Visualising statistical significance thresholds on EEG data

MNE-Python provides a range of tools for statistical hypothesis testing and the visualisation of the results. Here, we show a few options for exploratory and confirmatory tests - e.g., targeted t-tests, cluster-based permutation approaches (here with Threshold-Free Cluster Enhancement); and how to visualise the results.

The underlying data comes from 1; we contrast long vs. short words. TFCE is described in 2.

import numpy as np
import matplotlib.pyplot as plt
from scipy.stats import ttest_ind

import mne
from mne.channels import find_ch_adjacency, make_1020_channel_selections
from mne.stats import spatio_temporal_cluster_test

np.random.seed(0)

# Load the data
path = mne.datasets.kiloword.data_path() + '/kword_metadata-epo.fif'
epochs = mne.read_epochs(path)
name = "NumberOfLetters"

# Split up the data by the median length in letters via the attached metadata
median_value = str(epochs.metadata[name].median())
long_words = epochs[name + " > " + median_value]
short_words = epochs[name + " < " + median_value]

Out:

Reading /home/circleci/mne_data/MNE-kiloword-data/kword_metadata-epo.fif ...
Isotrak not found
    Found the data of interest:
        t =    -100.00 ...     920.00 ms
        0 CTF compensation matrices available
Adding metadata with 8 columns
Replacing existing metadata with 8 columns
960 matching events found
No baseline correction applied
0 projection items activated

If we have a specific point in space and time we wish to test, it can be convenient to convert the data into Pandas Dataframe format. In this case, the mne.Epochs object has a convenient mne.Epochs.to_data_frame() method, which returns a dataframe. This dataframe can then be queried for specific time windows and sensors. The extracted data can be submitted to standard statistical tests. Here, we conduct t-tests on the difference between long and short words.

time_windows = ((.2, .25), (.35, .45))
elecs = ["Fz", "Cz", "Pz"]
index = ['condition', 'epoch', 'time']

# display the EEG data in Pandas format (first 5 rows)
print(epochs.to_data_frame(index=index)[elecs].head())

report = "{elec}, time: {tmin}-{tmax} s; t({df})={t_val:.3f}, p={p:.3f}"
print("\nTargeted statistical test results:")
for (tmin, tmax) in time_windows:
    long_df = long_words.copy().crop(tmin, tmax).to_data_frame(index=index)
    short_df = short_words.copy().crop(tmin, tmax).to_data_frame(index=index)
    for elec in elecs:
        # extract data
        A = long_df[elec].groupby("condition").mean()
        B = short_df[elec].groupby("condition").mean()

        # conduct t test
        t, p = ttest_ind(A, B)

        # display results
        format_dict = dict(elec=elec, tmin=tmin, tmax=tmax,
                           df=len(epochs.events) - 2, t_val=t, p=p)
        print(report.format(**format_dict))

Out:

channel                     Fz  ...        Pz
condition epoch time            ...
film      0     -100  0.421970  ...  0.398182
                -96   0.453939  ...  0.222424
                -92   0.465606  ...  0.018485
                -88   0.468182  ... -0.173485
                -84   0.483485  ... -0.312121

[5 rows x 3 columns]

Targeted statistical test results:
Fz, time: 0.2-0.25 s; t(958)=-0.572, p=0.568
Cz, time: 0.2-0.25 s; t(958)=-2.836, p=0.005
Pz, time: 0.2-0.25 s; t(958)=-3.938, p=0.000
Fz, time: 0.35-0.45 s; t(958)=5.192, p=0.000
Cz, time: 0.35-0.45 s; t(958)=5.555, p=0.000
Pz, time: 0.35-0.45 s; t(958)=6.353, p=0.000

Absent specific hypotheses, we can also conduct an exploratory mass-univariate analysis at all sensors and time points. This requires correcting for multiple tests. MNE offers various methods for this; amongst them, cluster-based permutation methods allow deriving power from the spatio-temoral correlation structure of the data. Here, we use TFCE.

