0254 Fronto-Temporo-Occipital Cortical Thickness Measures Predict Poor Sleep Quality in At-Risk Youth

Abstract

Introduction: Poor subjective sleep quality (SQ) is a prominent risk factor for most forms of psychiatric illness, yet objective biomarkers of SQ have remained elusive. Our goal was to identify neural markers of SQ using a combination of structural and functional neuroimaging assessments in youth expressing a range of psychopathology. Methods: A transdiagnostic sample of 40 youth (8-17yr) completed an MRI assessment and rated past-week SQ with a modified Pittsburgh Sleep Quality Index (N=22 good SQ [PSQI≤5]; N=18 poor SQ [PSQI>;5]). Group-lasso logistic regression identified non-zero predictors of SQ from cortical thickness measures; BOLD response to reward and emotion fMRI tasks; sleep history; and demographic/clinical features. Results: Poor SQ was associated with higher depression severity and cortical thickness in sensory regions (thinner right superior temporal sulcus and left temporal pole, thicker right lateral occipital cortex). Age interacted with right superior frontal (SFC) cortical thickness to predict SQ, such that SFC thickness and age were positively associated in youth with good SQ and negatively related in those with poor SQ. Anxiety severity interacted with right rostral anterior cingulate (rACC) cortical thickness to predict SQ, such that rACC thickness and anxiety were negatively related in youth with good SQ and positively related among youth with poor SQ. Predictors explained 51.2% of the variance in SQ and correctly classified 85% of cases. Conclusion: Age, internalizing symptoms, and cortical thickness in sensory and frontal midline regions, were useful classifiers of SQ. A combination of measures may be necessary to understand the neural basis of poor SQ and its role in psychiatric illness. .