Physiologic attributes of the luteinizing hormone pulse signal in the human. Cross-validation studies in men

The performance of two new statistically based, independently formulated, endocrine peak detection methodologies (Cluster and Detect) were tested on the same large set of physiologic LH pulsations. Serum LH concentrations were determined in blood samples withdrawn at 5-minute intervals for 24 hours in eight healthy young men. The subsequent comprehensive LH time series were subjected to Cluster and Detect analysis, each constrained to an alpha level of 0.01 or 0.05. Under both circumstances, computer estimates of physiologic LH pulse frequency by the two separate algorithms were statistically indistinguishable. Moreover, the two distinct methodologies were concordant not only in relation to the original 5-minute sampling series but also with respect to the constituent 10-, 15-, 20-, 30-, and 45-minute sampling data. The cross-validation of these two mathematically independent methodologies has permitted delineation of the detailed spectrum of spontaneous LH pulse properties within individual men and among different males under basal physiologic conditions.