The overall goal of this study is to examine thalamocortical connectivity in albinism and controls using diffusion imaging and to compare the output optic radiation reconstruction of two tracking algorithms. Diffusion MRI and tractography can help answer key questions in vision research such as the effect of axonal misrouting on structural organization of the visual pathway in human albinism. The main advantage of this technique is that it allows non-invasive mapping of large, white matter pathways in the living brain and has shown promising advances in neurosurgical planning.
Acquire all imaging as described in this protocol. On a 3 Tesla MRI scanner equipped with a 32 channel head coil. Prior to imaging, first thoroughly screen every participant for MRI safety and have them sign a consent form describing the protocol.
Provide the subjects with ear plugs for hearing protection before positioning them supine and head first on the scanner table. Provide an alert squeeze bulb. Then place cushions to reduce head motion.
Landmark above the eyes at eyebrow level. Before sending the subject into the scanner. Begin imaging by acquiring a high resolution T1-weighted image using a 3D MP-RAGE sequence covering the entire brain.
Use the parameters seen on screen here. With a one millimeter isotropic voxel size. Next, acquire a DTI sequence covering the cortex with 64 directions in two millimeter slices.
Position slices in a transverse orientation following the anterior and posterior commissure line. Also, acquire 30 to 40 PD-weighted proton density images using a turbo spin echo pulse sequence. Set this up in a coronal orientation parallel to the brain stem covering from the exterior extent of the pons to the posterior portion of the inferior colliculus.
LGN delineation should be performed while blind to the subject's group membership. Begin by loading the high resolution PD image in FSL view. Then, click on the tools tab to select a single option to enlarge the image.
Next, select the file tab to select the create mask option and use the tool bar to trace the LGN in each slice. If desired, change the contrast of the image in the tool bar to facilitate LGN detection. Manually trace both right and left LGN masks three times each on averaged PD images that have been interpolated to twice the resolution, and therefore half the original voxel size.
To perform VI segmentation, first run the recon all command in free surfer on T1-weighted images in native anatomical space for automated processing. Then, convert the output to V1 parcellation into a volumetric mask using the label2surf and surf2volume commands. Before performing probabilistic tracking, first run flirt linear registration to bring the brain images that are in free surfer and anatomical space into diffusion space.
Select the free surface space output of recon all, or a subject's brain extracted T1 as the input image. Then, an eddy corrected and brain extracted diffusion weighted image as the reference image. Similarly, for deterministic tracking, use flirt linear registration to bring proton density brains to diffusion space.
Also, in preparation for probabilistic tracking, run this linear registration to bring participants'PD brains to free surface space and native anatomical space for LGN mask transformation. Note that this step creates two outputs. The input brain registered to the reference image and a transformation matrix.
Next, apply flirt transformation to prepare seed masks for tractography. For probabilistic tractography, use the dot mat output from linear registration of PD to free surfer or anatomical T1 as the transformation matrix. The original LGN mask as the input and the brain in free surfer space or anatomical space as the reference volume.
Be sure to use the nearest neighbor interpolation selection from advanced options. Repeat this for deterministic tractography only this time, with the brain in diffusion space as the reference volume. To normalize the LGN, use FSL maths to create an ROI point with the coordinates of the appropriate individual LGN mask in native anatomical space for probabilistic tractography or diffusion space for deterministic tractography.
Then, use FSL maths to apply the radius of the mean mask in MNI space calculated across all participants to create a sphere around the ROI point in native anatomical, or diffusion space. At this point, using free surface space files only, prepare target masks for probabilistic tractography. Register free surfer brains to native anatomical space.
Then, create target masks by applying transformation to V1 masks using trilinear interpolation. To run probabilistic tractography, first use eddy current correction to correct distortions in diffusion weighted images. Then brain extract the images.
Next, select the bedpost X option. Then choose probabilistic tracking and run this for each hemisphere separately. Keep the default basic options, but for increased accuracy, select modified oiler for computing probabilistic streamlines under advanced options.
Select single mask as the seed space. Then, load the transformed LGN mask as the seed image in native anatomical space along with the anatomical T1 to diffusion transformation matrix as the seed to diffusion transform. Finally, select V1 in anatomical space from optional targets as the target.
Repeat using the standard spherical ROIs and then again using non-normalized seed and target masks in free surface space. To perform deterministic tractography, first open the eddy corrected diffusion weighted images in DSI studio. Then load bvec and bval files onto a B table window that is automatically opened to create a source file.
Then select DTI as the reconstruction method and run this on the source files to produce fiber information files. Open the fiber information files in the program's tracking window and run tracking for each hemisphere separately. Use the LGN in diffusion space as the seed and region 17 from the DSI studio Brodmann atlas as determinative region.
In each run, set the contralateral white matter mask from the free surfer segmentation atlas as a region of avoidance. Repeat tracking using spherical ROIs in diffusion space instead of individual LGN as seed regions for tractography. An averaged coronal proton density image of a patient with albinism is seen here.
Manually traced right and left LGN areas of interest are depicted in red. LGN masks transform to free surfer space using nearest neighbor, red, and trilinear, blue, interpolations are shown here. Voxel wise statistical analysis with tract based statistics show no areas of significance in an albinism greater than control contrast due to reduced FA in albinism compared to controls.
However, in the control, greater than albinism contrast significant difference is seen between the groups. Here we see a thickened skeletonized version of these results. DSI studio fiber tracking output indicates reduced LGN to V1 connectivity in an albinism patient as compared to control subjects.
Similarly, probabilistic tracking output shows reduced LGN to V1 connectivity and albinism compared to control subjects. Here, mean tracked masks for both probabilistic and deterministic methods are overlaid for comparison. LGN, blue, and V1, pink masks illustrate the seed and target regions.
Once mastered, data collection and full analysis of three participants can be performed in two to three days while tractography time depends on the size of the seed. While running tractography, carefully choose the algorithm and analysis approach depending on the research question and brain area under investigation and check output files after each step. Do not wait until you achieve the end result to check your work.
Albinism is associated with the increased risk of skin cancer, and with syndromes affecting additional cell types beyond the monocytes. Imagine techniques combined with molecular techniques will assist in investigating the mechanisms of development in albinism and improve the understanding of structure function relationship. After its development, this technique paved the way for researchers in the field of neuroscience to explore brain connectivity in healthy and clinical human populations in vivo.
After watching this video, you should have a good understanding of how to perform white matter reconstruction using deterministic and probabilistic algorithms to examine differences in optic radiation connectivity between patient populations and controls. Don't forget that working with a powerful magnet can be extremely hazardous, and proper screening of participants for MRI safety should always be performed.
