Note
Go to the end to download the full example code
Decoding sensor space data with generalization across time and conditions#
This example runs the analysis described in [1]. It illustrates how one can fit a linear classifier to identify a discriminatory topography at a given time instant and subsequently assess whether this linear model can accurately predict all of the time samples of a second set of conditions.
# Authors: Jean-Remi King <jeanremi.king@gmail.com>
# Alexandre Gramfort <alexandre.gramfort@inria.fr>
# Denis Engemann <denis.engemann@gmail.com>
#
# License: BSD-3-Clause
# Copyright the MNE-Python contributors.
import matplotlib.pyplot as plt
from sklearn.linear_model import LogisticRegression
from sklearn.pipeline import make_pipeline
from sklearn.preprocessing import StandardScaler
import mne
from mne.datasets import sample
from mne.decoding import GeneralizingEstimator
print(__doc__)
# Preprocess data
data_path = sample.data_path()
# Load and filter data, set up epochs
meg_path = data_path / "MEG" / "sample"
raw_fname = meg_path / "sample_audvis_filt-0-40_raw.fif"
events_fname = meg_path / "sample_audvis_filt-0-40_raw-eve.fif"
raw = mne.io.read_raw_fif(raw_fname, preload=True)
picks = mne.pick_types(raw.info, meg=True, exclude="bads") # Pick MEG channels
raw.filter(1.0, 30.0, fir_design="firwin") # Band pass filtering signals
events = mne.read_events(events_fname)
event_id = {
"Auditory/Left": 1,
"Auditory/Right": 2,
"Visual/Left": 3,
"Visual/Right": 4,
}
tmin = -0.050
tmax = 0.400
# decimate to make the example faster to run, but then use verbose='error' in
# the Epochs constructor to suppress warning about decimation causing aliasing
decim = 2
epochs = mne.Epochs(
raw,
events,
event_id=event_id,
tmin=tmin,
tmax=tmax,
proj=True,
picks=picks,
baseline=None,
preload=True,
reject=dict(mag=5e-12),
decim=decim,
verbose="error",
)
Opening raw data file /home/circleci/mne_data/MNE-sample-data/MEG/sample/sample_audvis_filt-0-40_raw.fif...
Read a total of 4 projection items:
PCA-v1 (1 x 102) idle
PCA-v2 (1 x 102) idle
PCA-v3 (1 x 102) idle
Average EEG reference (1 x 60) idle
Range : 6450 ... 48149 = 42.956 ... 320.665 secs
Ready.
Reading 0 ... 41699 = 0.000 ... 277.709 secs...
Filtering raw data in 1 contiguous segment
Setting up band-pass filter from 1 - 30 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: 30.00 Hz
- Upper transition bandwidth: 7.50 Hz (-6 dB cutoff frequency: 33.75 Hz)
- Filter length: 497 samples (3.310 s)
[Parallel(n_jobs=1)]: Done 17 tasks | elapsed: 0.0s
[Parallel(n_jobs=1)]: Done 71 tasks | elapsed: 0.1s
[Parallel(n_jobs=1)]: Done 161 tasks | elapsed: 0.3s
[Parallel(n_jobs=1)]: Done 287 tasks | elapsed: 0.4s
We will train the classifier on all left visual vs auditory trials and test on all right visual vs auditory trials.
clf = make_pipeline(
StandardScaler(),
LogisticRegression(solver="liblinear"), # liblinear is faster than lbfgs
)
time_gen = GeneralizingEstimator(clf, scoring="roc_auc", n_jobs=None, verbose=True)
