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:    0.3s
[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,   29.10it/s]
  9%|▊         | Fitting GeneralizingEstimator : 3/35 [00:00<00:00,   43.51it/s]
 14%|█▍        | Fitting GeneralizingEstimator : 5/35 [00:00<00:00,   48.93it/s]
 20%|██        | Fitting GeneralizingEstimator : 7/35 [00:00<00:00,   51.59it/s]
 29%|██▊       | Fitting GeneralizingEstimator : 10/35 [00:00<00:00,   58.79it/s]
 37%|███▋      | Fitting GeneralizingEstimator : 13/35 [00:00<00:00,   64.37it/s]
 46%|████▌     | Fitting GeneralizingEstimator : 16/35 [00:00<00:00,   68.24it/s]
 54%|█████▍    | Fitting GeneralizingEstimator : 19/35 [00:00<00:00,   70.98it/s]
 60%|██████    | Fitting GeneralizingEstimator : 21/35 [00:00<00:00,   69.44it/s]
 69%|██████▊   | Fitting GeneralizingEstimator : 24/35 [00:00<00:00,   71.83it/s]
 74%|███████▍  | Fitting GeneralizingEstimator : 26/35 [00:00<00:00,   70.41it/s]
 80%|████████  | Fitting GeneralizingEstimator : 28/35 [00:00<00:00,   69.22it/s]
 89%|████████▊ | Fitting GeneralizingEstimator : 31/35 [00:00<00:00,   70.79it/s]
 94%|█████████▍| Fitting GeneralizingEstimator : 33/35 [00:00<00:00,   69.69it/s]
100%|██████████| Fitting GeneralizingEstimator : 35/35 [00:00<00:00,   70.61it/s]
100%|██████████| Fitting GeneralizingEstimator : 35/35 [00:00<00:00,   69.38it/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,  330.41it/s]
  2%|▏         | Scoring GeneralizingEstimator : 22/1225 [00:00<00:03,  313.46it/s]
  3%|▎         | Scoring GeneralizingEstimator : 34/1225 [00:00<00:03,  327.66it/s]
  4%|▍         | Scoring GeneralizingEstimator : 46/1225 [00:00<00:03,  335.07it/s]
  5%|▍         | Scoring GeneralizingEstimator : 58/1225 [00:00<00:03,  339.72it/s]
  6%|▌         | Scoring GeneralizingEstimator : 70/1225 [00:00<00:03,  342.82it/s]
  7%|▋         | Scoring GeneralizingEstimator : 82/1225 [00:00<00:03,  344.92it/s]
  8%|▊         | Scoring GeneralizingEstimator : 94/1225 [00:00<00:03,  327.57it/s]
  8%|▊         | Scoring GeneralizingEstimator : 104/1225 [00:00<00:03,  323.49it/s]
  9%|▉         | Scoring GeneralizingEstimator : 115/1225 [00:00<00:03,  323.55it/s]
 10%|█         | Scoring GeneralizingEstimator : 127/1225 [00:00<00:03,  327.21it/s]
 11%|█▏        | Scoring GeneralizingEstimator : 139/1225 [00:00<00:03,  330.25it/s]
 12%|█▏        | Scoring GeneralizingEstimator : 151/1225 [00:00<00:03,  332.85it/s]
 13%|█▎        | Scoring GeneralizingEstimator : 163/1225 [00:00<00:03,  335.00it/s]
 14%|█▍        | Scoring GeneralizingEstimator : 175/1225 [00:00<00:03,  336.83it/s]
 15%|█▌        | Scoring GeneralizingEstimator : 187/1225 [00:00<00:03,  338.38it/s]
 16%|█▌        | Scoring GeneralizingEstimator : 199/1225 [00:00<00:03,  339.89it/s]
 17%|█▋        | Scoring GeneralizingEstimator : 211/1225 [00:00<00:02,  341.22it/s]
 18%|█▊        | Scoring GeneralizingEstimator : 223/1225 [00:00<00:02,  342.33it/s]
 19%|█▉        | Scoring GeneralizingEstimator : 235/1225 [00:00<00:02,  343.41it/s]
 20%|██        | Scoring GeneralizingEstimator : 247/1225 [00:00<00:02,  344.36it/s]
 21%|██        | Scoring GeneralizingEstimator : 259/1225 [00:00<00:02,  345.14it/s]
 22%|██▏       | Scoring GeneralizingEstimator : 272/1225 [00:00<00:02,  345.95it/s]
