[Neuroimaging] [DIPY] Re: DKI free water elimination

[Rafael Henriques]

Hi Ping,

The code that you find in my pull request was intended to be integrated in
Dipy. However, I decided to stop developing a free water DKI model fitting
in Dipy, because I noticed that this dramatically compromises both
precision and accuracy of kurtosis estimates.

If you are interested in suppressing free water partial volume effect in
your kurtosis measures, my alternative suggestion is to use the standard
free water DTI Dipy's module to obtain a free water volume fraction map.
Then you can use this map to remove the voxels of your kurtosis images that
contain mostly free water (let's say f>0.5). This procedure will suppress
the major free water partial volume contamination without compromising
kurtosis precision and accuracy.

More information: [Re] Optimization of a free water elimination
two-compartment model for diffusion tensor imaging
<https://github.com/ReScience-Archives/Henriques-Rokem-Garyfallidis-St-Jean-Peterson-Correia-2017/raw/master/article/Henriques-Rokem-Garyfallidis-St-Jean-Peterson-Correia-2017.pdf>
.

Best,

Rafael

>> Hi DIPY users,
>>
>> I would like to know whether there is a FW-DKI module in DIPY.
>> I came across  Rafael's free water DKI,
>>
>> https://github.com/RafaelNH/free-water-DKI/blob/master/fwDKI.py
>>
>> Is this compatible with the current DIPY package? Any comments in using
this >> module?
>>
>> Thank you.
>>
>> Ping
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/neuroimaging/attachments/20180831/6a4e10ae/attachment.html>