Incorrect Epoch Labelling

💬 Support & Discussions
, ,
1

I am processing an EEG signal acquired by showing 10 image of each of 40 different classes. I want to label the signals converted to epochs to their corresponding class labels.

raw = mne.io.read_raw_bdf(raw_file,  preload =True)
events=mne.find_events(raw, stim_channel = None,
                       initial_event = True, consecutive='increasing', output='step',
                       uint_cast=True, verbose=None)

The events return

array([[      0,       0,   65280],
       [   3085,   65280,   65533],
       [  44751,   65280,   65281],
       ...,
       [4962462,   65280,   65281],
       [4974818,   65280,   65281],
       [5021696,   65280,       0]], dtype=int64)

The third column from events is the event code, that are the same for all events 65281, except for initial event.
I made an event dict, by combining the events[2], and the class labels which were shown:

event_dictt = []
for event, class_label in zip(events, class_labels):

    event_dictt.append((class_label,event[2]))

print(event_dictt)

The event_dictt is a list that contains data

[('initial event', 65280), ('initial event', 65533), ('airliner', 65281), ('watch', 65281), ('folding chair', 65281), ('radio telescope', 65281), ('jack-o-lantern', 65281),

As i plot the epochs to view the class labelling

epochs.plot(n_channels=10, events=events, event_id=event_dictt)
I gives the error
event_id[0] must be an int, got <class ‘tuple’>

if i convert the event_dictt to a dictionary, the epochs.plot works but the epoch labelling is incorrect,The epochs are plotted but the label from the second last class label ‘Pool table’ is used only:

Screenshot 2023-08-29 121209
Screenshot 2023-08-29 1212091911×1018 329 KB

All initial class labels are ignored.

The epochs are not correctly labelled

2

Hello,

I assume the problem is that you re-use the exact same event code (65281) multiple times. There can only be exactly one event name per event code.

Best wishes,
Richard

2 Likes
3

@richard
yes, as my events array returns the same exact code 65281 for all the 400 events.
and the images are shown 10 ×40 from each 40 classes.
if i convert the data to dictionary, it eliminates the multiple copies, keeping only the last one.
how can i do the correct labelling then, i also have 400 diffent names of images shown to subject, can they be used.
please help

4

You need to modify your events array to reflect that. Different event types need to have different values in the third column of the array.

2 Likes