par Van Cauter, Eve 
Référence American Journal of Physiology Endocrinology Metabolism and Gastrointestinal Physiology, 6, 3, page (E255-E264)
Publication Publié, 1979
Référence American Journal of Physiology Endocrinology Metabolism and Gastrointestinal Physiology, 6, 3, page (E255-E264)
Publication Publié, 1979
Article révisé par les pairs
| Résumé : | A statistical method developed specifically for the characterization of the variations of hormones and other serum constituents over the 24-h span is presented and illustrated. The relatively reduced length of such series of data and the coexistence of periodic (such as circadian rhythm) and nonperiodic (such as episodic variations) components forbid the use of classical spectral methods or of procedures based on models including a single periodicity. In the proposed method, the significance of the observed fluctuations is first tested against the hypothesis of their pure random occurrence. Then, periodogram calculations provide a best-fit theoretical pattern describing the low-frequency variation of the profile and accounting for its possible asymmetries. The amplitude and the acrophase are defined. Estimations of the relative contributions of low (e.g., circadian) and high (e.g., ultradian) components in the time dependence of the profile as well as indications regarding the frequency range and the periodicity or nonperiodicity of ultradian components are obtained. The method is applied to observations over the 24-h span of corticotropin, cortisol, thyrotropin, prolactin, β-melanocyte-stimulating hormone (β-MSH), and dopamine-β-hydroxylase (DBH) levels in plasma. Its ability to summarize the global time properties of the profiles and to provide standards of normalcy as well as objective bases of comparison between profiles is illustrated. The analysis of the profile of DBH observed in control subjects and in depressed patients examplifies how alterations associated with pathology may be quantified. |
