It has been hypothesized that sleep changes in endogenous depression are caused by a deficiency in the buildup of Process S (a sleep/wake-dependent component of sleep thought to be reflected in EEG power), and a resulting disinhibition of the REM sleep system. To test this hypothesis, we have reexamined the results of two studies in which sleep episodes of normal subjects could be considered to be similarly deficient in Process S, as a result of reduced prior wakefulness. Structural components of these sleep episodes, particularly REM sleep measures, showed remarkable similarities to those observed in depressed patients. Moreover, there was a strong relationship between REM latency and preceding slow-wave sleep. The findings are interpreted as providing support for the Process S deficiency hypothesis.