Bradley Voytek

Whew! Okay I looked through _all_ the diffs, and this looks great. The only remaining question I have relates to your original PR comment about the name of `PSD`, which...

On the surface this looks good. I can't quite tell, but does this construction allow for something like the dual-knee fitting approach we've been playing with, for example?

Does our conversation on Slack #specparam affect this at all? I think it looks good.

I'd say go ahead and merge this. This is a very well done implementation of this. The plotting / reporting is especially good.

This is the tension, right? Giving users more options and not constraining fit options versus simplicity that _does_ constrain fit options within "often good enough" bounds. I can't see a...

That's very steep! Can you share some plots of the spectra and the fits, as well as the fitting settings?

I'm not sure what you mean by "transplanted". This looks like a normal EEG spectrum with a large alpha (8-12 Hz or so) peak. Though you're likely overfitting – normally...

That doesn't exactly look like overfitting. R**2 tend to be quite high. But have you looked at our tuning and troubleshooting guide? https://fooof-tools.github.io/fooof/auto_tutorials/plot_07-TroubleShooting.html#sphx-glr-auto-tutorials-plot-07-troubleshooting-py

It's more complex than the visual "bending point" due to so many interacting factors – including the oscillatory "bumps" – so it's hard to give you any concrete advice. The...