I have polysomnographic EEG data where each recording consists of about 6-8 hours worth of continuous EEG signals. I am running into issues when I try process these files, it seems to freeze at about 5% when I try run a band-pass filter on the raw data, for example. I assume this is because of the large file size. I would be happy to only use about 1 or 2 hours of the data for each recording, and am simply trying to find a way to reduce the file size prior to pre-processing. Is this possible?
You probably have ways to export smaller segments of recordings from the acquisition software.
Saving smaller files in the first place would be the simpler approach.
If you don't have access to the computer on which the data was recorded, you have some option for doing it in Brainstorm. You may face memory errors as well for 1 or 2 hours of recordings, in that case, try working with smaller blocks. Open a resource monitor on your computer to see at which point it starts filling all your memory: try to find solutions so that the memory usage never goes over 90% of the physical memory.
The procedure could be the following:
- Link the long 6-8hrs recordings to the database
- Right-click on the "Link to raw file" > Import MEG/EEG, select a time window of 1hr (from 0s to 3600s)
- Do not try to open the imported file, it would take forever to display it. Right-click on it > Review as raw.
- Repeat for all the segments you are interested in. Then delete the files imported at step 2 and the link created at step 1.
Thank you kindly for your assistance, Francois!