Native to fsLR32k surface mapping for electrode coordinates

[Sumito Sato]

Dear experts,

I have a question about surface-based registration from native space to fsLR32k in the HCP pipeline.

My electrode locations were originally defined on each subject’s native pial surface
(subXX.L.pial.native.surf.gii, using native vertex indices).
I mapped them to fsaverage_LR32k using the standard MSM-based pipeline and can also obtain MNI152-aligned xyz coordinates with the help of ChatGPT.

While the transformation is technically correct, the relative geometry of the electrode arrays looks distorted compared to the native configuration (e.g., straight arrays become curved, spacing changes, some overlap).

I was wondering:

• Is this expected behavior of MSM-based surface registration?

• Are there recommended approaches to better preserve electrode array geometry?

• For cortical electrodes, is fsLR/Conte69 surface space usually preferred over MNI volume space?

Thanks in advance for any advice.

Best,
Sumito

[Tim Coalson]

Mesh resolutions are not a "space", the resampled surfaces represent the same contours as the surface they were resampled from, but with a different number of vertices.  The pipelines generate individual 32k surfaces in the T1w/fsaverage_LR32k/ folder that are in the same space as the T1w/ native-mesh surfaces.  If you display your resampled electrode files on the 32k surfaces that are in the T1w/ folder, they should be largely unchanged compared to the T1w/Native/ surface versions.

The surfaces inside the MNINonLinear/ folder are warped using the same nonlinear template-based volume registration as is used for the volume files in that folder.

If you want to look at data the way it existed in a subject without distortions, use the files inside the T1w/ folder, which haven't been deformed to MNI space.  Surface registration does not cause anatomical deformations, it is instead choosing where to put the standardized vertices onto the individual's cortex (to try to make each vertex index represent the same part of the same functional area in all subjects, for instance).

Tim

--
You received this message because you are subscribed to the Google Groups "HCP-Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hcp-users+...@humanconnectome.org.
To view this discussion visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/660d0fee-1b38-42b8-9c2d-57580bc79d8bn%40humanconnectome.org.

[Sumito Sato]

Dear Dr. Tim,

Thank you very much for your prompt and helpful response.

Sumito

2026年1月13日(火) 15:32 Tim Coalson <tim.c...@gmail.com>:

To view this discussion visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/CANm3QU4MOswykjHarh-NQ-3Xy0HvYxdy0KnC0NVb0cO3ZrSbQg%40mail.gmail.com.