What exactly is Goodness of fit in Dipole fitting model in MNE?

@agramfort @richard

I wanted to find out what exactly goodness of fit meant in Dipole fitting model on MNE so I plotted a graph between the Goodness of fit values given by MNE for each dipole and the euclidean distance between the dipoles fitted against the actual position of dipoles (both in MNI coordinates) in the Phantom head.

I expected the relationship to be inverse ( i.e as predicted distance from actual dipole increases the goodness of fit decreases), however that was not always the case. The scattered plot I obtained was very random

Does this suggest that the Goodness of fit not only rely on the distance but also the orientation of the dipoles? What exactly is the definition of Goodness of fit?

Also how can I get the angle values of the dipole orientation from dipole fitting model?

Hello, please don’t tag people just because you’re eager for a response. I did see your message and I don’t know the answer.


1 Like

goodness of fit is : “1 - explained variance”

you can read about it here Coefficient of determination - Wikipedia