Hi Richard,
Thanks for your help. I used this:
noiseCov = mne.compute_covariance(epochsList, rank = ‘info’, tmin=None, tmax=0, method=‘shrunk’,return_estimators=False, verbose=True, n_jobs=1,projs=projList)
My next step for looking at this is the whitened evoked and GFP. But with the new rank, the whitened GFP looks “weird.”
Original:
Rank=‘info’:
Per this old thread, here is the info for my evoked data (it is bandpass filtered):
In [42]: evokeds[‘central’].info
Out[42]:
<Info | 26 non-empty values
acq_pars: ACQch001 110113 ACQch002 110112 ACQch003 110111 ACQch004 110122 …
bads: 9 items (MEG2032, MEG2323, MEG2343, MEG2312, MEG1632, MEG2341, …)
ch_names: MEG0113, MEG0112, MEG0111, MEG0122, MEG0123, MEG0121, MEG0132, …
chs: 204 Gradiometers, 102 Magnetometers
custom_ref_applied: False
description: Vectorview system at MRN
dev_head_t: MEG device → head transform
dig: 90 items (3 Cardinal, 4 HPI, 83 Extra)
events: 1 item (list)
experimenter: MEG (meg)
file_id: 4 items (dict)
highpass: 1.0 Hz
hpi_meas: 1 item (list)
hpi_results: 1 item (list)
hpi_subsystem: 2 items (dict)
line_freq: 60.0
lowpass: 30.0 Hz
maxshield: False
meas_date: 2019-06-06 13:27:32 UTC
meas_id: 4 items (dict)
nchan: 306
proc_history: 1 item (list)
proj_id: 1 item (ndarray)
proj_name: dquinn_navigate
projs: planar-998–0.500-0.500-PCA-01: on, planar-998-- …
sfreq: 1000.0 Hz
subject_info: 6 items (dict)
Anything else I should check to make sure this covariance is ok, or to figure out why Whitened GFP looks so high?
Thank you,
Megan