ICA

Introductory notes:

This notebook presents ICA artifact repairment functionality:

  • Fitting ICA

  • Exploring components

  • Excluding components from the raw data

Recommended reading:

  1. MNE: Repairing artifacts with ICA

  2. Learning eeg: artifacts

Import data

Import module

from sleepeegpy.pipeline import ICAPipe
from os import makedirs, path
import mne

mne.set_log_level("ERROR")

Initialize ICAPipe object

If you wish to change the path for output_dir ot input dir, change it below. If no such folders, they will be created automatically.

output_dir = "output_folder"  # Output path and name can be changed here
input_dir = "input_files"  # input files dir can be changed here
makedirs(input_dir, exist_ok=True)
makedirs(output_dir, exist_ok=True)

Add required files

  • Put all your files in the input folder.

  • Modify your eeg file name below. The file can be any format supported by the mne.read_raw() function.

  • For more information about the supported formats, see mne documentation

eeg_file_name = "resampled_raw.fif"
ica_file_name = (
    None  # If you want to load already fitted ICA - replace None with the file name.
)

path_to_eeg = path.join(input_dir, eeg_file_name)
path_to_ica = path.join(input_dir, ica_file_name) if ica_file_name else None

ica_pipe = ICAPipe(
    # can be any type of eeg file that MNE's read_raw() function supports.
    path_to_eeg=path_to_eeg,
    output_dir=output_dir,
    method="fastica",  # ICA method, can be either 'fastica', 'infomax' or 'picard'
    fit_params=None,  # Additional parameters passed to the ICA estimator as specified by 'method'.
    n_components=40,  # was 40,  Number of principal components that are passed to the ICA algorithm during fitting.
    random_state=120,  # A seed for the NumPy random number generator (RNG).
    path_to_ica=path_to_ica,
)

Fit

Run the ICA decomposition on 1 Hz high-pass filtered data

ica_pipe.fit(filter_kwargs=dict(l_freq=1.0, h_freq=None, n_jobs=-1))

Explore

Plot independent components from raw data, you can select here components you want to exclude.

ica_pipe.plot_sources()

Plot components on interpolated sensor topography.

_ = ica_pipe.plot_components()
../_images/1f2d679846b137c365160f667bdc95d456f1d3dccca37e60636c7328cc36f6dc.png ../_images/699e3af77579e9b455b5fc72a2068ee1bb92957831c3c76f300fd3ea254e402a.png

Display component properties.

_ = ica_pipe.plot_properties(picks=[7], psd_args=dict(fmin=0, fmax=40))
../_images/3510f059eaf547c7bf2d1d412c564e9a2fbe774ea9c40a16df7af5b38a8f0a76.png

Exclude

Apply to the raw signal ICA component exclusion.

ica_pipe.apply()

Save repaired EEG signal and ica solution

ica_pipe.save_raw("after_ica_raw.fif", overwrite=True)

ica_pipe.save_ica(fname="exclude-ica.fif", overwrite=True)