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]
3%|▎ | Fitting GeneralizingEstimator : 1/35 [00:00<00:01, 29.20it/s]
9%|▊ | Fitting GeneralizingEstimator : 3/35 [00:00<00:00, 44.44it/s]
14%|█▍ | Fitting GeneralizingEstimator : 5/35 [00:00<00:00, 49.58it/s]
20%|██ | Fitting GeneralizingEstimator : 7/35 [00:00<00:00, 52.14it/s]
26%|██▌ | Fitting GeneralizingEstimator : 9/35 [00:00<00:00, 53.70it/s]
34%|███▍ | Fitting GeneralizingEstimator : 12/35 [00:00<00:00, 60.31it/s]
40%|████ | Fitting GeneralizingEstimator : 14/35 [00:00<00:00, 60.13it/s]
49%|████▊ | Fitting GeneralizingEstimator : 17/35 [00:00<00:00, 64.39it/s]
57%|█████▋ | Fitting GeneralizingEstimator : 20/35 [00:00<00:00, 67.69it/s]
63%|██████▎ | Fitting GeneralizingEstimator : 22/35 [00:00<00:00, 66.62it/s]
71%|███████▏ | Fitting GeneralizingEstimator : 25/35 [00:00<00:00, 69.20it/s]
77%|███████▋ | Fitting GeneralizingEstimator : 27/35 [00:00<00:00, 68.11it/s]
86%|████████▌ | Fitting GeneralizingEstimator : 30/35 [00:00<00:00, 70.22it/s]
91%|█████████▏| Fitting GeneralizingEstimator : 32/35 [00:00<00:00, 69.15it/s]
100%|██████████| Fitting GeneralizingEstimator : 35/35 [00:00<00:00, 72.09it/s]
100%|██████████| Fitting GeneralizingEstimator : 35/35 [00:00<00:00, 69.69it/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, 350.34it/s]
2%|▏ | Scoring GeneralizingEstimator : 29/1225 [00:00<00:02, 426.18it/s]
4%|▎ | Scoring GeneralizingEstimator : 44/1225 [00:00<00:02, 429.72it/s]
5%|▍ | Scoring GeneralizingEstimator : 60/1225 [00:00<00:02, 441.45it/s]
6%|▌ | Scoring GeneralizingEstimator : 76/1225 [00:00<00:02, 447.17it/s]
8%|▊ | Scoring GeneralizingEstimator : 92/1225 [00:00<00:02, 451.45it/s]
9%|▉ | Scoring GeneralizingEstimator : 108/1225 [00:00<00:02, 455.22it/s]
10%|█ | Scoring GeneralizingEstimator : 124/1225 [00:00<00:02, 456.88it/s]
11%|█▏ | Scoring GeneralizingEstimator : 140/1225 [00:00<00:02, 459.16it/s]
13%|█▎ | Scoring GeneralizingEstimator : 156/1225 [00:00<00:02, 460.92it/s]
14%|█▍ | Scoring GeneralizingEstimator : 173/1225 [00:00<00:02, 465.54it/s]
16%|█▌ | Scoring GeneralizingEstimator : 190/1225 [00:00<00:02, 466.03it/s]
17%|█▋ | Scoring GeneralizingEstimator : 207/1225 [00:00<00:02, 469.87it/s]
18%|█▊ | Scoring GeneralizingEstimator : 224/1225 [00:00<00:02, 472.77it/s]
20%|█▉ | Scoring GeneralizingEstimator : 241/1225 [00:00<00:02, 474.89it/s]
21%|██ | Scoring GeneralizingEstimator : 257/1225 [00:00<00:02, 474.68it/s]
22%|██▏ | Scoring GeneralizingEstimator : 274/1225 [00:00<00:01, 477.17it/s]
24%|██▎ | Scoring GeneralizingEstimator : 290/1225 [00:00<00:01, 476.82it/s]
25%|██▌ | Scoring GeneralizingEstimator : 307/1225 [00:00<00:01, 478.66it/s]
26%|██▋ | Scoring GeneralizingEstimator : 324/1225 [00:00<00:01, 480.59it/s]
28%|██▊ | Scoring GeneralizingEstimator : 340/1225 [00:00<00:01, 480.12it/s]
29%|██▉ | Scoring GeneralizingEstimator : 356/1225 [00:00<00:01, 479.70it/s]
30%|███ | Scoring GeneralizingEstimator : 372/1225 [00:00<00:01, 479.15it/s]
32%|███▏ | Scoring GeneralizingEstimator : 389/1225 [00:00<00:01, 480.86it/s]
33%|███▎ | Scoring GeneralizingEstimator : 405/1225 [00:00<00:01, 480.20it/s]
34%|███▍ | Scoring GeneralizingEstimator : 421/1225 [00:00<00:01, 479.79it/s]
36%|███▌ | Scoring GeneralizingEstimator : 438/1225 [00:00<00:01, 481.34it/s]
37%|███▋ | Scoring GeneralizingEstimator : 454/1225 [00:00<00:01, 480.84it/s]
38%|███▊ | Scoring GeneralizingEstimator : 471/1225 [00:00<00:01, 481.98it/s]
40%|███▉ | Scoring GeneralizingEstimator : 487/1225 [00:01<00:01, 480.78it/s]
41%|████ | Scoring GeneralizingEstimator : 503/1225 [00:01<00:01, 479.84it/s]
42%|████▏ | Scoring GeneralizingEstimator : 520/1225 [00:01<00:01, 481.12it/s]
44%|████▍ | Scoring GeneralizingEstimator : 537/1225 [00:01<00:01, 482.16it/s]
