非官方中文翻译 本页译自 MNE-Python 官方文档： 查看英文原文 ↗。 文档版权归 MNE Developers 所有（BSD-3-Clause）。

带运动补偿的 Maxwell 滤波

在模拟数据上演示运动补偿。模拟数据包含双侧听觉皮层激活，在 14 种不同的头部旋转下重复（头中心固定）。详见：

mne-tools/mne-misc-data

# Authors: Eric Larson <larson.eric.d@gmail.com>
#
# License: BSD-3-Clause
# Copyright the MNE-Python contributors.

import mne
from mne.preprocessing import maxwell_filter

print(__doc__)

data_path = mne.datasets.misc.data_path(verbose=True) / "movement"

head_pos = mne.chpi.read_head_pos(data_path / "simulated_quats.pos")
raw = mne.io.read_raw_fif(data_path / "simulated_movement_raw.fif")
raw_stat = mne.io.read_raw_fif(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.

可视化“受试者”的头部运动。通过提供测量信息，可用蓝色显示各方向上到最近传感器的距离（如 X 方向为左右、Y 方向为前后），并可用红色显示目标位置（若给出）。

mne.viz.plot_head_positions(
    head_pos, mode="traces", destination=raw.info["dev_head_t"], info=raw.info
)

也可以用箭头图（quiver）可视化。

mne.viz.plot_head_positions(
    head_pos, mode="field", destination=raw.info["dev_head_t"], info=raw.info
)

Getting helmet for system 306m

处理模拟的原始数据（考虑头部运动）。

# 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", vlim=(-500, 500))

Finding events on: STI 014
14 events found on stim channel STI 014
Event IDs: [ 1  2  3  4  5  6  7  8  9 10 11 12 13 14]

首先，对静止数据（双侧听觉模式）求平均。

evoked_stat = mne.Epochs(raw_stat, events, 1, -0.2, 0.8).average()
fig = evoked_stat.plot_topomap(**topo_kwargs)
fig.suptitle("Stationary")

Not setting metadata
14 matching events found
Setting baseline interval to [-0.19979521315838786, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated

其次，做朴素平均：对模拟为不同头部位置与朝向的分段求平均，这会从空间上模糊活动。

epochs = mne.Epochs(raw, events, 1, -0.2, 0.8)
evoked = epochs.average()
fig = evoked.plot_topomap(**topo_kwargs)
fig.suptitle("Moving: naive average")

Not setting metadata
14 matching events found
Setting baseline interval to [-0.19979521315838786, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated

第三，使用原始数据运动补偿（可恢复模式）。

raw_sss = maxwell_filter(raw, head_pos=head_pos, mc_interp="hann")
evoked_raw_mc = mne.Epochs(raw_sss, events, 1, -0.2, 0.8).average()
fig = evoked_raw_mc.plot_topomap(**topo_kwargs)
fig.suptitle("Moving: movement compensated (raw)")

Maxwell filtering raw data
    No bad MEG channels
    Processing 203 gradiometers and 102 magnetometers
    Automatic origin fit: head of radius 91.2 mm
    Using origin -4.2, 16.4, 51.8 mm in the head frame
    Appending head position result channels and loading raw data from disk
        Using 90/95 harmonic components for    0.000  (75/80 in, 15/15 out)
    Processing    1 data chunk of (at least) 10.0 s with 0.0 s overlap and boxcar windowing
    The final 4.000899143496717 s will be lumped into the final window
        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)
[done]
Not setting metadata
14 matching events found
Setting baseline interval to [-0.19979521315838786, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated

第四，使用诱发数据运动补偿。对于这些包含很大旋转的数据，它恢复模式的效果不那么干净。

evoked_evo_mc = mne.epochs.average_movements(epochs, head_pos=head_pos)
fig = evoked_evo_mc.plot_topomap(**topo_kwargs)
fig.suptitle("Moving: movement compensated (evoked)")

    Automatic origin fit: head of radius 91.2 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