Wavelet meets actogram.

A variety of methods to determine phase markers and period length from experimental data sets have traditionally assumed a rhythm of fixed period and amplitude. But most biological oscillations exhibit fluctuations in both period and amplitude, leading to the recent interest in the application of wavelet transforms that can measure how rhythms vary over time. Here we examine how wavelet-based methods can be extended to the analysis of conventional actograms, including the detection of onsets in circadian activity and temperature rhythms of rodents.