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, 29.15it/s]
11%|█▏ | Fitting GeneralizingEstimator : 4/35 [00:00<00:00, 59.43it/s]
17%|█▋ | Fitting GeneralizingEstimator : 6/35 [00:00<00:00, 59.36it/s]
23%|██▎ | Fitting GeneralizingEstimator : 8/35 [00:00<00:00, 59.32it/s]
31%|███▏ | Fitting GeneralizingEstimator : 11/35 [00:00<00:00, 65.81it/s]
43%|████▎ | Fitting GeneralizingEstimator : 15/35 [00:00<00:00, 75.66it/s]
49%|████▊ | Fitting GeneralizingEstimator : 17/35 [00:00<00:00, 72.95it/s]
57%|█████▋ | Fitting GeneralizingEstimator : 20/35 [00:00<00:00, 75.30it/s]
66%|██████▌ | Fitting GeneralizingEstimator : 23/35 [00:00<00:00, 77.11it/s]
74%|███████▍ | Fitting GeneralizingEstimator : 26/35 [00:00<00:00, 78.43it/s]
80%|████████ | Fitting GeneralizingEstimator : 28/35 [00:00<00:00, 76.21it/s]
89%|████████▊ | Fitting GeneralizingEstimator : 31/35 [00:00<00:00, 77.57it/s]
94%|█████████▍| Fitting GeneralizingEstimator : 33/35 [00:00<00:00, 75.68it/s]
100%|██████████| Fitting GeneralizingEstimator : 35/35 [00:00<00:00, 77.75it/s]
100%|██████████| Fitting GeneralizingEstimator : 35/35 [00:00<00:00, 76.38it/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 : 12/1225 [00:00<00:03, 351.29it/s]
2%|▏ | Scoring GeneralizingEstimator : 27/1225 [00:00<00:03, 398.17it/s]
3%|▎ | Scoring GeneralizingEstimator : 42/1225 [00:00<00:02, 413.35it/s]
4%|▍ | Scoring GeneralizingEstimator : 55/1225 [00:00<00:02, 405.29it/s]
6%|▌ | Scoring GeneralizingEstimator : 70/1225 [00:00<00:02, 413.43it/s]
7%|▋ | Scoring GeneralizingEstimator : 86/1225 [00:00<00:02, 424.69it/s]
8%|▊ | Scoring GeneralizingEstimator : 101/1225 [00:00<00:02, 427.06it/s]
10%|▉ | Scoring GeneralizingEstimator : 117/1225 [00:00<00:02, 433.90it/s]
11%|█ | Scoring GeneralizingEstimator : 132/1225 [00:00<00:02, 434.58it/s]
12%|█▏ | Scoring GeneralizingEstimator : 148/1225 [00:00<00:02, 439.09it/s]
13%|█▎ | Scoring GeneralizingEstimator : 164/1225 [00:00<00:02, 442.25it/s]
15%|█▍ | Scoring GeneralizingEstimator : 179/1225 [00:00<00:02, 441.70it/s]
16%|█▌ | Scoring GeneralizingEstimator : 195/1225 [00:00<00:02, 444.97it/s]
17%|█▋ | Scoring GeneralizingEstimator : 210/1225 [00:00<00:02, 444.66it/s]
18%|█▊ | Scoring GeneralizingEstimator : 226/1225 [00:00<00:02, 447.39it/s]
19%|█▉ | Scoring GeneralizingEstimator : 238/1225 [00:00<00:02, 439.15it/s]
20%|██ | Scoring GeneralizingEstimator : 250/1225 [00:00<00:02, 431.12it/s]
22%|██▏ | Scoring GeneralizingEstimator : 266/1225 [00:00<00:02, 434.61it/s]
23%|██▎ | Scoring GeneralizingEstimator : 282/1225 [00:00<00:02, 437.59it/s]
24%|██▍ | Scoring GeneralizingEstimator : 297/1225 [00:00<00:02, 438.04it/s]
26%|██▌ | Scoring GeneralizingEstimator : 313/1225 [00:00<00:02, 440.54it/s]
27%|██▋ | Scoring GeneralizingEstimator : 328/1225 [00:00<00:02, 440.70it/s]
28%|██▊ | Scoring GeneralizingEstimator : 344/1225 [00:00<00:01, 443.02it/s]
29%|██▉ | Scoring GeneralizingEstimator : 359/1225 [00:00<00:01, 442.95it/s]
31%|███ | Scoring GeneralizingEstimator : 374/1225 [00:00<00:01, 443.00it/s]
32%|███▏ | Scoring GeneralizingEstimator : 389/1225 [00:00<00:01, 442.87it/s]
33%|███▎ | Scoring GeneralizingEstimator : 404/1225 [00:00<00:01, 442.92it/s]
34%|███▍ | Scoring GeneralizingEstimator : 419/1225 [00:00<00:01, 442.97it/s]
35%|███▌ | Scoring GeneralizingEstimator : 430/1225 [00:00<00:01, 435.38it/s]
36%|███▌ | Scoring GeneralizingEstimator : 442/1225 [00:01<00:01, 430.31it/s]
37%|███▋ | Scoring GeneralizingEstimator : 457/1225 [00:01<00:01, 430.90it/s]
39%|███▊ | Scoring GeneralizingEstimator : 472/1225 [00:01<00:01, 431.64it/s]
40%|███▉ | Scoring GeneralizingEstimator : 487/1225 [00:01<00:01, 432.03it/s]
41%|████ | Scoring GeneralizingEstimator : 502/1225 [00:01<00:01, 432.59it/s]
42%|████▏ | Scoring GeneralizingEstimator : 517/1225 [00:01<00:01, 433.08it/s]
