[Mne_analysis] Covariance and noise density in simulation

[Taylor Williams]

        External Email - Use Caution        

So, how do folks typically go about simulating realistic noise conditions
from scratch? Do they just always assume a unit ratio of passband bandwidth
: nave?

If I wanted to simulate a particular noise density (sigma_n) with a
different ratio of passband bandwidth (B) to nave (nave), could I get away
with doing that by scaling my covariance strength from the default values,
then using simulate_evoked to reduce the noise for the number of epochs?
E.g.

cov = mne.make_ad_hoc_cov(info)
cov['data'] *= ((sigma_n*np.sqrt(B))/20e-15)**2
evoked = mne.simulation.simulate_evoked(fwd, stc, info, cov, nave=nave)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20190819/a604a8c2/attachment.html