Smoothing Timetrends

Moving average

• CSPM (from SPM12) includes a function that will create a moving average.
• Start CSPM (to add the folder to the path)
• Load the variable into matlab (see here for details).
• Get the variable name of the timetrend
• Set the number of samples over which to calculate the average

>> num_samples4window = 20;

• Run the following command
  • "t" is the name of the timetrend variable in matlab workspace

>> tma = movingaverage( t, num_samples4window )

The filtered timetrend is saved in the variable "tma"

To view the results, run the following commands:

>> figure('Name','Before/after smoothing')

>> plot(t,'b')

>> hold on

>> plot(tma,'r')

>> legend {'Original' 'Smoothed'}

To save, either copy the smoother timetrend values from the command window, or use the save command with the -ascii option:

>> save smoothed_tt.txt tma -ascii