Time-Series Analysis

The figures that we generated in our three processes that the results and discussion section refer to are in Appendix A5.

Threshold Analysis Results

The textual output from this method highlights the number and the percentage of points above our given threshold in each channel. In the graphs produced, the x-axis in this graph represents the samples provided by our data, with each data point representing one sample. Since the samples were collected at 250 Hz, then the time interval in between each data point represents 1/250th of a second. The y-axis represents the voltage of the sample in microvolts.

Autocorrelation Analysis Results

The graphs produced by this method show the different autocorrelation values of a given channel of a given patient for different lag values.

 

Wavelet Analysis Results

The upper graph in Figure 7 represents the signal that we are analyzing over time. Each unit in the x axis in the upper graph represents one sample. Since the samples were taken at 250 Hz, then 250 samples represent one second of data. Consequently, the sample displayed in the above figure represents a little less than one half of one second. The y-axis of the upper graph represents the voltage of each sample in microvolts.

The lower graph in Figure 7 is a spectrogram created by our Morlet’s wavelet transform code. The x-axis also represents the samples taken in this channel, with the same time intervals as the upper graph. The y-axis represents the different measures of ‘scale’. The coloration value at each x, y coordinate pair indicates the strength of the Morlet’s wavelet coefficient at that scale and time. The lighter the color, the stronger the coefficient. The significance of this is discussed in Appendix A4.