Hi there!
Will be grateful for your advice and I believe you will know the answer.
I have two sets of data from two separate scans. The previous scan I used the Bold data in +orig space I got from preprocessing the data and I used this in a deconvolution along with a mask I obtained from 3dAutomask on the preprocessed data. The resulting SPMs were then laid on top of the tlrc anatomy with excellent fit. We did a second scan on the same person at a later date and tried to analyze the same way but this time using the tlrc volume from the first scan since we didn’t collect a second volume because we thought we could get away with using the +tlrc Volume from the first. They didn’t overlay well and the masks from each were very different. My question is is there an accurate way to overlay the second scan +orig data onto the second scan +tlrc volume and how would I do that?
I would appreciate your advice.
Thanks,
-Linda
If you have the native anatomical data from the first scan, you should use that instead of one that is aligned to a standard template. It’s still possible, but it’s trickier than the simpler case.
I am thinking about what you said. Are you saying I should use adwarp to register my SPMs to the Native +orig space of the first scan instead of the +tlrc Volume of the first scan? Or what are you saying? Also will this be a problem for group data which I need to be in +tlrc space?
Here are my analysis procedures:
3dDeconvolve -input catrunregblurob+orig -polort 2 \ catrunregblurob +orig is the Bold which has been tshifted,registered,blurred and deobliqued
-num_stimts 9
-stim_file 1 s_PULSE.1D -stim_label 1 “PULSE” -stim_maxlag 1 6 \ s_PULSE.1D, s_PPI.1D and s_PPI2.1D are the stimulus timing files
-stim_file 2 s_PPI.1D -stim_label 2 “PPI” -stim_maxlag 2 6
-stim_file 3 s_PPI2.1D -stim_label 3 “PPI2” -stim_maxlag 3 6
-stim_file 4 ‘motionrun[1]’ -stim_label 4 “Roll” \ motionrun is from registration.
-stim_base 4
-stim_file 5 ‘motionrun[2]’ -stim_label 5 “Pitch”
-stim_base 5
-stim_file 6 ‘motionrun[3]’ -stim_label 6 “Yaw”
-stim_base 6
-stim_file 7 ‘motionrun[4]’ -stim_label 7 “IS”
-stim_base 7
-stim_file 8 ‘motionrun[5]’ -stim_label 8 “RL”
-stim_base 8
-stim_file 9 ‘motionrun[6]’ -stim_label 9 “AP”
-stim_base 9
-concat concat.1D \ concat.1D is gives onsets of the beginning of runs.
-iresp 1 Pulse.IRF \ *.IRF is the output of impulse response functions
-iresp 2 PPI.IRF
-iresp 3 PPI2.IRF
-mask maskrun+orig \ maskrun+orig is found by: 3dAutomask -prefix maskrun catrunregblurob+orig
-bucket Deconmotionrun
-fout -tout -vout -rout -bout
The Volume+orig is deobliqued as follows: 3dWarp -deoblique -prefix Volumeob Volume+orig
Then the Volumeob+orig is converted to Talaraich space the old way using markers obtaining Volumeob+tlrc
Next we convert Deconmotionrun to +tlrc space by: adwarp -apar Volumeob+tlrc -force -dpar Deconmotionrun+orig -dxyz 2.0
PLEASE let me know what to change in the above analysis to allow us to use an old +tlrc Volume of the same subject at another date since we couldn’t collect the Volume of the subject in this last scan. What are your suggestions?
Thanks!
-Linda
My response was based on the assumption that you have EPI data for each session and a single anatomical dataset and that you would reanalyze the data with similar pipelines. Alignment of statistical results is almost impossible and not worthwhile if you have the original data. For the second session, you could align the EPI data to the anatomical data of the first run with “align_epi_anat.py -giant_move” or “-ginormous_move”.
In most FMRI processing pipelines, however, you would probably use afni_proc.py to guide you instead. There are numerous examples in the program help that are probably useful. The alignment to a standard space template is now often done with our nonlinear warping tools. The typical processing pipelines will handle the obliquity of the datasets internally, so you won’t have that as a separate step that introduces addition blurring from interpolation. A single anatomical dataset can be transformed to a standard template space for each subject and then the affine and nonlinear transformations can be applied to the motion-corrected and aligned EPI datasets. The @SSwarper script takes care of the alignment to the template and skullstripping. See these presentations for descriptions and examples:
See around slide 18 here:
https://afni.nimh.nih.gov/pub/dist/edu/latest/afni_handouts/afni15_templates_atlases.pdf
See last few slides here:
https://afni.nimh.nih.gov/pub/dist/edu/latest/afni_handouts/afni_proc.pdf