Colin27_4NIRS fluences missing for specific voxels: optodes outside coverage?

Discussions
,
1

Hi,

I am trying to compute a head model from fluences using the
Colin27_4NIRS pre-computed fluence template in NIRSTORM.

My setup:

  • Brainstorm 23-Mar-2026 with NIRSTORM plugin
  • Colin27_4NIRS anatomy correctly assigned
  • 1154 fluence files successfully downloaded from
    neuroimage.usc.edu/resources/nst_data/fluence/MRI__Colin27_4NIRS/
  • fNIRS optodes registered on Colin27_4NIRS head model
  • 18 optodes, frontal bilateral configuration

The error I get is 18 missing fluences for both wavelengths
(756nm and 853nm). The missing voxel indices are:

000458, 000628, 001905, 002026, 002027, 002159, 002346,
007024, 007030, 007482, 008704, 009059, 009060, 009465,
009876, 009990, 010010

These files do not exist on the server.

My optode coordinates are in CapTrak space in meters
(x range: -0.075 to 0.075, y range: 0 to 0.06,
z=0 for all optodes — flat 2D grid with no height).

My questions:

  1. Are these missing voxels outside the pre-computed
    fluence coverage region?
  2. Is the flat z=0 causing incorrect registration to
    voxels outside the fluence patch?
  3. Do I need to compute fluences from scratch, or is
    there a way to fix the registration?

Thank you!

2

Hello,

Thanks a lot for the report. Could you export the subject to a zip and host it somewhere I can download it to perform a few checks?

On the server, we only provide the fluences for the 685 wavelength. So you would need to recompute them.

Does the montage look well-aligned to the head ?

3

Thank you for your response. Please find the subject export here: https://drive.google.com/file/d/11g-A7KPcHHCaJpzm5NCpXHVjgR1RQzd8/view?usp=drive_link

I am attaching a screenshot of the montage and my optodes.tsv file for your review.

image
image982×1010 163 KB

I have investigated the alignment issue further and found the following:

The NEMO BIDS dataset provides optode positions in CapTrak space with z=0.04m for all optodes (they are on a flat plane). When I attempt to project these onto the Colin27 scalp surface in Brainstorm using "Project to default anatomy", the optodes scatter incorrectly around the entire head perimeter rather than staying on the frontal area. This suggests the CapTrak to Colin27 coordinate transformation is not being handled correctly.

I have not saved any of these incorrect projections my current registration is the one shown in the attached screenshot, which visually appears frontal and reasonable.

Could you advise on the correct way to transform CapTrak coordinates to Colin27 MRI space in Brainstorm/NIRSTORM? Or if you are able to check the subject zip, that would help identify exactly what is wrong.

Regarding the fluences, my dataset uses 756nm and 853nm. I do not have a GPU to recompute them locally. Would it be possible for you to compute the missing fluences for these wavelengths? The 18 missing voxel indices are listed in my original message above.

Thank you very much for your time.

4

This is important. We need to find a way to align the cap with the subject anatomy before moving further. Do you have a link to this dataset? I recently heard about it in a nirx video and wanted to explore it a bit. That might be the occasion. It’s a bit strange / sad that they don’t provide the 3D coordinates of the optodes. I will try to see if i can find anything about that next week.

Yes; no problem. The only issue is that the missing voxel indicated are wrong since the position of the optodes is not correct. We need to figure out the correct coordinate before going further.

5

The dataset is publicly available at https://osf.io/pd9rv the optode file is sub-101_optodes.tsv in the nirs folder. The coordsystem.json confirms CapTrak space in meters with z=0 for all optodes.