# ERP XDAWN Decoding

Dear MNE experts:

I have one issue while computing an XDAWN. I have tried XDAWN on ERP datasets. I thought the problem is due to gbk encoding. But I can not add it in the fname = os.path.join(eeg_path,‘010101_1.set’).
May I ask how to solve this ‘gbk’ problem?

Best wishes
YuTong`

``````type or paste code here
import mne
import os
import numpy as np
import matplotlib.pyplot as plt

from sklearn.model_selection import StratifiedKFold
from sklearn.pipeline import make_pipeline
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import classification_report, confusion_matrix
from sklearn.preprocessing import MinMaxScaler

from mne import io, pick_types, read_events, Epochs, EvokedArray, create_info
from mne.preprocessing import Xdawn
from mne.decoding import Vectorizer

print(__doc__)

data_path = data_path = 'C:\WM_DATA\DATA\DATA'
eeg_path = data_path
fname = os.path.join(eeg_path,'010101_1.set')
tmin, tmax = -0.1, 0.3
event_id = {'B9(size4_left_noc)/71': 46, 'B5(size2_left_noc)/61': 55, 'B3(size1_right_noc)/53': 56, 'B2(size1_left_c)/52': 59, 'B1(size1_left_noc)/51': 59, 'B8(size2_right_c)/64': 48, 'B4(size1_right_c)/54': 56, 'B6(size2_left_c)/62': 56, 'B12(size4_right_c)/74': 19, 'B10(size4_left_c)/72': 49}
n_filter = 3

# Setup for reading the raw data
raw.filter(1, 20, fir_design='firwin')

picks = pick_types(raw.info, meg=False, eeg=True, stim=False, eog=False,

epochs = Epochs(raw, events, event_id, tmin, tmax, proj=False,
verbose=False)

# Create classification pipeline
clf = make_pipeline(Xdawn(n_components=n_filter),
Vectorizer(),
MinMaxScaler(),
LogisticRegression(penalty='l1', solver='liblinear',
multi_class='auto'))

# Get the labels
labels = epochs.events[:, -1]

# Cross validator
cv = StratifiedKFold(n_splits=10, shuffle=True, random_state=42)

# Do cross-validation
preds = np.empty(len(labels))
for train, test in cv.split(epochs, labels):
clf.fit(epochs[train], labels[train])
preds[test] = clf.predict(epochs[test])

# Classification report
target_names = ['aud_l', 'aud_r', 'vis_l', 'vis_r']
report = classification_report(labels, preds, target_names=target_names)
print(report)

# Normalized confusion matrix
cm = confusion_matrix(labels, preds)
cm_normalized = cm.astype(float) / cm.sum(axis=1)[:, np.newaxis]

# Plot confusion matrix
fig, ax = plt.subplots(1)
im = ax.imshow(cm_normalized, interpolation='nearest', cmap=plt.cm.Blues)
ax.set(title='Normalized Confusion matrix')
fig.colorbar(im)
tick_marks = np.arange(len(target_names))
plt.xticks(tick_marks, target_names, rotation=45)
plt.yticks(tick_marks, target_names)
fig.tight_layout()
ax.set(ylabel='True label', xlabel='Predicted label')
``````

Module created for script run in IPython
Extracting parameters from C:\WM_DATA\DATA\DATA\010101_1.set…
613 matching events found
c:\users\yyt.spyder-py3\xdawn.py:27: RuntimeWarning: At least one epoch has multiple events. Only the latency of the first event will be retained.
C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\mne\io\eeglab\eeglab.py:149: DeprecationWarning: The truth value of an empty array is ambiguous. Returning False, but in future this will result in an error. Use `array.size > 0` to check that an array is not empty.
if d.get(“type”, None) != ‘FID’:
C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\mne\io\eeglab\eeglab.py:149: DeprecationWarning: The truth value of an empty array is ambiguous. Returning False, but in future this will result in an error. Use `array.size > 0` to check that an array is not empty.
if d.get(“type”, None) != ‘FID’:
No baseline correction applied
0 projection items activated
Setting up band-pass filter from 1 - 20 Hz

## FIR filter parameters

Designing a one-pass, zero-phase, non-causal bandpass filter:

• Windowed time-domain design (firwin) method
• Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
• Lower passband edge: 1.00
• Lower transition bandwidth: 1.00 Hz (-6 dB cutoff frequency: 0.50 Hz)
• Upper passband edge: 20.00 Hz
• Upper transition bandwidth: 5.00 Hz (-6 dB cutoff frequency: 22.50 Hz)
• Filter length: 3301 samples (3.301 sec)

c:\users\yyt.spyder-py3\xdawn.py:28: RuntimeWarning: filter_length (3301) is longer than the signal (1400), distortion is likely. Reduce filter length or filter a longer signal.
raw.filter(1, 20, fir_design=‘firwin’)
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done 1 out of 1 | elapsed: 0.0s remaining: 0.0s
[Parallel(n_jobs=1)]: Done 2 out of 2 | elapsed: 0.0s remaining: 0.0s
[Parallel(n_jobs=1)]: Done 3 out of 3 | elapsed: 0.0s remaining: 0.0s
[Parallel(n_jobs=1)]: Done 4 out of 4 | elapsed: 0.0s remaining: 0.0s
[Parallel(n_jobs=1)]: Done 17777 out of 17777 | elapsed: 7.4s finished
c:\users\yyt.spyder-py3\xdawn.py:29: RuntimeWarning: This filename (C:\WM_DATA\DATA\DATA\010101_1.set) does not conform to MNE naming conventions. All events files should end with .eve, -eve.fif, -eve.fif.gz, -eve.lst, -eve.txt, _eve.fif, _eve.fif.gz, _eve.lst, _eve.txt or -annot.fif
Traceback (most recent call last):

