[andreashorn]

Great. But again, SPM Segment is just as good and much quicker:
These two papers show it:
https://www.researchgate.net/publication/327908494_Optimization_and_comparative_evaluation_of_nonlinear_deformation_algorithms_for_atlas-based_segmentation_of_DBS_target_nuclei

https://www.researchgate.net/publication/327380509_Lead-DBS_v2_Towards_a_comprehensive_pipeline_for_deep_brain_stimulation_imaging

Probably will make SPM default soon again.

Best, Andy

[andreashorn]

Hi there, this means that Matlab cannot write into the lead-dbs folder. (The error message even says so: Please set write permissions to Lead-DBS install directory accordingly.).
Maybe simply put Lead-DBS into your home directory instead?

Best, Andy

[andreashorn]

Hi, apparently ANTs normalization crashes on your system (probably because of a lack of memory).
Use SPM Segment instead – it’s as good and much faster.

Best, Andy

[andreashorn]

Hi Anthony,

you can do so by averaging the vertices of the atlas surface I think:

lead path
load([ea_space([],'atlases'),'DISTAL Minimal (Ewert 2017)',filesep,'atlas_index.mat']);
centroid=mean(atlases.fv{1,1}.vertices);

would calculate the centroid of the right STN (11.2668 -12.6726 -7.7385) of the DISTAL atlas.
With
atlases.names
you can see the entries of the atlas set.
In this case, STN is the first one.
In Lead-DBS, right hemisphere is always entry number one, left is number two.
So replacing the third line with
centroid=mean(atlases.fv{5,2}.vertices);
would give you the left GPi (GPi is entry number 5, 2 stands for the left side.

Hope this helps,

best, Andy

[andreashorn]

Great!

[andreashorn]

Thanks!

[andreashorn]

Hi mapleck,

I’m sorry for the late reply.. I cannot reproduce the error here. Can you visualize the DISTAL atlas in MNI space?

Best, Andy

[andreashorn]

Hi rubenscury,

re the cerebellum – see the labeling folder which has the suit atlas and the yeo cerebellar functional zones atlas.
You can try to use ea_labeling2atlas(‘Cerebellum_SUIT (Diedrichsen 2006)’) to convert the labeling into an atlas – but I’m not 100% sure if it will work. If not need to do manually (just put nifti files in the folders, ignore the atlas_index.mat and gm.nii files that will be built automatically.

Re toolboxes I think you need stats & machinelearning as well as image processing optimally. Most things should work without but some functions need them.

Best, Andy

[andreashorn]

…Yes – as already mentioned on twitter – you can use SPM coregister: write to do so.

[andreashorn]

Hi Rubens,

corr is a built-in Matlab function but part of the Stats & ML toolbox (https://www.mathworks.com/help/stats/corr.html) needed for some functions of Lead-DBS.
You can probably use an alternative function for correlation and replace the call in
ea_permcorr (line 17) if you have no access to the toolbox.

Best, Andy

[andreashorn]

Hi Rubens,

could you please send the full error message? The first (most crucial) line is missing..

If you haven’t done so, it might also make sense to watch the walkthrough video to check if things look the same on your side:

Walkthrough-Videos

Best, Andy

[andreashorn]

Hey Rubens,

thanks for the nice words – it was all Chen Yao though..
There is a button in Lead-Group, called something like prepare DBS stats.
That would create overlap VTA/STN of the selected atlas.

You can then correlate a variable with that – in theory directly from within Lead group. It’s not very intuitive, best documentation are these slides: http://www.lead-dbs.org/docs/tutorials/group_analyses_horn.pdf

Best, Andy

[andreashorn]

Hi there,

unfortunately, the convis tool is pretty old code and I haven’t touched it in a while. I do agree it’s not really useful since one needs to recompute things quite often.
I do think Ari Kappel made some edits to it at some point if you run in developer mode. Feel free to reach out to him on the slack user channel.

Re your last question, that’s not possible, lead connectome mapper is only giving you .nii output (i.e. volumetric), the name suggests so :)
I guess you can maybe use tools like trackvis or dsi studio to isolate fibers connected to nifti files. Lead connectome should also export .trk files that you can visualize there..

I’m sorry since I feel this answer doesn’t help you that much. Just for the record I do agree that there are many different areas in which one can do “connectomics” inside Lead-DBS with very different purposes, formats, etc. It must be a jungle to find out how/what to do given it’s not well documented..

So potentially could help better if you’re able to formulate *very precisely* what you want to do.
If you say, you want to compare patient specific tracts, that’s not so precise. Tracts connected to what? Or whole connectomes? Compare how? Number of tracts, their anatomical courses, specific connections between regions?
All can be done but I’d need a very very specific scientific question + a good idea of your data (i.e. ~how many subjects and is patient specific dMRI data available, what kind?) to be of good help.

Best, Andy

[andreashorn]

nAtlasintersection = AtlasIntersection / VTA volume.
So it’s normalized to VTA volume.

Best, Andy

[andreashorn]

Hi Kailaing,

good to hear from you.
The downloads to work, must be something in your Matlab internet connection (institutional firewall or similar?).

For Flip-LR you can use https://filedn.com/lsPIJ4ragTWjjmV6PvlDQLu/data/fliplr.zip

Best, Andy

[Author]

Posts