Structural Connectivity Analyses

Structural connectivity maps were derived from the diffusion weighted imaging data of 83 healthy adult participants.  Motion artifacts were removed and eddy-current distortions were corrected with DTIPrep (Oguz et al., 2014). Using the FMRIB Diffusion Toolbox, diffusion tensors were then fitted at each voxel within a cortical mask. DTI volumes were coregistered with subjects’ anatomical T1-weighted volumes using the FreeSurfer image analysis suite (Dale, Fischl, and Sereno, 1999; Fischl et al., 1999; Fischl, Sereno, and Dale, 1999). FreeSurfer was also used to identify white matter voxels that lie 2 mm below the cortical gray-white boundary of the anatomically-defined seed regions (Kang, Herron, Turken, and Woods, 2012). For each subject, probablistic tractography was performed from each seed region to the rest of the cortical surface using the FMRIB diffusion toolbox.  5000 streamlines traced probabilistic paths from each seed region voxel to the brain surface.  Streamlines were terminated if they exceeded 2000 steps (corresponding to a 1m long tract). To compare connectivity across individuals, a surface voxel was considered “connected” to a seed region if at least 10 streamlines from the seed region reached that voxel in a given subject.  These data were projected to an average brain surface using Freesurfer and averaged across subjects.  At each vertex, figures show the percentage of subjects for whom that vertex is connected to the seed region; blue represents 50% of subjects connected to the seed region at that vertex, and green represents 80% of subjects.