File “C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\spyder_kernels\py3compat.py”, line 356, in compat_exec
exec(code, globals, locals)

File “c:\users\yyt.spyder-py3\xdawn.py”, line 29, in

File “”, line 12, in read_events

File “C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\mne\event.py”, line 266, in read_events

File “C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\numpy\lib\npyio.py”, line 1098, in loadtxt
first_line = next(fh)

UnicodeDecodeError: ‘gbk’ codec can’t decode byte 0xca in position 188: illegal multibyte sequence

Hello @hp20072929 and welcome to the forum!

`read_events()` cannot work with a `.set` file.

The EEGLAB reader automatically creates annotations, which can be converted to events:

``````raw = mne.io.read_raw_eeglab(fname)
events, event_id = mne.events_from_annotations(raw)
``````

I’m also not sure why you’re using `read_epochs_eeglab()` to read raw data? I’m surprised it even works…

Best wishes,
Richard

Thank you so much, Richard.
After changing, I have a new problem with mat file. And may I ask if would it be possible to run XDAWN on Event-related potential dataset?

Best regards
YuTong

``````import mne
import os
import numpy as np
import matplotlib.pyplot as plt

from sklearn.model_selection import StratifiedKFold
from sklearn.pipeline import make_pipeline
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import classification_report, confusion_matrix
from sklearn.preprocessing import MinMaxScaler

from mne import io, pick_types, read_events, Epochs, EvokedArray, create_info
from mne.preprocessing import Xdawn
from mne.decoding import Vectorizer

print(__doc__)

data_path = data_path = 'C:\WM_DATA\\raw'
eeg_path = data_path
fname = os.path.join(eeg_path,'010101_1.vhdr')
tmin, tmax = -0.1, 0.3
event_id = {'B9(size4_left_noc)/71': 46, 'B5(size2_left_noc)/61': 55, 'B3(size1_right_noc)/53': 56, 'B2(size1_left_c)/52': 59, 'B1(size1_left_noc)/51': 59, 'B8(size2_right_c)/64': 48, 'B4(size1_right_c)/54': 56, 'B6(size2_left_c)/62': 56, 'B12(size4_right_c)/74': 19, 'B10(size4_left_c)/72': 49}
n_filter = 3

# Setup for reading the raw data

events, event_id = mne.events_from_annotations(raw)

picks = pick_types(raw.info, meg=False, eeg=True, stim=False, eog=False,

epochs = Epochs(raw, events, event_id, tmin, tmax, proj=False,
verbose=False)

# Create classification pipeline
clf = make_pipeline(Xdawn(n_components=n_filter),
Vectorizer(),
MinMaxScaler(),
LogisticRegression(penalty='l1', solver='liblinear',
multi_class='auto'))

# Get the labels
labels = epochs.events[:, -1]

# Cross validator
cv = StratifiedKFold(n_splits=10, shuffle=True, random_state=42)

# Do cross-validation
preds = np.empty(len(labels))
for train, test in cv.split(epochs, labels):
clf.fit(epochs[train], labels[train])
preds[test] = clf.predict(epochs[test])

# Classification report
target_names = ['aud_l', 'aud_r', 'vis_l', 'vis_r']
report = classification_report(labels, preds, target_names=target_names)
print(report)

# Normalized confusion matrix
cm = confusion_matrix(labels, preds)
cm_normalized = cm.astype(float) / cm.sum(axis=1)[:, np.newaxis]

# Plot confusion matrix
fig, ax = plt.subplots(1)
im = ax.imshow(cm_normalized, interpolation='nearest', cmap=plt.cm.Blues)
ax.set(title='Normalized Confusion matrix')
fig.colorbar(im)
tick_marks = np.arange(len(target_names))
plt.xticks(tick_marks, target_names, rotation=45)
plt.yticks(tick_marks, target_names)
fig.tight_layout()
ax.set(ylabel='True label', xlabel='Predicted label')
``````
``````runfile('C:/Users/yyt/.spyder-py3/11_15.py', wdir='C:/Users/yyt/.spyder-py3')
Module created for script run in IPython
Traceback (most recent call last):

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\spyder_kernels\py3compat.py", line 356, in compat_exec
exec(code, globals, locals)

File "c:\users\yyt\.spyder-py3\11_15.py", line 27, in <module>

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\mne\io\eeglab\eeglab.py", line 259, in read_raw_eeglab

File "<decorator-gen-277>", line 12, in __init__

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\mne\io\eeglab\eeglab.py", line 358, in __init__

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\mne\io\eeglab\eeglab.py", line 60, in _check_load_mat

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\mne\io\eeglab\_eeglab.py", line 82, in _readmat

mjv, _ = matfile_version(fid)

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\scipy\io\matlab\_miobase.py", line 223, in matfile_version
return _get_matfile_version(fileobj)

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\scipy\io\matlab\_miobase.py", line 251, in _get_matfile_version
raise ValueError('Unknown mat file type, version %s, %s' % ret)

ValueError: Unknown mat file type, version 10, 68
``````

