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, 29.18it/s]
11%|█▏ | Fitting GeneralizingEstimator : 4/35 [00:00<00:00, 59.39it/s]
17%|█▋ | Fitting GeneralizingEstimator : 6/35 [00:00<00:00, 59.34it/s]
23%|██▎ | Fitting GeneralizingEstimator : 8/35 [00:00<00:00, 59.29it/s]
31%|███▏ | Fitting GeneralizingEstimator : 11/35 [00:00<00:00, 65.79it/s]
40%|████ | Fitting GeneralizingEstimator : 14/35 [00:00<00:00, 70.12it/s]
49%|████▊ | Fitting GeneralizingEstimator : 17/35 [00:00<00:00, 73.15it/s]
57%|█████▋ | Fitting GeneralizingEstimator : 20/35 [00:00<00:00, 75.41it/s]
63%|██████▎ | Fitting GeneralizingEstimator : 22/35 [00:00<00:00, 73.16it/s]
71%|███████▏ | Fitting GeneralizingEstimator : 25/35 [00:00<00:00, 75.04it/s]
77%|███████▋ | Fitting GeneralizingEstimator : 27/35 [00:00<00:00, 73.18it/s]
83%|████████▎ | Fitting GeneralizingEstimator : 29/35 [00:00<00:00, 71.62it/s]
91%|█████████▏| Fitting GeneralizingEstimator : 32/35 [00:00<00:00, 73.35it/s]
97%|█████████▋| Fitting GeneralizingEstimator : 34/35 [00:00<00:00, 71.94it/s]
100%|██████████| Fitting GeneralizingEstimator : 35/35 [00:00<00:00, 73.38it/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, 320.54it/s]
2%|▏ | Scoring GeneralizingEstimator : 26/1225 [00:00<00:03, 383.12it/s]
3%|▎ | Scoring GeneralizingEstimator : 41/1225 [00:00<00:02, 402.12it/s]
5%|▍ | Scoring GeneralizingEstimator : 57/1225 [00:00<00:02, 420.95it/s]
6%|▌ | Scoring GeneralizingEstimator : 72/1225 [00:00<00:02, 425.53it/s]
7%|▋ | Scoring GeneralizingEstimator : 88/1225 [00:00<00:02, 434.51it/s]
8%|▊ | Scoring GeneralizingEstimator : 103/1225 [00:00<00:02, 435.21it/s]
10%|▉ | Scoring GeneralizingEstimator : 119/1225 [00:00<00:02, 440.83it/s]
11%|█ | Scoring GeneralizingEstimator : 134/1225 [00:00<00:02, 441.28it/s]
12%|█▏ | Scoring GeneralizingEstimator : 150/1225 [00:00<00:02, 445.21it/s]
14%|█▎ | Scoring GeneralizingEstimator : 166/1225 [00:00<00:02, 447.91it/s]
15%|█▍ | Scoring GeneralizingEstimator : 181/1225 [00:00<00:02, 447.26it/s]
16%|█▌ | Scoring GeneralizingEstimator : 197/1225 [00:00<00:02, 450.03it/s]
17%|█▋ | Scoring GeneralizingEstimator : 212/1225 [00:00<00:02, 448.85it/s]
19%|█▊ | Scoring GeneralizingEstimator : 227/1225 [00:00<00:02, 448.45it/s]
20%|█▉ | Scoring GeneralizingEstimator : 243/1225 [00:00<00:02, 450.22it/s]
21%|██ | Scoring GeneralizingEstimator : 259/1225 [00:00<00:02, 452.15it/s]
22%|██▏ | Scoring GeneralizingEstimator : 274/1225 [00:00<00:02, 450.97it/s]
24%|██▎ | Scoring GeneralizingEstimator : 290/1225 [00:00<00:02, 452.66it/s]
25%|██▍ | Scoring GeneralizingEstimator : 306/1225 [00:00<00:02, 454.28it/s]
26%|██▋ | Scoring GeneralizingEstimator : 322/1225 [00:00<00:01, 455.56it/s]
