Hey all,
We are creating a the Pearson's coefficient, correlation matrix created between time series from several ROI's. To do this we are using the 1ddot with the -dem option. The hope was you might be able to fill us in on how the program deals with missing data. And how to best represent those removed time points so they don't throw the results off.
For instance when leaving blank spaces in our text file for censored time points the results are slightly different from those derived via Matlab's processing the same data with "NaN's". Any suggestions or better approaches to ensure results are accurate?
~`Dane