Luigi,
For a long time I have wanted a routine that would look up anatomic
locations directly in MNI coordinates, without the need to convert to
Talairach coordinates, with the known weakness of the single-subject
Talairach atlas.
So I wrote a simple matlab script, included below, which is based on
the more accurate probabilistic Hammers_mith_atlas_n30r83 atlas. The
script essentially loads the atlas image, and the list of labels, and
then "looks up" the label at the voxel coordinates requested by the
user.
Thought you might find this useful!
Jim
% Script to find anatomic location in MNI coords using Hammers_mith n30r83 atlas
% Written Mar 19, 2013 by James Lee, Imaging and Neuroscience center,
% University of Utah
% Known to run with spm5, not tested on any other version
%
% Run this from the directory that has the Hammers_mith_atlab_n30r83_SPM5.img
addpath('/Users/jim/spm5_1782');
% Load the atlas image
disp('Loading Hammers_mith atlas n30r83 ...')
P = char('Hammers_mith_atlas_n30r83_SPM5.img');
V = spm_vol(P);
for k=1:V.dim(3),
d = V.dim(1:2);
M = spm_matrix([0 0 k]);
img = spm_slice_vol(V,M,d,0);
ham(:,:,k)=img; % THE slice index can go first OR last
end
% Enter array of all labels -- better check this for typos
% ODD NUMBERS = RIGHT SIDE
labels{1}='R hippocampus';
labels{2}='L hippocampus';
labels{3}='R amygdala';
labels{4}='L amygdala';
labels{5}='R anterior temp lobe, medial part';
labels{6}='L anterior temp lobe, medial part';
labels{7}='R anterior temp lobe, lateral part';
labels{8}='L anterior temp lobe, lateral part';
labels{9}='R parahippocampal and ambient gyri';
labels{10}='L parahippocampal and ambient gyri';
labels{11}='R superior temp gyrus, posterior part';
labels{12}='L superior temp gyrus, posterior part';
labels{13}='R middle and inferior temporal gyrus';
labels{14}='L middle and inferior temporal gyrus';
labels{15}='R fusiform gyrus';
labels{16}='L fusiform gyrus';
labels{17}='R cerebellum';
labels{18}='L cerebellum';
labels{19}='brainstem';
labels{20}='L insula';
labels{21}='R insula';
labels{22}='L lateral remainder of occipital lobe';
labels{23}='R lateral remainder of occipital lobe';
labels{24}='L cingulate gyrus, anterior part';
labels{25}='R cingulate gyrus, anterior part';
labels{26}='L cingulate gyrus, posterior part';
labels{27}='R cingulate gyrus, posterior part';
labels{28}='L middle frontal gyrus';
labels{29}='R middle frontal gyrus';
labels{30}='L posterior temporal lobe';
labels{31}='R posterior temporal lobe';
labels{32}='L inferiolateral remainder of parietal lobe';
labels{33}='R inferiolateral remainder of parietal lobe';
labels{34}='L caudate nucleus';
labels{35}='R caudate nucleus';
labels{36}='L nucleus accumbens';
labels{37}='R nucleus accumbens';
labels{38}='L putamen';
labels{39}='R putamen';
labels{40}='L thalamus';
labels{41}='R thalamus';
labels{42}='L pallidum';
labels{43}='R pallidum';
labels{44}='corpus callosum';
labels{45}='R lateral ventricle (excluding temporal horn)';
labels{46}='L lateral ventricle (excluding temporal horn)';
labels{47}='R lateral ventricle, temporal horn';
labels{48}='L lateral ventricle, temporal horn';
labels{49}='third ventricle';
labels{50}='L precentral';
labels{51}='R precentral';
labels{52}='L straight gyrus';
labels{53}='R straight gyrus';
labels{54}='L anterior orbital gyrus';
labels{55}='R anterior orbital gyrus';
labels{56}='L inferior frontal gyrus';
labels{57}='R inferior frontal gyrus';
labels{58}='L superior frontal gyrus';
labels{59}='R superior frontal gyrus';
labels{60}='L postcentral gyrus';
labels{61}='R postcentral gyrus';
labels{62}='L superior parietal gyrus';
labels{63}='R superior parietal gyrus';
labels{64}='L lingual gyrus';
labels{65}='R lingual gyrus';
labels{66}='L cuneus';
labels{67}='R cuneus';
labels{68}='L medial orbital gyrus';
labels{69}='R medial orbital gyrus';
labels{70}='L lateral orbital gyrus';
labels{71}='R lateral orbital gyrus';
labels{72}='L posterior orbital gyrus';
labels{73}='R posterior orbital gyrus';
labels{74}='L substantia nigra';
labels{75}='R substantia nigra';
labels{76}='L subgenual frontal cortex';
labels{77}='R subgenual frontal cortex';
labels{78}='L subcallosal area';
labels{79}='R subcallosal area';
labels{80}='L pre-subgenual frontal cortex';
labels{81}='R pre-subgenual frontal cortex';
labels{82}='L superior temp gyrus, anterior part';
labels{83}='R superior temp gyrus, anterior part';
addr(1) = 0;
while addr(1) ~= 999
addr(1:3)=input('Enter X Y Z voxel coordinates in brackets...or 999
for exit >> ');
if (addr(1) ~= 999) answer = ham(addr(1), addr(2), addr(3));
else answer=0;
end
if (answer > 0) disp(labels(answer));
else disp('No value found');
end % End of if
end % End of while
On 3/12/13, John Ashburner <[log in to unmask]> wrote:
>> I did custom pre-processing, so, of course, I normalized functional
>> images
>> with normalize fiunction of SPM.
