Hi!
I am recently working with EEG signals. I need to divide my signal (which is 30 minutes) into 2-second parts with an overlap of 50 present. Since EEG data is recorded in the rest state, there are no events reported in the raw data. I wonder if the epoching tool in the MNE package is the one that I should exploit or maybe there is another option. In fact, I’m in two minds if “windowing” and “epoching” are the same things.
Also, I have another question about how to feed the multi-patient EEG data to a neural network (take into account that the raw signal is multi-channel as well).
For epoching continuous data without events in overlapping windows, have a look at make_fixed_length_epochs().
Your second question seems a little vague… how you prepare the data depends a lot on what you want to learn from it. Maybe you can post a new question (with appropriate title) that includes more details about what your data is like and what you want the neural network to do with it.
1 Like
@drammock, many thanks for your response. I do appreciate it.
To clarify the second question I should say that In fact, I want to use a deep learning model to make a decision if patients have MCI or not according to the EEG data. I have 61 patient’s data and the recorded EEG signal has 19 channels.
I want to feed the data to a kind of 1D convolutional network. Actually, I’m not sure about the preparation way of the input dataset to my model. Should it be the data frame of all epoched data of 61 patients or maybe there is another option.
This increases your chances of receiving helpful responses.