I’m currently working on stereotactic electrode localization using Brainstorm and wanted to share a few challenges we’ve encountered, in case you might have suggestions or clarifications:
Missing Microdeep Model
We were unable to locate the DIXI D08-15CM microdeep electrode model in the system. Is there any way to import or define custom electrode models?
Manual Tip & Skull Entry Alignment
When using the “Set tip” and “Set skull entry” method manually, we’ve found it difficult to align the electrode precisely in the center. The resulting trajectories, when compared to Curry-based clinical results, show considerable deviation. do we have better solution to localize precisely?
Automatic Contact Localization
We also tested the automatic contact localization solutions, but noticed that the computed contact spacing along the same electrode is often uneven. In some cases, contact positions deviate substantially, even though they should follow a regular spacing.
We would greatly appreciate any guidance or suggestions on improving the accuracy of contact localization — whether through additional settings, model editing, or external alignment workflows.
In the last few months we have been developing newer features oriented to facilitate and improve the localization of SEEG electrodes.
The tutorial you have been following do not capture the recent changes and will be removed soon.
All the latest features w.r.t. localization can be found in the section Electrode labeling and contact localization in the new Seizure Fingerprinting tutorial.
DIXI D08-15CM microdeep is an electrode with non-uniform spacing and currently Brainstorm does not support creation of such electrodes. It is in our pipeline for future development.
However, you can define your own "uniformly" spaced custom electrode using the Actions section in the iEEG panel:
Please retry the implantation following the new tutorial link I provided above. That should give you a more precise localization.
Did you use the prerelease version mentioned in the tutorial for this or some other solution ?
On a side note, while there is no direct way to handle SEEG contact localization for non-uniform electrodes in Brainstorm out of the box as of now, it can however be achieved by doing the following:
Create an electrode and set the model as any of default uniform electrode from the list which has the same number of contacts as your desired non-uniform electrode or as mentioned above just create your own custom electrode.
Set the tip and entry to have the electrode rendered on your screen and also updated in the panel.
Thank you so much, chinara. I followed the new section tutorial you sent to me, and redo the same pipeline to set the electrodes manually. The automatic center localization contacts is truly well-designed. when compared to Curry-based clinical results, the euclidean distance still have considerable deviation(average 5mm). And we plot the MNI coordinates between curry(blue) and brainstorm(red) pipeline in the MNI space, the contacts of brainstorm show distort trajectory(red) as the following screenshot. Do you have any ideas? we have checked trying our best, but still couldn't figure out why.
Thank you for your reply. After overnight's debug, we have fixed the issues about the distorted contacts in brainstorm. and we compare them between curry and brainstorm, we found that Euclidean distance difference is 3.5 mm. Do you have any ideas that could minimize the difference ? Thank you again.~