45%|████▌ | Scoring GeneralizingEstimator : 554/1225 [00:01<00:01, 483.23it/s]
47%|████▋ | Scoring GeneralizingEstimator : 570/1225 [00:01<00:01, 482.55it/s]
48%|████▊ | Scoring GeneralizingEstimator : 587/1225 [00:01<00:01, 483.61it/s]
49%|████▉ | Scoring GeneralizingEstimator : 604/1225 [00:01<00:01, 484.55it/s]
51%|█████ | Scoring GeneralizingEstimator : 621/1225 [00:01<00:01, 485.68it/s]
52%|█████▏ | Scoring GeneralizingEstimator : 637/1225 [00:01<00:01, 484.62it/s]
53%|█████▎ | Scoring GeneralizingEstimator : 654/1225 [00:01<00:01, 485.45it/s]
55%|█████▍ | Scoring GeneralizingEstimator : 671/1225 [00:01<00:01, 486.29it/s]
56%|█████▌ | Scoring GeneralizingEstimator : 688/1225 [00:01<00:01, 487.24it/s]
58%|█████▊ | Scoring GeneralizingEstimator : 705/1225 [00:01<00:01, 488.02it/s]
59%|█████▉ | Scoring GeneralizingEstimator : 721/1225 [00:01<00:01, 487.13it/s]
60%|██████ | Scoring GeneralizingEstimator : 738/1225 [00:01<00:00, 487.98it/s]
62%|██████▏ | Scoring GeneralizingEstimator : 755/1225 [00:01<00:00, 488.63it/s]
63%|██████▎ | Scoring GeneralizingEstimator : 771/1225 [00:01<00:00, 487.67it/s]
64%|██████▍ | Scoring GeneralizingEstimator : 788/1225 [00:01<00:00, 488.41it/s]
66%|██████▌ | Scoring GeneralizingEstimator : 804/1225 [00:01<00:00, 487.38it/s]
67%|██████▋ | Scoring GeneralizingEstimator : 821/1225 [00:01<00:00, 488.25it/s]
68%|██████▊ | Scoring GeneralizingEstimator : 837/1225 [00:01<00:00, 487.16it/s]
70%|██████▉ | Scoring GeneralizingEstimator : 854/1225 [00:01<00:00, 488.02it/s]
71%|███████ | Scoring GeneralizingEstimator : 871/1225 [00:01<00:00, 488.72it/s]
72%|███████▏ | Scoring GeneralizingEstimator : 887/1225 [00:01<00:00, 487.70it/s]
74%|███████▍ | Scoring GeneralizingEstimator : 904/1225 [00:01<00:00, 488.27it/s]
75%|███████▍ | Scoring GeneralizingEstimator : 917/1225 [00:01<00:00, 482.80it/s]
76%|███████▌ | Scoring GeneralizingEstimator : 932/1225 [00:01<00:00, 480.77it/s]
77%|███████▋ | Scoring GeneralizingEstimator : 948/1225 [00:01<00:00, 480.21it/s]
79%|███████▊ | Scoring GeneralizingEstimator : 964/1225 [00:02<00:00, 479.89it/s]
80%|████████ | Scoring GeneralizingEstimator : 980/1225 [00:02<00:00, 479.23it/s]
81%|████████▏ | Scoring GeneralizingEstimator : 996/1225 [00:02<00:00, 478.90it/s]
83%|████████▎ | Scoring GeneralizingEstimator : 1012/1225 [00:02<00:00, 478.64it/s]
84%|████████▍ | Scoring GeneralizingEstimator : 1028/1225 [00:02<00:00, 478.41it/s]
85%|████████▌ | Scoring GeneralizingEstimator : 1044/1225 [00:02<00:00, 478.15it/s]
87%|████████▋ | Scoring GeneralizingEstimator : 1060/1225 [00:02<00:00, 477.85it/s]
88%|████████▊ | Scoring GeneralizingEstimator : 1075/1225 [00:02<00:00, 475.87it/s]
89%|████████▉ | Scoring GeneralizingEstimator : 1092/1225 [00:02<00:00, 476.62it/s]
91%|█████████ | Scoring GeneralizingEstimator : 1109/1225 [00:02<00:00, 477.76it/s]
92%|█████████▏| Scoring GeneralizingEstimator : 1125/1225 [00:02<00:00, 477.02it/s]
93%|█████████▎| Scoring GeneralizingEstimator : 1140/1225 [00:02<00:00, 475.26it/s]
94%|█████████▍| Scoring GeneralizingEstimator : 1157/1225 [00:02<00:00, 476.39it/s]
96%|█████████▌| Scoring GeneralizingEstimator : 1173/1225 [00:02<00:00, 475.67it/s]
97%|█████████▋| Scoring GeneralizingEstimator : 1190/1225 [00:02<00:00, 477.06it/s]
98%|█████████▊| Scoring GeneralizingEstimator : 1202/1225 [00:02<00:00, 470.86it/s]
99%|█████████▉| Scoring GeneralizingEstimator : 1212/1225 [00:02<00:00, 461.82it/s]
100%|██████████| Scoring GeneralizingEstimator : 1225/1225 [00:02<00:00, 460.88it/s]
100%|██████████| Scoring GeneralizingEstimator : 1225/1225 [00:02<00:00, 474.67it/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 7.343 seconds)
Estimated memory usage: 175 MB