Participants (N=47) are patients with major depression. Participants are assessed with clinical interview, polysomnographic evaluation and screening with Insomnia Severity Index (ISI), Dysfunctional Beliefs About Sleep-scale (DBAS), HDRS, MINI and Sleep Diary pre and post treatment.

Participants are randomly assigned to either treatment as usual or CBT-I. The latter receives individual as well as group therapy with CBT-I during 6 sessions.

Post-treatment: The investigators will compare changes in sleep diary parameters, HamD17, ISI and DBAS. Data are analyzed by using Chi-square test, Wilcoxon-Mann-Whitney U-test and ANOVA. STATA is used for data-processing.

Participants are included from an Outpatient clinic and GPs in Aarhus. Initial assessment is carried out through a personal interview with the participants who are screened with Hamilton Depression Rating Scale, MINI (Mini International Neuropsychiatric Interview) and instructed to fill out an ISI-form (Insomnia Severity Index (Morin, 1993)), DBAS-16, Dysfunctional Beliefs and Attitudes about Sleep Scale, (Morin, 1993) and to keep a sleep diary for two weeks.

All participants go through an objective examination of their sleep through a polysomnography measuring brain activity (EEG), eye activity (EOG), muscular activity (EMG), heart rate, respiration a.o. during sleep. This happens on an outpatient basis with assistance of an experienced sleep technician. Data is analysed by an experienced sleep physician.

After assessment participants are randomized receiving either treatment as usual or CBT-I.

CBT-I includes: Sleep Restriction Therapy, Stimulus Control Therapy, Relaxation Training, Identification and analyzation of negative thoughts and cognitive distortion, Modification of negative thoughts about sleep, Reattribution, Implementation of "Scheduled worry", Relapse prevention.

Post-treatment participants undergo polysomnographic evaluation and screening with HDRS, ISI and DBAS. Again they keep a sleep diary.

Categorical variables are analysed with a chi-square test. Ordinal variables are analysed with non-parametrical tests (Wilcoxon-Mann-Whitney U-test) and variance analysis (ANOVA). Possible confounders as differences in demographics or other treatment, which may affect the target outcome, will be analysed in part with non-parametrical tests and in part in a logistic regression analysis. As secondary effect targets total sleep time and changes in polysomnography as well as changes in HDRS, dysfunctional assumptions on sleep and life quality are measured. STATA is used for data processing with SPSS as a possible alternative solution.

Inclusion Criteria

• Unipolar depression, either as a single episode or as periodic depression.
• HAM-D17 > 18, i.e. moderate to severe depression.
• Time spent on falling asleep: more than 30 minutes at least three nights a week or
• Interrupted sleep with wake periods lasting more than 30 minutes at least three nights a week.
• Symptoms must have lasted for at least a month.

Exclusion Criteria:

• Other sleep disorder (e.g. moderat to severe Sleep apnoea and Periodic Limb Movements)
• Physical disorder that affects sleep to a considerable degree.
• Schizophrenia and bipolar disorder.
• Subject to other ongoing psychological treatment besides standard treatment.
• Suicidal to the equivalent of level 3 on HAM-D 17.
• Active substance abuse problem.
• Pregnancy.
• Shift work.
• Do not speak or understand danish