Study Objectives:To perform a within-subject analysis of the changes in absolute and normalized rates of non-rapid eye movement (NREM) delta power production across daytime naps and baseline sleep, specifically to determine whether these rates change linearly or exponentially.

Design: Subjects were studied in a 2-day protocol after screening for sleep disorders. The first night was a baseline sleep recording that was followed by a nap the next day at one of the following times: 900, 1200, 1500, 1800 with the nap order varied between subjects. Sleep was then recorded on the postnap night with the baseline bedtime schedule. Subjects completed all 4 nap conditions.

Setting:A 4-bed sleep research laboratory. Each subject slept in a separate bedroom.

Interventions:Not applicable.

Participants:Subjects were young normal (YNs, mean age 22.4 yrs) and elderly normal (ENs, mean age 71.4 yrs) adults.

Measurements and Results:Electroencephalography was recorded continuously during baseline, nap, and postnap sleep. Fast Fourier transform and period-amplitude analysis and visual scoring were performed. Delta power/min and period-amplitude integrated amplitude/min were significantly lower in EN than YN subjects. Both measures declined linearly across NREM periods and increased linearly across naps in the two age groups. The linear slopes for absolute power and integrated amplitude were significantly flatter in the elderly. However, when normalized to each individual subject’s baseline rate of delta production, the slopes of the ENs closely approximated those of the YNs.

Conclusions:The observation of linear changes in delta across NREM periods and daytime naps contradicts the exponential assumptions of the two-process model. The similarity of the normalized slopes shows that waking and sleep exert the same proportionate change in NREM delta in YNs and ENs. This raises the possibility that homeostatic drive is similar in the two groups despite different rates of absolute delta production. The data here, taken in association with other findings, lead us to hypothesize that NREM delta is specifically homeostatic for the neuronal systems that underlie consciousness.