What data should be used for source analysis? Assume that you are interested in analyzing the spectrum of a single source. Because of the large point spread function at low energies (e.g., 68% of the counts will be within a 3.5 degree radius at 100 MeV), you want to use the counts within a region around your source. Nearby sources will contribute counts to this region, and you will probably want to model them. That is, to model a single source you may be forced to model a handful of sources. And therefore you may need to include counts from an even larger region.
For the greatest accuracy possible in modeling a single source you should model the entire sky! This will usually not be feasible. And in reality, the influence of sources a very great distance away from your source will be greatly attenuated. Thus, you should include sources from a large 'Source Region' and counts from a smaller 'Region of Interest' (ROI). The positions and spectra of sources in the Source Region outside of the ROI must have been obtained previously, for example from a catalog; you include these sources for their contribution to the counts in the ROI. How you treat the sources in the ROI is under your control. You may wish to fix the parameters of the sources other than the one you are studying at their catalog values, or you might want to perform a fit on the parameters of all these sources. This will be discussed at greater length below.
Therefore, you will use all sources in the Source Region. You will determine the size of the Source Region appropriate for your needs from experience and experimentation. A list of the suggested data selection cuts can be found in the data preparation section of the Cicerone.