0376 The Hyperarousal Subtype of Insomnia Disorder Is Associated with Depressive Symptoms

Abstract Introduction Both the hyperarousal state and depressive symptoms are associated with insomnia disorder (ID). However, it is unknown whether the characteristics of hyperarousal and depressive symptoms vary across ID subtypes. Methods This study (NCT05985512) included 136 participants aged 18 to 65, diagnosed with ID by DSM-5 criteria. We employed psychological assessments of depression using Hamilton Depression Rating Scale (HAMD-17) and one-night polysomnography (PSG) to collect psychological and electroencephalographic (EEG) characteristics including macro sleep structures and spectral features including peak frequency (frequency of the local maximum peak of mean power). Uniform Manifold Approximation and Projection (UMAP) and Variational Bayesian Gaussian Mixture Model (VB-GMM) were applied for data dimensionality reduction and cluster analysis. Dunn’s test were conducted to compare depressive and EEG features across subtypes. Significance was set at P 0.05. Results VB-GMM clustering identified three subtypes of ID: (i) Light Sleep Subtype: This group exhibited moderate HAMD-17 scores and the least amount of N3 sleep (Dunn's test p 0.001). (ii) Least Depression Subtype: Participants reported the lowest levels of depression and had the most N3 sleep (Dunn's test p 0.05) with the lowest alpha and beta EEG peak frequencies during NREM sleep (Dunn's test p 0.05). (iii) Hyperarousal with Depression Subtype: This subtype had the highest level of depressive symptoms (HAMD-17 scores, Dunn's test p 0.001). Although the third subtype's N3 sleep was more than that in Subtype 1 (Dunn's test p 0.001), the higher alpha and beta EEG peak frequencies (Dunn's test p 0.05) during NREM sleep suggested a state of hyperarousal. Conclusion The subtype of ID with the highest level of depressive symptoms is associated with cortical hyperarousal. Future research is needed to study the causal relationship between insomnia and depression under the hyperarousal model. Support (if any) National Natural Science Foundation of China (81971261); Shanghai Shenkang Development Center (SHDC2020CR4074); Shanghai Science and Technology Innovation Action Plan (20Y11906600)