In iEEG/ECoG
study, it is very common to localize the
electrodes using ‘CT’ scans. The CT images have different orientations
to the MR images, hence co-registration is often needed. This package
does not provide co-registration functions, as different labs might have
different approaches. threeBrain
only takes co-registered
CT scans and visualize them directly. Therefore if you have already
co-registered CT, skip the next part.
The rest part of this article uses N27
sample files
mentioned by the previous vignette. If you are using this example,
please execute the following R code to set up. Alternatively, you can
substitute variables subject_code
and
subject_path
accordingly.
library(threeBrain)
<- "N27"
subject_code <- "~/Downloads/N27"
subject_path <- freesurfer_brain2(subject_path, subject_code) brain
(skip if you have generated co-registered CT in ‘Nifti’ formats)
This part provides one working method via dcm2niix
and
FLIRT
packages. Again, other methods will work as long as
they output CT aligned to MRI in ‘Nifti’ formats.
To start, please download and install dcm2niix
(link),
FLIRT
packages (link).
Step 1: Merge all DICOM
images to
Nifti
format:
Open your terminal (bash
or power shell
),
change directories to where CT
DICOM
images
are, and run
dcm2niix [folder with DICOM images]
Do the same to T1
MR images too.
Step 2: Copy the two .nii
files
generated in the previous step to a same folder, rename them to be
ct.nii
and t1.nii
flirt -in ct.nii -ref t1.nii -out ct_in_t1.nii -omat ct2t1.mat -interp trilinear -cost mutualinfo -dof 6 -searchcost mutualinfo -searchrx -180 180 -searchry -180 180 -searchrz -180 180
There will be a ct_in_t1.nii
file generated.
Assuming you have generated co-registered CT scans in file
ct_in_t1.nii
, please copy it to the subject folder. The
file should be located at
mri/ct_in_t1.nii
Next, type in and change the file path to ct_in_t1.nii
just created
<- file.path(subject_path, "mri/ct_in_t1.nii")
ct_path ::localization_module(subject_code = subject_code,
threeBrainfs_path = subject_path,
ct_path = ct_path)
You will see the threshold CT displayed along with the pial surfaces.
Make sure Edit Mode
(in the
Electrode Localization
panel) is CT/Volume
.
Now move your mouse cursor to the green electrodes, the localization
control panel will display the tkrRAS
(FreeSurfer
coordinate), T1
(native RAS
coordinate), and MNI305
(‘MNI’ template coordinate).
Double-click on the green areas, and a new electrode
will be created.
The default threshold of CT scans is set to 3000. However, this threshold might need to be adjusted in order to achieve high accuracy. The following figures demonstrate the results of different threshold.
If the threshold is too low, then other irrelevant structures will appear. If the threshold is too high, some electrodes might disappear.
To adjust the CT threshold, open the Volume Settings
from the control panel, slide Voxel Min
.
When electrodes are too small, or the CT is unavailable, electrode localization can be done on the MR images. Localization via MRI can be done by the following steps:
Edit Mode
(in the
Electrode Localization
panel) to
MRI slices
Volume Settings
panel, enable
Overlay Coronal
, Overlay Axial
, or
Overlay Sagittal
.Coronal
, Axial
,
Sagittal
plane to desired positionIn most of the scenarios, electrodes form grids or strips, The
threeBrain
provides interpolation functions.
For example, we want to localize the electrodes circled in orange
color (top-left). You could of course mark them one by one. More
efficiently, you can simply localize the electrodes at both ends, set
the Interpolate Size
to 6 (there are 6 more electrodes to
be added), make sure there is no other objects blocking the sight, and
hit the Interpolate from Recently Added
button (circled in
the figure). All other electrodes will be automatically generated.
The newly added electrodes might be inaccurate. Don’t worry, you can adjust them later!
Once all the electrodes are marked, change Edit Mode
to
refine
. In the refine mode, you can
Enable/Disable Electrode
, the electrode will be
disabled/enabled. The disabled electrode will be colored in gray.To automatically adjust all electrodes, simply click
Auto-Adjust All
. The new electrode locations will be
calculated via a weighted average.
To automatically adjust a specific electrode,
double-click one electrode. The highlighted electrode
will be colored in red. Then click on
Auto-Adjust Highlighted
.
To manually adjust a electrode, double-click one
electrode. The highlighted electrode will be colored in red. Then hover
the mouse above the canvas, press and hold keyboard 1
or
shift+1
to move the electrode along Right/Left
axis. Press and hold 2
or shift+2
to move the
electrode along Anterior/Posterior
axis. Press and hold
3
or shift+3
to move the electrode along
Superior/Inferior
axis.
Once Finished electrode localization, you can take screenshots by
opening Default
panel, and click on
Screenshot
.
To download electrode table, click on Download as csv
in
the Electrode Localization
panel. The electrode table
should look as follows:
You will notice that the Electrode
column is left blank.
This is because localization order might be different than the recording
channel number. Please fill out Electrode
column as the
actual channel number (integers).