Group grand averages¶
Grand-average group contrasts for sensor space, dSPM, and LCMV.
# sphinx_gallery_thumbnail_number = 1
import os.path as op
import sys
import matplotlib.pyplot as plt
import mne
sys.path.append(op.join('..', '..', 'processing'))
from library.config import (meg_dir, subjects_dir, set_matplotlib_defaults,
l_freq, tmax, annot_kwargs) # noqa: E402
evokeds = mne.read_evokeds(op.join(meg_dir,
'grand_average_highpass-%sHz-ave.fif' % l_freq))[:3]
Out:
Reading /tsi/doctorants/data_gramfort/dgw_faces_reproduce/MEG/grand_average_highpass-NoneHz-ave.fif ...
Read a total of 1 projection items:
Average EEG reference (1 x 70) active
Found the data of interest:
t = -200.00 ... 2900.00 ms (0.062 * famous + 0.062 * famous + 0.062 * famous + 0.062 * famous + 0.062 * famous + 0.062 * famous + 0.062 * famous + 0.062 * famous + 0.062 * famous + 0.062 * famous + 0.062 * famous + 0.062 * famous + 0.062 * famous + 0.062 * famous + 0.062 * famous + 0.062 * famous)
0 CTF compensation matrices available
nave = 2830 - aspect type = 100
Projections have already been applied. Setting proj attribute to True.
No baseline correction applied
Read a total of 1 projection items:
Average EEG reference (1 x 70) active
Found the data of interest:
t = -200.00 ... 2900.00 ms (0.062 * scrambled + 0.062 * scrambled + 0.062 * scrambled + 0.062 * scrambled + 0.062 * scrambled + 0.062 * scrambled + 0.062 * scrambled + 0.062 * scrambled + 0.062 * scrambled + 0.062 * scrambled + 0.062 * scrambled + 0.062 * scrambled + 0.062 * scrambled + 0.062 * scrambled + 0.062 * scrambled + 0.062 * scrambled)
0 CTF compensation matrices available
nave = 2702 - aspect type = 100
Projections have already been applied. Setting proj attribute to True.
No baseline correction applied
Read a total of 1 projection items:
Average EEG reference (1 x 70) active
Found the data of interest:
t = -200.00 ... 2900.00 ms (0.062 * unfamiliar + 0.062 * unfamiliar + 0.062 * unfamiliar + 0.062 * unfamiliar + 0.062 * unfamiliar + 0.062 * unfamiliar + 0.062 * unfamiliar + 0.062 * unfamiliar + 0.062 * unfamiliar + 0.062 * unfamiliar + 0.062 * unfamiliar + 0.062 * unfamiliar + 0.062 * unfamiliar + 0.062 * unfamiliar + 0.062 * unfamiliar + 0.062 * unfamiliar)
0 CTF compensation matrices available
nave = 2841 - aspect type = 100
Projections have already been applied. Setting proj attribute to True.
No baseline correction applied
Read a total of 1 projection items:
Average EEG reference (1 x 70) active
Found the data of interest:
t = -200.00 ... 2900.00 ms (0.062 * contrast + 0.062 * contrast + 0.062 * contrast + 0.062 * contrast + 0.062 * contrast + 0.062 * contrast + 0.062 * contrast + 0.062 * contrast + 0.062 * contrast + 0.062 * contrast + 0.062 * contrast + 0.062 * contrast + 0.062 * contrast + 0.062 * contrast + 0.062 * contrast + 0.062 * contrast)
0 CTF compensation matrices available
nave = 1832 - aspect type = 100
Projections have already been applied. Setting proj attribute to True.
No baseline correction applied
Read a total of 1 projection items:
Average EEG reference (1 x 70) active
Found the data of interest:
t = -200.00 ... 2900.00 ms (0.062 * faces + 0.062 * faces + 0.062 * faces + 0.062 * faces + 0.062 * faces + 0.062 * faces + 0.062 * faces + 0.062 * faces + 0.062 * faces + 0.062 * faces + 0.062 * faces + 0.062 * faces + 0.062 * faces + 0.062 * faces + 0.062 * faces + 0.062 * faces)
0 CTF compensation matrices available
nave = 5672 - aspect type = 100
Projections have already been applied. Setting proj attribute to True.
No baseline correction applied
Read a total of 1 projection items:
Average EEG reference (1 x 70) active
Found the data of interest:
t = -200.00 ... 2900.00 ms (0.062 * faces_eq + 0.062 * faces_eq + 0.062 * faces_eq + 0.062 * faces_eq + 0.062 * faces_eq + 0.062 * faces_eq + 0.062 * faces_eq + 0.062 * faces_eq + 0.062 * faces_eq + 0.062 * faces_eq + 0.062 * faces_eq + 0.062 * faces_eq + 0.062 * faces_eq + 0.062 * faces_eq + 0.062 * faces_eq + 0.062 * faces_eq)
0 CTF compensation matrices available
nave = 2702 - aspect type = 100
Projections have already been applied. Setting proj attribute to True.