44%|████▎ | Scoring GeneralizingEstimator : 533/1225 [00:01<00:01, 434.96it/s]
45%|████▍ | Scoring GeneralizingEstimator : 548/1225 [00:01<00:01, 435.52it/s]
46%|████▌ | Scoring GeneralizingEstimator : 564/1225 [00:01<00:01, 437.46it/s]
47%|████▋ | Scoring GeneralizingEstimator : 579/1225 [00:01<00:01, 437.76it/s]
48%|████▊ | Scoring GeneralizingEstimator : 594/1225 [00:01<00:01, 438.05it/s]
50%|████▉ | Scoring GeneralizingEstimator : 609/1225 [00:01<00:01, 438.21it/s]
51%|█████ | Scoring GeneralizingEstimator : 625/1225 [00:01<00:01, 440.24it/s]
52%|█████▏ | Scoring GeneralizingEstimator : 641/1225 [00:01<00:01, 441.79it/s]
54%|█████▎ | Scoring GeneralizingEstimator : 657/1225 [00:01<00:01, 443.32it/s]
55%|█████▍ | Scoring GeneralizingEstimator : 672/1225 [00:01<00:01, 443.22it/s]
56%|█████▌ | Scoring GeneralizingEstimator : 688/1225 [00:01<00:01, 444.90it/s]
57%|█████▋ | Scoring GeneralizingEstimator : 704/1225 [00:01<00:01, 446.14it/s]
59%|█████▉ | Scoring GeneralizingEstimator : 720/1225 [00:01<00:01, 447.62it/s]
60%|██████ | Scoring GeneralizingEstimator : 736/1225 [00:01<00:01, 448.92it/s]
61%|██████▏ | Scoring GeneralizingEstimator : 752/1225 [00:01<00:01, 450.17it/s]
63%|██████▎ | Scoring GeneralizingEstimator : 768/1225 [00:01<00:01, 451.41it/s]
64%|██████▍ | Scoring GeneralizingEstimator : 784/1225 [00:01<00:00, 452.61it/s]
65%|██████▌ | Scoring GeneralizingEstimator : 800/1225 [00:01<00:00, 453.73it/s]
67%|██████▋ | Scoring GeneralizingEstimator : 816/1225 [00:01<00:00, 454.71it/s]
68%|██████▊ | Scoring GeneralizingEstimator : 833/1225 [00:01<00:00, 457.01it/s]
69%|██████▉ | Scoring GeneralizingEstimator : 849/1225 [00:01<00:00, 457.84it/s]
71%|███████ | Scoring GeneralizingEstimator : 865/1225 [00:01<00:00, 458.60it/s]
72%|███████▏ | Scoring GeneralizingEstimator : 881/1225 [00:01<00:00, 459.37it/s]
73%|███████▎ | Scoring GeneralizingEstimator : 897/1225 [00:02<00:00, 459.65it/s]
75%|███████▍ | Scoring GeneralizingEstimator : 913/1225 [00:02<00:00, 460.24it/s]
76%|███████▌ | Scoring GeneralizingEstimator : 929/1225 [00:02<00:00, 460.85it/s]
77%|███████▋ | Scoring GeneralizingEstimator : 945/1225 [00:02<00:00, 461.52it/s]
78%|███████▊ | Scoring GeneralizingEstimator : 961/1225 [00:02<00:00, 462.02it/s]
80%|███████▉ | Scoring GeneralizingEstimator : 977/1225 [00:02<00:00, 462.44it/s]
81%|████████ | Scoring GeneralizingEstimator : 993/1225 [00:02<00:00, 463.06it/s]
82%|████████▏ | Scoring GeneralizingEstimator : 1009/1225 [00:02<00:00, 463.63it/s]
84%|████████▎ | Scoring GeneralizingEstimator : 1025/1225 [00:02<00:00, 464.01it/s]
85%|████████▍ | Scoring GeneralizingEstimator : 1041/1225 [00:02<00:00, 464.40it/s]
86%|████████▋ | Scoring GeneralizingEstimator : 1057/1225 [00:02<00:00, 464.76it/s]
88%|████████▊ | Scoring GeneralizingEstimator : 1073/1225 [00:02<00:00, 465.19it/s]
89%|████████▉ | Scoring GeneralizingEstimator : 1090/1225 [00:02<00:00, 467.18it/s]
90%|█████████ | Scoring GeneralizingEstimator : 1106/1225 [00:02<00:00, 467.35it/s]
92%|█████████▏| Scoring GeneralizingEstimator : 1122/1225 [00:02<00:00, 467.71it/s]
93%|█████████▎| Scoring GeneralizingEstimator : 1138/1225 [00:02<00:00, 468.07it/s]
94%|█████████▍| Scoring GeneralizingEstimator : 1154/1225 [00:02<00:00, 468.34it/s]
96%|█████████▌| Scoring GeneralizingEstimator : 1170/1225 [00:02<00:00, 468.65it/s]
97%|█████████▋| Scoring GeneralizingEstimator : 1186/1225 [00:02<00:00, 468.96it/s]
98%|█████████▊| Scoring GeneralizingEstimator : 1202/1225 [00:02<00:00, 469.07it/s]
99%|█████████▉| Scoring GeneralizingEstimator : 1218/1225 [00:02<00:00, 469.29it/s]
100%|██████████| Scoring GeneralizingEstimator : 1225/1225 [00:02<00:00, 454.41it/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 7.556 seconds)
Estimated memory usage: 128 MB