Figure 2.

MSE analysis. Sample entropy is calculated by counting the number of sequences of (m + 1) data points will be similar to each other given that they were similar for the first m points. This reflects the unpredictability of the time series and the information that is generated by the underlying system. (A) An example of the sample entropy algorithm (m = 1), for 2 different starting points. (B) A multiscale representation of the signal is achieved by downsampling the original time series to progressively coarser time scales. (C) Two example signals and (D) their MSE curves.