Identification and classification of voxels of human brain for rewardless-related decision making using ANN technique

Functional magnetic resonance imaging (fMRI) data analysis is developing rapidly in a fast emerging society, because of temporal and spatial resolution and the inoffensive feature of their acquisition in human brains. Spatial resolution governs how “sharp” the image is in appearance, whereas temporal resolution denotes the precision of a measurement with respect to time. The goal of fMRI technique is to identify the activation pattern and functional connectivity in the brain regions. In our study, artificial neural network technique was used for the identification and classification of decision-making voxels of fMRI data based on Brodmann areas 10 and 47 from the prefrontal cortex of human brain. The total number of voxels of Brodmann areas was 159, and we determined that some particular voxels played dominant role for decision-making process while performing a visual task. We also analyzed true positive and false positive classification rates between two decisions in the context of a well-known receiver operating characteristic curve (ROC).