Visualize channel over epochs as an image

This will produce what is sometimes called an event related potential / field (ERP/ERF) image.

Two images are produced, one with a good channel and one with a channel that does not show any evoked field.

It is also demonstrated how to reorder the epochs using a 1D spectral embedding as described in [1].

# Authors: Alexandre Gramfort <alexandre.gramfort@inria.fr>
#
# License: BSD-3-Clause
# Copyright the MNE-Python contributors.

import matplotlib.pyplot as plt
import numpy as np

import mne
from mne import io
from mne.datasets import sample

print(__doc__)

data_path = sample.data_path()

Set parameters

meg_path = data_path / "MEG" / "sample"
raw_fname = meg_path / "sample_audvis_filt-0-40_raw.fif"
event_fname = meg_path / "sample_audvis_filt-0-40_raw-eve.fif"
event_id, tmin, tmax = 1, -0.2, 0.4

# Setup for reading the raw data
raw = io.read_raw_fif(raw_fname)
events = mne.read_events(event_fname)

# Set up pick list: EEG + MEG - bad channels (modify to your needs)
raw.info["bads"] = ["MEG 2443", "EEG 053"]

# Create epochs, here for gradiometers + EOG only for simplicity
epochs = mne.Epochs(
    raw,
    events,
    event_id,
    tmin,
    tmax,
    proj=True,
    picks=("grad", "eog"),
    baseline=(None, 0),
    preload=True,
    reject=dict(grad=4000e-13, eog=150e-6),
)

Show event-related fields images

# and order with spectral reordering
# If you don't have scikit-learn installed set order_func to None
from sklearn.manifold import spectral_embedding  # noqa
from sklearn.metrics.pairwise import rbf_kernel  # noqa


def order_func(times, data):
    this_data = data[:, (times > 0.0) & (times < 0.350)]
    this_data /= np.sqrt(np.sum(this_data**2, axis=1))[:, np.newaxis]
    return np.argsort(
        spectral_embedding(
            rbf_kernel(this_data, gamma=1.0), n_components=1, random_state=0
        ).ravel()
    )


good_pick = 97  # channel with a clear evoked response
bad_pick = 98  # channel with no evoked response

# We'll also plot a sample time onset for each trial
plt_times = np.linspace(0, 0.2, len(epochs))

plt.close("all")
mne.viz.plot_epochs_image(
    epochs,
    [good_pick, bad_pick],
    sigma=0.5,
    order=order_func,
    vmin=-250,
    vmax=250,
    overlay_times=plt_times,
    show=True,
)

References

[1]

Alexandre Gramfort, Renaud Keriven, and Maureen Clerc. Graph-based variability estimation in single-trial event-related neural responses. IEEE Transactions on Biomedical Engineering, 57(5):1051–1061, 2010. doi:10.1109/tbme.2009.2037139.

Estimated memory usage: 0 MB