the calculation process of function source_induced_power()

hello, everyone, we use two ways to compute time-frequency maps in a label of the source space:
1.Using source_ Induced_Power() to calculate the time-frequency map of the label;
2.Firstly, using apply_inverse_epochs() calculate the source, then using mne.extract_label_time_course() to extract the label’s time courses, finally, computing time-frequency maps using the time courses.
But there are some differences in our calculation results,Did I make a calculation mistake somewhere? what’s more, I want to ask the calculation process of function source_induced_power() ?
Thank you very much!