Shape analysis of corpus callosum in autism subtype using planar conformal mapping

Qing He, Ye Duan, Xiaotian Yin, Xianfeng Gu, Kevin Karsch, Judith H. MilesProceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE 2009 · DOI

TL;DR

This paper introduces a novel method for analyzing subtle 3D shape differences in the corpus callosum (CC) of individuals with essential autism. By using planar conformal mapping to align 3D CC models, they identified significant shape abnormalities in the anterior regions of the CC in the autism group, offering a more precise understanding of neurobiological differences.

What Problem Does This Paper Address?

Autism is a complex neurodevelopmental disorder with a known neurobiological basis, but the precise anatomical extent of these abnormalities remains largely unknown. Previous studies on the corpus callosum (CC), a crucial white matter tract connecting the brain hemispheres, have yielded inconsistent results regarding its size and shape in individuals with autism. This inconsistency might stem from the heterogeneity within the autism diagnosis. To address this, the authors focus on a specific subtype, "essential autism," aiming to find more homogeneous results. They tackle the challenge of comparing the 3D shapes of CCs from different individuals, which requires establishing point-to-point correspondence across these complex structures.

Methods

The core of this paper lies in its innovative shape analysis methodology.

  • 3D CC Modeling: They begin by extracting 2D contours of the CC from MR images using a semiautomatic segmentation method. These contours are then stacked to reconstruct a 3D model of the CC.
  • Planar Conformal Mapping: A key contribution is the development of a new planar conformal mapping technique. This method maps the 3D CC surface, which has two open boundaries, onto a planar domain (a rectangle). This process is crucial for establishing point correspondence between different individuals' CC models. The mapping ensures that the surface is deformed without stretching or tearing, preserving local angles.
  • Surface Correspondence: After mapping each CC model to a planar domain, they align these planar domains. This alignment allows for a one-to-one correspondence of points between different CC surfaces. The source surfaces are then re-triangulated to match the connectivity of a target surface, ensuring that corresponding points exist on both.
  • Statistical Shape Analysis: Finally, they employ statistical shape analysis, specifically using Hotelling's T² test, to compare the 3D shapes point by point between the autism and control groups. This method is chosen for its ability to detect localized shape differences. They also apply False Discovery Rate (FDR) correction to account for multiple comparisons across thousands of surface points.

Datasets

  • Essential Autism Patients: 30 children diagnosed with essential autism, meeting DSM-IV criteria and confirmed by ADI-R and ADOS algorithm.
  • Controls: 5 healthy patients from the NIH Human Brain Project and 19 normal individuals from the IBSR database.
  • MR Images: T1-weighted MR images acquired using a Siemens Symphony 1.5 T scanner with specific parameters (TR=35ms, NEX=1, flip-angle=30 degrees, thickness=1.5mm, FOV=22cm, matrix=512x512). IBSR data had 3.1mm thickness and was resampled to 1mm isotropic voxels.

Key Insights and Ideas

  • Focusing on Essential Autism: By isolating the "essential autism" subtype, the researchers aimed to reduce variability and uncover more consistent neurobiological markers, a strategy that proved fruitful.
  • Planar Conformal Mapping for High Genus Surfaces: The development of a conformal mapping technique specifically for surfaces with boundaries (like the CC) is a significant methodological advancement, enabling robust 3D shape comparisons.
  • Localized Shape Abnormalities: The study moves beyond simple volumetric measurements to pinpoint specific regions of the CC that exhibit shape differences, offering a more nuanced understanding of CC morphology in autism.
  • Anterior CC Differences: The finding of significant shape differences in the anterior most and anterior body of the CC is a key insight, suggesting these regions may be particularly affected in essential autism.

Key Results

  • Significant Shape Differences: The statistical shape analysis revealed significant differences in the shape of the corpus callosum between individuals with essential autism and controls.
  • Anterior Region Abnormalities: Specifically, significant shape differences were found in the anterior most and anterior body of the CC. The raw significance map also indicated differences in the posterior top, but these were not retained after FDR correction.
  • Qualitative Observations: Visualizations showed that in the autism group, the anterior most part of the CC was more projected, the anterior body was thicker, and the posterior body was thinner compared to controls.
  • Global Shape Difference: A global shape analysis also indicated a significant difference in the overall shape of the CC between the two groups (p-value of 0.008).

Limitations and Future Work

  • Age Range: The authors acknowledge that the broad age range of the participants (3-18 years) could influence brain size and thus CC shape. While Procrustes alignment was used to remove size effects, further investigation with brain size as a covariate is suggested.
  • Subgroup Inconsistencies: The study notes some inconsistencies with previous findings, particularly with study [16], which might be due to differences in sample size, population age, and methodology.
  • Future Studies: The authors recommend further studies on different autistic subgroups (e.g., male vs. female, complex vs. essential autism) to better explain inconsistent results and gain a more comprehensive understanding. They also suggest further validation of their findings for clinical diagnosis.

Summary updated

Abstract

A number of studies have documented that autism has a neurobiological basis, but the anatomical extent of these neurobiological abnormalities is largely unknown. In this study, we aimed at analyzing highly localized shape abnormalities of the corpus callosum in a homogeneous group of autism children. Thirty patients with essential autism and twenty-four controls participated in this study. 2D contours of the corpus callosum were extracted from MR images by a semiautomatic segmentation method, and the 3D model was constructed by stacking the contours. The resulting 3D model had two openings at the ends, thus a new conformal parameterization for high genus surfaces was applied in our shape analysis work, which mapped each surface onto a planar domain. Surface matching among different individual meshes was achieved by re-triangulating each mesh according to a template surface. Statistical shape analysis was used to compare the 3D shapes point by point between patients with autism and their controls. The results revealed significant abnormalities in the anterior most and anterior body in essential autism group.
BibTeX
@inproceedings{he2009shape,
  author = {Qing He and Ye Duan and Xiaotian Yin and Xianfeng Gu and Kevin Karsch and Judith H. Miles},
  title = {{Shape analysis of corpus callosum in autism subtype using planar conformal mapping}},
  booktitle = {Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE},
  year = {2009},
  doi = {10.1117/12.812285},
}
Figure