# Calculate adjacency matrix between sensors from their locations
adjacency, _ = find_ch_adjacency(epochs.info, "eeg")

# Extract data: transpose because the cluster test requires channels to be last
# In this case, inference is done over items. In the same manner, we could
# also conduct the test over, e.g., subjects.
X = [long_words.get_data().transpose(0, 2, 1),
     short_words.get_data().transpose(0, 2, 1)]
tfce = dict(start=.2, step=.2)

# Calculate statistical thresholds
t_obs, clusters, cluster_pv, h0 = spatio_temporal_cluster_test(
    X, tfce, adjacency=adjacency,
    n_permutations=100)  # a more standard number would be 1000+
significant_points = cluster_pv.reshape(t_obs.shape).T < .05
print(str(significant_points.sum()) + " points selected by TFCE ...")

Out:

Could not find a adjacency matrix for the data. Computing adjacency based on Delaunay triangulations.
-- number of adjacent vertices : 29
stat_fun(H1): min=0.000000 max=81.298503
Running initial clustering
Using 406 thresholds from 0.20 to 81.20 for TFCE computation (h_power=2.00, e_power=0.50)
Found 7424 clusters
Permuting 99 times...

  0%|          |  : 0/99 [00:00<?,       ?it/s]
  1%|1         |  : 1/99 [00:00<01:02,    1.57it/s]
  2%|2         |  : 2/99 [00:00<01:00,    1.60it/s]
  3%|3         |  : 3/99 [00:01<00:59,    1.61it/s]
  4%|4         |  : 4/99 [00:02<00:58,    1.62it/s]
  5%|5         |  : 5/99 [00:02<00:58,    1.61it/s]
  6%|6         |  : 6/99 [00:03<00:57,    1.61it/s]
  7%|7         |  : 7/99 [00:03<00:56,    1.63it/s]
  8%|8         |  : 8/99 [00:04<00:54,    1.67it/s]
  9%|9         |  : 9/99 [00:04<00:53,    1.69it/s]
 10%|#         |  : 10/99 [00:05<00:51,    1.71it/s]
 11%|#1        |  : 11/99 [00:05<00:50,    1.75it/s]
 12%|#2        |  : 12/99 [00:05<00:48,    1.78it/s]
 13%|#3        |  : 13/99 [00:06<00:48,    1.78it/s]
 14%|#4        |  : 14/99 [00:06<00:47,    1.80it/s]
 15%|#5        |  : 15/99 [00:07<00:47,    1.77it/s]
 16%|#6        |  : 16/99 [00:08<00:46,    1.77it/s]
 17%|#7        |  : 17/99 [00:08<00:45,    1.80it/s]
 18%|#8        |  : 18/99 [00:08<00:44,    1.83it/s]
 19%|#9        |  : 19/99 [00:09<00:43,    1.85it/s]
 20%|##        |  : 20/99 [00:09<00:42,    1.86it/s]
 21%|##1       |  : 21/99 [00:10<00:42,    1.83it/s]
 22%|##2       |  : 22/99 [00:10<00:41,    1.85it/s]
 23%|##3       |  : 23/99 [00:11<00:41,    1.85it/s]
 24%|##4       |  : 24/99 [00:12<00:40,    1.85it/s]
 25%|##5       |  : 25/99 [00:12<00:39,    1.89it/s]
 26%|##6       |  : 26/99 [00:12<00:38,    1.92it/s]
 27%|##7       |  : 27/99 [00:13<00:37,    1.91it/s]
 28%|##8       |  : 28/99 [00:13<00:36,    1.93it/s]
 29%|##9       |  : 29/99 [00:14<00:36,    1.93it/s]
 30%|###       |  : 30/99 [00:14<00:35,    1.93it/s]
 31%|###1      |  : 31/99 [00:15<00:34,    1.95it/s]
 32%|###2      |  : 32/99 [00:15<00:33,    1.98it/s]
 33%|###3      |  : 33/99 [00:15<00:32,    2.01it/s]
 34%|###4      |  : 34/99 [00:16<00:31,    2.05it/s]
 35%|###5      |  : 35/99 [00:16<00:31,    2.04it/s]
 36%|###6      |  : 36/99 [00:17<00:30,    2.06it/s]
 37%|###7      |  : 37/99 [00:17<00:30,    2.06it/s]
 38%|###8      |  : 38/99 [00:18<00:29,    2.07it/s]
 39%|###9      |  : 39/99 [00:18<00:29,    2.06it/s]
 40%|####      |  : 40/99 [00:19<00:28,    2.04it/s]