28%|██▊ | Scoring GeneralizingEstimator : 338/1225 [00:00<00:01, 456.82it/s]
29%|██▉ | Scoring GeneralizingEstimator : 354/1225 [00:00<00:01, 457.95it/s]
30%|███ | Scoring GeneralizingEstimator : 370/1225 [00:00<00:01, 458.74it/s]
32%|███▏ | Scoring GeneralizingEstimator : 386/1225 [00:00<00:01, 459.51it/s]
33%|███▎ | Scoring GeneralizingEstimator : 401/1225 [00:00<00:01, 458.19it/s]
34%|███▍ | Scoring GeneralizingEstimator : 417/1225 [00:00<00:01, 458.93it/s]
35%|███▌ | Scoring GeneralizingEstimator : 433/1225 [00:00<00:01, 459.59it/s]
37%|███▋ | Scoring GeneralizingEstimator : 448/1225 [00:00<00:01, 458.47it/s]
38%|███▊ | Scoring GeneralizingEstimator : 464/1225 [00:01<00:01, 459.46it/s]
39%|███▉ | Scoring GeneralizingEstimator : 480/1225 [00:01<00:01, 459.80it/s]
40%|████ | Scoring GeneralizingEstimator : 495/1225 [00:01<00:01, 458.79it/s]
42%|████▏ | Scoring GeneralizingEstimator : 509/1225 [00:01<00:01, 456.05it/s]
43%|████▎ | Scoring GeneralizingEstimator : 523/1225 [00:01<00:01, 453.47it/s]
44%|████▍ | Scoring GeneralizingEstimator : 538/1225 [00:01<00:01, 452.61it/s]
45%|████▍ | Scoring GeneralizingEstimator : 551/1225 [00:01<00:01, 448.49it/s]
46%|████▌ | Scoring GeneralizingEstimator : 560/1225 [00:01<00:01, 436.71it/s]
47%|████▋ | Scoring GeneralizingEstimator : 571/1225 [00:01<00:01, 429.84it/s]
48%|████▊ | Scoring GeneralizingEstimator : 585/1225 [00:01<00:01, 426.00it/s]
49%|████▉ | Scoring GeneralizingEstimator : 599/1225 [00:01<00:01, 425.01it/s]
50%|████▉ | Scoring GeneralizingEstimator : 611/1225 [00:01<00:01, 420.87it/s]
51%|█████ | Scoring GeneralizingEstimator : 624/1225 [00:01<00:01, 418.74it/s]
52%|█████▏ | Scoring GeneralizingEstimator : 638/1225 [00:01<00:01, 418.46it/s]
53%|█████▎ | Scoring GeneralizingEstimator : 653/1225 [00:01<00:01, 419.75it/s]
54%|█████▍ | Scoring GeneralizingEstimator : 667/1225 [00:01<00:01, 419.35it/s]
56%|█████▌ | Scoring GeneralizingEstimator : 682/1225 [00:01<00:01, 420.71it/s]
57%|█████▋ | Scoring GeneralizingEstimator : 696/1225 [00:01<00:01, 420.02it/s]
58%|█████▊ | Scoring GeneralizingEstimator : 707/1225 [00:01<00:01, 414.77it/s]
59%|█████▊ | Scoring GeneralizingEstimator : 718/1225 [00:01<00:01, 409.75it/s]
60%|█████▉ | Scoring GeneralizingEstimator : 732/1225 [00:01<00:01, 409.81it/s]
61%|██████ | Scoring GeneralizingEstimator : 746/1225 [00:01<00:01, 410.07it/s]
62%|██████▏ | Scoring GeneralizingEstimator : 760/1225 [00:01<00:01, 410.03it/s]
63%|██████▎ | Scoring GeneralizingEstimator : 773/1225 [00:01<00:01, 408.46it/s]
64%|██████▍ | Scoring GeneralizingEstimator : 787/1225 [00:01<00:01, 408.80it/s]
65%|██████▌ | Scoring GeneralizingEstimator : 802/1225 [00:01<00:01, 410.62it/s]
