Hi Johnny,

If you used a different TE for the B0-only images, then you have to take into account the 
T2-weighting in your data.  That is, the normal DTI fit: S = S0 e^(-bD) assumes that TE 
(and TR) is the same for all images.  If not, you should replace S0 with S0 e^(-TE/T2), 
and fit also for T2 (as well as S0 and D).

I'm not including any T1 component because if TR is long enough (say, greater than 5 s), 
as a first approximation all the longitudinal magnetization would have recovered no 
matter what the exact TR is.

Therefore, you could use your data, if you fit them to the right model.

I hope this helps,

-Pablo


On Wed, 18 Feb 2009 09:27:30 +0000, Steve Smith <[log in to unmask]> wrote:

>Hi - this would depend on exactly what caused the differences in the
>first place, and that is worth looking into. However it is quite
>likely the case that the B0 images taken with different acquisition
>parameters than the rest of the diffusion dataset are not going to be
>safely usable.
>
>Cheers.
>
>
>On 17 Feb 2009, at 12:54, Johnny Zhuang wrote:
>
>> Hi all
>>
>> Recently, we have been collecting and processing our DTI data with 32
>> directions and several separate b0s images (it's recommended that
>> every 8
>> directions should have one B0 image). Initially, we want to average
>> the
>> separate b0s together with the b0 of the 4D DTI data. However, after
>> some
>> analyses, we found that the grayscale intensity range of the
>> separate b0s is
>> quite different from that of the b0s acquired together with b1s,
>> which maybe
>> due to the different scanning parameters (eg. TR and TE). Further,
>> it is found
>> that the relationship between the intensity range of these two kinds
>> of b0s is
>> nonlinear, which means that we can not simply average them. Is there
>> anyone
>> who also came across this sort of problem and knows how to fix it?
>> Thanks for your help
>>
>> Many thanks and regards
>>
>> Johnny
>>
>> School of Psychiatry
>> University of New South Wales