Related Experiment Videos
Unified structural equation modeling approach for the analysis of multisubject, multivariate functional MRI data
Jieun Kim1, Wei Zhu, Linda Chang
1Department of Applied Mathematics and Statistics, State University of New York at Stony Brook, Stony Brook, New York, USA. kimjieun@nidcd.nih.gov
Human Brain Mapping
|May 24, 2006
Summary
This study introduces a novel two-stage unified structural equation modeling (SEM) and General Linear Model (GLM) approach for analyzing functional magnetic resonance imaging (fMRI) data. This method enhances the understanding of brain connectivity by integrating longitudinal and contemporaneous pathways, considering subject-level covariates.
Area of Science:
- Neuroscience
- Statistics
- Cognitive Science
Background:
- Brain connectivity studies aim to map directional brain pathways.
- Structural Equation Modeling (SEM) is a suitable statistical method for path analysis.
- Functional Magnetic Resonance Imaging (fMRI) provides multivariate time series data essential for these studies.
Purpose of the Study:
- To propose a novel two-stage unified SEM plus General Linear Model (GLM) approach.
- To analyze multisubject, multivariate fMRI time series data incorporating subject-level covariates.
- To compare the proposed method with conventional SEM and Dynamic Causal Modeling (DCM).
Main Methods:
- Stage 1: Unified SEM for individual subject fMRI analysis, combining multivariate autoregressive (MAR) models for longitudinal pathways and conventional SEM for contemporaneous pathways.
- Stage 2: GLM to integrate subject-level path coefficients with covariates (gender, age, IQ) to assess their impact on effective connectivity.
- Application to fMRI data from a visual attention experiment.
Main Results:
- The unified SEM approach effectively models both longitudinal and contemporaneous brain connectivity.
- Subject-level covariates significantly influence effective connectivity pathways.
- Comparison demonstrated the advantages of the proposed method over conventional SEM and DCM in capturing network dynamics.
Conclusions:
- The proposed two-stage unified SEM plus GLM approach offers a robust framework for analyzing complex fMRI data.
- This method provides deeper insights into effective brain connectivity by incorporating temporal dynamics and individual differences.
- The findings advance the statistical methodologies available for neuroimaging research.