 41%|####1     |  : 41/99 [00:19<00:28,    2.03it/s]
 42%|####2     |  : 42/99 [00:20<00:27,    2.05it/s]
 43%|####3     |  : 43/99 [00:20<00:27,    2.07it/s]
 44%|####4     |  : 44/99 [00:20<00:26,    2.10it/s]
 45%|####5     |  : 45/99 [00:21<00:25,    2.12it/s]
 46%|####6     |  : 46/99 [00:21<00:24,    2.14it/s]
 47%|####7     |  : 47/99 [00:22<00:24,    2.11it/s]
 48%|####8     |  : 48/99 [00:22<00:24,    2.06it/s]
 49%|####9     |  : 49/99 [00:23<00:24,    2.05it/s]
 51%|#####     |  : 50/99 [00:23<00:23,    2.08it/s]
 52%|#####1    |  : 51/99 [00:24<00:22,    2.09it/s]
 53%|#####2    |  : 52/99 [00:24<00:22,    2.11it/s]
 54%|#####3    |  : 53/99 [00:25<00:21,    2.11it/s]
 55%|#####4    |  : 54/99 [00:25<00:21,    2.11it/s]
 56%|#####5    |  : 55/99 [00:25<00:20,    2.14it/s]
 57%|#####6    |  : 56/99 [00:26<00:20,    2.12it/s]
 58%|#####7    |  : 57/99 [00:27<00:20,    2.07it/s]
 59%|#####8    |  : 58/99 [00:27<00:20,    2.00it/s]
 60%|#####9    |  : 59/99 [00:28<00:20,    1.99it/s]
 61%|######    |  : 60/99 [00:28<00:19,    2.02it/s]
 62%|######1   |  : 61/99 [00:29<00:18,    2.06it/s]
 63%|######2   |  : 62/99 [00:29<00:17,    2.08it/s]
 64%|######3   |  : 63/99 [00:29<00:17,    2.09it/s]
 65%|######4   |  : 64/99 [00:30<00:16,    2.09it/s]
 66%|######5   |  : 65/99 [00:30<00:15,    2.14it/s]
 67%|######6   |  : 66/99 [00:31<00:15,    2.15it/s]
 68%|######7   |  : 67/99 [00:31<00:15,    2.11it/s]
 69%|######8   |  : 68/99 [00:32<00:14,    2.13it/s]
 70%|######9   |  : 69/99 [00:32<00:14,    2.05it/s]
 71%|#######   |  : 70/99 [00:33<00:14,    1.99it/s]
 72%|#######1  |  : 71/99 [00:34<00:14,    1.92it/s]
 73%|#######2  |  : 72/99 [00:35<00:14,    1.90it/s]
 74%|#######3  |  : 73/99 [00:35<00:13,    1.87it/s]
 75%|#######4  |  : 74/99 [00:36<00:13,    1.86it/s]
 76%|#######5  |  : 75/99 [00:37<00:12,    1.86it/s]
 77%|#######6  |  : 76/99 [00:37<00:12,    1.82it/s]
 78%|#######7  |  : 77/99 [00:38<00:11,    1.84it/s]
 79%|#######8  |  : 78/99 [00:38<00:11,    1.81it/s]
 80%|#######9  |  : 79/99 [00:39<00:10,    1.82it/s]
 81%|########  |  : 80/99 [00:40<00:10,    1.82it/s]
 82%|########1 |  : 81/99 [00:40<00:09,    1.82it/s]
 83%|########2 |  : 82/99 [00:41<00:09,    1.84it/s]
 84%|########3 |  : 83/99 [00:41<00:08,    1.85it/s]
 85%|########4 |  : 84/99 [00:42<00:08,    1.82it/s]
 86%|########5 |  : 85/99 [00:42<00:07,    1.84it/s]
 87%|########6 |  : 86/99 [00:43<00:07,    1.80it/s]
 88%|########7 |  : 87/99 [00:44<00:06,    1.77it/s]
 89%|########8 |  : 88/99 [00:44<00:06,    1.77it/s]
 90%|########9 |  : 89/99 [00:45<00:05,    1.77it/s]
 91%|######### |  : 90/99 [00:45<00:05,    1.80it/s]
 92%|#########1|  : 91/99 [00:46<00:04,    1.81it/s]
 93%|#########2|  : 92/99 [00:46<00:03,    1.80it/s]
 94%|#########3|  : 93/99 [00:47<00:03,    1.75it/s]
 95%|#########4|  : 94/99 [00:48<00:02,    1.77it/s]
 96%|#########5|  : 95/99 [00:48<00:02,    1.82it/s]
 97%|#########6|  : 96/99 [00:48<00:01,    1.83it/s]
 98%|#########7|  : 97/99 [00:49<00:01,    1.87it/s]
 99%|#########8|  : 98/99 [00:49<00:00,    1.90it/s]
100%|##########|  : 99/99 [00:49<00:00,    1.93it/s]
100%|##########|  : 99/99 [00:49<00:00,    1.98it/s]
Computing cluster p-values
Done.
1433 points selected by TFCE ...

