Is there a way to refine periodic fit ?

[Clemencegdx]

Hello, I have a little problem with the fit. Here are two examples of fits I made on 2 different electrodes. Here you can see that for one psd it fits the peak at 12hz, the one at 21 and a big one at 29-30. for the second one there is only the one at 29. How can I refine all this?

(I wanted to add that i've got 1407 PSD and in 80% the fit is good but 20% there is this problem..)

Thanks you very much.

Clémence

[voytek]

Is this scalp EEG? That's a really big beta / gamma bump. Are you sure that's neural in origin and not a muscle artifact?

If so, have you reviewed the tutorials on tuning and troubleshooting fits?

https://specparam-tools.github.io/auto_tutorials/plot_08-TroubleShooting.html#sphx-glr-auto-tutorials-plot-08-troubleshooting-py

You need to define some success / acceptability criteria and adjust your parameters accordingly. I'm sure with the right set of fitting parameters (peak_threshold, max_n_peaks) you can get the results you want.

[Clemencegdx]

Hello, thanks you very much for your answer.

It's intracranial EEG. Yes we are sure that is neural.
I tried to work with the fitting parameters it's a bite better but not yet good enough. Do you think i can smooth a little bit my psd to have a better fit ?