forked from NeuroTechX/moabb
-
Notifications
You must be signed in to change notification settings - Fork 0
/
p300.py
361 lines (293 loc) · 11.7 KB
/
p300.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
"""P300 Paradigms"""
import abc
import logging
import mne
import numpy as np
import pandas as pd
from moabb.datasets import utils
from moabb.datasets.fake import FakeDataset
from moabb.paradigms.base import BaseParadigm
log = logging.getLogger(__name__)
class BaseP300(BaseParadigm):
"""Base P300 paradigm.
Please use one of the child classes
Parameters
----------
filters: list of list (defaults [[7, 35]])
bank of bandpass filter to apply.
events: List of str | None (default None)
event to use for epoching. If None, default to all events defined in
the dataset.
tmin: float (default 0.0)
Start time (in second) of the epoch, relative to the dataset specific
task interval e.g. tmin = 1 would mean the epoch will start 1 second
after the beginning of the task as defined by the dataset.
tmax: float | None, (default None)
End time (in second) of the epoch, relative to the beginning of the
dataset specific task interval. tmax = 5 would mean the epoch will end
5 second after the beginning of the task as defined in the dataset. If
None, use the dataset value.
baseline: None | tuple of length 2
The time interval to consider as “baseline” when applying baseline
correction. If None, do not apply baseline correction.
If a tuple (a, b), the interval is between a and b (in seconds),
including the endpoints.
Correction is applied by computing the mean of the baseline period
and subtracting it from the data (see mne.Epochs)
channels: list of str | None (default None)
list of channel to select. If None, use all EEG channels available in
the dataset.
resample: float | None (default None)
If not None, resample the eeg data with the sampling rate provided.
"""
def __init__(
self,
filters=([1, 24],),
events=None,
tmin=0.0,
tmax=None,
baseline=None,
channels=None,
resample=None,
):
super().__init__()
self.filters = filters
self.events = events
self.channels = channels
self.baseline = baseline
self.resample = resample
if tmax is not None:
if tmin >= tmax:
raise (ValueError("tmax must be greater than tmin"))
self.tmin = tmin
self.tmax = tmax
def is_valid(self, dataset):
ret = True
if not (dataset.paradigm == "p300"):
ret = False
# check if dataset has required events
if self.events:
if not set(self.events) <= set(dataset.event_id.keys()):
ret = False
# we should verify list of channels, somehow
return ret
@abc.abstractmethod
def used_events(self, dataset):
pass
def process_raw( # noqa: C901
self, raw, dataset, return_epochs=False, return_raws=False
):
"""
Process one raw data file.
This function apply the preprocessing and eventual epoching on the
individual run, and return the data, labels and a dataframe with
metadata.
metadata is a dataframe with as many row as the length of the data
and labels.
Parameters
----------
raw: mne.Raw instance
the raw EEG data.
dataset : dataset instance
The dataset corresponding to the raw file. mainly use to access
dataset specific information.
return_epochs: boolean
This flag specifies whether to return only the data array or the
complete processed mne.Epochs
return_raws: boolean
To return raw files and events, to ensure compatibility with braindecode.
Mutually exclusive with return_epochs
returns
-------
X : Union[np.ndarray, mne.Epochs]
the data that will be used as features for the model
Note: if return_epochs=True, this is mne.Epochs
if return_epochs=False, this is np.ndarray
labels: np.ndarray
the labels for training / evaluating the model
metadata: pd.DataFrame
A dataframe containing the metadata
"""
if return_epochs and return_raws:
message = "Select only return_epochs or return_raws, not both"
raise ValueError(message)
# get events id
event_id = self.used_events(dataset)
# find the events, first check stim_channels then annotations
stim_channels = mne.utils._get_stim_channel(None, raw.info, raise_error=False)
if len(stim_channels) > 0:
events = mne.find_events(raw, shortest_event=0, verbose=False)
else:
try:
events, _ = mne.events_from_annotations(
raw, event_id=event_id, verbose=False
)
except ValueError:
log.warning(f"No matching annotations in {raw.filenames}")
return
# picks channels
if self.channels is None:
picks = mne.pick_types(raw.info, eeg=True, stim=False)
else:
picks = mne.pick_channels(
raw.info["ch_names"], include=self.channels, ordered=True
)
# pick events, based on event_id
try:
if "Target" in event_id and "NonTarget" in event_id:
if (
type(event_id["Target"]) is list
and type(event_id["NonTarget"]) == list
