Importing EGI Netstation Raw Data

Bugs
1

Hi Francois!

Help! We are trying to import Netstation Raw Data but brainstorm fails to detect events properly. It identifies >50,000 events, even though only a few dozens exist. The reason is that a single event is detected multiple times with time markers 1ms apart.

Editing events is extremely slow at this high number, and also saving an event file does not work.

An example dataset is available here:
http://dl.dropbox.com/u/4202951/Yuri%2020100728%201723001.raw.zip

Thank you!
Dimitrios

2

Hello Dimitrios,

I added the support for extended events (=more than one time sample) in EGI .raw files.
I added also the ability to import those extended events in the brainstorm database.
I started to work on the definition on bad segments of data in the raw viewer, that can be then tagged as bad trials once imported in the Brainstorm database.

I have not finished this work, but the part that was crashing your viewer is fixed (I hope).

Can you try it out and let me know if it works better?
Francois

3

Snapshot of what it looks like

test_dimi.jpg820×461 90.2 KB

4

Hi Francois,

Thank you for working on this, you are terrific!

I had a look at the updated brainstorm, and you can indeed load the extended events now and create related blocks. However, I am not sure this is what the students wanted here. I have a feeling they treat this data as normal triggers, considering only the event onset and not the event offset. In such case they create trials with baseline and post-stimulus period as all other events. However, I need to confirm this with them.

I thought to export these events and convert them to regular events by removing the event offset. However, the event exportation still does not work (File->Export all events), giving me this error:
??? Reference to non-existent field ‘raw’.

Error in ==> bst_events2track at 79
samples = samples -
double(sFile.header.raw.first_samp);

Error in ==> out_events_eve at 27
track = bst_events2track( sFile, 1 );

Error in ==> export_events at 119
out_events_eve(sFile, OutputFile);

Error in ==> panel_event>@(h,ev)export_events() at 95
gui_component(‘MenuItem’, jMenu, [], ‘Export all
events’, IconLoader.ICON_SAVE, [],
@(h,ev)export_events());

Any thoughts?

Also, I have noticed that when events are exported (from other data files), they do not maintain their name. For example, if you have 3 events: ‘FAC1’, ‘FAC2’, ‘FAC3’, then the exported event file will instead have numbers ‘1’, ‘2’, ‘3’. Is this a feature?

Thank you :slight_smile:
Dimitrios

5

Hey Dimitrios
I’m sorry, there is still a lot of debugging to be done with this viewer…

For you, today:

  1. I fixed the bug
  2. I added a menu to convert extended events to simple (in the events menu)
  3. This export you are talking about uses the .eve neuromag/mne format, which is too simple to represent the full set of parameters of an “event type”. It describes only one channel of trigger values, without names/colors, etc… It’s also impossible to have many segments intersecting.

In the next few days, while waiting for your following bug reports, I will add support for defining and managing bad segments of data.

Cheers
Francois

6

Hi Francois,

Super! Everything works great now. I was able to convert extended to single events, and then import the data into the database. Fantastic. I cannot thank you enough.

Best,
Dimitrios

7

Dear Francois,
We at the Sinha Lab, MIT- BCS, would like to thank you for your effort to fix the bug on the net station data label format.
We are currently expecting to provide further detailed feedback by next week.
Once Again, many thanks!
Best Regards,
Leo