The file you’re trying to load ends with .vhdr. This is a BrainVision, not an EEGLAB file.

thanks for the awesome information.

You are still trying to load a BrainVision file (`fname = os.path.join(eeg_path,'010101_1.vhdr')`) with the EEGLab reader `raw = mne.io.read_raw_eeglab(fame)`, which can not work. Please read it with the correct reader, in this case: `raw = mne.io.read_raw_brainvision(fname)` or with the general reader `raw = mne.io.read_raw(fname)` which infers the data format from the file extension.

1 Like

Thanks so much.
After solving the format.
I went into a problem: MemoryError: Unable to allocate 9.88 GiB for an array with shape (401, 3307889) and data type float64

``````runfile('C:/Users/yyt/.spyder-py3/11_15.py', wdir='C:/Users/yyt/.spyder-py3')
Module created for script run in IPython
Used Annotations descriptions: ['S  5', 'S  6', 'S 11', 'S 12', 'S 21', 'S 22', 'S 23', 'S 24', 'S 31', 'S 32', 'S 33', 'S 34', 'S 41', 'S 42', 'S 43', 'S 44', 'S 51', 'S 52', 'S 53', 'S 54', 'S 61', 'S 62', 'S 63', 'S 64', 'S 71', 'S 72', 'S 73', 'S 74', 'S 77', 'S 88', 'S 99', 'boundary']
c:\users\yyt\.spyder-py3\11_15.py:27: RuntimeWarning: The data contains 'boundary' events, indicating data discontinuities. Be cautious of filtering and epoching around these events.
C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\sklearn\model_selection\_split.py:684: UserWarning: The least populated class in y has only 1 members, which is less than n_splits=10.
warnings.warn(
Computing rank from data with rank='full'
EEG: rank 31 from info
Reducing data rank from 31 -> 31
Estimating covariance using EMPIRICAL
Done.
Traceback (most recent call last):

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\spyder_kernels\py3compat.py", line 356, in compat_exec
exec(code, globals, locals)

File "c:\users\yyt\.spyder-py3\11_15.py", line 54, in <module>
clf.fit(epochs[train], labels[train])

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\sklearn\pipeline.py", line 378, in fit
Xt = self._fit(X, y, **fit_params_steps)

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\sklearn\pipeline.py", line 336, in _fit
X, fitted_transformer = fit_transform_one_cached(

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\joblib\memory.py", line 349, in __call__
return self.func(*args, **kwargs)

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\sklearn\pipeline.py", line 870, in _fit_transform_one
res = transformer.fit_transform(X, y, **fit_params)

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\mne\decoding\mixin.py", line 33, in fit_transform
return self.fit(X, y, **fit_params).transform(X)

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\mne\preprocessing\xdawn.py", line 460, in fit
filters, patterns, evokeds = _fit_xdawn(

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\mne\preprocessing\xdawn.py", line 171, in _fit_xdawn
evokeds, toeplitzs = _least_square_evoked(

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\mne\preprocessing\xdawn.py", line 87, in _least_square_evoked
toeplitz.append(linalg.toeplitz(trig[0:window], trig))

File "C:\Users\yyt\mne-python\1.1.1_0\lib\site-packages\scipy\linalg\_special_matrices.py", line 199, in toeplitz
return as_strided(vals[len(c)-1:], shape=out_shp, strides=(-n, n)).copy()

MemoryError: Unable to allocate 9.88 GiB for an array with shape (401, 3307889) and data type float64
``````

Best regards
YuTong

The error is self-explanatory. You ran out of RAM thus it could not allocate 9.88 Gb of RAM for this operation. Either you optimize your code/RAM usage to decrease the memory consumption (for instance making sure you don’t store large arrays you don’t use in a variable), or you run your code on a computer with more RAM.

Thanks.
In fact, this computer has 16GB RAM.
I have two ideas:

1. Modify virtual memory (not working)
2. Numpy uses lower precision when defining arrays. Reduced from float64 to float32.
My concern is whether would this affect the results a lot.
Best wishes
YuTong

You should just use a computer with more RAM or use less data. Anything else will just eat up much of your time without a promise of success.

Best wishes,
Richard

there is also a `preload=False` option to `read_raw_brainvision()`. It allows you to load the raw metadata, then crop to something shorter (or perform epoching), then load the data samples.

1 Like

Thanks, everyone.
It is working perfectly.
Best wishes

What did you do to fix it?

Generally, I re-run the preprocessing.