The insula is a complex region of the human brain, the functions of which are still only partially understood. Literature implicates the insula in a wide and diverse range of activities including sensory, motor, language, behavior and cognitive functions. Several insulo-cortical connections have been reported yet no precise, integrated map of functional and anatomic connectivity of the human insula exists. In this study we identify specific connections between the human insula and the ipsilateral cortex using a Q-ball residual bootstrap tractography algorithm, and to provide a parcellation atlas of the insula based on structural white matter connections.17 healthy subjects (42±12 years, M/F=10/7) underwent a 3T MRI protocol at the University of California, San Francisco (UCSF) including a structural 3D T1w MPRAGE and a high angular resolution diffusion (HARDI) imaging acquisition (2.2 mm3 isotropic voxel, b-value=2000 s/mm2, 55 diffusion gradients, 1 b0). HARDI datasets were corrected for movement and eddy-current distortions using FSL. Fractional Anisotropy maps were computed. Freesurfer software was used to parcellate cortical areas of the brain in the T1 images using the Desikan-Kylliany Atlas. Linear and non-linear transformations were used to register b0 volumes to T1 images and then T1 images to MNI standard template. The insulae were extracted bilaterally from the T1 images parcellated with Freesurfer and then registered in each diffusion subject’s space. Whole brain tractography was performed with a previously described q-ball residual bootstrap algorithm1 with 33 seeding density for each voxel. First we extracted the tractography of the insula by using the region as a first-end ROI. Then each cortical region of the brain obtained from the Freesurfer parcellation was used as a second either-end ROI to identify any connection with the ipsilateral insula. The resulting density maps (indicating how many streamlines pass through a voxel) were obtained from the tracts and then registered in the MNI space. Average density maps across subjects for each connection between the insula and each cortical region were finally calculated. Visual inpecting the average density maps, a threshold of 85% of the maximum number of streamlines connecting the insular with each cortex region and falling in the insula region, was chosen to minimize the overlap between the regions in which the parcellation resulted.Tractography of the insula was successfully identified in all the subjects (Figure 1). A total of 22 (out of 32) cortical connections with the insula for each hemisphere were found, the majority being with frontal and parietal lobes. The results of the structural-connectivity-based parcellation of the insula are consistent with results of previous structural and functional studies2,3. A caudal-rostral gradient of connection is evident displaying insular connections from posterior to anterior cortex. Anterior dorsal insula shows a high connectivity with frontopercularis areas and precentral gyrus. The most inferior insular portion connects with lateral orbitofrontal, rostral middle frontal, and superior frontal regions. The anterior ventral insula is most connected with medial orbitofrontal cortex. The posterior insula has connections with the postcentral gyrus, the superior parietal gyrus and the inferior parietal lobule. Specifically the posterior long insular gyrus associates with the occipital lobe and the pole of the posterior insular lobule with the temporal lobe. Figure 2 show an example of the insular parcellation based on 12 connected cortex regions extracted with Freesurfer.Q-ball residual bootstrap tractography provides an anatomical parcellation of human insula based on homolateral cortical connectivity. These results are consistent with functional areas described in fMRI studies. Availability of a template atlas of insular anatomic connectivity would contribute to a better understanding of its multiple functions and would have a clinical impact on insular lesion studies, on neurosurgical application in pre-intervention mapping, and in predicting post-intervention deficits.