It could also mean shorter N1 or N3 phases of non-REM sleep. From my limited understanding it seems like the subjects falling asleep faster would correspond to a shorter N1 phase, but I haven't read the paper yet.