# Fit classifiers on the epochs where the stimulus was presented to the left.
# Note that the experimental condition y indicates auditory or visual
time_gen.fit(X=epochs["Left"].get_data(copy=False), y=epochs["Left"].events[:, 2] > 2)
0%| | Fitting GeneralizingEstimator : 0/35 [00:00<?, ?it/s]
3%|▎ | Fitting GeneralizingEstimator : 1/35 [00:00<00:01, 29.12it/s]
11%|█▏ | Fitting GeneralizingEstimator : 4/35 [00:00<00:00, 59.02it/s]
17%|█▋ | Fitting GeneralizingEstimator : 6/35 [00:00<00:00, 59.10it/s]
23%|██▎ | Fitting GeneralizingEstimator : 8/35 [00:00<00:00, 59.15it/s]
31%|███▏ | Fitting GeneralizingEstimator : 11/35 [00:00<00:00, 65.66it/s]
40%|████ | Fitting GeneralizingEstimator : 14/35 [00:00<00:00, 69.88it/s]
46%|████▌ | Fitting GeneralizingEstimator : 16/35 [00:00<00:00, 68.12it/s]
54%|█████▍ | Fitting GeneralizingEstimator : 19/35 [00:00<00:00, 71.17it/s]
60%|██████ | Fitting GeneralizingEstimator : 21/35 [00:00<00:00, 69.50it/s]
69%|██████▊ | Fitting GeneralizingEstimator : 24/35 [00:00<00:00, 71.88it/s]
74%|███████▍ | Fitting GeneralizingEstimator : 26/35 [00:00<00:00, 70.38it/s]
80%|████████ | Fitting GeneralizingEstimator : 28/35 [00:00<00:00, 69.14it/s]
89%|████████▊ | Fitting GeneralizingEstimator : 31/35 [00:00<00:00, 71.13it/s]
97%|█████████▋| Fitting GeneralizingEstimator : 34/35 [00:00<00:00, 72.74it/s]
100%|██████████| Fitting GeneralizingEstimator : 35/35 [00:00<00:00, 71.32it/s]
Score on the epochs where the stimulus was presented to the right.
scores = time_gen.score(
X=epochs["Right"].get_data(copy=False), y=epochs["Right"].events[:, 2] > 2
)
0%| | Scoring GeneralizingEstimator : 0/1225 [00:00<?, ?it/s]
1%| | Scoring GeneralizingEstimator : 11/1225 [00:00<00:03, 321.55it/s]
2%|▏ | Scoring GeneralizingEstimator : 26/1225 [00:00<00:03, 383.54it/s]
3%|▎ | Scoring GeneralizingEstimator : 41/1225 [00:00<00:02, 403.75it/s]
5%|▍ | Scoring GeneralizingEstimator : 56/1225 [00:00<00:02, 414.24it/s]
6%|▌ | Scoring GeneralizingEstimator : 70/1225 [00:00<00:02, 414.05it/s]
7%|▋ | Scoring GeneralizingEstimator : 85/1225 [00:00<00:02, 418.43it/s]
8%|▊ | Scoring GeneralizingEstimator : 100/1225 [00:00<00:02, 421.92it/s]
9%|▉ | Scoring GeneralizingEstimator : 114/1225 [00:00<00:02, 419.91it/s]
11%|█ | Scoring GeneralizingEstimator : 130/1225 [00:00<00:02, 426.17it/s]
12%|█▏ | Scoring GeneralizingEstimator : 145/1225 [00:00<00:02, 428.23it/s]
13%|█▎ | Scoring GeneralizingEstimator : 161/1225 [00:00<00:02, 432.57it/s]
14%|█▍ | Scoring GeneralizingEstimator : 176/1225 [00:00<00:02, 433.65it/s]
16%|█▌ | Scoring GeneralizingEstimator : 191/1225 [00:00<00:02, 434.64it/s]
17%|█▋ | Scoring GeneralizingEstimator : 206/1225 [00:00<00:02, 435.54it/s]
18%|█▊ | Scoring GeneralizingEstimator : 221/1225 [00:00<00:02, 435.87it/s]
19%|█▉ | Scoring GeneralizingEstimator : 235/1225 [00:00<00:02, 433.99it/s]
20%|██ | Scoring GeneralizingEstimator : 249/1225 [00:00<00:02, 431.91it/s]
21%|██▏ | Scoring GeneralizingEstimator : 262/1225 [00:00<00:02, 427.86it/s]
22%|██▏ | Scoring GeneralizingEstimator : 275/1225 [00:00<00:02, 424.31it/s]
24%|██▎ | Scoring GeneralizingEstimator : 288/1225 [00:00<00:02, 421.22it/s]
24%|██▍ | Scoring GeneralizingEstimator : 300/1225 [00:00<00:02, 416.04it/s]
26%|██▌ | Scoring GeneralizingEstimator : 313/1225 [00:00<00:02, 413.64it/s]
27%|██▋ | Scoring GeneralizingEstimator : 327/1225 [00:00<00:02, 413.65it/s]
