Dear SPM users,
I am writing to you because I encounter some bug that I have trouble to fix with the msp group inversion,
the ReML iteration give NaN results (Matrix is singular, close to singular or badly scaled. Results may be inaccurate. RCOND = NaN.
> In spm_reml_sc (line 206)
In spm_eeg_invert (line 840))
I found that this is caused by hyperparameters (h) calculated in spm_sp_reml that are infinite and I don't find why this happens,
the inversion msp runs well on a patient to patient basis and when combining some patients, but on the group it gives this error,
did you ever have such issue?
I have run the headmodels with the patients' specific electrode positions and BEM (btw, the ft_read_headshape function of fieldtrip did not work properly from the spm12 updates (r6732) so I used a former version of this function and it worked well)
Thank you much for any indication you might have on this issue,
Astrid
|