No. 1052

Title:

Innovation approach to the modeling of spatial time series
with application to fMRI connectivity studies

Author(s):

Ozaki, T. (The Institute of Statistical Mathematics)

Key words:

Spatial temporal correlations; spatial stochastic process; fMRI data; SPM; NN-ARX model, brain connectivity, causality. ch word or phrase.

Abstract:

    The statistical analysis of spatial and temporal data is discussed from the viewpoint of an fMRI connectivity study. The limitations of the well-known SPM method for the characterization of fMRI connectivity study are pointed out. A use of the innovation approach with NN-ARX is suggested to overcome the limitations of the SPM modelling. The maximum likelihood method is presented for the NN-ARX model estimation. The exploratory use of innovations for the identification of brain connectivity between remote voxels is discussed.