Morphometry of human insular cortex and insular volume reduction in Williams syndrome

Abstract

Functional imaging in humans and anatomical data in monkeys have implicated the insula as a multimodal sensory integrative brain region. The topography of insular connections is organized by its cytoarchitectonic regions. Previous attempts to measure the insula have utilized either indirect or automated methods. This study was designed to develop a reliable method for obtaining volumetric magnetic resonance imaging (MRI) measurements of the human insular cortex, and to validate that method by examining the anatomy of insular cortex in adults with Williams syndrome (WS) and healthy age-matched controls. Statistical reliability was obtained among three raters for this method, supporting its reproducibility not only across raters, but within different software packages. The procedure described here utilizes native-space morphometry as well as a method for dividing the insula into connectivity-based sub-regions estimated from cytoarchitectonics. Reliability was calculated in both ANALYZE (N=3) and BrainImageJava (N=10) where brain scans were measured once in each hemisphere by each rater. This highly reliable method revealed total, anterior, and posterior insular volume reduction bilaterally (all p's<.002) in WS, after accounting for reduced total brain volumes in these participants. Although speculative, the reduced insular volumes in WS may represent a neural risk for the development of hyperaffiliative social behavior with increased specific phobias, and implicate the insula as a critical limbic integrative region. Native-space quantification of the insula may be valuable in the study of neurodevelopmental or neuropsychiatric disorders related to anxiety and social behavior.

Figure 1

Insular boundaries. Depicted are the major boundaries used to delinate the insula cortex. The top panels (A & B) show the insula in the sagittal view, which is best for visualizing the anterior limiting sulcus and the posterior pole, but also shows the circular sulci (all labeled in panel A). The bottom panels (C & D) show the insula in the coronal view, which is best for visualizing the circular sulci most accurately and delineating the insula from neighboring opercular cortex (all labeled in panel C). Panels A and C show the total insular ROI, while panels B and D show the insula divided into anterior (red 1) and posterior (green 2).

Figure 2

Rationale for connectivity-based sub-regions. This figure recreates insular topography set forth by Mesulam and Mufson (1982), and schematically depicts the rationale and methodology used in the creation of the connectivity-based sub-regions. The red rectangle represents the most anterior, posterior, inferior and superior points of the insula tracing. The solid diagonal black line across the rectangle is the approximation of the Mesulam’s and Mufson sub-regions, and separates the insula into anteroventral (red) and posterodorsal (green). The cytoarchitechtonic regions are also marked: Ia-p, agranular-periallocortical; Idg, dysgranular; Ig, granular.

Figure 3

Scatterplots of intra-rater reliability. The graphs depict measurement 1 and 2 made by the first author for each insular ROI. The left column shows the left hemisphere and the right hemisphere is in the right column. Insular ROIs are organized as total, anterior, and posterior from top to bottom. Measurement 1 is on the x-axis and measurement 2 is on the y-axis. All measurements are in cm³.

Figure 4

Scatterplot graphs of reliability measurements by rater 1 and rater 2 using BIJ. The left column shows the left hemisphere and the right hemisphere is in the right column. Insular ROIs are organized as total anterior, and posterior from top to bottom. Rater 1 is on the x-axis and rater 2 is on the y-axis with all measurements displayed in cm³.

Figure 5

Insular Raw Volume Scatter Plots. Insular volumes plotted by group and hemisphere for (5A) total insula, (5B) anterior insula, and (5C) posterior insula for each subject in cm³.