Investigating Cuff Inflation Effects on Sleep Arousal and Sleep Stage Changes

Jeonghoon Kim1, Jongshill Lee1, DongHun Kang1, Junhee Han2, Minjae Cho3, Jeyeon Lee1, Myung-Kyu Yi1, In Young Kim1
1Hanyang university, 2Department of Biomedical Engineering, Hanyang University, Korea, 3Department of Electronic Engineering, Hanyang University, Korea


Abstract

Ambulatory blood pressure monitoring (ABPM) enables repeated nocturnal measurements during polysomnography (PSG) but may itself disturb sleep and electroencephalographic (EEG) signals. We investigated cuff-centered EEG activation, AASM-defined arousal, and sleep-stage stability using laboratory-acquired two-night PSG data. Seventy-two participants wore ABPM on Night 1 or Night 2 (n = 36 per randomized sequence); demographics and two-night mean sleep indices were comparable (all p ≥ 0.07). To reduce first-night effects, acute EEG analysis was restricted to 34 participants monitored on Night 2 and 322 measurements initiated during stable sleep without baseline arousal. Mean C3-A2/C4-A1 EEG was compared between the 30-s pre-cuff epoch and the approximately 2-min cuff interval after measurements were averaged within participants. Relative delta power decreased from 0.719 to 0.688 (mean difference, -0.0301; p = 0.0061; dz = -0.50), whereas theta did not change significantly (0.119 to 0.110; p = 0.0632). Alpha, beta, and gamma increased from 0.0696 to 0.0916, 0.0593 to 0.0880, and 0.00951 to 0.01890, respectively (all p < 0.0001; dz = 0.86, 0.96, and 0.77). Stage/arousal analysis included all 72 participants and 893 cuff measurements, of which 655 began during sleep. Arousal accompanied 395/655 sleep-onset measurements (60.3%), while the starting stage was preserved in 561/655 (85.6%). Arousal was most frequent in REM (71.5%) and least frequent in N3 (52.2%); 344/395 arousal-positive measurements (87.1%) retained the underlying stage. Thus, cuff inflation was associated with transient high-frequency EEG activation and frequent arousal, usually without a scored stage transition. Because no sham-inflation control was available and transition proportions were event-weighted, causal interpretation and between-person generalizability remain limited. Cuff-related intervals should be annotated or modeled in PSG analyses, and cuffless monitoring should be considered when uninterrupted sleep physiology is essential.