleepers, with no group difference in response threshold. Across both groups, longer TFD to either sleep-related or neutral images interacted with probe location to predict higher response thresholds: prolonged fixation was associated with increased threshold when the probe required a response away from the originally fixated location. No robust effects emerged for drift rate or for other eye-movement predictors. Conclusion Computational modelling of eye-movement patterns reveals underlying attentional and decision-making processes that are not apparent in conventional attentional-bias indices. Specifically, youth with insomnia showed a reduced efficiency in global information processing. However, both youth with insomnia and healthy sleepers showed cautious response during disengagement, which revealed no specific effects in insomnia. Future studies should adopt other computational models to further delineate the underlying cognitive process in youth with insomnia. Support (if any) This work was supported by Hong Kong Research Grants Council under the General Research Fund (17613321).