Abstract Although several studies reveal that cognitive therapy effectively remediates depressive symptoms in many unipolar nonpsychotic depressed outpatients, the question as to which depressions respond to cognitive therapy remains unanswered. We hypothesized that patients with reduced rapid eye movement (REM) latency (⩽65.0 min) before treatment would be less likely than those with nonreduced REM latency (>65.0 min) to respond to cognitive therapy. The rationale for this prediction was that endogenous depressions are more likely to exhibit this abnormality and also tend to respond to tricyclic antidepressant medication. Thus, we queried whether these depressions might also respond less to a psychosocial intervention. To date, 39 outpatients with nonpsychotic, unipolar major depression (by the Schedule for Affective Disorders and Schizophrenia-Lifetime Version and Research Diagnostic Criteria) who score at least 14 on the 17-item Hamilton Rating Scale for Depression have completed this project, which is still in process. Preliminary findings do not suggest a systematic relationship between pretreatment REM latency and response to cognitive therapy. Further, these results suggests that at least some patients with biological dysregulation, as indicated by reduced REM latency, show a favorable response to an acute trial of cognitive therapy. Study limitations include a small sample of patients who exhibit extremely reduced REM latencies (</51.0 min) and a small number of endogenous depressions. Data collection continues.