Construction of Spatiotemporal Infant Cortical Surface Functional Templates

the UNC/UMN Baby Connectome Project Consortium

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Infant cortical surface templates play an essential role in spatial normalization of cortical surfaces across individuals in pediatric neuroimaging analysis. However, existing infant surface templates have two major limitations in functional MRI analysis. First, they are constructed by co-registration of cortical surfaces based on structural attributes, which cannot lead to accurate functional alignment, due to the highly variable relationship between cortical folds and functions. Second, they are constructed by simply averaging co-registered cortical attributes, which is sensitive to registration errors and lead to blurred attribute patterns on templates, thus deteriorating the accuracy in spatial normalization. Therefore, construction of infant cortical functional templates encoding sharp functional architectures is critical for infant fMRI analysis. To this end, we construct the first set of spatiotemporal infant cortical surface functional templates using Wasserstein barycenter and a state-of-the-art functional feature, namely the gradient density of functional connectivity. To address the first issue, we leverage functional gradient density to drive surface registration to improve inter-individual functional correspondences. To address the second issue, we compute templates based on the Wasserstein barycenter of functional gradient density maps across individuals. The motivation is that Wasserstein barycenter represents a meaningful mean under the Wasserstein distance metric, which takes into account the alignment of local spatial distribution of cortical attributes and thus is robust to registration errors, leading to sharp and detailed patterns on templates. Experiments on a dataset with 207 fMRI scans between 0 and 2 years of age show the validity and accuracy of our constructed infant cortical functional templates.

Original languageEnglish
Title of host publicationMedical Image Computing and Computer Assisted Intervention – MICCAI 2020 - 23rd International Conference, Proceedings
EditorsAnne L. Martel, Purang Abolmaesumi, Danail Stoyanov, Diana Mateus, Maria A. Zuluaga, S. Kevin Zhou, Daniel Racoceanu, Leo Joskowicz
PublisherSpringer Science and Business Media Deutschland GmbH
Pages238-248
Number of pages11
ISBN (Print)9783030597276
DOIs
Publication statusPublished - 2020
Event23rd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2020 - Lima, Peru
Duration: 2020 Oct 42020 Oct 8

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12267 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference23rd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2020
Country/TerritoryPeru
CityLima
Period20/10/420/10/8

Keywords

  • Functional templates
  • Infant
  • Wasserstein barycenter

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint

Dive into the research topics of 'Construction of Spatiotemporal Infant Cortical Surface Functional Templates'. Together they form a unique fingerprint.

Cite this