The authors conducted a study of 18 depressed patients to see whether
EEG sleep measurements might provide a predictive tool for response to
antidepressant medication. They found that although the sedative
characteristics of amitriptyline did not differentiate good responders from
poor responders until the third week of drug treatment, the good responders
showed significant increases in REM latency, decreases in REM sleep time,
decreases in REM sleep percent, and decreases in REM activity after only 2
nights of drug treatment.