Callosum to Temporal Lobe

From VISTA LAB WIKI

Jump to: navigation, search

This describes the method Alison Kevan and Michael Perry used to identify where the region of the callosum correlated with reading (in the temporal projection zone) projects on to the cortical surface.

The idea was to identify regions on the cortical surface using functional activity (MT, and a STG zone responsive during rhyming - as measured by MBS) with anywhere in the callosum. We discovered that the projections from MT to the callosum do not overlap with the callosal zone identified for reading; the STG fibers did overlap with the callosum. Notes on the process are described here.

Alison's tutorial slides should be linked here.

Contents

[edit] Setting ROIs

[edit] Free Surfer segmentation and parcellation

FreeSurfer finds labeled regions; nFind the shell script and check it in to the SVN.

dtiFindCallosum - Explain

dti_ROTS_talDefinedSphereRois.m

dtiFindBroderBetweenRois - finds border and dilates

Find Elena's script also. Merge and make it all general so that someone who might want to

dtiFindROISkeleton - this finds the skeleton (path) of the ROI. We need to describe this algorithm

dtiResampleFiberGroup.m - Select sample points along the fiber track

[edit] Using ConTrack to identify paths

Where is a ConTrack script? Go to the ConTrack section to see how to run it.

ctrInit - Set up the diffusion data and the two regions of interest. This creates the candidate fiber tracks, but it does not score them. We typically track 100K and the time for this depends on the data (region of the brain). It could take a couple of hours. In a lousy region, it could take a day. Try a smaller number first to see what's happening.

ctr_makeConTrackFiles.m - this is a script that batch processes the creation of these ROIs

Scoring issues

The track returns a set of candidate fibers. Many of these have kind of lousy scores. These are in a .pdb file. fg_projectname_logName.pdb

This file can be opened in Quench.

ctr_conTrackBaatchScore.m

The scoring is contrack_score.glxa65 -i ctrSampler.txt

Quench won't let you have some fibers with and some fibers without.

[edit] Measuring along the identified paths

Personal tools