Open dlewis326 opened 10 months ago
Hi @dlewis326, thanks for the request!
cc: @alejoe91 @samuelgarcia
I think this is the first formal request for NeuroNexus format here on NeuroConv, but it's been in discussion for a while on the upstream projects (SpikeInterface and Neo)
SpikeInterface: https://github.com/SpikeInterface/spikeinterface/discussions/1719
Neo: https://github.com/NeuralEnsemble/python-neo/issues/1296
In both cases it looks like what we need most is some example data - thanks for being willing to share some!
It also looks like aside from that the problem might be that the reader through their API is not completely open (need a license with them to initialize the client) which makes it really difficult or impossible for projects like ours to integrate this.
As a customer, you could request that they make such readers completely open. Other than that, we can try our best to figure it out once we see some example data but no guarantees that it will work.
Your system won't let me attach the .xdat files -- is there another way to send them to you?
The most common way is through Google Drive or Dropbox (or w/e similar platform you and your institution prefer); GitHub has pretty strict size and file type limits
@CodyCBakerPhD Unfortunately the xdat file requires a license fron NeuroNexus...https://github.com/SpikeInterface/spikeinterface/discussions/1719#discussioncomment-6183762
No example recordings yet (just setting it up) but it looks like _data.xdat is float32 (with last 6 chans reserved for aux or digital IO) and _timestamp.xdat is int64, plus there's a .json metadata file. Their matlab getAsKilosort2Matrix() just casts to int16 with the OpenEphys 0.195 uV/bit, +/- 6.38976mV range. I don't see the requirement for a license. Code is from: https://www.neuronexus.com/radiens-allego-download-page/ allego_file_reader.py.txt allegoXDatFileReaderR2019b.m.txt
I've started a reader for this format on Neo. I'm just waiting for permission to add the files shared with me to our gin testing suite for final compliance testing, but feel free to test the IO if you want while I wait for final testing.
What format would you like to see added to NeuroConv?
Acquisition using NeuroNexus Smartbox Pro and NeuroNexus Radiens Allego software. The software produces 3 files: a data file (.xdat), a timestamp file (.xdat), and a text file with metadata (.json).
Does the format have any documentation?
From the Radiens website: https://nnx.mcoutput.com/1376106/Content/Home-Radiens.htm Radiens saves primary time-series data sets (e.g., neural recordings) in a simple, open format. Each dataset (i.e. recording) is saved in a file set named ‘XDAT’ that consists of three files: (i) flat binary file containing the matrix of all sample values (typical in uV) across the multiple channels, (ii) flat binary file containing all timestamps, (iii) json text file containing enriched metadata.
Existing APIs for format
No. Of note, the Radiens website claims that .xdat files can be read by MatLab and Python, but I have been unsuccessful at reading .xdat files into open source spike sorting programs run through MatLab.
Do you have any example files you are willing to share?
Your system won't let me attach the .xdat files -- is there another way to send them to you? allego_6__uid1211-12-22-59.xdat.json
Do you have any interest in helping implement the feature?
No.
Code of Conduct