We are not familiar with the Open Ephys binary format, our files are saved in .continuous, .spikes format.
Is there a way to covert them to binary?
Thanks.
This is the file format we record at Open Ephys:
https://open-ephys.github.io/gui-docs/User-Manual/Recording-data/Open-Ephys-format.html
My other question is whether files that contain spike sorted data can be processed.
Here is an example:
https://we.tl/t-tU8p2Dgokm
It looks like there is a Matlab function readily available for reading this file format (analysis-tools/load_open_ephys_data.m at master · open-ephys/analysis-tools · GitHub), so it should not be too complicated to add the support for it in Brainstorm easily.
Please share one or more (short) example files that include continuous signals and event markers.
If possible, try first that you can actually load them with the Matlab function above.
My other question is whether files that contain spike sorted data can be processed.
The Plexon .plx are supported for continuous signals only.
At the moment, it is not possible to import anything but continuous signals into the Brainstorm database. Spike waveforms and clustering results are not supported.
Here is an example file .continuous that I have already loaded in Matlab using the load_ephys_data.m function:
After saving it as .mat file I was able to import it in Brainstorm, but only as timestamps and when I try to view the content it looks like it's empty.
I managed to load the data part of the .mat file and I am doing spike sorting using WaveClus.
It is not clear how the threshold value can be modified in the set parameters file.
The red line running along my raw data is set at +2 , whereas is negative (~-3) in the tutorial example.
One more thing, this time regarding the events file:
I have opened your example .csv file from the Pack lab paper and it seems that there are 96 times for each of the 9 experiment conditions, but is not clear to what they correspond and they seem completely random.
Thanks.
Unfortunately, I can't help you with WavClus.
I have opened your example .csv file from the Pack lab paper and it seems that there are 96 times for each of the 9 experiment conditions, but is not clear to what they correspond and they seem completely random.
"9 new types of events at the end of the events list (Stim On 1-9) that correspond to the screen location of the presented stimulus"
https://neuroimage.usc.edu/brainstorm/e-phys/Introduction#Import_events
I have tried also the the other two spike sorting algorithms but I get the following errors:
1)Ultramegasort2000
"Error: Line 406: plot_waveforms
** Too many input arguments."
2)Kilosort
When using it wihout GPU acivated
Error: [process_spikesorting_kilosort] Electrophysiology > Unsupervised Spike Sorting > KiloSort
** Error with Kilosort while training on the CPU
Ok, that but in the link's page, the numbers in the events matrix what do they correspond to?
Adding the reading for this file format is more complicated than reading only this single .continuous file, as according to the description of the file format, the entire folder should be read as one entity, one file per channel.
The simplest approach is probably to convert these recordings to the multiplexed "binary" format.
Have you tried the library documented here?
https://groups.google.com/g/open-ephys/c/OAvY0Ln0YA4?pli=1
Sorry, I don't understand.
I read the single .continuous file that I shared here and then I was able to to import it as raw data file in Brainstorm.
Yes, one file can be read independently, one channel per file, and this single signal could be made accessible in the Brainstorm database.
But one .continuous file does not represent a recording session. A recording session is the entire folder, with multiple .continuous files, a .events file, a .openephys file...
If implementing the support for this file format in Brainstorm, we should include all this, including the possibility to import the demultiplexed .continuous data as a single multiplexed file. Otherwise it wouldn't be possible to process or display multiple channels together in Brainstorm.
This requires more work than what I can dedicate to this topic at the moment. Unfortunately, the person who coded this e-phys toolbox is not working with us anymore.
Can you please try to convert your dataset to a binary file that can be imported in Brainstorm?
Ok, I will try to convert them in .dat.
If you know any MatLab library that does the conversion, I would be grateful.
Thanks.
If you know any MatLab library that does the conversion, I would be grateful.
Unfortunately I don't.
If you find any, please let us know.
Hi,
I am doing some preprocessing of the e-phys raw .dat file that has many channels and I was wondering whether there is the option to substract the mean of all channels from the signal.
Thanks.
You can do this with an average reference montage: