Feature identification in circadian rhythms of mice strains using in vivo information.

The objective of this work was to identify strain-specific characteristics from real-time measurements of circadian rhythms of two inbred mouse strains. In particular, heart rate, temperature, and activity data collected from A/J and C57BL/6J (B6) mice using telemetry are analyzed. The influence of activity on heart rate and temperature is minimized by correlation analysis followed by regression analysis. The correlation analysis is used to determine the length of the activity data filter that results in the best correlation between activity data and heart rate or temperature. After the activity data are filtered, they are used in regression analysis. The temperature and heart rate rhythms obtained as the intercepts of the regression analysis are interpreted as the zero-activity rhythms and consequently are good estimates of the circadian rhythms. The circadian temperature rhythms for the B6 mice follow a smoother cosine-like time waveform, whereas those for the A/J mice follow a more square-wave-like waveform. To quantify the difference between these two temperature rhythms, a feature based on Fourier analysis of the time-series data is used. Detrended fluctuation analysis is used to identify features in the heart rate rhythms. The results of this work show that the features for the circadian temperature and heart rate rhythms can be used as distinguishing characteristics of the A/J and B6 strains. This work provides the foundation for future studies directed at investigating the influence of chromosomal substitutions on the regulation of circadian rhythms in these two strains.