Thank you very much for your input Gwenaëlle. It is extremely helpful.
I hope you have a wonderful day! Best, Catherine

On 3/22/13, Gwenaëlle DOUAUD <[log in to unmask]> wrote:
> Hi  Catherine,
>
> I would anticipate that a "loss of integrity" would lead to higher
> dispersion and lower FA, so would not explain your FA increase...?
>
> There are indicative values for the dispersion of the PDD in the Fig 3 of
> the paper referenced below. Not sure about a correlation between dispersion
> and FA, but it's worth trying. You could also divide the subjects in two
> groups according to their FA and see if they have significant differences in
> dispersion of the PDD.
>
> To help with the interpretation of the results, you need first to make sure
> that you indeed have a clear second fibre population in the region of
> increased FA. As Matt indicated, having a high f2 (say >0.1) is a good
> indication - but again, the variation of this value across subjects is
> difficult to interpret.
>
> Cheers,
> Gwenaëlle
>
>
> --------------------------------------------------------------------
> Gwenaëlle Douaud, PhD
> FMRIB Centre, University of Oxford
> John Radcliffe Hospital, Headington
> OX3 9DU  Oxford  UK
> Tel: +44 (0) 1865 222 523
> Fax: +44 (0) 1865 222 717
> www.fmrib.ox.ac.uk/~douaud
> --------------------------------------------------------------------
>
>
> ________________________________
>  De : Nucleos P <[log in to unmask]>
> À : [log in to unmask]
> Envoyé le : Jeudi 21 mars 2013 21h05
> Objet : Re: [FSL] bedpostx dyads_dispersion results
>
>
> Addition to previous message--
>
> Or should I just be concerned with the correlation between the
> dyads1_dispersion and FA values (e.g a negative correlation)??
>
> What exactly does the negative correlation imply??
> e.g.1: A loss of integrity...?
> e.g.2: Crossing fibers...?
>
> Thank you!!-
> -Catherine
> On Thu, Mar 21, 2013 at 3:31 PM, Gwenaëlle DOUAUD <[log in to unmask]>
> wrote:
>
> Hi Both,
>>
>>
>>in our own experience (in 3 different datasets), it was unfortunately not
>> possible to interpret across subjects the variations of f1 and f2 in
>> regions of crossing fibres. This is mainly due to the constraints imposed
>> on the two "sticks" of the bedpostx biophysical model (they have the same
>> diffusivity).
>>
>>
>>
>>On the other hand, the estimation by the bedpostx model of the orientation
>> of the (crossing) fibres is very robust, so I would be inclined to think
>> that looking at the dispersion of, say, the principal diffusion direction
>> (PDD), would be a quite sensitive approach with reasonably interpretable
>> results.
>>
>>
>>
>>For instance, if the hypothesis is that an increase of FA is due to a
>> selective degeneration of one of the fibre population crossing in a given
>> region (i.e this region is apparently "more organised"), then one should
>> also see a decrease in the dispersion of the PDD (more organised ~ less
>> dispersed) -- shameless auto-reference:
>> http://www.sciencedirect.com/science/article/pii/S1053811909002973.
>>
>>
>>
>>Hope this helps,
>>Gwenaëlle
>>--------------------------------------------------------------------
>>Gwenaëlle Douaud, PhD
>>FMRIB Centre, University of Oxford
>>John Radcliffe Hospital, Headington
>>OX3 9DU Oxford UK
>>Tel: +44 (0) 1865 222 523
>>Fax: +44 (0) 1865 222 717
>>www.fmrib.ox.ac.uk/~douaud
>>--------------------------------------------------------------------
>>
>>
>>________________________________
>> De : Matt Glasser <[log in to unmask]>
>>À : [log in to unmask]
>>Envoyé le : Jeudi 21 mars 2013 18h37
>>Objet : Re: [FSL] bedpostx dyads_dispersion results
>>
>>
>>
>>You cannot directly count the number of fibers, however you could measure
>> the two volume fractions in the ROI.  Perhaps the measure that would be
>> most useful would be f1-f2.  The higher that number is, the less crossing
>> is going on.  The closer the number is to zero, the closer to equal
>> crossings you are.
>>
>>
>>Peace,
>>
>>
>>Matt.
>>
>>From: Nucleos P <[log in to unmask]>
>>Date: Thursday, March 21, 2013 1:13 PM
>>To: FSL - FMRIB's Software Library <[log in to unmask]>
>>Cc: Matt Glasser <[log in to unmask]>
>>Subject: Re: [FSL] bedpostx dyads_dispersion results
>>
>>
>>
>>Hello Matt,
>>
>>How does one go about counting the number of crossing fibers in a given
>> ROI?
>>
>>Thanks much,
>>Catherine
>>
>>
>>On Thu, Mar 21, 2013 at 1:36 PM, Matt Glasser <[log in to unmask]> wrote:
>>
>>In that case, I would compare the volume fraction (mean_f#samples) of the
>> first and second fibers in the region you are interested in.  If f1 and f2
>> are similar in value, there is a strong crossing there.  The dispersion is
>> related to the amount of uncertainty in the estimates, rather than the
>> amount of crossing.
>>>
>>>
>>>Peace,
>>>
>>>
>>>Matt.
>>>
>>>From: Nucleos P <[log in to unmask]>
>>>Reply-To: FSL - FMRIB's Software Library <[log in to unmask]>
>>>Date: Thursday, March 21, 2013 11:57 AM
>>>To: <[log in to unmask]>
>>>Subject: [FSL] bedpostx dyads_dispersion results
>>>
>>>
>>>
>>>Hello experts,
>>>
>>>I am trying to interpret the numbers.
>>>Can someone please explain what is meant by the high dispersion index and
>>> low dispersion index?
>>>I want to see if FA results are due to crossing fibers or not. I have the
>>> values but am not sure how to correctly interpret them.
>>>
>>>Thank you,
>>>Catherine
>>
>>
>>