bad_sources = [17, 18, 19]
X_filtered = smica.filter(
raw._data[picks], bad_sources=bad_sources, method="wiener"
)
raw_filtered = raw.copy()
raw_filtered._data[picks] = X_filtered
raw.filter(1, 70)
ica = ICA_mne(n_components=20, method="fastica", random_state=0)
ica.fit(raw, picks=picks)
sources = ica.get_sources(raw).get_data()
ica_mne = transfer_to_ica(
raw, picks, freqs, ica.get_sources(raw).get_data(), ica.get_components()
)
# ica_mne.plot_extended(sources, sort=False)
bads_infomax = [0, 1, 2]
X_ifmx = ica_mne.filter(
raw._data[picks], bad_sources=bads_infomax, method="pinv"
)
raw_ifmx = raw.copy()
raw_ifmx._data[picks] = X_ifmx
# We identify that clusters 6, 7, 8, 9 correspond to noise
max_raw = raw.copy()
max_raw = mne.preprocessing.maxwell_filter(max_raw)
setups = []
setups.append({"raw": raw, "proj": False, "name": "Unfiltered"})
smica.fit(raw, picks=picks, verbose=100, tol=1e-10, em_it=100000)
# Plot powers
noise_sources = [6, 8, 9]
muscle_source = [7]
f, ax = plt.subplots(figsize=(4, 2))
plot_powers(smica.powers, noise_sources, muscle_source, ax, 'smica')
plt.show()
ica = ICA_mne(n_components=n_components, method='picard', random_state=0)
ica.fit(raw, picks=picks)
ica_mne = transfer_to_ica(raw, picks, freqs,
ica.get_sources(raw).get_data(),
ica.mixing_matrix_)
noise_sources = [1, 2]
muscle_source = [4]
f, ax = plt.subplots(figsize=(4, 2))
plot_powers(ica_mne.powers, noise_sources, muscle_source, ax, 'infomax')
plt.show()
sobi = SOBI_mne(p=2000, n_components=n_components, freqs=freqs, rng=0)
sobi.fit(raw, picks=picks, verbose=True, tol=1e-7, max_iter=10000)
# Plot the powers
noise_sources = [1, 8]