Sleep abnormalities such as insomnia, nightmares, hyper-arousal, and difficulty initiating or maintaining sleep, are diagnostic criteria of post-traumatic stress disorder (PTSD). The vivid dream state, rapid eye movement (REM) sleep, has been implicated in processing emotional memories. We have hypothesized that REM sleep is maladaptive in those suffering from PTSD. However, the precise neurobiological mechanisms regulating these sleep disturbances following trauma exposure are poorly understood. Using single prolonged stress (SPS), a well-validated rodent model of PTSD, we measured sleep alterations in response to stress exposure and over a subsequent 7-day isolation period during which the PTSD-like phenotype develops in rats. SPS resulted in acutely increased REM sleep, transition to REM sleep, and decreased waking in addition to alterations in sleep architecture. The severity of the PTSD-like phenotype was later assessed by measuring freezing levels on a fear-associated memory test. Interestingly, the change in REM sleep following SPS was significantly correlated with freezing behavior during extinction recall assessed more than a week later. We also report reductions in theta (4–10 Hz) and sigma (10–15 Hz) band power during transition to REM sleep which also correlated with impaired fear-associated memory processing. These data reveal that changes in REM sleep, transition to REM sleep, waking, and theta and sigma power may serve as sleep biomarkers to identify individuals with increased susceptibility to PTSD following trauma exposure.