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: 0.5s
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]
6%|▌ | Fitting GeneralizingEstimator : 2/35 [00:00<00:00, 57.58it/s]
11%|█▏ | Fitting GeneralizingEstimator : 4/35 [00:00<00:00, 58.28it/s]
17%|█▋ | Fitting GeneralizingEstimator : 6/35 [00:00<00:00, 58.62it/s]
26%|██▌ | Fitting GeneralizingEstimator : 9/35 [00:00<00:00, 66.65it/s]
34%|███▍ | Fitting GeneralizingEstimator : 12/35 [00:00<00:00, 71.49it/s]
43%|████▎ | Fitting GeneralizingEstimator : 15/35 [00:00<00:00, 74.73it/s]
51%|█████▏ | Fitting GeneralizingEstimator : 18/35 [00:00<00:00, 77.05it/s]
60%|██████ | Fitting GeneralizingEstimator : 21/35 [00:00<00:00, 78.81it/s]
69%|██████▊ | Fitting GeneralizingEstimator : 24/35 [00:00<00:00, 80.15it/s]
77%|███████▋ | Fitting GeneralizingEstimator : 27/35 [00:00<00:00, 81.19it/s]
83%|████████▎ | Fitting GeneralizingEstimator : 29/35 [00:00<00:00, 78.65it/s]
91%|█████████▏| Fitting GeneralizingEstimator : 32/35 [00:00<00:00, 79.76it/s]
100%|██████████| Fitting GeneralizingEstimator : 35/35 [00:00<00:00, 81.34it/s]
100%|██████████| Fitting GeneralizingEstimator : 35/35 [00:00<00:00, 79.86it/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 : 12/1225 [00:00<00:03, 350.76it/s]
2%|▏ | Scoring GeneralizingEstimator : 27/1225 [00:00<00:03, 397.61it/s]
4%|▎ | Scoring GeneralizingEstimator : 43/1225 [00:00<00:02, 421.62it/s]
5%|▍ | Scoring GeneralizingEstimator : 59/1225 [00:00<00:02, 434.94it/s]
6%|▌ | Scoring GeneralizingEstimator : 75/1225 [00:00<00:02, 443.41it/s]
7%|▋ | Scoring GeneralizingEstimator : 91/1225 [00:00<00:02, 448.66it/s]
9%|▊ | Scoring GeneralizingEstimator : 107/1225 [00:00<00:02, 452.10it/s]
10%|█ | Scoring GeneralizingEstimator : 123/1225 [00:00<00:02, 454.73it/s]
11%|█▏ | Scoring GeneralizingEstimator : 139/1225 [00:00<00:02, 456.47it/s]
13%|█▎ | Scoring GeneralizingEstimator : 155/1225 [00:00<00:02, 457.90it/s]
14%|█▍ | Scoring GeneralizingEstimator : 171/1225 [00:00<00:02, 459.34it/s]
15%|█▌ | Scoring GeneralizingEstimator : 187/1225 [00:00<00:02, 460.40it/s]
17%|█▋ | Scoring GeneralizingEstimator : 203/1225 [00:00<00:02, 461.43it/s]
18%|█▊ | Scoring GeneralizingEstimator : 219/1225 [00:00<00:02, 462.28it/s]
19%|█▉ | Scoring GeneralizingEstimator : 235/1225 [00:00<00:02, 463.10it/s]
21%|██ | Scoring GeneralizingEstimator : 252/1225 [00:00<00:02, 466.51it/s]
22%|██▏ | Scoring GeneralizingEstimator : 268/1225 [00:00<00:02, 467.01it/s]
23%|██▎ | Scoring GeneralizingEstimator : 284/1225 [00:00<00:02, 467.61it/s]
25%|██▍ | Scoring GeneralizingEstimator : 301/1225 [00:00<00:01, 470.23it/s]
26%|██▌ | Scoring GeneralizingEstimator : 317/1225 [00:00<00:01, 470.50it/s]
27%|██▋ | Scoring GeneralizingEstimator : 333/1225 [00:00<00:01, 470.52it/s]
29%|██▊ | Scoring GeneralizingEstimator : 350/1225 [00:00<00:01, 472.76it/s]
30%|██▉ | Scoring GeneralizingEstimator : 366/1225 [00:00<00:01, 472.82it/s]
31%|███▏ | Scoring GeneralizingEstimator : 383/1225 [00:00<00:01, 474.96it/s]
33%|███▎ | Scoring GeneralizingEstimator : 399/1225 [00:00<00:01, 474.91it/s]
34%|███▍ | Scoring GeneralizingEstimator : 416/1225 [00:00<00:01, 476.85it/s]
35%|███▌ | Scoring GeneralizingEstimator : 432/1225 [00:00<00:01, 476.43it/s]
37%|███▋ | Scoring GeneralizingEstimator : 449/1225 [00:00<00:01, 477.97it/s]
38%|███▊ | Scoring GeneralizingEstimator : 466/1225 [00:00<00:01, 479.64it/s]
39%|███▉ | Scoring GeneralizingEstimator : 482/1225 [00:01<00:01, 479.04it/s]
41%|████ | Scoring GeneralizingEstimator : 499/1225 [00:01<00:01, 480.50it/s]
42%|████▏ | Scoring GeneralizingEstimator : 516/1225 [00:01<00:01, 481.56it/s]
