Conductivity estimation

1

Dear Mrtrix community,

I have a question regarding the possibility to model the conduction delays within a given tract. I know that the conduction of action potentials along myelinated axons is affected by its structural features, such as the axonal g-ratio, which can be used to predict conduction delays in ms. The g-ratio could be calculate with Automated Fiber Quantification (AFQ) toolbox. Unfortunately, I don’t have the 3D Spoiled gradient echo images. The only things that I have are the dwi image and the T1 weighted.

I was wondering if you could please help me to solve this problem in as best as possibile.

Thanks in advance

2

Hi Davide,

I suspect that in the absence of image contrasts over and above DWI and T1, the best you will be able to do is assume that conduction delay is proportional to streamline length. I’m not aware of anyone genuinely purporting to be able to estimate features such as axonal diameters in the presence of complex fibre orientation distributions, and current DWI acquisitions are more-or-less insensitive to the presence of myelin and hence g-ratio.

Rob

3

Thanks Rob! I immagined given that my DWI data doesn’t allow for a more advanced diffusion model (e.g. noddi). Another quick question, do you have any suggestions on the best way to estimate the streamline length. I was thinking to import the tract volume in MATLAB and see the size of the matrix but I’m open to other solutions

Thanks again

4

tckstats provides statistics on streamlines lengths; even providing a histogram if you want that level of detail.

5

Thanks Rob! I got my stats and It’s actually more immediate/fast that doing that in MATLAB.
I was wondering what are the meaning of the outputs? Does the mean reflect the path length in mm/cm? Or is everything referring to voxel size?

thanks again

D.

6

All streamlines lengths are calculated intrinsically in mm units, which is consistent with the rest of MRtrix3 where we operate in “real” / “scanner” space coordinates rather than being intrinsically tied to any particular image voxel grid.

If your application demands accurate quantification of streamlines lengths I would suggest reading this issue listing and related linked pages. There’s also updated code in an open Pull Request that can be downloaded if desired.

Rob