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: 0.2s
[Parallel(n_jobs=1)]: Done 161 tasks | elapsed: 0.4s
[Parallel(n_jobs=1)]: Done 287 tasks | elapsed: 0.6s
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, 29.16it/s]
11%|█▏ | Fitting GeneralizingEstimator : 4/35 [00:00<00:00, 59.35it/s]
17%|█▋ | Fitting GeneralizingEstimator : 6/35 [00:00<00:00, 59.14it/s]
23%|██▎ | Fitting GeneralizingEstimator : 8/35 [00:00<00:00, 59.00it/s]
29%|██▊ | Fitting GeneralizingEstimator : 10/35 [00:00<00:00, 59.02it/s]
37%|███▋ | Fitting GeneralizingEstimator : 13/35 [00:00<00:00, 64.57it/s]
46%|████▌ | Fitting GeneralizingEstimator : 16/35 [00:00<00:00, 68.55it/s]
54%|█████▍ | Fitting GeneralizingEstimator : 19/35 [00:00<00:00, 71.54it/s]
60%|██████ | Fitting GeneralizingEstimator : 21/35 [00:00<00:00, 69.84it/s]
69%|██████▊ | Fitting GeneralizingEstimator : 24/35 [00:00<00:00, 72.16it/s]
74%|███████▍ | Fitting GeneralizingEstimator : 26/35 [00:00<00:00, 70.65it/s]
80%|████████ | Fitting GeneralizingEstimator : 28/35 [00:00<00:00, 69.39it/s]
86%|████████▌ | Fitting GeneralizingEstimator : 30/35 [00:00<00:00, 68.35it/s]
94%|█████████▍| Fitting GeneralizingEstimator : 33/35 [00:00<00:00, 70.31it/s]
100%|██████████| Fitting GeneralizingEstimator : 35/35 [00:00<00:00, 71.30it/s]
100%|██████████| Fitting GeneralizingEstimator : 35/35 [00:00<00:00, 70.14it/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 : 9/1225 [00:00<00:04, 263.14it/s]
2%|▏ | Scoring GeneralizingEstimator : 23/1225 [00:00<00:03, 339.12it/s]
3%|▎ | Scoring GeneralizingEstimator : 36/1225 [00:00<00:03, 353.20it/s]
4%|▍ | Scoring GeneralizingEstimator : 50/1225 [00:00<00:03, 368.93it/s]
5%|▌ | Scoring GeneralizingEstimator : 65/1225 [00:00<00:03, 383.80it/s]
6%|▋ | Scoring GeneralizingEstimator : 78/1225 [00:00<00:02, 383.92it/s]
7%|▋ | Scoring GeneralizingEstimator : 90/1225 [00:00<00:03, 378.04it/s]
8%|▊ | Scoring GeneralizingEstimator : 104/1225 [00:00<00:02, 382.84it/s]
10%|▉ | Scoring GeneralizingEstimator : 118/1225 [00:00<00:02, 387.17it/s]
11%|█ | Scoring GeneralizingEstimator : 131/1225 [00:00<00:02, 386.82it/s]
12%|█▏ | Scoring GeneralizingEstimator : 144/1225 [00:00<00:02, 385.91it/s]
13%|█▎ | Scoring GeneralizingEstimator : 158/1225 [00:00<00:02, 388.76it/s]
14%|█▍ | Scoring GeneralizingEstimator : 172/1225 [00:00<00:02, 391.33it/s]
15%|█▍ | Scoring GeneralizingEstimator : 183/1225 [00:00<00:02, 384.91it/s]
16%|█▌ | Scoring GeneralizingEstimator : 194/1225 [00:00<00:02, 378.97it/s]
17%|█▋ | Scoring GeneralizingEstimator : 209/1225 [00:00<00:02, 384.78it/s]
18%|█▊ | Scoring GeneralizingEstimator : 224/1225 [00:00<00:02, 389.51it/s]
20%|█▉ | Scoring GeneralizingEstimator : 240/1225 [00:00<00:02, 396.31it/s]
21%|██ | Scoring GeneralizingEstimator : 255/1225 [00:00<00:02, 400.13it/s]
22%|██▏ | Scoring GeneralizingEstimator : 267/1225 [00:00<00:02, 396.35it/s]
23%|██▎ | Scoring GeneralizingEstimator : 280/1225 [00:00<00:02, 395.44it/s]
24%|██▍ | Scoring GeneralizingEstimator : 292/1225 [00:00<00:02, 392.40it/s]
25%|██▌ | Scoring GeneralizingEstimator : 308/1225 [00:00<00:02, 398.05it/s]