43%|████▎ | Scoring GeneralizingEstimator : 532/1225 [00:01<00:01, 481.07it/s]
45%|████▍ | Scoring GeneralizingEstimator : 549/1225 [00:01<00:01, 482.43it/s]
46%|████▌ | Scoring GeneralizingEstimator : 566/1225 [00:01<00:01, 483.70it/s]
48%|████▊ | Scoring GeneralizingEstimator : 583/1225 [00:01<00:01, 484.60it/s]
49%|████▉ | Scoring GeneralizingEstimator : 599/1225 [00:01<00:01, 483.83it/s]
50%|█████ | Scoring GeneralizingEstimator : 616/1225 [00:01<00:01, 484.92it/s]
52%|█████▏ | Scoring GeneralizingEstimator : 632/1225 [00:01<00:01, 484.32it/s]
53%|█████▎ | Scoring GeneralizingEstimator : 649/1225 [00:01<00:01, 485.16it/s]
54%|█████▍ | Scoring GeneralizingEstimator : 666/1225 [00:01<00:01, 486.23it/s]
56%|█████▌ | Scoring GeneralizingEstimator : 683/1225 [00:01<00:01, 487.02it/s]
57%|█████▋ | Scoring GeneralizingEstimator : 699/1225 [00:01<00:01, 486.24it/s]
58%|█████▊ | Scoring GeneralizingEstimator : 716/1225 [00:01<00:01, 487.19it/s]
60%|█████▉ | Scoring GeneralizingEstimator : 732/1225 [00:01<00:01, 486.42it/s]
61%|██████ | Scoring GeneralizingEstimator : 749/1225 [00:01<00:00, 487.33it/s]
63%|██████▎ | Scoring GeneralizingEstimator : 766/1225 [00:01<00:00, 487.91it/s]
64%|██████▍ | Scoring GeneralizingEstimator : 782/1225 [00:01<00:00, 487.09it/s]
65%|██████▌ | Scoring GeneralizingEstimator : 799/1225 [00:01<00:00, 487.96it/s]
67%|██████▋ | Scoring GeneralizingEstimator : 816/1225 [00:01<00:00, 488.41it/s]
68%|██████▊ | Scoring GeneralizingEstimator : 832/1225 [00:01<00:00, 487.50it/s]
69%|██████▉ | Scoring GeneralizingEstimator : 849/1225 [00:01<00:00, 488.24it/s]
71%|███████ | Scoring GeneralizingEstimator : 865/1225 [00:01<00:00, 487.35it/s]
72%|███████▏ | Scoring GeneralizingEstimator : 882/1225 [00:01<00:00, 488.21it/s]
73%|███████▎ | Scoring GeneralizingEstimator : 899/1225 [00:01<00:00, 488.67it/s]
75%|███████▍ | Scoring GeneralizingEstimator : 915/1225 [00:01<00:00, 487.86it/s]
76%|███████▌ | Scoring GeneralizingEstimator : 932/1225 [00:01<00:00, 488.68it/s]
77%|███████▋ | Scoring GeneralizingEstimator : 948/1225 [00:01<00:00, 487.73it/s]
79%|███████▉ | Scoring GeneralizingEstimator : 965/1225 [00:02<00:00, 488.58it/s]
80%|████████ | Scoring GeneralizingEstimator : 982/1225 [00:02<00:00, 489.07it/s]
81%|████████▏ | Scoring GeneralizingEstimator : 998/1225 [00:02<00:00, 488.17it/s]
83%|████████▎ | Scoring GeneralizingEstimator : 1015/1225 [00:02<00:00, 488.97it/s]
84%|████████▍ | Scoring GeneralizingEstimator : 1032/1225 [00:02<00:00, 489.30it/s]
86%|████████▌ | Scoring GeneralizingEstimator : 1048/1225 [00:02<00:00, 488.22it/s]
87%|████████▋ | Scoring GeneralizingEstimator : 1065/1225 [00:02<00:00, 489.00it/s]
88%|████████▊ | Scoring GeneralizingEstimator : 1081/1225 [00:02<00:00, 488.12it/s]
90%|████████▉ | Scoring GeneralizingEstimator : 1098/1225 [00:02<00:00, 488.89it/s]
91%|█████████ | Scoring GeneralizingEstimator : 1114/1225 [00:02<00:00, 487.97it/s]
92%|█████████▏| Scoring GeneralizingEstimator : 1131/1225 [00:02<00:00, 488.72it/s]
94%|█████████▎| Scoring GeneralizingEstimator : 1148/1225 [00:02<00:00, 489.03it/s]
95%|█████████▌| Scoring GeneralizingEstimator : 1164/1225 [00:02<00:00, 488.17it/s]
96%|█████████▋| Scoring GeneralizingEstimator : 1181/1225 [00:02<00:00, 488.96it/s]
98%|█████████▊| Scoring GeneralizingEstimator : 1198/1225 [00:02<00:00, 489.57it/s]
99%|█████████▉| Scoring GeneralizingEstimator : 1214/1225 [00:02<00:00, 488.72it/s]
100%|██████████| Scoring GeneralizingEstimator : 1225/1225 [00:02<00:00, 489.44it/s]
100%|██████████| Scoring GeneralizingEstimator : 1225/1225 [00:02<00:00, 483.58it/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 7.222 seconds)
Estimated memory usage: 175 MB