Hi,
Unless you are absolutely sure that there is zero partial volume with white matter
or CSF *within the atlas*, which because of the blurring in the atlas and the
substantial partial volumes present near borders at standard resolutions in MRI,
means that it is unlikely that things will sum close to 100.
I hope this helps.
All the best,
Mark
On 7 Feb 2012, at 10:55, Maite Termenon wrote:
> Hi Mark,
>
> thanks for your response. I am particularly interested on Caudate and Putamen, that is why I choose MNI Atlas.
>
> My mask is only for GM. It should then sum 100 or still not?
>
> Thanks,
>
> --
> Maite Termenón - http://www.ehu.es/ccwintco/index.php/Usuario:Mai.termenon
> Computational Intelligence Group - http://www.ehu.es/ccwintco
> UPV/EHU
>
>
> Mark Jenkinson <[log in to unmask]> escribió:
>
>> Hi Maite,
>>
>> The MNI Structural Atlas does not label all possible tissues/structures.
>> For instance, white matter and the lateral ventricles are not included.
>> This explains why the numbers do not add up to 100%.
>>
>> As for your resolution - as long as the sform/qform is set correctly
>> then it should not matter.
>>
>> Your mask however is not particularly useful for this application though,
>> as it seems to include the frontal lobe, the cerebellum, the caudate and
>> the occipital lobe, so you might want to try running it on a more specific
>> mask, as I don't think you'll gain much from the atlasquery tool on such
>> a disperse mask.
>>
>> All the best,
>> Mark
>>
>>
>>
>> On 6 Feb 2012, at 11:18, Maite Termenon wrote:
>>
>>> Hi all,
>>>
>>> I'm trying to use atlasquery using the option -m, where I place a mask image where my activations are. The command line I'm using is:
>>>
>>> "atlasquery -a "MNI Structural Atlas" -m feat_mask.nii >> locations.txt"
>>>
>>> Results are:
>>> ---------------------------------------------------------------------------------
>>> Using atlas: MNI Structural Atlas
>>> Working from mask : feat_mask_FA_C_EA_fastsym_45_99dot5.nii
>>>
>>> Caudate:10.1344
>>> 1
>>> Cerebellum:0.095793
>>> 2
>>> Frontal Lobe:2.17663
>>> 3
>>> Insula:4.36143
>>> 4
>>> Occipital Lobe:1.45434
>>> 5
>>> Parietal Lobe:2.13323
>>> 6
>>> Putamen:0.120506
>>> 7
>>> Temporal Lobe:1.05662
>>> 8
>>> Thalamus:0.770808
>>> ---------------------------------------------------------------------------
>>>
>>> Is this correct? If I understood, if the atlas is a label atlas, then the value is the overlap percentage (between 0 and 100). My question is, why all these percentages do not sum 100?
>>>
>>> If my mask is registered to MNI but it is in 1.5x1.5x1.5mm. Do I need to transform it to 1x1x1mm?
>>>
>>> Thanks in advance,
>>>
>>> Maite
>>>
>>
>
|