Maxwell filter data with movement compensation#

Demonstrate movement compensation on simulated data. The simulated data contains bilateral activation of auditory cortices, repeated over 14 different head rotations (head center held fixed). See the following for details:

# Authors: Eric Larson <larson.eric.d@gmail.com>
#
# License: BSD-3-Clause
from os import path as op

import mne
from mne.preprocessing import maxwell_filter

print(__doc__)

data_path = op.join(mne.datasets.misc.data_path(verbose=True), 'movement')

head_pos = mne.chpi.read_head_pos(op.join(data_path, 'simulated_quats.pos'))
raw = mne.io.read_raw_fif(op.join(data_path, 'simulated_movement_raw.fif'))
raw_stat = mne.io.read_raw_fif(op.join(data_path,
                                       'simulated_stationary_raw.fif'))
Opening raw data file /home/circleci/mne_data/MNE-misc-data/movement/simulated_movement_raw.fif...
    Range : 25800 ... 34208 =     42.956 ...    56.955 secs
Ready.
Opening raw data file /home/circleci/mne_data/MNE-misc-data/movement/simulated_stationary_raw.fif...
    Range : 25800 ... 34208 =     42.956 ...    56.955 secs
Ready.

Visualize the “subject” head movements. By providing the measurement information, the distance to the nearest sensor in each direction (e.g., left/right for the X direction, forward/backward for Y) can be shown in blue, and the destination (if given) shown in red.

mne.viz.plot_head_positions(
    head_pos, mode='traces', destination=raw.info['dev_head_t'], info=raw.info)
Position (mm), Rotation (quat)

This can also be visualized using a quiver.

mne.viz.plot_head_positions(
    head_pos, mode='field', destination=raw.info['dev_head_t'], info=raw.info)
movement compensation
Getting helmet for system 306m

Process our simulated raw data (taking into account head movements).

# extract our resulting events
events = mne.find_events(raw, stim_channel='STI 014')
events[:, 2] = 1
raw.plot(events=events)

topo_kwargs = dict(times=[0, 0.1, 0.2], ch_type='mag', vmin=-500, vmax=500,
                   time_unit='s')
Raw plot
14 events found
Event IDs: [ 1  2  3  4  5  6  7  8  9 10 11 12 13 14]
Opening raw-browser...

First, take the average of stationary data (bilateral auditory patterns).

evoked_stat = mne.Epochs(raw_stat, events, 1, -0.2, 0.8).average()
evoked_stat.plot_topomap(title='Stationary', **topo_kwargs)
Stationary, 0.000 s, 0.100 s, 0.200 s, fT
Not setting metadata
14 matching events found
Setting baseline interval to [-0.19979521315838786, 0.0] sec
Applying baseline correction (mode: mean)
0 projection items activated

Second, take a naive average, which averages across epochs that have been simulated to have different head positions and orientations, thereby spatially smearing the activity.

epochs = mne.Epochs(raw, events, 1, -0.2, 0.8)
evoked = epochs.average()
evoked.plot_topomap(title='Moving: naive average', **topo_kwargs)
Moving: naive average, 0.000 s, 0.100 s, 0.200 s, fT
Not setting metadata
14 matching events found
Setting baseline interval to [-0.19979521315838786, 0.0] sec
Applying baseline correction (mode: mean)
0 projection items activated

Third, use raw movement compensation (restores pattern).

raw_sss = maxwell_filter(raw, head_pos=head_pos)
evoked_raw_mc = mne.Epochs(raw_sss, events, 1, -0.2, 0.8).average()
evoked_raw_mc.plot_topomap(title='Moving: movement compensated (raw)',
                           **topo_kwargs)
Moving: movement compensated (raw), 0.000 s, 0.100 s, 0.200 s, fT
Maxwell filtering raw data
    No bad MEG channels
    Processing 203 gradiometers and 102 magnetometers
    Automatic origin fit: head of radius 91.0 mm
    Using origin -4.1, 16.0, 51.7 mm in the head frame
        Using 90/95 harmonic components for    0.000  (75/80 in, 15/15 out)
    Appending head position result channels and loading raw data from disk
    Processing 1 data chunk
        Using 87/95 harmonic components for    0.000  (72/80 in, 15/15 out)
        Using 88/95 harmonic components for    1.001  (73/80 in, 15/15 out)
        Using 90/95 harmonic components for    2.000  (75/80 in, 15/15 out)
        Using 88/95 harmonic components for    3.000  (73/80 in, 15/15 out)
        Using 88/95 harmonic components for    3.999  (73/80 in, 15/15 out)
        Using 88/95 harmonic components for    5.000  (73/80 in, 15/15 out)
        Using 89/95 harmonic components for    6.001  (74/80 in, 15/15 out)
        Using 93/95 harmonic components for    6.999  (78/80 in, 15/15 out)
        Using 88/95 harmonic components for    8.000  (73/80 in, 15/15 out)
        Using 91/95 harmonic components for    9.001  (76/80 in, 15/15 out)
        Using 93/95 harmonic components for   10.000  (78/80 in, 15/15 out)
        Using 93/95 harmonic components for   11.000  (78/80 in, 15/15 out)
        Using 89/95 harmonic components for   11.999  (74/80 in, 15/15 out)
        Using 88/95 harmonic components for   13.000  (73/80 in, 15/15 out)
        Used  14 head positions for    0.000 -   13.999 sec (#1/1)
[done]
Not setting metadata
14 matching events found
Setting baseline interval to [-0.19979521315838786, 0.0] sec
Applying baseline correction (mode: mean)
0 projection items activated

Fourth, use evoked movement compensation. For these data, which contain very large rotations, it does not as cleanly restore the pattern.

Moving: movement compensated (evoked), 0.000 s, 0.100 s, 0.200 s, fT
    Automatic origin fit: head of radius 91.0 mm
Aligning and averaging up to 14 epochs
    No bad MEG channels
    Processing 203 gradiometers and 102 magnetometers
    Processing epoch 1 (device location: 0.0, 0.0, 40.0 mm)
    Processing epoch 2 (device location: 0.0, 0.0, 40.0 mm)
    Processing epoch 3 (device location: 0.0, 0.0, 40.0 mm)
    Processing epoch 4 (device location: 0.0, 0.0, 40.0 mm)
    Processing epoch 5 (device location: 0.0, 0.0, 40.0 mm)
    Processing epoch 6 (device location: 0.0, 0.0, 40.0 mm)
    Processing epoch 7 (device location: 0.0, 0.0, 40.0 mm)
    Processing epoch 8 (device location: 0.0, 0.0, 40.0 mm)
    Processing epoch 9 (device location: 0.0, 0.0, 40.0 mm)
    Processing epoch 10 (device location: 0.0, 0.0, 40.0 mm)
    Processing epoch 11 (device location: 0.0, 0.0, 40.0 mm)
    Processing epoch 12 (device location: 0.0, 0.0, 40.0 mm)
    Processing epoch 13 (device location: 0.0, 0.0, 40.0 mm)
    Processing epoch 14 (device location: 0.0, 0.0, 40.0 mm)
Created Evoked dataset from 14 epochs

Total running time of the script: ( 0 minutes 18.877 seconds)

Estimated memory usage: 48 MB

Gallery generated by Sphinx-Gallery