Spikeglx with onebox file format

[samuelgarcia]

Spikeglx also handle the onebox device.
The auxilary file looks like this

<run name>_g<gate index>_t<trial index>.obx<OneBox index>.obx.bin

acqXaDwSy=12,1,1
appVersion=20250201
fileCreateTime=2025-04-04T18:30:39
fileName=C:/SGL_DATA/myRun_g0/myRun_g0_t0.obx0.obx.bin
fileSHA1=DBD1CC02EDEE0F63AAF881869E1F3633631FA999
fileSizeBytes=6537300
fileTimeSecs=7.704682704682705
firstSample=252273
gateMode=Immediate
imDatApi=3.70.3
imDatBs_fw=3.3.134
imDatBsc_fw=0.0.0
imDatBsc_hw=4.3
imDatBsc_pn=NPOB_ACQ_00-04-03
imDatBsc_sn=21470089
imDatObx_slot=21
imTrgRising=true
imTrgSource=0
nDataDirs=1
nSavedChans=14
obAiRangeMax=5
obAiRangeMin=-5
obMaxInt=32768
obSampRate=30303
snsSaveChanSubset=all
snsXaDwSy=12,1,1
syncSourceIdx=0
syncSourcePeriod=1
trigMode=Immediate
typeImEnabled=0
typeNiEnabled=0
typeObEnabled=1
typeThis=obx
userNotes=
~snsChanMap=(12,1,1)(XA0;0:0)(XA1;1:1)(XA2;2:2)(XA3;3:3)(XA4;4:4)(XA5;5:5)(XA6;6:6)(XA7;7:7)(XA8;8:8)(XA9;9:9)(XA10;10:10)(XA11;11:11)(XD0;12:12)(SY0;13:13)

@jenniferColonell
@alejoe91
@h-mayorquin

[h-mayorquin]

Yes, it would be great to support this. We need testing data.

[jenniferColonell]

Hi -- Here's the same small test run I sent to Alessio via slack.

myRun_g0.zip

[h-mayorquin]

Thanks, any info that you could provide for the file so we can add that in the provenance readme of the testing data repo?

[jenniferColonell]

It's just a few seconds of data, collecting from all 12 OneBox ADC channels with inputs floating.
I could make you a more meaningful dataset (e.g. with signals). If anything feels unclear about reading these data, I will do that.

[h-mayorquin]

Hey, sorry this slipped through the cracks. I’ll upload the data to Gin, but I have one question before I do, so I can add this to the documentation and also for myself:

The run above doesn’t have any probes connected to the box, right? If it did, it would include IMEC entries like imDatPrb_type and imDatPrb_pn and there would also be ap.bin and lf.bin files as usual, correct?

[h-mayorquin]

Added the data:

https://gin.g-node.org/NeuralEnsemble/ephy_testing_data/pulls/156#issuecomment-5909

@alejoe91 and @samuelgarcia if you can take a brief look at the gin PR so we can merge.

[h-mayorquin]

Hey, sorry this slipped through the cracks. I’ll upload the data to Gin, but I have one question before I do, so I can add this to the documentation and also for myself:

The run above doesn’t have any probes connected to the box, right? If it did, it would include IMEC entries like imDatPrb_type and imDatPrb_pn and there would also be ap.bin and lf.bin files as usual, correct?

It seems this would be the case. Quoting from the documentation:

Imec OneBoxes are compact and inexpensive alternatives to PXI chassis. Each OneBox connects via USB. A box has two ports for neural headstages. The probes you plug into a OneBox are treated as additional imecj data streams, as if those probes were plugged into PXI modules. OneBoxes can also read up to 12 analog channels and those channels can be thresholded, making 12 pseudo digital channels. These nonneural inputs are referred to as OneBox streams, with labels obx0, obx1, and so on.

https://billkarsh.github.io/SpikeGLX/Sgl_help/UserManual.html

[h-mayorquin]

I added a PR now #1719

It would be great to have a data example that contains both neuropixels, analog and digital signals though for testing though.

[jenniferColonell]

Hi Heberto, Sorry it has been a busy week for me. Here's a OneBox data sample with a 1.0 probe (saving 30 channels), plus analog and digital read through the OneBox. See the notes in the metadata for other details.

OneBox_sample_g0.zip

[h-mayorquin]

Thanks @jenniferColonell

I just downloaded your data and the current PR reads it just fine:

Does the data looks correct? 30 channels of neural data, 5 channels in the one box plus the sync channels.

@zm711 @alejoe91 @samuelgarcia note that the sampling rate of one box is higher than then neural channels and we use the sampling rates to order the streams:
https://github.com/h-mayorquin/python-neo/blob/90721d38b0466888c250f38b6fe6ddb4fc8ca621/neo/rawio/spikeglxrawio.py#L133-L134

This means that someone relying on the first stream_index to be AP or LF will get a surprise here. This is the problem described on #1604

[jenniferColonell]

Those are the correct streams. The obx streams have real signals, if you want to check the data loading.
Bear in mind that the sample rates for ni streams in SGLX data are set by the user, and can be either higher or lower than the OneBox, AP, and LF streams. Ordering the streams by the type would more reliable than sample rate.