The results of these mass univariate analyses can be visualised by plotting mne.Evoked objects as images (via mne.Evoked.plot_image) and masking points for significance. Here, we group channels by Regions of Interest to facilitate localising effects on the head.

# We need an evoked object to plot the image to be masked
evoked = mne.combine_evoked([long_words.average(), short_words.average()],
                            weights=[1, -1])  # calculate difference wave
time_unit = dict(time_unit="s")
evoked.plot_joint(title="Long vs. short words", ts_args=time_unit,
                  topomap_args=time_unit)  # show difference wave

# Create ROIs by checking channel labels
selections = make_1020_channel_selections(evoked.info, midline="12z")

# Visualize the results
fig, axes = plt.subplots(nrows=3, figsize=(8, 8))
axes = {sel: ax for sel, ax in zip(selections, axes.ravel())}
evoked.plot_image(axes=axes, group_by=selections, colorbar=False, show=False,
                  mask=significant_points, show_names="all", titles=None,
                  **time_unit)
plt.colorbar(axes["Left"].images[-1], ax=list(axes.values()), shrink=.3,
             label="µV")

plt.show()
  • 0.212 s, 0.324 s, 0.440 s
  • Left (8 channels, 83.6% of points masked), Left (13 channels, 79.3% of points masked), Left (8 channels, 80% of points masked)

Out:

No projector specified for this dataset. Please consider the method self.add_proj.

References

1

Dufau, S., Grainger, J., Midgley, KJ., Holcomb, PJ. A thousand words are worth a picture: Snapshots of printed-word processing in an event-related potential megastudy. Psychological Science, 2015

2

Smith and Nichols 2009, “Threshold-free cluster enhancement: addressing problems of smoothing, threshold dependence, and localisation in cluster inference”, NeuroImage 44 (2009) 83-98.

Total running time of the script: ( 0 minutes 57.457 seconds)

Estimated memory usage: 117 MB

Gallery generated by Sphinx-Gallery