26%|██▋ | Scoring GeneralizingEstimator : 322/1225 [00:00<00:02, 399.12it/s]
27%|██▋ | Scoring GeneralizingEstimator : 334/1225 [00:00<00:02, 395.73it/s]
28%|██▊ | Scoring GeneralizingEstimator : 343/1225 [00:00<00:02, 387.01it/s]
29%|██▊ | Scoring GeneralizingEstimator : 351/1225 [00:00<00:02, 376.85it/s]
29%|██▉ | Scoring GeneralizingEstimator : 360/1225 [00:00<00:02, 369.53it/s]
30%|███ | Scoring GeneralizingEstimator : 371/1225 [00:00<00:02, 366.54it/s]
31%|███▏ | Scoring GeneralizingEstimator : 385/1225 [00:01<00:02, 369.61it/s]
32%|███▏ | Scoring GeneralizingEstimator : 397/1225 [00:01<00:02, 368.57it/s]
33%|███▎ | Scoring GeneralizingEstimator : 409/1225 [00:01<00:02, 367.67it/s]
34%|███▍ | Scoring GeneralizingEstimator : 421/1225 [00:01<00:02, 366.89it/s]
35%|███▌ | Scoring GeneralizingEstimator : 434/1225 [00:01<00:02, 367.97it/s]
37%|███▋ | Scoring GeneralizingEstimator : 448/1225 [00:01<00:02, 370.75it/s]
38%|███▊ | Scoring GeneralizingEstimator : 462/1225 [00:01<00:02, 373.29it/s]
39%|███▊ | Scoring GeneralizingEstimator : 473/1225 [00:01<00:02, 370.33it/s]
40%|███▉ | Scoring GeneralizingEstimator : 485/1225 [00:01<00:02, 369.38it/s]
41%|████ | Scoring GeneralizingEstimator : 498/1225 [00:01<00:01, 370.10it/s]
42%|████▏ | Scoring GeneralizingEstimator : 510/1225 [00:01<00:01, 369.00it/s]
43%|████▎ | Scoring GeneralizingEstimator : 521/1225 [00:01<00:01, 366.48it/s]
44%|████▎ | Scoring GeneralizingEstimator : 534/1225 [00:01<00:01, 367.21it/s]
45%|████▍ | Scoring GeneralizingEstimator : 546/1225 [00:01<00:01, 366.19it/s]
46%|████▌ | Scoring GeneralizingEstimator : 559/1225 [00:01<00:01, 367.13it/s]
47%|████▋ | Scoring GeneralizingEstimator : 574/1225 [00:01<00:01, 371.05it/s]
48%|████▊ | Scoring GeneralizingEstimator : 588/1225 [00:01<00:01, 373.20it/s]
49%|████▉ | Scoring GeneralizingEstimator : 598/1225 [00:01<00:01, 368.99it/s]
50%|████▉ | Scoring GeneralizingEstimator : 608/1225 [00:01<00:01, 364.98it/s]
51%|█████ | Scoring GeneralizingEstimator : 619/1225 [00:01<00:01, 362.72it/s]
52%|█████▏ | Scoring GeneralizingEstimator : 634/1225 [00:01<00:01, 367.07it/s]
53%|█████▎ | Scoring GeneralizingEstimator : 650/1225 [00:01<00:01, 372.57it/s]
54%|█████▍ | Scoring GeneralizingEstimator : 664/1225 [00:01<00:01, 374.78it/s]
56%|█████▌ | Scoring GeneralizingEstimator : 680/1225 [00:01<00:01, 380.06it/s]
57%|█████▋ | Scoring GeneralizingEstimator : 695/1225 [00:01<00:01, 383.08it/s]
58%|█████▊ | Scoring GeneralizingEstimator : 710/1225 [00:01<00:01, 386.23it/s]
59%|█████▉ | Scoring GeneralizingEstimator : 725/1225 [00:01<00:01, 389.28it/s]
60%|██████ | Scoring GeneralizingEstimator : 739/1225 [00:01<00:01, 390.58it/s]
61%|██████▏ | Scoring GeneralizingEstimator : 753/1225 [00:01<00:01, 391.71it/s]
63%|██████▎ | Scoring GeneralizingEstimator : 767/1225 [00:02<00:01, 392.54it/s]
64%|██████▎ | Scoring GeneralizingEstimator : 780/1225 [00:02<00:01, 392.11it/s]
65%|██████▍ | Scoring GeneralizingEstimator : 795/1225 [00:02<00:01, 394.78it/s]
66%|██████▌ | Scoring GeneralizingEstimator : 810/1225 [00:02<00:01, 397.03it/s]
67%|██████▋ | Scoring GeneralizingEstimator : 821/1225 [00:02<00:01, 392.78it/s]