>>
>> So, the coordinates I see in activations are MNI coordinates?
>
> If you spatially normalised the images by matching them to one of the
> MNI-space template images supplied with SPM, then they should be in
> MNI space, with co-ordinates in MNI space.
>
>
>> In MNI space, are positive values of x coordinate related to right
>> emisphere?
>
> They should be.
>
>
>> I'm asking this, because, for example, for right hand task, I see
>> more activations in right emisphere (positive value of x coordinate). Is
>> it
>> possible?
>
> This will depend on the file format of the original images. If you
> have used NIfTI throughout, then the left and right sides should not
> get mixed up. If they do come out wrong, then this would indicate a
> bug (probably one of mine) somewhere in the code.
>
> If the original data were the old ANALYZE format, as would be used
> with older versions of SPM (before SPM5), then there is a chance of
> the laterality getting mixed up. However, if you converted your
> original DICOM files using software that knows about NIfTI (and the
> way it saves orientations in the headers), and all image manipulations
> were with software that is NIfTI compatible, then there should not be
> a problem.
>
> Note that the slices option of the stats results sometimes displays
> the left side of the brain on the right side of the image. However,
> the x coordinates reported should be correct.
>
> Best regards,
> -John
>
> On 11 March 2013 20:08, Luigi Pavone <[log in to unmask]> wrote:
>> Thanks Dr. Lee, thanks Dr. Ashburner., thanks Dr. Hammers.
>>
>> I did custom pre-processing, so, of course, I normalized functional
>> images
>> with normalize fiunction of SPM.
>>
>> So, the coordinates I see in activations are MNI coordinates?
>>
>> In MNI space, are positive values of x coordinate related to right
>> emisphere? I'm asking this, because, for example, for right hand task, I
>> see
>> more activations in right emisphere (positive value of x coordinate). Is
>> it
>> possible?
>>
>> Thanks,
>>
>> Luigi
>>
>>
>> 2013/3/11 James Lee <[log in to unmask]>
>>>
>>> Luigi,
>>>
>>> If you want the coordinates in Talairach space you can use mni2tal.
>>> You may want to consider using icbm2tal, which is evidently more
>>> accurate. See the comments at http://www.brainmap.org/icbm2tal.
>>>
>>> Jim
>>>
>>> On 3/11/13, John Ashburner <[log in to unmask]> wrote:
>>> > If your data are spatially normalised to MNI space, then the reported
>>> > coordinates will also be in MNI space.
>>> > Best regards,
>>> > -John
>>> >
>>> >
>>> > On 11 March 2013 16:05, Luigi Pavone <[log in to unmask]> wrote:
>>> >> Hi to all.
>>> >>
>>> >> I did fMRI analysis with spm v.8. Now I need to translate coordinates
>>> >> coming
>>> >> from fMRI results (button "whole brain" in panels "p value") in
>>> >> tailarach
>>> >> coordinates.(please see tha attached file)
>>> >>
>>> >> Are they MNI coordinates, so i can use mni2tal m file?
>>> >>
>>> >> Thanks,
>>> >>
>>> >> Luigi
>>> >
>>>
>>
>
|
