def run_subject(name):
loader = DataManager(name, parameters=parameters_meg)
loader.load_epochs(recompute=True)
print(f'Done {name}.')
bands = dict(raw=(0.1, 13),
delta=(0.1, 3),
theta=(3.5, 7.5),
alpha=(7.5, 13),
lower=(0.1, 7.5))
n_jobs = 48
# %%
for name in [
'MEG_S01', 'MEG_S02', 'MEG_S03', 'MEG_S04', 'MEG_S05', 'MEG_S06',
'MEG_S07', 'MEG_S08', 'MEG_S09', 'MEG_S10'
]:
# Load MEG data
# name = 'MEG_S02'
dm = DataManager(name)
dm.load_epochs(recompute=False)
# Init cross validation
cv = CV_split(dm)
cv.reset()
# MVPA parameters
n_components = 6
# Cross validation
# y_pred and y_true will be stored in [labels]
labels = []
while cv.is_valid():
# Recursive
parameters_meg = dict(
picks='eeg',
stim_channel='from_annotations',
l_freq=0.1,
h_freq=15.0,
tmin=-0.2,
tmax=1.2,
decim=10,
detrend=1,
reject=dict(
eeg=150e-6, # 150 $mu$V
),
baseline=None)
# %%
name = 'EEG_S02'
loader = DataManager(name)
loader.load_raws()
print('Done')
# %%
raw = loader.raws[0]
raw
# %%
mne.find_events(raw)
# %%
mne.events_from_annotations(raw)
# %%
bands = dict(raw=(0.1, 13),
delta=(0.1, 3),
theta=(3.5, 7.5),
alpha=(7.5, 13),
lower=(0.1, 7.5))
n_jobs = 48
# %%
for subject in [
'S01', 'S02', 'S03', 'S04', 'S05', 'S06', 'S07', 'S08', 'S09', 'S10'
]:
name_eeg = f'EEG_{subject}'
name_meg = f'MEG_{subject}'
# Load data
dm_eeg = DataManager(name_eeg)
dm_eeg.load_epochs(recompute=False)
dm_meg = DataManager(name_meg)
dm_meg.load_epochs(recompute=False)
# Init cross validation
cv_eeg = CV_split(dm_eeg)
cv_eeg.reset()
cv_meg = CV_split(dm_meg)
cv_meg.reset()
# MVPA parameters
n_components = 6
# Cross validation
# y_pred and y_true will be stored in [labels]
from sklearn import manifold
from tools.data_manager import DataManager
from tools.epochs_tools import epochs_get_MVPA_data
from tools.figure_toolbox import Drawer
# %%
drawer = Drawer()
events_ID = ['1', '2', '4']
n_components = 6
n_jobs = 48
# %%
for name in ['MEG_S02', 'EEG_S02']:
# -----------------------------------------------
# Init data manager
dm = DataManager(name)
# Load epochs
dm.load_epochs(recompute=False)
# -----------------------------------------------
# Separate epochs
epochs_1, epochs_2 = dm.leave_one_session_out(includes=[1, 3, 5],
excludes=[2, 4, 6])
# Xdawn enhancemen
xdawn = mne.preprocessing.Xdawn(n_components=n_components)
# Fit
xdawn.fit(epochs_1)
# Baseline correction
# %%
name = 'MEG_S02'
parameters_meg = dict(
picks='mag',
stim_channel='UPPT001',
l_freq=0.1,
h_freq=20, # 7,
tmin=-0.2,
tmax=1.2,
decim=12,
detrend=1,
reject=dict(mag=4e-12, ),
baseline=None)
dm = DataManager(name, parameters=parameters_meg)
dm.load_epochs(recompute=True)
# %%
epochs_1, epochs_2 = dm.leave_one_session_out(includes=[1, 3, 5],
excludes=[2, 4, 6])
def relabel(_epochs):
# Relabel _epochs events from 4 to 2
_epochs.events[_epochs.events[:, -1] == 4, -1] = 2
# Remove 4 in event_id pool
_epochs.event_id.pop('4')
relabel(epochs_1)
plotly.offline.init_notebook_mode(connected=True)
plt.style.use('tableau-colorblind10')
# %%
drawer = Drawer()
events_ID = ['1', '2', '4']
n_jobs = 48
# %%
name = 'MEG_S02'
# for name in [f'MEG_S{j+1:02d}' for j in range(2, 3)]:
# -----------------------------------------------
# Init data manager
dm = DataManager(name)
# Load epochs
dm.load_epochs(recompute=False)
# -----------------------------------------------
# Separate epochs
epochs, epochs_2 = dm.leave_one_session_out(includes=[1, 2, 3, 4, 5, 6, 7],
excludes=[0])
epochs = epochs[events_ID]
# Xdawn enhancemen
xdawn = mne.preprocessing.Xdawn(n_components=12)
# Fit
xdawn.fit(epochs)
# Baseline correction