Dear developers & community,
We are recording EEG with “NeurOne” amplifiers (Mega Electronics Ltd, Finland) which has its own output format. I would love to analyze the data with Brainstorm, however there seems to be no easy way to import the data into brainstorm currently.
Do you have ideas how to do that, or if it is possible to include the format into the toolbox’s import function?
I uploaded a short example measurement with EEG, EMG, EOG, and ECG data with electrical stimulation triggers (note: only ECG and triggers are really connected here, the rest is noise):
https://www.dropbox.com/s/gfa2ogjx7eu5zct/BrainstormECGwithTriggers.zip?dl=1
The folder structure is as followed:
The outer folder is named after experiment & subject (in this case “BrainstormECGwithTriggers”). In that, folders with date and time (e.g. “2015-09-14T161619”) are created by the software for each session, i.e. each time the measurement is stopped completely (e.g. for a lunch break). Again in there, the individual runs are stored within numbered folders (“1”, “2” etc.); these are created each time the recording is paused, e.g. for small breaks.
It would be important to be able to import all data of one subject together, i.e. all data within the “BrainstormECGwithTriggers” folder including the two sessions.
Thanks in advance for any help!
Best,
Martin
There is also a general Matlab import plugin for NeurOne data (attached here), which could help solving the problem.
However, one problem with this function is that it only reads individual session folders. To import all data from one experiment with multiple session folders, one has to concatenate them manually.
Best,
Martin
Hi Martin,
The existing reading functions are equipped to read recordings split in multiple .bin files (one folder containing 1.bin, 2.bin, 3.bin, etc).
Is this something that you need?
If so, could you send me an example of such a session (with at least two files)?
Thanks,
Francois
Hi Francois,
I don’t even know how more than 1 bin file per folder can occur … every time we pause the recording, we get a new numbered folder. So, we always have only one bin-file per folder: 1.bin.
Thus, we would not need this function.
Best,
Martin
Hi Martin,
It’s ready, you can update Brainstorm and try it out (menu Help > Update Brainstorm).
Use the menu “Review raw file” and file format “EEG: NeurOne session folder”.
The reading of the XML is awfully slow but works. The function to optimize to make this faster would be brainstorm3/toolbox/io/in_xml.m
The various sessions and phases are handled as different files because it was to complicated to contactenate them on the fly. Time gaps were going to make the files incompatible with most pre-processing functions in Brainstorm.
The different types of channels are recognized separately (EEG, EMG, ECG, EOG).
The new reading functions are: in_fopen_neurone.m, in_fread_neurone.m
They are distantly inspired from the Matlab toolbox, but I recoded almost everything.
Therefore the result is probably full of bugs. Please take some time to make sure that the events and recordings are read correctly.
Cheers,
Francois
Hi Francois,
first of all: Thanks so much for solving this so quickly! I really appreciate your work.
Using the option “Review raw file” and epoching it afterwards, I found no bug so far, everything seems to work, the data looks fine. It’s also fine that the files stay seperate, because they are still loaded automatically and I can group the events later on.
One thing I noticed: If I instead use the “Import MEG/EEG” option and choose which event types to import, it only loads the number of events which are present in the first .bin-file for each file, even if there are more events in the subsequent .bin-files. This might be causing some confusion if somebody tries to import the data in that way. For me, it doesn’t matter, because I anyway like to use the “Review raw file” option first.
Best,
Martin
[QUOTE=MartinFR;9265]Hi Francois,
I don’t even know how more than 1 bin file per folder can occur … every time we pause the recording, we get a new numbered folder. So, we always have only one bin-file per folder: 1.bin.
Thus, we would not need this function.
Best,
Martin[/QUOTE]
Hi, this happens when one specifiers a maximal file size.
Martin: Thank you for the feedback. Ideally, these recordings should always be manipulated phase by phase (folders “1”, “2”, etc) instead of session by session (including multiple phases). These are in fact separate recording sessions. Brainstorm is not equipped to handle non-continuous recordings.
Mahelita: If you or someone else are interested in reading recordings split in multiple .bin files, please send me an example. If it is not urgent, we can leave it like this for now until it’s needed.
Cheers,
Francois
[QUOTE=Francois;9277]
Mahelita: If you or someone else are interested in reading recordings split in multiple .bin files, please send me an example. If it is not urgent, we can leave it like this for now until it’s needed.
Francois[/QUOTE]
I guess with 64-bit systems splitting files is obsolete?
We do not use this feature and do not plan to.
Thanks a lot for the super fast implementation!