Tamura Toshiyo, Kitamura Kei-Ichiro, Nemoto Tetsu, Kanaya Shigehiko
Annu Int Conf IEEE Eng Med Biol Soc. 2014;2014:4220-3. doi: 10.1109/EMBC.2014.6944555.
This study was motivated by the needs of precise characterization for the ultradian and circadian rhythmicity of human core body temperature (CBT). The CBT data, two-whole-days' data of two female bed-ridden old aged suffering from cerebral infarction sequelae, was detrended to eliminate the long-term components with periods longer than two days and normalized at first. It was then analyzed by the stationary wavelets transform (SWT) to get the time-frequency information. In the step of SWT, symlet 6 was used, and the approximation waveforms in the 5th, 6th and 7th levels were used to reveal the targeted rhythmicity. The results of the SWT show that SWT can faithfully reveal the time-frequency information of feature elements (peaks and troughs) of waveforms and rhythmicity can be characterized by analyzing temporal information of feature elements.
本研究旨在精确表征人体核心体温（CBT）的超日节律和昼夜节律。CBT数据来自两名患有脑梗死后遗症的卧床老年女性，记录了整整两天的数据。首先对数据进行去趋势处理，以消除周期超过两天的长期成分，并进行归一化。然后通过平稳小波变换（SWT）进行分析，以获取时频信息。在SWT步骤中，使用了symlet 6，并用第5、6和7级的近似波形来揭示目标节律。SWT结果表明，SWT能够忠实地揭示波形特征元素（峰值和谷值）的时频信息，并且可以通过分析特征元素的时间信息来表征节律。