 23%|██▎       | Scoring GeneralizingEstimator : 284/1225 [00:00<00:02,  346.60it/s]
 24%|██▍       | Scoring GeneralizingEstimator : 295/1225 [00:00<00:02,  345.21it/s]
 25%|██▌       | Scoring GeneralizingEstimator : 308/1225 [00:00<00:02,  347.78it/s]
 26%|██▌       | Scoring GeneralizingEstimator : 319/1225 [00:00<00:02,  346.34it/s]
 27%|██▋       | Scoring GeneralizingEstimator : 331/1225 [00:00<00:02,  346.98it/s]
 28%|██▊       | Scoring GeneralizingEstimator : 343/1225 [00:00<00:02,  347.50it/s]
 29%|██▉       | Scoring GeneralizingEstimator : 355/1225 [00:01<00:02,  347.99it/s]
 30%|███       | Scoring GeneralizingEstimator : 368/1225 [00:01<00:02,  348.66it/s]
 31%|███       | Scoring GeneralizingEstimator : 380/1225 [00:01<00:02,  349.09it/s]
 32%|███▏      | Scoring GeneralizingEstimator : 392/1225 [00:01<00:02,  349.48it/s]
 33%|███▎      | Scoring GeneralizingEstimator : 405/1225 [00:01<00:02,  351.56it/s]
 34%|███▍      | Scoring GeneralizingEstimator : 416/1225 [00:01<00:02,  341.97it/s]
 35%|███▍      | Scoring GeneralizingEstimator : 426/1225 [00:01<00:02,  339.32it/s]
 35%|███▌      | Scoring GeneralizingEstimator : 432/1225 [00:01<00:02,  330.09it/s]
 36%|███▌      | Scoring GeneralizingEstimator : 439/1225 [00:01<00:02,  322.59it/s]
 37%|███▋      | Scoring GeneralizingEstimator : 448/1225 [00:01<00:02,  319.40it/s]
 38%|███▊      | Scoring GeneralizingEstimator : 460/1225 [00:01<00:02,  321.19it/s]
 39%|███▊      | Scoring GeneralizingEstimator : 473/1225 [00:01<00:02,  324.78it/s]
 40%|███▉      | Scoring GeneralizingEstimator : 485/1225 [00:01<00:02,  326.45it/s]
 41%|████      | Scoring GeneralizingEstimator : 498/1225 [00:01<00:02,  329.70it/s]
 42%|████▏     | Scoring GeneralizingEstimator : 510/1225 [00:01<00:02,  331.14it/s]
 43%|████▎     | Scoring GeneralizingEstimator : 523/1225 [00:01<00:02,  333.97it/s]
 44%|████▎     | Scoring GeneralizingEstimator : 535/1225 [00:01<00:02,  335.15it/s]
 45%|████▍     | Scoring GeneralizingEstimator : 549/1225 [00:01<00:01,  338.52it/s]
 46%|████▌     | Scoring GeneralizingEstimator : 561/1225 [00:01<00:01,  339.47it/s]
 47%|████▋     | Scoring GeneralizingEstimator : 576/1225 [00:01<00:01,  345.09it/s]
 48%|████▊     | Scoring GeneralizingEstimator : 591/1225 [00:01<00:01,  350.38it/s]
 49%|████▉     | Scoring GeneralizingEstimator : 606/1225 [00:01<00:01,  355.41it/s]
 51%|█████     | Scoring GeneralizingEstimator : 621/1225 [00:01<00:01,  360.15it/s]
 52%|█████▏    | Scoring GeneralizingEstimator : 637/1225 [00:01<00:01,  365.17it/s]
 53%|█████▎    | Scoring GeneralizingEstimator : 652/1225 [00:01<00:01,  369.37it/s]
 55%|█████▍    | Scoring GeneralizingEstimator : 668/1225 [00:01<00:01,  374.91it/s]
 56%|█████▌    | Scoring GeneralizingEstimator : 682/1225 [00:01<00:01,  376.95it/s]
 57%|█████▋    | Scoring GeneralizingEstimator : 697/1225 [00:01<00:01,  380.52it/s]
 58%|█████▊    | Scoring GeneralizingEstimator : 711/1225 [00:01<00:01,  382.30it/s]
 59%|█████▉    | Scoring GeneralizingEstimator : 724/1225 [00:02<00:01,  382.41it/s]
 60%|██████    | Scoring GeneralizingEstimator : 738/1225 [00:02<00:01,  383.51it/s]
 61%|██████▏   | Scoring GeneralizingEstimator : 752/1225 [00:02<00:01,  385.13it/s]
