Sleep, dreaming, and adaptation to a stressful intellectual activity.

Sixteen male subjects slept in the laboratory for 4 consecutive nights. Night 1 was an adaptation night and night 2 was a baseline night of uninterrupted sleep. On nights 3 and 4 subjects were asked to complete "intelligence" tests prior to sleep. One half of the subjects attempted to complete difficult versions of "intelligence" tests without knowing that they could not be completed in te time allotted. The other subjects were given easier versions of the same tests that they were able to complete in the time allotted. Night 3 was a night of uninterrupted sleep and night 4 involved REM period awakenings for the purpose of dream collection. Stressful manipulation consisted of telling the subjects before the administration of either set of tests that an average university student should complete most of the items within the allotted time. On night 3 both groups showed a significant increase in sleep latency and a significant decrease in REM density compared to the baseline night. On night 4 subjects in the difficult condition experienced significantly more anxiety in their dreams and somewhat higher levels of incorporation of the presleep material than subjects in the easy condition. Our findings suggest that following a stressful experience uninterrupted sleep has more short term adaptive value than a procedure which enhances dream recall and that subjects who incorporate elements from a presleep stressful event into their remembered dreams show less adaptation on awakening than subjects who do not.