| Abstract: | The concurrent validity of different definitions of REM latency has been tested by comparing the ability of each definition to discriminate between primary depressives (outpatients and inpatients) and normal controls. In outpatients the percentage of cases correctly identified ranged from 62.5% to 70.8%; in inpatients, from 64.6% to 70.8%. REM latency definitions with the least stringent sleep-onset criteria yielded the lowest specificity. In contrast, the range of sensitivities yielded by different definitions was narrower and not clearly affected by sleep-onset criterion or exclusion/inclusion of wakefulness between sleep onset and first REM period. Furthermore, different definitions of REM latency correlated equally well (p less than 0.01) with Hamilton depression ratings. The shorter REM latencies in both outpatients and inpatients were associated with a later time of NREM sleep onset than in controls, rather than with an earlier REM sleep-onset time. |