28%|██▊ | Scoring GeneralizingEstimator : 338/1225 [00:00<00:02, 407.42it/s]
29%|██▉ | Scoring GeneralizingEstimator : 353/1225 [00:00<00:02, 409.45it/s]
30%|██▉ | Scoring GeneralizingEstimator : 366/1225 [00:00<00:02, 407.78it/s]
31%|███ | Scoring GeneralizingEstimator : 378/1225 [00:00<00:02, 404.13it/s]
32%|███▏ | Scoring GeneralizingEstimator : 388/1225 [00:00<00:02, 396.88it/s]
33%|███▎ | Scoring GeneralizingEstimator : 400/1225 [00:00<00:02, 394.16it/s]
34%|███▎ | Scoring GeneralizingEstimator : 413/1225 [00:01<00:02, 393.34it/s]
35%|███▍ | Scoring GeneralizingEstimator : 428/1225 [00:01<00:02, 396.23it/s]
36%|███▌ | Scoring GeneralizingEstimator : 441/1225 [00:01<00:01, 395.06it/s]
37%|███▋ | Scoring GeneralizingEstimator : 451/1225 [00:01<00:01, 388.70it/s]
38%|███▊ | Scoring GeneralizingEstimator : 462/1225 [00:01<00:01, 384.73it/s]
39%|███▊ | Scoring GeneralizingEstimator : 472/1225 [00:01<00:01, 379.39it/s]
39%|███▉ | Scoring GeneralizingEstimator : 481/1225 [00:01<00:01, 372.46it/s]
40%|████ | Scoring GeneralizingEstimator : 491/1225 [00:01<00:01, 367.92it/s]
41%|████ | Scoring GeneralizingEstimator : 502/1225 [00:01<00:01, 365.17it/s]
42%|████▏ | Scoring GeneralizingEstimator : 514/1225 [00:01<00:01, 364.53it/s]
43%|████▎ | Scoring GeneralizingEstimator : 524/1225 [00:01<00:01, 360.48it/s]
44%|████▍ | Scoring GeneralizingEstimator : 539/1225 [00:01<00:01, 365.22it/s]
45%|████▌ | Scoring GeneralizingEstimator : 552/1225 [00:01<00:01, 366.09it/s]
46%|████▌ | Scoring GeneralizingEstimator : 566/1225 [00:01<00:01, 368.75it/s]
47%|████▋ | Scoring GeneralizingEstimator : 579/1225 [00:01<00:01, 369.59it/s]
48%|████▊ | Scoring GeneralizingEstimator : 592/1225 [00:01<00:01, 370.17it/s]
49%|████▉ | Scoring GeneralizingEstimator : 604/1225 [00:01<00:01, 369.08it/s]
50%|█████ | Scoring GeneralizingEstimator : 614/1225 [00:01<00:01, 365.07it/s]
51%|█████ | Scoring GeneralizingEstimator : 627/1225 [00:01<00:01, 366.15it/s]
52%|█████▏ | Scoring GeneralizingEstimator : 641/1225 [00:01<00:01, 368.73it/s]
53%|█████▎ | Scoring GeneralizingEstimator : 652/1225 [00:01<00:01, 366.37it/s]
54%|█████▍ | Scoring GeneralizingEstimator : 667/1225 [00:01<00:01, 370.56it/s]
56%|█████▌ | Scoring GeneralizingEstimator : 682/1225 [00:01<00:01, 374.48it/s]
57%|█████▋ | Scoring GeneralizingEstimator : 694/1225 [00:01<00:01, 373.46it/s]
58%|█████▊ | Scoring GeneralizingEstimator : 707/1225 [00:01<00:01, 374.07it/s]
59%|█████▉ | Scoring GeneralizingEstimator : 721/1225 [00:01<00:01, 376.16it/s]
60%|█████▉ | Scoring GeneralizingEstimator : 731/1225 [00:01<00:01, 371.94it/s]
61%|██████ | Scoring GeneralizingEstimator : 744/1225 [00:01<00:01, 372.52it/s]
62%|██████▏ | Scoring GeneralizingEstimator : 759/1225 [00:01<00:01, 376.04it/s]
63%|██████▎ | Scoring GeneralizingEstimator : 770/1225 [00:02<00:01, 373.40it/s]
64%|██████▍ | Scoring GeneralizingEstimator : 783/1225 [00:02<00:01, 373.97it/s]
65%|██████▌ | Scoring GeneralizingEstimator : 797/1225 [00:02<00:01, 376.09it/s]
66%|██████▌ | Scoring GeneralizingEstimator : 808/1225 [00:02<00:01, 372.97it/s]
67%|██████▋ | Scoring GeneralizingEstimator : 823/1225 [00:02<00:01, 376.58it/s]
68%|██████▊ | Scoring GeneralizingEstimator : 837/1225 [00:02<00:01, 378.42it/s]
69%|██████▉ | Scoring GeneralizingEstimator : 847/1225 [00:02<00:01, 374.11it/s]
70%|██████▉ | Scoring GeneralizingEstimator : 856/1225 [00:02<00:01, 368.54it/s]
