Comparition of raw and pre-processed data

💬 Support & Discussions
1

Hello! there is a option for comparation of raw and pre-processed data in the same window or graph ? i want to see raw data with artifacts and data processed with ICA.

Thnks! :blush:

2

@wmvanvliet worked very hard on a system that allows for linking figures together. I’m not sure if what you’re asking for is possible yet, but in theory we should be able to support it.

3

You may also want to take a look at ica.plot_overlay()

2 Likes
4

You can open two separate raw browsers simultaneously, e.g.

raw.plot()
raw_cleaned.plot()

You might have to type %matplotlib in the interactive Python console for this to work, or maybe it works out of the box.

1 Like
5

thanks to both of you!

6

Hi Nicole,

Perhaps something like this (currently just for muscle artifacts, not eye-blinks - 90% sure it works):

import os
import mne
from mne.preprocessing import ICA, create_ecg_epochs
import matplotlib.pyplot as plt
import matplotlib.widgets as widgets
import numpy as np

os.chdir(“your path to your data”)

file1 = mne.read_epochs(‘subject1-epo.fif’, preload=True)
file1.pick_types(eeg=True)

ica = mne.preprocessing.ICA(
n_components=19, method=‘infomax’, max_iter=“auto”, fit_params=dict(extended=True), random_state=97
)

ica.fit(file1)

muscle_idx_auto, scores = ica.find_bads_muscle(file1)
ica.plot_scores(scores, exclude=muscle_idx_auto)

ica.plot_properties(file1, picks=muscle_idx_auto, log_scale=False)
print(
f"Automatically found muscle artifact ICA components: {muscle_idx_auto}"
)

epochs_corrected = ica.apply(file1.copy(), exclude=muscle_idx_auto)

Create a time vector (assumes your epochs are 4 seconds long and your sampling rate is 250 Hz)

time_vector = np.arange(0, 4, 1 / 250)

Get the number of channels and epochs

num_channels = file1.get_data().shape[1]
num_epochs = file1.get_data().shape[0]

Get channel names

channel_names = file1.ch_names

Create the initial plot

fig, axs = plt.subplots(num_channels, 1, figsize=(10, 20))
fig.suptitle(‘Comparison of Uncorrected and ICA Corrected Epochs’)

Define a function to update the plot

def update_plot(epoch_idx):
for channel_idx in range(num_channels):
axs[channel_idx].clear()

    # Get the data before and after ICA correction
    epoch_uncorrected = file1.get_data()[epoch_idx, channel_idx, :]
    epoch_corrected = epochs_corrected.get_data()[epoch_idx, channel_idx, :]

    # Create the plot
    axs[channel_idx].plot(time_vector, epoch_uncorrected, color='blue', label='Uncorrected')
    axs[channel_idx].plot(time_vector, epoch_corrected, color='red', label='ICA Corrected')
    axs[channel_idx].set(ylabel='{}'.format(channel_names[channel_idx]))

    # Add labels and legend to the last subplot of each figure
    axs[-1].set_xlabel('Time (s)')
    axs[0].legend(loc='upper right')

plt.draw()

update_plot(0)

Add next and previous buttons

ax_prev = plt.axes([0.7, 0.05, 0.1, 0.075])
ax_next = plt.axes([0.81, 0.05, 0.1, 0.075])
btn_prev = widgets.Button(ax_prev, ‘Previous’)
btn_next = widgets.Button(ax_next, ‘Next’)

Define what happens when the buttons are clicked

current_epoch = [0]

def prev_clicked(event):
current_epoch[0] = max(current_epoch[0] - 1, 0)
update_plot(current_epoch[0])

btn_prev.on_clicked(prev_clicked)

def next_clicked(event):
current_epoch[0] = min(current_epoch[0] + 1, num_epochs - 1)
update_plot(current_epoch[0])

btn_next.on_clicked(next_clicked)

plt.show()

All the best
Soren

1 Like
7

thank you!,
i will try with this code

Regards :blush: