Figure 1.
Phase Markers from Six Method Types of Analyzing Melatonin Data.
Diagram of the phase markers from six methods used to analyze a single melatonin profile (subject 1849v). The upper panel includes the assayed melatonin values, plotted as melatonin concentration (pmol/L). The lower panel groups the various phase markers by method and indicates their position with respect to upper panel. Fourier-based analysis methods F(2) and F(3) have been combined in this diagram because of their similarity.
Table 1.
Description of melatonin phase estimate methods and abbreviations used throughout text. DLMO refers to Dim Light Melatonin Onset.
Figure 2.
A. Variability by Study With Drift and Drift Removed.
Box plots of the variability with and without drift of three circadian phase assessment methods by study. Note difference in y-axis scales between With Drift and Drift Removed. The Box plots show Maximum, Mean, Minimum, and Percentiles 99, 75, 50, 25, and 1 of the data. Figure 2B. Variability by age group and sex. Scatter plots of each subject's variability (with drift removed) of three of the methods using data from all subjects. Also indicated as (#not fit) is the number of subjects for whom that method did not yield a phase estimate.
Figure 3.
Variability of all methods With Drift and Drift Removed.
The Box plots show Maximum, Mean, Minimum, and Percentiles 99, 75, 50, 25, and 1 of the data from all subjects. Black are onset methods, red are offset methods, and blue are maximum or midpoint methods. Note difference in y-axis scales between With Drift and Drift Removed.
Figure 4.
Goodness of fit using Adjusted-R2 for four curve-fitting methods.
The Box plots show Maximum, Mean, Minimum, and Percentiles 99, 75, 50, 25, and 1 of the data from all subjects using the four curve-fitting methods: F(2), F(3), Skewed and Physiological.
Figure 5.
Change in phase estimates (hours) for subjects with 2-hour gaps of data.
Change in phase estimates for simulations of 2-hour missing data at different times relative to the melatonin phase marker for five subjects total from Studies 1, 2 and 3. Each panel plots the change in phase estimates for a different method. Data gaps are referenced as the time of midpoint of each gap relative to the DLMOpost-threshold computed from the complete data set for each subject. Positive changes in phase estimate indicate that the estimated phase from the missing data set is later relative to the estimate from the complete data set, while negative values indicate that the estimate from the missing data set was earlier.