Multi-task fMRI Data Fusion Using Independent Vector Analysis and tensor decompositions

With the goal of identifying novel biomarkers for complex neurological disorders, data fusion of medical imaging data has recently received particular attention in the biomedical research area. Especially important is multi-task functional Magnetic Resonance Imaging (fMRI) data, i.e., data collected from the same subjects while they are performing different tasks.
In this talk, we study Independent Vector Analysis (IVA), i.e., extension of Independent Component Analysis (ICA) to multiple datasets, and PARAFAC2, a tensor factorization approach, for data fusion. First, our simulations reveal that both methods can accurately capture the underlying latent components, albeit with certain differences in capturing the corresponding subject scores. We then apply both methods for the analysis of fMRI datasets collected from subjects that perform 3 different tasks with well-defined relationship among them. Both methods are able to achieve two important goals at once, namely capturing group differences between patients with schizophrenia and healthy controls with interpretable components, as well as understanding the relationship across multiple tasks.