Slow wave sleep abnormalities have long been described in depression but were considered to be nonspecific indicators of psychopathology. Computerized techniques, including amplitude frequency measures and spectral analyses, are permitting new approaches to the examination of delta sleep. Early studies suggested that many depressed patients demonstrate lower delta wave intensity during the first non-rapid eye movement period than the second one. This finding, prominent in middle-aged depressed patients, has led to an examination of the ratio between the first and second non-rapid eye movement periods. This delta sleep measure seems to be a more robust predictor of recurrence than rapid eye movement latency. Analysis of data on 74 patients in a long-term maintenance treatment study for a minimum of 24 months demonstrates that the delta sleep ratio can predict survival time following discontinuation of drug treatment. Individuals with a high delta sleep ratio remain clinically remitted five times longer than those with a low delta sleep ratio.