[Mristudio-users] Slice thickness of DTI

[susumu mori]

When you consider effects of imaging parameters on the final outcomes, you
need to know variability and bias.
If the outcomes have simple Gaussian noises, less scans (thus lower SNR)
would lead to higher variability. In your case, suppose you draw an ROI on
the internal capsule and measure Trace of 100 voxels in the ROI, your
second time points would have larger standard deviations. However, you
could argue that the average values are still comparable.

FA is known to have bias; as SNR goes down (e.g. less signal averaging), FA
goes up. Depending on tensor fitting algorithm, completely random noise
seem to converge at around 0.8 - 1.0 or even higher. As SNR goes up, the
measured FA values converge to the real FA value.

If your image protocol have sufficient SNR with 1 signal averaging such
that the measured FA values are already close to the real FA values, then
you won't see much impact between 2 and 4 signal averaging. If your image
protocol has sufficient SNR with 4 averaging but not with 2 averaging, you
would see uniformly higher FA with 2 signal averaging. In this case, you
can't combine the data. Practically, if you are using 12-orientation / 3T /
2.5mm protocol, you may find FA increase with 2 averaging compared to 4. If
you are using protocols with less SNR (e.g. 2mm resolution or 1.5T), the
effect would be worse.

On Mon, Nov 26, 2012 at 3:56 PM, Santosh Yadav
<santoshyadav20076 at gmail.com>wrote:

>  Thanks a lot Sir,
>
>
> I have another question regarding DTI data; I have longitudinal DTI data
> of 3 time points. First time point have 4 averages second time point have 2
> averages and third time point have 4 averages, can I combine these data
> sets?
>
> So finally the question is what the effect of averaging of DTI data, is it
> going to the change the DTI matrix such as FA and Trace?
>
>
> With best Regards,
>
>
>
> On Mon, Nov 26, 2012 at 2:31 AM, susumu mori <susumu at mri.jhu.edu> wrote:
>
>> Hi Santosh,
>>
>> Practically, the slice thickness, which is the thinner the better, is
>> determined by SNR.
>> If you use 2.5 mm, you need about 50-55 slices to cover the entire brain.
>> 2.2 mm needs 55-60 and 2mm needs more than 60. These are all without gaps.
>>
>> With 5-6 min scans, you can acquire about 40-50 DWIs for 2.5mm/50-slice
>> 3D data. These numbers decrease as the thickness become thinner and thus
>> more slices are required and takes longer time.
>>
>> 40-50DWI means about 4 repetitions of 12-orientations.
>>
>> If you use a 1.5T, you may need more than 60 DWIs with 2.5mm, which takes
>> about 7-8min. With thinner slices (e.g. 2.2mm), you need longer scans to
>> achieve good SNR.
>>
>> With a 3T, 30-40 DWIs would give decent SNR with 2.5mm. With 2-2.2mm
>> slice, you want more than 50-60 DWIs (5-8min).
>>
>> These numbers are to achieve a good SNR, which depends on my subjective
>> judgement. So different people could recommend different numbers, but they
>> may not be that much different.
>>
>> As for image distortion, the most important factor is the size of the
>> image matrix; the smaller the better (I assume that you already use
>> parallel imaging, which reduces the matrix size into half).
>>
>> For example, 128x128points / 256x256mm and 96x96/192x192mm both give the
>> same 2mm resolution, but 96x96 would give better (less) distortion and
>> better SNR.
>>
>> On Mon, Nov 26, 2012 at 1:28 AM, Santosh Yadav <
>> santoshyadav20076 at gmail.com> wrote:
>>
>>> Hi DTI experts,
>>>
>>> Thanks for great supports!
>>>
>>> I would like know the recommended slice thickness and gap between the
>>> slice for DTI data acquisition for 12 directions. Is higher
>>> Slice thickness( more that 3 mm) have partial volume effect?
>>>
>>> Second question: Is eye open during the DTI acquisition have any impact
>>> on DTI distortion or image quality?
>>>
>>> With Best Regards,
>>>
>>> Santosh
>>>
>>> _______________________________________________
>>> mristudio-users mailing list
>>> mristudio-users at mristudio.org
>>> http://lists.mristudio.org/mailman/listinfo/
>>> Unsubscribe, send a blank email to:
>>> mristudio-users-unsubscribe at mristudio.org
>>>
>>>
>>
>> _______________________________________________
>> mristudio-users mailing list
>> mristudio-users at mristudio.org
>> http://lists.mristudio.org/mailman/listinfo/
>> Unsubscribe, send a blank email to:
>> mristudio-users-unsubscribe at mristudio.org
>>
>>
>
>
> --
> With Best Regards,
>
> Neuroscience and MR Research Program
>
> Department of Medicine, Queen's Medical Center
>
> University of Hawaii
>
> 1356 Lusitana Street, 7th Floor, Honolulu, HI 96813
>
> Phone No. 808-220-7152
>
> Fax No.  808-545-8970
>
>
>
>
> _______________________________________________
> mristudio-users mailing list
> mristudio-users at mristudio.org
> http://lists.mristudio.org/mailman/listinfo/
> Unsubscribe, send a blank email to:
> mristudio-users-unsubscribe at mristudio.org
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.mristudio.org/pipermail/mristudio-users/attachments/20121127/10f6c632/attachment-0001.html