How to generate subjects' labels in source localization

In the source location of example data, the labels of auditory or visual areas are directly imported. In practical application, how to generate labels of the selected areas of subjects in my experiment, such as auditory areas