Kurtosis Anisotropy

[Ehsan Misaghi]

Hi everyone,

I have a DKI dataset that I wanted to calculate the Kurtosis Anisotropy (KA) values for. I know ExploreDTI is able to extract the map in a NII file and is also able to extract the values for tracts after doing tractography, but I was wondering if there is a way for me to calculate KA, myself, using other maps? More specifically, I want to know how ExploreDTI calculates the KA values, so that I can use the same equation.

Thanks a lot,

Ehsan

[stijn m]

Dear Ehsan,

Maybe the paper by Tax et al 2015; DOI: 10.1002/mrm.25165 can be useful to figure out the calculations. This paper in turn refers to Poot et al. 2010; DOI: 10.1109/TMI.2009.2037915 in which you can find the formula for KA:

So, nothing very complicated here. This is the Kurtosis version of the FA measure.

Although, the confusion in the paper comes with the proposal of new KA parameter that would be more in line with FA.

If you try and solve both formula, you can know which one was applied in ExploreDTI. Please report back :)

HTH.
BW.
Stijn

[Ehsan Misaghi]

Thanks Stijn! Is there a way to check the ExploreDTI code for KA and find out about it that way? That would make things much easier! :)

Also, in order to find avg(K), K1, K2, and K3, I can use the kurtosis tensor exported from ExploreDTI (Plugins -> Export stuff to *.nii files), right?

Thanks and I'll report back as soon as I have an answer!

Ehsan

[stijn m]

Dear Ehsan,

Sorry, but there is no way to check in the p-code.
Yes, you can export the nii files as described.

HTH.
BW.
Stijn

[Samuel St-Jean]

You could try asking nicely either chantal or alexander, they can probably point you to the correct equations, I do not have a copy of the code on hand myself at the moment to check them out though.

[Alexander Leemans]

Hi Shsan,

The KA is defined according to Poot et al. In ExploreDTI, you could use "KA = E_DTI_Kurtosis_Anisotropy(KT,DT)" (with "KT" and "DT" from the *.mat file) to compute it from the command line.

Cheers,

Alexander  

[stijn m]

Dear Alexander,

Thanks for helping out. Still the question remains which of the two formulas described in Poot et al. has been used in EDTI.
The new command is useful!

BW.
Stijn

[Alexander Leemans]

Hi Stijn,

It's the "KA" definition (your image 2).

Cheers,

-A