The authors compared total night sleep measures and REM sleep
architecture values for normal control subjects (N = 36), unipolar
depressed patients (N = 36), and bipolar depressed patients (N = 22). The
unipolar and bipolar patients had significantly greater fragmentation of
REM periods than control subjects, and bipolar patients showed greater
fragmentation of REM periods than unipolar patients. In both the unipolar
and bipolar samples, the duration of successive REM periods was related to
the total number of REM periods during sleep.Abstract Teaser