Multivariate autoregressive modeling of fMRI time series

L Harrison, W D Penny, K Friston

Research output: Contribution to journalArticlepeer-review

294 Citations (Scopus)


We propose the use of multivariate autoregressive (MAR) models of functional magnetic resonance imaging time series to make inferences about functional integration within the human brain. The method is demonstrated with synthetic and real data showing how such models are able to characterize interregional dependence. We extend linear MAR models to accommodate nonlinear interactions to model top-down modulatory processes with bilinear terms. MAR models are time series models and thereby model temporal order within measured brain activity. A further benefit of the MAR approach is that connectivity maps may contain loops, yet exact inference can proceed within a linear framework. Model order selection and parameter estimation are implemented by using Bayesian methods.

Original languageEnglish
Pages (from-to)1477-1491
Number of pages15
Issue number4
Early online date23 Jul 2003
Publication statusPublished - Aug 2003


  • Algorithms
  • Arousal
  • Attention
  • Brain Mapping
  • Cerebral Cortex
  • Ocular Fixation
  • Humans
  • Linear Models
  • Magnetic Resonance Imaging
  • Mathematical Computing
  • Motivation
  • Multivariate Analysis
  • Nerve Net
  • Neural Networks (Computer)
  • Nonlinear Dynamics
  • Parietal Lobe
  • Visual Pattern Recognition
  • Prefrontal Cortex
  • Regression Analysis
  • Synaptic Transmission

Cite this