Why data imported into Brainstorm was lost?

Bugs
1

Hi, all,
I recorded SEEG data in Neuroport system, the recording data's time duration and the data size were consistent, about 190 s.
However, the size of data (*.ns format) that I imported into Brainstorm was only 54 s. I can read this same data file in Neuroexplorer and the size is 190 s. What was wrong with it? Could someone give me any suggestions? Thanks a lot.

Clara

2

For reading the Blackrock NeuroPort files, we rely 100% on the NPMK toolbox.
https://github.com/BlackrockMicrosystems/NPMK/releases

Make sure you have the latest version of the library.
Then check what you obtain if you try to read the file directly with openNSx() function.
Example from the Brainstorm code:

If you can find ways to access the complete 190s-long signals with the NPMK function openNSx, there is probably something we can do on the Brainstorm side. In this case, please upload an example dataset somewhere (as small as possible) with which we can replicate the issue, and post the download link here.

If you can't, I recommend you contact the developers of the NPMK library by creating a new issue on their github repository, providing an example dataset to replicate the problem:
https://github.com/BlackrockMicrosystems/NPMK/issues

3

Hi, Francois,

I have updated the latest version of NPMK. I tried the 'openNSx' function, and it said as followed:
Screen Shot 2020-09-24 at 9.08.11 AM
Screen Shot 2020-09-24 at 9.08.11 AM936×233 25.5 KB

eeg = openNSx('/Users/panna/Desktop/Qiu_Rui_003.ns3')
The data is in unit of 1/4 �V. This mean that 100 in the NSx file equals to 25 �V. All values must be divided by 4.
To read the data in unit of �V, use openNSx('uv'). For more information type: help openNSx
Do you want NPMK to continue to ask you about this every time? yes

eeg =

struct with fields:

      MetaTags: [1×1 struct]
          Data: {[128×109340 int16]  [128×181 int16]  [128×836773 int16]}
       RawData: [1×1 struct]
ElectrodesInfo: [1×128 struct]

In the 'Data' information, the first matrix '[128×109340 int16]' is what I get in Brainstorm, about 54 s-data (sampling f=2K), however, the third matrix '[128×836773 int16]' is exactly the raw data size, about 418 s (not 190 s mentioned above, sorry for that). How can I get access of the 418 s-data in brainstorm? Why the data division happened?

Would you please offer me a e-mail address to upload you the *.ns data?

looking forward to your reply. Thanks a lot.

Clara

4

Why the data division happened?

I have no idea, this is something you could ask either to the manufacturer's customer service, or to the NPMK developers.

How can I get access of the 418 s-data in brainstorm?

I can probably find a way to identify these multiple data blocks from the header, and then read all of them as if they were continuous recordings (if you have jumps with no data in between blocks, you would lose them).

Would you please offer me a e-mail address to upload you the *.ns data?

You can send me a private message from this forum, by clicking on my username.

Thanks

5

Hi, Francois,

I have send the raw data to you via message from this form.

The data seemed to be segmented into three parts (three matrix viewed above) and it only could be read the first segment in brainstorm. I tried to merge the three parts via 'openNSx' function, but it failed. Would you please give me some suggestions? Perhaps could the brainstorm read only the third part of the raw data, since the third part contained most of the task information of eeg.

Thanks a lot and looking for your reply.

Clara

6

Hello,
I can't directly download this file without installing a client on my computer. Maybe there is a solution, but it's all in Chinese, so it's difficult to find.
Please upload your file on a website that allows a direct direct without any software installation (or tell me where I should click to download the file)
Thanks

7

Thank you for the example dataset.
It seems that the openNSx function reads seamlessly across multiple data blocks, it just returns the values in cell arrays when the requested segment spans over multiple blocks. I just added a test to concatenate the recordings from the different blocks if a cell-array is returned.

I also added events to indicate the beginning of the different blocks (Block01, Block02...)
And I fixed this annoying question that was asked in the command window.

Please update Brainstorm and try again linking this file to the database.

8

PS: Please double-check carefully the timings in the blocks >= 2, as I'm not very confident about what I did here.

9

Hi, Francois,

Sorry for the late reply. Thank you so much for fixing the problem. I updated the version of Brainstorm and I could import the full-length rawdata. Thanks for your consideration for labeled the three blocks as individual event.

Before this, I also consulted the engineer of BlackRock, it is replied that the data seemed to be interrupted during the recording due to some reason happed to the hardware, that is why there were three blocks of rawdata. Whaterver, it had been solved in Brainstorm.

Thanks again for your help.

Clara