A hybrid inter and intra subject model for FMRI analysis

Nowadays fMRI studies are generally conducted over cohorts of individuals in order to study brain activity across different areas and using various stimuli. This paper generalizes a joint parcellation detection estimation (JPDE) model in order to perform an inter and intra subject analysis. The proposed approach allows one to estimate group and subject-level parcellations as well as the corresponding underlying dynamics. Moreover, it allows detecting the evoked activity for each individual. The proposed model is evaluated using synthetic and real fMRI data from a group study.