Dear SPM users,
I'm trying to use the robust average option in spm_eeg_average, but I
get this warning for half of the conditions:
Warning: Robust averaging could not converge. Maximal number of
iterations exceeded.
> In spm_robust_average at 64
In spm_eeg_average at 165
In SPM8preproc at 853
In sublibabyPreProc at 166
Robust averaging finished after 201 iterations.
What is the consequence of this warning? I'm including more noise,
less data? Because I debugged and the distribution of the weights are
very similar to the ones of the conditions that converged.
I noticed that if I increase the S.robust.ks parameter to 4 or 5 it
converges every time, but what is the consequence of this? I found in
the manual that
"Finally, you will have to choose an o set for the weighting function.
This value, default value 3, defines the weighting function used for
averaging the data. The value
3 will roughly preserve 95% of data points drawn randomly from a
Gaussian distribution."
So I'm guessing that a value 4 or 5 will preserve less data from my trials...
Should I keep ks = 3 even if it is not converging for half of the
trials or should I increase it to 4 or 5?
Thanks in advance for any help!
Best Regards,
Leonardo Barbosa
LSCP Research Assistant
École normale supérieure - Paris
|