Skip to main content
Fig. 1 | Molecular Autism

Fig. 1

From: Altered functional organization within the insular cortex in adult males with high-functioning autism spectrum disorder: evidence from connectivity-based parcellation

Fig. 1

The procedure for the functional connectivity-based parcellation of the insula. We first identified voxels in the left and right insula of each participant and generated a set of functional connectivity maps by correlating the resting-state time-series of each voxel with voxels in a whole gray matter mask (excluding the insula) for each hemisphere. Following a Fisher’s z-transformation for functional connectivity maps, we constructed the individual-level similarity matrix using eta-squared, which is a measure of similarity between a pair of functional connectivity maps (see the “Connectivity-based functional parcellation” section). We applied the spectral clustering algorithm to the set of individual-level similarity matrices in order to cluster voxels with similar time-series of the resting-state signal fluctuations. For group-level analysis, we first calculated a binary adjacency matrix for each participant. Adjacency matrices of all participants were averaged separately for TD and ASD individuals to generate a group-level similarity matrix. Finally, we applied the spectral clustering algorithm to the group-level similarity matrix to assign a k clustering label to each voxel (see the “Connectivity-based functional parcellation” section)

Back to article page