Saad, what do you mean by a "secondary" fiber? Stefano Marenco
-----Original Message-----
From: Saad Jbabdi [mailto:[log in to unmask]]
Sent: Tuesday, September 27, 2011 11:05 AM
To: [log in to unmask]
Subject: Re: [FSL] RE : [FSL] Correlation between FA and tractography
Hi Hervé
Your explanation would only make sense if:
- You looked at FA within your tracts instead of the whole skeleton
- There were other cortical targets where the effect was in the opposite direction (tracts that don't reach a target must go somewhere else)
Note also that FA correlates (negatively) with dispersion (uncertainty) for the primary fibre, but not necessarily for the secondary one.
Saad
On 26 Sep 2011, at 10:50, LEMAITRE Hervé Université Paris Sud wrote:
> Hi Matt and Saad,
>
> To clarify my point, we found that as the mean FA of the entire skeleton increased, the number of samples decreased between an ROI in the corpus callosum and cortical regions. We did this correlation for different cortical regions (seeds_to_...) and it is almost systematic.
> One explanation was that the higher is your anisotropy the smaller is your dispersion, and so, the lower is the number of streamlines that go from one seed to the targets. Could you tell me if this is an over-interpretation.
> Which kind of data would you see to have a good sense of what is going on?
>
> Thanks again,
>
> Hervé
>
> ________________________________________
> De : FSL - FMRIB's Software Library [[log in to unmask]] de la part de Matt Glasser [[log in to unmask]]
> Date d'envoi : vendredi 23 septembre 2011 17:45
> À : [log in to unmask]
> Objet : Re: [FSL] Correlation between FA and tractography
>
> To clarify, you found that as FA increased the number of streamlines
> decreased and vice versa?
>
> Increased amounts of crossing fibers could lead to reduced FA, but more
> connectedness between many regions.
>
> You would probably have to show more data for me to be able to better
> interpret your findings, however.
>
> Peace,
>
> Matt.
>
> -----Original Message-----
> From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
> Of Herve Lemaitre
> Sent: Friday, September 23, 2011 8:17 AM
> To: [log in to unmask]
> Subject: [FSL] Correlation between FA and tractography
>
> Dear FSL users,
>
> we ran tbss on several subjects and also bedpostx/probtrackx to do
> tractography between one part of the corpus callosum and different regions
> of the cortex.
> Computing the average number of samples between the seed and each cortical
> region, we found a negative correlation between global mean skeleton FA and
> number of samples for each cortical region.
> Is it normal to have this kind of negative correlation and how can we
> explain it?
>
> Thanks,
>
>
> Hervé Lemaître
>
--
Saad Jbabdi
University of Oxford, FMRIB Centre
JR Hospital, Headington, OX3 9DU, UK
(+44)1865-222466 (fax 717)
www.fmrib.ox.ac.uk/~saad
|