):
event_id_new = dict(Target=1, NonTarget=0)
events = mne.merge_events(events, event_id["Target"], 1)
events = mne.merge_events(events, event_id["NonTarget"], 0)
event_id = event_id_new
events = mne.pick_events(events, include=list(event_id.values()))
except RuntimeError:
# skip raw if no event found
return
if return_raws:
raw = raw.pick(picks)
else:
# get interval
tmin = self.tmin + dataset.interval[0]
if self.tmax is None:
tmax = dataset.interval[1]
else:
tmax = self.tmax + dataset.interval[0]
X = []
for bandpass in self.filters:
fmin, fmax = bandpass
# filter data
raw_f = raw.copy().filter(
fmin, fmax, method="iir", picks=picks, verbose=False
)
# epoch data
baseline = self.baseline
if baseline is not None:
baseline = (
self.baseline[0] + dataset.interval[0],
self.baseline[1] + dataset.interval[0],
)
bmin = baseline[0] if baseline[0] < tmin else tmin
bmax = baseline[1] if baseline[1] > tmax else tmax
else:
bmin = tmin
bmax = tmax
epochs = mne.Epochs(
raw_f,
events,
event_id=event_id,
tmin=bmin,
tmax=bmax,
proj=False,
baseline=baseline,
preload=True,
verbose=False,
picks=picks,
event_repeated="drop",
on_missing="ignore",
)
if bmin < tmin or bmax > tmax:
epochs.crop(tmin=tmin, tmax=tmax)
if self.resample is not None:
epochs = epochs.resample(self.resample)
# rescale to work with uV
if return_epochs:
X.append(epochs)
else:
X.append(dataset.unit_factor * epochs.get_data())
# overwrite events in case epochs have been dropped:
# (assuming all filters produce the same number of epochs...)
events = epochs.events
inv_events = {k: v for v, k in event_id.items()}
labels = np.array([inv_events[e] for e in events[:, -1]])
if return_epochs:
X = mne.concatenate_epochs(X)
elif return_raws:
X = raw
elif len(self.filters) == 1:
# if only one band, return a 3D array
X = X[0]
else:
# otherwise return a 4D
X = np.array(X).transpose((1, 2, 3, 0))
metadata = pd.DataFrame(index=range(len(labels)))
return X, labels, metadata
@property
def datasets(self):
if self.tmax is None:
interval = None
else:
interval = self.tmax - self.tmin
return utils.dataset_search(
paradigm="p300", events=self.events, interval=interval, has_all_events=True
)
@property
def scoring(self):
return "roc_auc"
class SinglePass(BaseP300):
"""Single Bandpass filter P300
P300 paradigm with only one bandpass filter (default 1 to 24 Hz)
Parameters
----------
fmin: float (default 1)
cutoff frequency (Hz) for the high pass filter
fmax: float (default 24)
cutoff frequency (Hz) for the low pass filter
events: List of str | None (default None)
event to use for epoching. If None, default to all events defined in
the dataset.
tmin: float (default 0.0)
Start time (in second) of the epoch, relative to the dataset specific
task interval e.g. tmin = 1 would mean the epoch will start 1 second
after the beginning of the task as defined by the dataset.
tmax: float | None, (default None)
End time (in second) of the epoch, relative to the beginning of the
dataset specific task interval. tmax = 5 would mean the epoch will end
5 second after the beginning of the task as defined in the dataset. If
None, use the dataset value.
baseline: None | tuple of length 2
The time interval to consider as “baseline” when applying baseline
correction. If None, do not apply baseline correction.
If a tuple (a, b), the interval is between a and b (in seconds),
including the endpoints.
Correction is applied by computing the mean of the baseline period
and subtracting it from the data (see mne.Epochs)
channels: list of str | None (default None)
list of channel to select. If None, use all EEG channels available in
the dataset.
resample: float | None (default None)
If not None, resample the eeg data with the sampling rate provided.
"""
def __init__(self, fmin=1, fmax=24, **kwargs):
if "filters" in kwargs.keys():
raise (ValueError("P300 does not take argument filters"))
super().__init__(filters=[[fmin, fmax]], **kwargs)
@property
def fmax(self):
return self.filters[0][1]
@property
def fmin(self):
return self.filters[0][0]
class P300(SinglePass):
"""P300 for Target/NonTarget classification
Metric is 'roc_auc'
"""
def __init__(self, **kwargs):
if "events" in kwargs.keys():
raise (ValueError("P300 dont accept events"))
super().__init__(events=["Target", "NonTarget"], **kwargs)
def used_events(self, dataset):
return {ev: dataset.event_id[ev] for ev in self.events}
@property
def scoring(self):
return "roc_auc"
class FakeP300Paradigm(P300):
"""Fake P300 for Target/NonTarget classification."""
@property
def datasets(self):
return [FakeDataset(["Target", "NonTarget"], paradigm="p300")]
def is_valid(self, dataset):
return True