Hi,
Thank you for your answer - in the list archive it was not THAT
straight-forward.
So can I rely on FA values and just ignore those greater then 1 (or
assume it is 1) ?
George
2008/8/28 Mark Jenkinson <[log in to unmask]>:
> Hi,
>
> The values greater than 1.0 result from noise and
> the fact that this can perturb the eigenvalues
> of the reconstructed tensor outside their theoretical
> range. If you want more info I believe that this
> has been discussed on the list before.
>
> All the best,
> Mark
>
>
>
>
>
> Jiøí Keller, M.D. wrote:
>>
>> Dear FSLers,
>> I am planning to do some ROI fractional anisotropy (FA) comparisons
>> using FSL and I came across one strange problem. The results of dtifit
>> (FA) are values between 0 and 1.2 (approximately). This range is/was
>> described in tbss manual as being OK. But my question is weather it is
>> really the value of FA, because in theory - as far as I know - FA
>> should not be greater then 1.Is there some "conversion correction" how
>> to convert these values to "real FA" or how to handle those values.
>>
>> thank you for answers
>> George
>>
>>
>
|