No baseline correction applied
Read a total of 1 projection items:
Average EEG reference (1 x 70) active
Found the data of interest:
t = -200.00 ... 2900.00 ms (0.062 * scrambled_eq + 0.062 * scrambled_eq + 0.062 * scrambled_eq + 0.062 * scrambled_eq + 0.062 * scrambled_eq + 0.062 * scrambled_eq + 0.062 * scrambled_eq + 0.062 * scrambled_eq + 0.062 * scrambled_eq + 0.062 * scrambled_eq + 0.062 * scrambled_eq + 0.062 * scrambled_eq + 0.062 * scrambled_eq + 0.062 * scrambled_eq + 0.062 * scrambled_eq + 0.062 * scrambled_eq)
0 CTF compensation matrices available
nave = 2702 - aspect type = 100
Projections have already been applied. Setting proj attribute to True.
No baseline correction applied
Sensor-space. See 11. Group average at the sensor level We use the same sensor EEG065 as in Wakeman et al.
idx = evokeds[0].ch_names.index('EEG065')
assert evokeds[1].ch_names[idx] == 'EEG065'
assert evokeds[2].ch_names[idx] == 'EEG065'
mapping = {'Famous': evokeds[0], 'Scrambled': evokeds[1],
'Unfamiliar': evokeds[2]}
Let us apply baseline correction now. Here we are dealing with a single sensor
for evoked in evokeds:
evoked.apply_baseline(baseline=(-100, 0))
Out:
Applying baseline correction (mode: mean)
Applying baseline correction (mode: mean)
Applying baseline correction (mode: mean)
We could have used the one-line MNE function for the comparison.
# mne.viz.plot_compare_evokeds(mapping, [idx],
# title='EEG065 (Baseline from -200ms to 0ms)',)
But here we prefer a slightly more involved plotting script to make a publication ready graph.
set_matplotlib_defaults()
fig, ax = plt.subplots(1, figsize=(3.3, 2.3))
scale = 1e6
ax.plot(evoked.times * 1000, mapping['Scrambled'].data[idx] * scale,
'r', label='Scrambled')
ax.plot(evoked.times * 1000, mapping['Unfamiliar'].data[idx] * scale,
'g', label='Unfamiliar')
ax.plot(evoked.times * 1000, mapping['Famous'].data[idx] * scale, 'b',
label='Famous')
ax.grid(True)
ax.set(xlim=[-100, 1000 * tmax], xlabel='Time (in ms after stimulus onset)',
ylim=[-12.5, 5], ylabel=u'Potential difference (μV)')
ax.axvline(800, ls='--', color='k')
if l_freq == 1:
ax.legend(loc='lower right')
ax.annotate('A' if l_freq is None else 'B', (-0.2, 1), **annot_kwargs)
fig.tight_layout(pad=0.5)
fig.savefig(op.join('..', 'figures',
'grand_average_highpass-%sHz.pdf' % l_freq))
plt.show()
Source-space. See 14. Group average on source level
fname = op.join(meg_dir, 'contrast-average_highpass-%sHz' % (l_freq,))
stc = mne.read_source_estimate(fname, subject='fsaverage').magnitude()
lims = (1, 3, 5) if l_freq is None else (0.5, 1.5, 2.5)
brain_dspm = stc.plot(
views=['ven'], hemi='both', subject='fsaverage', subjects_dir=subjects_dir,
initial_time=0.17, time_unit='s', figure=1, background='w',
clim=dict(kind='value', lims=lims), foreground='k')
brain_dspm.save_image(op.join('..', 'figures',
'dspm-ave_highpass-%sHz.png' % (l_freq,)))
LCMV beamformer
fname = op.join(meg_dir, 'contrast-average-lcmv_highpass-%sHz' % (l_freq,))
stc = mne.read_source_estimate(fname, subject='fsaverage')
lims = (0.015, 0.03, 0.045) if l_freq is None else (0.01, 0.02, 0.03)
brain_lcmv = stc.plot(
views=['ven'], hemi='both', subject='fsaverage', subjects_dir=subjects_dir,
initial_time=0.17, time_unit='s', figure=2, background='w',
clim=dict(kind='value', lims=lims), foreground='k')
brain_lcmv.save_image(op.join('..', 'figures',
'lcmv-ave_highpass-%sHz.png' % (l_freq,)))
Total running time of the script: ( 1 minutes 52.565 seconds)