68%|██████▊ | Scoring GeneralizingEstimator : 834/1225 [00:02<00:00, 391.78it/s]
69%|██████▉ | Scoring GeneralizingEstimator : 847/1225 [00:02<00:00, 391.40it/s]
70%|███████ | Scoring GeneralizingEstimator : 860/1225 [00:02<00:00, 390.99it/s]
71%|███████ | Scoring GeneralizingEstimator : 871/1225 [00:02<00:00, 387.55it/s]
72%|███████▏ | Scoring GeneralizingEstimator : 882/1225 [00:02<00:00, 384.30it/s]
73%|███████▎ | Scoring GeneralizingEstimator : 895/1225 [00:02<00:00, 384.28it/s]
74%|███████▍ | Scoring GeneralizingEstimator : 906/1225 [00:02<00:00, 380.88it/s]
75%|███████▍ | Scoring GeneralizingEstimator : 914/1225 [00:02<00:00, 373.50it/s]
75%|███████▌ | Scoring GeneralizingEstimator : 924/1225 [00:02<00:00, 369.53it/s]
76%|███████▌ | Scoring GeneralizingEstimator : 934/1225 [00:02<00:00, 365.77it/s]
77%|███████▋ | Scoring GeneralizingEstimator : 943/1225 [00:02<00:00, 360.63it/s]
78%|███████▊ | Scoring GeneralizingEstimator : 952/1225 [00:02<00:00, 355.68it/s]
78%|███████▊ | Scoring GeneralizingEstimator : 961/1225 [00:02<00:00, 351.15it/s]
79%|███████▉ | Scoring GeneralizingEstimator : 970/1225 [00:02<00:00, 346.74it/s]
80%|███████▉ | Scoring GeneralizingEstimator : 979/1225 [00:02<00:00, 342.54it/s]
81%|████████ | Scoring GeneralizingEstimator : 992/1225 [00:02<00:00, 344.63it/s]
82%|████████▏ | Scoring GeneralizingEstimator : 1006/1225 [00:02<00:00, 347.85it/s]
83%|████████▎ | Scoring GeneralizingEstimator : 1017/1225 [00:02<00:00, 346.64it/s]
84%|████████▍ | Scoring GeneralizingEstimator : 1028/1225 [00:02<00:00, 345.59it/s]
85%|████████▍ | Scoring GeneralizingEstimator : 1040/1225 [00:02<00:00, 345.79it/s]
86%|████████▌ | Scoring GeneralizingEstimator : 1050/1225 [00:02<00:00, 343.23it/s]
86%|████████▋ | Scoring GeneralizingEstimator : 1059/1225 [00:02<00:00, 339.29it/s]
87%|████████▋ | Scoring GeneralizingEstimator : 1069/1225 [00:02<00:00, 336.86it/s]
88%|████████▊ | Scoring GeneralizingEstimator : 1080/1225 [00:02<00:00, 335.94it/s]
89%|████████▉ | Scoring GeneralizingEstimator : 1092/1225 [00:02<00:00, 336.87it/s]
90%|█████████ | Scoring GeneralizingEstimator : 1105/1225 [00:03<00:00, 339.28it/s]
91%|█████████▏| Scoring GeneralizingEstimator : 1119/1225 [00:03<00:00, 343.03it/s]
92%|█████████▏| Scoring GeneralizingEstimator : 1131/1225 [00:03<00:00, 343.65it/s]
93%|█████████▎| Scoring GeneralizingEstimator : 1145/1225 [00:03<00:00, 347.03it/s]
95%|█████████▍| Scoring GeneralizingEstimator : 1158/1225 [00:03<00:00, 348.47it/s]
95%|█████████▌| Scoring GeneralizingEstimator : 1167/1225 [00:03<00:00, 344.25it/s]
96%|█████████▌| Scoring GeneralizingEstimator : 1176/1225 [00:03<00:00, 340.29it/s]
97%|█████████▋| Scoring GeneralizingEstimator : 1187/1225 [00:03<00:00, 339.52it/s]
98%|█████████▊| Scoring GeneralizingEstimator : 1198/1225 [00:03<00:00, 338.80it/s]
99%|█████████▉| Scoring GeneralizingEstimator : 1212/1225 [00:03<00:00, 342.29it/s]
100%|█████████▉| Scoring GeneralizingEstimator : 1223/1225 [00:03<00:00, 341.41it/s]
100%|██████████| Scoring GeneralizingEstimator : 1225/1225 [00:03<00:00, 363.24it/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 8.677 seconds)
Estimated memory usage: 175 MB