66%|██████▋ | Scoring GeneralizingEstimator : 814/1225 [00:01<00:01, 407.45it/s]
68%|██████▊ | Scoring GeneralizingEstimator : 828/1225 [00:01<00:00, 407.81it/s]
69%|██████▉ | Scoring GeneralizingEstimator : 843/1225 [00:01<00:00, 409.30it/s]
70%|██████▉ | Scoring GeneralizingEstimator : 855/1225 [00:02<00:00, 406.37it/s]
71%|███████ | Scoring GeneralizingEstimator : 870/1225 [00:02<00:00, 408.36it/s]
72%|███████▏ | Scoring GeneralizingEstimator : 884/1225 [00:02<00:00, 408.24it/s]
73%|███████▎ | Scoring GeneralizingEstimator : 898/1225 [00:02<00:00, 408.53it/s]
74%|███████▍ | Scoring GeneralizingEstimator : 911/1225 [00:02<00:00, 407.09it/s]
76%|███████▌ | Scoring GeneralizingEstimator : 926/1225 [00:02<00:00, 408.62it/s]
77%|███████▋ | Scoring GeneralizingEstimator : 941/1225 [00:02<00:00, 410.45it/s]
78%|███████▊ | Scoring GeneralizingEstimator : 954/1225 [00:02<00:00, 409.03it/s]
79%|███████▉ | Scoring GeneralizingEstimator : 970/1225 [00:02<00:00, 412.27it/s]
80%|████████ | Scoring GeneralizingEstimator : 984/1225 [00:02<00:00, 412.15it/s]
82%|████████▏ | Scoring GeneralizingEstimator : 999/1225 [00:02<00:00, 413.70it/s]
83%|████████▎ | Scoring GeneralizingEstimator : 1014/1225 [00:02<00:00, 415.27it/s]
84%|████████▍ | Scoring GeneralizingEstimator : 1027/1225 [00:02<00:00, 413.44it/s]
85%|████████▌ | Scoring GeneralizingEstimator : 1043/1225 [00:02<00:00, 416.52it/s]
86%|████████▋ | Scoring GeneralizingEstimator : 1058/1225 [00:02<00:00, 417.74it/s]
88%|████████▊ | Scoring GeneralizingEstimator : 1073/1225 [00:02<00:00, 418.96it/s]
89%|████████▉ | Scoring GeneralizingEstimator : 1088/1225 [00:02<00:00, 420.26it/s]
90%|█████████ | Scoring GeneralizingEstimator : 1103/1225 [00:02<00:00, 421.49it/s]
91%|█████████▏| Scoring GeneralizingEstimator : 1118/1225 [00:02<00:00, 422.65it/s]
92%|█████████▏| Scoring GeneralizingEstimator : 1133/1225 [00:02<00:00, 423.76it/s]
94%|█████████▎| Scoring GeneralizingEstimator : 1148/1225 [00:02<00:00, 424.80it/s]
95%|█████████▍| Scoring GeneralizingEstimator : 1163/1225 [00:02<00:00, 425.70it/s]
96%|█████████▌| Scoring GeneralizingEstimator : 1179/1225 [00:02<00:00, 428.15it/s]
97%|█████████▋| Scoring GeneralizingEstimator : 1193/1225 [00:02<00:00, 427.10it/s]
98%|█████████▊| Scoring GeneralizingEstimator : 1205/1225 [00:02<00:00, 423.50it/s]
99%|█████████▉| Scoring GeneralizingEstimator : 1217/1225 [00:02<00:00, 420.04it/s]
100%|██████████| Scoring GeneralizingEstimator : 1225/1225 [00:02<00:00, 421.89it/s]
100%|██████████| Scoring GeneralizingEstimator : 1225/1225 [00:02<00:00, 425.93it/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 7.898 seconds)
Estimated memory usage: 129 MB