Postoperative delirium (POD), expressed from mild forms such as short cognitive impairments to chronic cognitive deficits or death, is relevant in up to 80% of patients after surgery under general anesthesia (GA). The systemic and mechanistic relationship between preoperative anxiety (POA) and POD is largely unknown. A preoperative anxious phenotype and the identification of sleep impairments due to it hold the strong potential to establish predictive parameters for POD. These discoveries may further elucidate the neuronal mechanisms underlying POA and POD. 37 C57BL/6N male mice (12-weeks-old) were used in the present study. Chronic electroencephalogram (EEG) and electromyogram recordings were performed to assess basal sleep/wake behavior. Following baseline recordings, a cued fear-conditioning (FC) was performed (120s-acclimation, 30s-tone (10kHz, 75dB/octave), 2s 0.6mA electrical foot-shock, repetition-rate=5) for anxiety-phenotype classification. After FC, high anxiety (HA) mice exhibited increased rapid eye movement sleep (REMS)% during the light phase and decreased REMS% during the dark phase. HA mice also displayed elevated WAKE% during dark phase and reduced non-REMS%. HA mice had a significant theta-power reduction during REMS over a 24-hour period after FC. Linear regression revealed that REMS theta-power before FC had a robust effect at distinguishing between low anxiety (LA) and HA mice, predicting the phenotype after FC. Our findings suggest basal EEG-parameters as a predictive tool for categorizing mice into distinct anxiety groups prior to any external stressors. By leveraging these EEG-derived biomarkers and combining them with cognitive tests and GA, we are presently establishing the first robust mouse model for POA and POD.