Unipolar and bipolar depressions show abnormal behavioral manifestations of ultradian (less than 24 h) rhythms, but abnormal rhythms of the central neurotransmitters thought to be important for depression pathophysiology (eg dopamine (DA) and serotonin (5-HT)) have not been shown in this time frame. Since antidepressant treatments normalize disrupted rhythms in depression (eg rapid-eyemovement sleep and hormonal rhythms), we hypothesized that depression-related changes in ultradian oscillations of DA and 5-HT might be revealed during antidepressant treatment. Cerebrospinal fluid (CSF) samples collected q10 min for 24 h in 13 patients experiencing major depressive episodes (MDE) before and after treatment for 5 weeks with sertraline or bupropion were assayed for levels of homovanillic acid (HVA) and 5-hydroxyindoleacetic acid (5-HIAA), and their ratio was calculated. Data were analyzed in the frequency domain using Fourier transforms and multivariate permutation testing. Antidepressant treatments were associated with decreased variance for 5-HIAA, increased variance for HVA, and markedly increased variance for the HVA : 5-HIAA ratio (po0.05, po0.02, and po0.003, respectively). With treatment, the correlations between 5-HIAA and HVA weakened (p ¼ 0.06). Power spectral density (PSDFthe Fourier magnitude squared) of the 5-HIAA signals at periods of 1.75 and 3.7 h (both po0.05) decreased, while circadian cycling of HVA levels (po0.05) and of the ratio (po0.005) increased after treatment. The PSD of the full-length HVA : 5-HIAA ratio series after treatment increased in rapid variability (20-103 min periods, po0.05). Spectrographic windowing demonstrated a focal span of enhanced HVA : 5-HIAA ratio variability following antidepressant treatment, in an approximately 84-min period through the evening (po0.05). Periodic neurotransmitter relationships in depressed patients were altered by treatment in this analysis of a small data set. This may represent a baseline abnormality in the regulation of periodic functions involved in the depression pathophysiology, but it could also be due to an unrelated antidepressant effect. Further studies including comparisons with healthy subject data are in progress.