Now-a-days a Conflict identification and categorization in brain functional MRI (fMRI) are inherently a toilsome in research. It is particularly because of the overlapping intensity distribution between the healthy and pathological tissues in the fMRI. The important features of that characterize the brain have to be diagnoses for efficient categorization and deblocking of contradiction from fMRI. Since MRI suffers from substantial grayscale contrast the categorized procedure should be done in a trained manner. This work proposes a Neuro-fuzzy based system for categorization and deblocking of abnormalities from Brain fMRI. The work consists of three major stages such as Feature deblocking, categorization and conflict detection. In the feature deblocking phase vital data that drive to categorization are analyzed. Texture and Wavelet features are used as discriminating features to diagnose the image class. The categorization phase discriminates the normal and pathological fMRI slices using feed forward Back propagation neural network. The categorized abnormal images are then applied for feature extraction and comparison of them with a ground truth data.