Importing .cfs files from Signal

💬 Support & Discussions
1
  • MNE version: 1.7.1_0
  • operating system: Windows 11

Hi, I’m trying to use MNE for time-frequency analysis. My data is in .cfs form and it is collected using Signal. I’m trying to figure out the best way to import my data in a usable way. Will it work just with this file type? Do I need to somehow change the file type? I can export it as .txt or .mat as well.

2

Hey @squidguy3400

Assuming that MNE-Python doesn’t have a reader for this filetype, you want to take a look at Creating MNE-Python data structures from scratch — MNE 1.8.0.dev133+g868746b42 documentation

The gist is that you will need to:

  1. Pass a list of the channel names, a list of the channel types (e.g. “eeg”), and the sampling frequency, to mne.create_info
  2. Get your data into a numpy array. If you export your data to .mat format, you may be able to use pymatreader for this. Just make sure your numpy array has a shape (n_channels, n_samples).
  3. pass the info object you created in step 1, and the numpy array to mne.io.RawArray
1 Like
3

Thanks for the help. If I may ask, how can I change my numpy array shape? the current shape is (samples[450641], channels[11]) and I can’t seem to reshape it using

my_array.reshape([11,450641])

4

No worries. The easiest way is probably just to use the transpose method (.T for short):

import numpy as np

my_data = np.random.rand(450641, 11)

my_data = my_data.T

Make sure you are assigning the transposed array back into a variable, like I did above. Numpy doesn’t modify your array “in-place”, it always returns a copy.

EDIT: I probably shouldn’t say always copies, I just mean that when you do my_array.T or my_array.reshape(...), the result is not persistent, so you need to pass that result into some variable.

5

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.