 63%|██████▎   | Scoring GeneralizingEstimator : 766/1225 [00:02<00:01,  386.70it/s]
 64%|██████▎   | Scoring GeneralizingEstimator : 779/1225 [00:02<00:01,  386.66it/s]
 65%|██████▍   | Scoring GeneralizingEstimator : 792/1225 [00:02<00:01,  384.78it/s]
 66%|██████▌   | Scoring GeneralizingEstimator : 805/1225 [00:02<00:01,  384.83it/s]
 67%|██████▋   | Scoring GeneralizingEstimator : 819/1225 [00:02<00:01,  386.40it/s]
 68%|██████▊   | Scoring GeneralizingEstimator : 833/1225 [00:02<00:01,  387.91it/s]
 69%|██████▉   | Scoring GeneralizingEstimator : 847/1225 [00:02<00:00,  389.24it/s]
 70%|███████   | Scoring GeneralizingEstimator : 859/1225 [00:02<00:00,  387.51it/s]
 71%|███████   | Scoring GeneralizingEstimator : 871/1225 [00:02<00:00,  385.82it/s]
 72%|███████▏  | Scoring GeneralizingEstimator : 886/1225 [00:02<00:00,  388.86it/s]
 74%|███████▎  | Scoring GeneralizingEstimator : 901/1225 [00:02<00:00,  391.64it/s]
 75%|███████▍  | Scoring GeneralizingEstimator : 915/1225 [00:02<00:00,  392.86it/s]
 76%|███████▌  | Scoring GeneralizingEstimator : 930/1225 [00:02<00:00,  393.82it/s]
 77%|███████▋  | Scoring GeneralizingEstimator : 944/1225 [00:02<00:00,  394.92it/s]
 78%|███████▊  | Scoring GeneralizingEstimator : 958/1225 [00:02<00:00,  395.91it/s]
 79%|███████▉  | Scoring GeneralizingEstimator : 971/1225 [00:02<00:00,  395.40it/s]
 80%|████████  | Scoring GeneralizingEstimator : 984/1225 [00:02<00:00,  394.93it/s]
 81%|████████▏ | Scoring GeneralizingEstimator : 996/1225 [00:02<00:00,  392.84it/s]
 82%|████████▏ | Scoring GeneralizingEstimator : 1009/1225 [00:02<00:00,  392.47it/s]
 84%|████████▎ | Scoring GeneralizingEstimator : 1024/1225 [00:02<00:00,  395.08it/s]
 85%|████████▍ | Scoring GeneralizingEstimator : 1038/1225 [00:02<00:00,  396.12it/s]
 86%|████████▌ | Scoring GeneralizingEstimator : 1051/1225 [00:02<00:00,  394.88it/s]
 87%|████████▋ | Scoring GeneralizingEstimator : 1062/1225 [00:02<00:00,  391.44it/s]
 88%|████████▊ | Scoring GeneralizingEstimator : 1077/1225 [00:02<00:00,  394.12it/s]
 89%|████████▉ | Scoring GeneralizingEstimator : 1091/1225 [00:02<00:00,  395.15it/s]
 90%|█████████ | Scoring GeneralizingEstimator : 1108/1225 [00:02<00:00,  399.41it/s]
 92%|█████████▏| Scoring GeneralizingEstimator : 1121/1225 [00:03<00:00,  398.74it/s]
 93%|█████████▎| Scoring GeneralizingEstimator : 1135/1225 [00:03<00:00,  399.44it/s]
 94%|█████████▍| Scoring GeneralizingEstimator : 1149/1225 [00:03<00:00,  400.23it/s]
 95%|█████████▍| Scoring GeneralizingEstimator : 1163/1225 [00:03<00:00,  401.02it/s]
 96%|█████████▌| Scoring GeneralizingEstimator : 1175/1225 [00:03<00:00,  398.69it/s]
 97%|█████████▋| Scoring GeneralizingEstimator : 1187/1225 [00:03<00:00,  396.53it/s]
 98%|█████████▊| Scoring GeneralizingEstimator : 1201/1225 [00:03<00:00,  395.88it/s]
 99%|█████████▉| Scoring GeneralizingEstimator : 1215/1225 [00:03<00:00,  396.78it/s]
100%|██████████| Scoring GeneralizingEstimator : 1225/1225 [00:03<00:00,  398.31it/s]
100%|██████████| Scoring GeneralizingEstimator : 1225/1225 [00:03<00:00,  373.36it/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()
Generalization across time and condition

References#

Total running time of the script: (0 minutes 5.806 seconds)

Gallery generated by Sphinx-Gallery