Automatic Head Model including 1010-System / Electrode Placement

[lucky_lin]

This post is deleted!

[lucky_lin]

@bryn Thank you very much for your help. I successfully completed the TI simulation, but I found that the simulation results for the two single-electrode pairs are not completely symmetrical. The maximum electric fields they generate in the GM are 0.0933 V/m and 0.0882 V/m, respectively. My method was to clone the two LF-Electro Ohmic Quasi-Stat simulations, modifying the frequency and boundary conditions for each. The electrode pairs I chose are F5-P5 and F6-P6. Is my approach correct, or should I consider using a multi-port simulation instead?

[lucky_lin]

This post is deleted!

[lucky_lin]

I encountered a grid mismatch error when calculating the max modulation, but I cloned the LF-Electro Ohmic Quasi-Stat. Why is it different?

[ofli]

Hi @lucky_lin Did you update the grid before running the first simulation? And did you make any changes to the model after setting the grid for the first simulation?
If possible, please share your project with us at s4l-support@zmt.swiss, and we will take a look.

[lucky_lin]

I updated the grid for the second LF-Electro Ohmic Quasi-Stat, and now I have chosen to add all electrodes in each simulation grid. The problem has been resolved, but I still have questions about the asymmetry between the two simulation results.

[bryn]

Did you check the segmentation? Is the head symmetric? Are there differences between left/right, e.g. thinner skull, holes in the CSF or Dura, etc.

If there are holes, and these seem to be the cause of the asymmetry, maybe reduce the output_spacing to the default 0.3mm. The value output_spacing=0.6 set above in the script could be too coarse to resolve certain thin tissues.

Do the the electrodes seem symmetrically placed wrt the anatomy?

Did you check the simulation voxels? Is there something "asymmetric" about the voxels?

[lucky_lin]

The segmentation is indeed asymmetric. I used the IXI025-Guys-0852-T1.nii.gz provided in the tutorial. Is it because the T1W is asymmetric? I don't know much about brain structure; is this normal?

[lucky_lin]

There are some holes in the other tissues.

[bryn]

@lucky_lin "Other tissues" is basically fat. I don't think this is an issue.

Note, that the head is not perfectly aligned with the axis, i.e. your XY plane may not be perpendicular to the "up" direction. You can also change the slice by using the interactive slice instead of axis-aligned.

Finally. Yes, it is normal that real subjects/people have asymmetrical features. If you are interested in the research question, you could test the impact of individual tissues (conductivity) by

• assigning the same tissue property to all tissues (except for Air?). Is this nearly/more symmetrical?
• assigning the same tissue property to all tissues, except for one tissue (and Air). Candidates are the thin layers around the brain, e.g. Dura, CSF, Brain grey matter, Skull cortical/cancellous, [Galea, Muscle, ...].

[lucky_lin]

I don't quite understand why the T1W I imported is not aligned with the grid. In the GUI, the grid can be adjusted, but how can I ensure they are aligned in a script simulation?

[lucky_lin]

@bryn I tried to set both the rotation and translation of T1W to 0 and then generate the reference point, but it reported an error: Modeler : [Error] Exception during import: Expecting 'Version 1.0' on first line Modeler : [Error] operation unsuccessful.

[bryn]

The T1w image is placed in world (scanner) coordinates. This is useful, e.g. to align different acquisitions (e.g. T1, T2 with different resolutions or field of view).

You could remove the rotation and translation, e.g., by setting an identity transform. However, in my experience, it is often useful to preserve the position in world coordinates.

img = XCoreModeling.Import("some_t1w_mri.nii.gz")
img.Transform = XCoreModeling.Transform()

[lucky_lin]

@bryn I set the rotation and translation of T1W and the grid to 0, but I cannot generate the landmarks. [Error] Exception during import: Expecting 'Version 1.0' on first line Modeler : [Error] operation unsuccessful.

[bryn]

I don't understand what you are doing. The error looks like Sim4Life cannot parse the .pts file produced by the landmark predictor.

• did you edit the .pts file manually?
• where/how did you set the rotation/translation to "0"?
• did you try to set the transform [to zero] (before running the prediction) as suggested in my last post?

[lucky_lin]

After importing T1W in the GUI, it is not aligned with the grid, so I adjusted the rotation and translation of T1W in the controller.
I did not modify the .pts file; I just made the changes above and then tried to generate the landmarks.

[bryn]

But previously you managed to generate them, i.e. before you edited the transform?

[lucky_lin]

No, I only did these things after creating a new file.

[bryn]

I can reproduce your issue. If I set the transform to "0", i.e. Identity, the predictor fails. The head40 segmentation is also less accurate! We need to investigate.

A workaround would be:

• load image
• predict landmarks & segmentation
• compute inverse image transform
• apply this inverse to landmarks/segmentation/surfaces/etc

# assumes verts and labelfield are already predicted (without setting transform to "0")
inv_tx = img.Transform.Inverse()

# transform segmentation
labelfield.ApplyTransform(inv_tx )

# transform landmarks
for v in verts:
    v.ApplyTransform(inv_tx )

[bryn]

The issue is that our neural network was trained with the data in the RAS orientation (with some deviation +- 15 degrees, and flipping in all axes). If you manually edit the transform, you break assumptions used to pre-orient the data into RAS ordering.

Since RAS is a widely used convention in neuroscience, and medical images are always acquired with a direction (rotation) matrix and offset (translation), I think it is best you don't modify the transform.

For instance, if you try to assign DTI-based conductivity maps - you will need to rotate the grid AND the tensors accordingly. It can be done, but it will be more effort...

If this is to investigate if the fields are (nearly) symmetric, I suggest you

• find an approximate symmetry plane (wrt to the brain or skull or ...)
• align the plane of a slice viewer perpendicular to the symmetry plane