71%|███████ | Scoring GeneralizingEstimator : 865/1225 [00:02<00:00, 363.26it/s]
71%|███████ | Scoring GeneralizingEstimator : 872/1225 [00:02<00:00, 355.11it/s]
72%|███████▏ | Scoring GeneralizingEstimator : 880/1225 [00:02<00:00, 348.99it/s]
73%|███████▎ | Scoring GeneralizingEstimator : 889/1225 [00:02<00:00, 344.68it/s]
74%|███████▎ | Scoring GeneralizingEstimator : 901/1225 [00:02<00:00, 344.81it/s]
75%|███████▍ | Scoring GeneralizingEstimator : 915/1225 [00:02<00:00, 348.22it/s]
76%|███████▌ | Scoring GeneralizingEstimator : 929/1225 [00:02<00:00, 351.38it/s]
77%|███████▋ | Scoring GeneralizingEstimator : 939/1225 [00:02<00:00, 348.49it/s]
78%|███████▊ | Scoring GeneralizingEstimator : 950/1225 [00:02<00:00, 347.34it/s]
78%|███████▊ | Scoring GeneralizingEstimator : 961/1225 [00:02<00:00, 346.26it/s]
79%|███████▉ | Scoring GeneralizingEstimator : 972/1225 [00:02<00:00, 344.90it/s]
81%|████████ | Scoring GeneralizingEstimator : 987/1225 [00:02<00:00, 349.87it/s]
82%|████████▏ | Scoring GeneralizingEstimator : 1002/1225 [00:02<00:00, 354.65it/s]
83%|████████▎ | Scoring GeneralizingEstimator : 1014/1225 [00:02<00:00, 354.58it/s]
84%|████████▍ | Scoring GeneralizingEstimator : 1026/1225 [00:02<00:00, 354.53it/s]
85%|████████▍ | Scoring GeneralizingEstimator : 1040/1225 [00:02<00:00, 357.55it/s]
86%|████████▌ | Scoring GeneralizingEstimator : 1052/1225 [00:02<00:00, 357.41it/s]
87%|████████▋ | Scoring GeneralizingEstimator : 1063/1225 [00:02<00:00, 355.64it/s]
88%|████████▊ | Scoring GeneralizingEstimator : 1076/1225 [00:02<00:00, 357.02it/s]
89%|████████▊ | Scoring GeneralizingEstimator : 1087/1225 [00:02<00:00, 355.43it/s]
90%|████████▉ | Scoring GeneralizingEstimator : 1098/1225 [00:02<00:00, 353.90it/s]
91%|█████████ | Scoring GeneralizingEstimator : 1110/1225 [00:02<00:00, 353.94it/s]
92%|█████████▏| Scoring GeneralizingEstimator : 1122/1225 [00:03<00:00, 353.59it/s]
93%|█████████▎| Scoring GeneralizingEstimator : 1134/1225 [00:03<00:00, 353.60it/s]
94%|█████████▎| Scoring GeneralizingEstimator : 1146/1225 [00:03<00:00, 353.59it/s]
94%|█████████▍| Scoring GeneralizingEstimator : 1157/1225 [00:03<00:00, 352.18it/s]
96%|█████████▌| Scoring GeneralizingEstimator : 1170/1225 [00:03<00:00, 353.75it/s]
96%|█████████▋| Scoring GeneralizingEstimator : 1182/1225 [00:03<00:00, 353.74it/s]
97%|█████████▋| Scoring GeneralizingEstimator : 1193/1225 [00:03<00:00, 352.21it/s]
98%|█████████▊| Scoring GeneralizingEstimator : 1204/1225 [00:03<00:00, 350.84it/s]
99%|█████████▉| Scoring GeneralizingEstimator : 1217/1225 [00:03<00:00, 352.14it/s]
100%|██████████| Scoring GeneralizingEstimator : 1225/1225 [00:03<00:00, 349.12it/s]
100%|██████████| Scoring GeneralizingEstimator : 1225/1225 [00:03<00:00, 368.37it/s]
Plot
fig, ax = plt.subplots(layout="constrained")
im = ax.matshow(
scores,
vmin=0,
vmax=1.0,
cmap="RdBu_r",
origin="lower",
extent=epochs.times[[0, -1, 0, -1]],
)
ax.axhline(0.0, color="k")
ax.axvline(0.0, color="k")
ax.xaxis.set_ticks_position("bottom")
ax.set_xlabel(
'Condition: "Right"\nTesting Time (s)',
)
ax.set_ylabel('Condition: "Left"\nTraining Time (s)')
ax.set_title("Generalization across time and condition", fontweight="bold")
fig.colorbar(im, ax=ax, label="Performance (ROC AUC)")
plt.show()
References#
Total running time of the script: (0 minutes 8.578 seconds)
Estimated memory usage: 153 MB