A genetically informed, group fMRI connectivity modeling approach: Application to schizophrenia

Aiping Liu, Xiaohui Chen, Z. Jane Wang, Qi Xu, Silke Appel-Cresswell, Martin J. McKeown

Research output: Contribution to journalArticlepeer-review

Abstract

While neuroimaging data can provide valuable phenotypic information to inform genetic studies, the opposite is also true: known genotypes can be used to inform brain connectivity patterns from fMRI data. Here, we propose a framework for genetically informed group brain connectivity modeling. Subjects are first stratified according to their genotypes, and then a group regularized regression model is employed for brain connectivity modeling utilizing the time courses from a priori specified regions of interest (ROIs). With such an approach, each ROI time course is in turn predicted from all other ROI time courses at zero lag using a group regression framework which also incorporates a penalty based on genotypic similarity. Simulations supported such an approach when, as previously studies have indicated to be the case, genetic influences impart connectivity differences across subjects. The proposed method was applied to resting state fMRI data from Schizophrenia and normal control subjects. Genotypes were based on D-amino acid oxidase activator (DAOA) single-nucleotide polymorphisms (SNPs) information. With DAOA SNPs information integrated, the proposed approach was able to more accurately model the diversity in connectivity patterns. Specifically, connectivity with the left putamen, right posterior cingulate, and left middle frontal gyri were found to be jointly modulated by DAOA genotypes and the presence of Schizophrenia. We conclude that the proposed framework represents a multimodal analysis approach for incorporating genotypic variability into brain connectivity analysis directly.

Original languageEnglish (US)
Article number6678714
Pages (from-to)946-956
Number of pages11
JournalIEEE Transactions on Biomedical Engineering
Volume61
Issue number3
DOIs
StatePublished - Mar 1 2014

Keywords

  • Brain connectivity modeling
  • Schizophrenia
  • fMRI
  • group inference
  • prior knowledge

ASJC Scopus subject areas

  • Biomedical Engineering

Fingerprint

Dive into the research topics of 'A genetically informed, group fMRI connectivity modeling approach: Application to schizophrenia'. Together they form a unique fingerprint.

Cite this