Fuzzy Relational Approach to Determine Functional Brain Connectivity in Learning Tasks for Dyslexia Children Using an f-NIRS Device

The paper deals with functional brain-connectivity analysis between pairs of brain lobes for a given learning task by utilizing the fuzzy implication relations between the extracted signal features acquired from selected channels of the lobes. The Dienes-Rescher type fuzzy implication relation is chosen for its closest similarity with propositional implication, resembling implication in the true sense of its logical semantics. The Diens-Recher type implication has successfully been employed to check similarity in functional brain-connectivity for healthy (normal) children (below 2 years) in a learning task of fruits and animals. It is noted that children with Dyslexia disease exhibit different brain-connectivity patterns with respect to those of healthy subjects. This very finding opens up a new vista of research to recognize dyslexia patients from their healthy counterpart. Additionally, the k-means clustering algorithm is employed to cluster children suffering from Dyslexia into groups, based on their similarity in possible functional brain-connectivity. Such similarities of Dyslexia patients indicate commonality in wrong terminations of neural pathways, which is a well-known phenomenon in Dyslexia. The performance analysis undertaken reveals that the proposed fuzzy relational approach outperforms classical Granger causality, convergence cross-mapping technique, probabilistic relative correlation adjacency matrix and transfer entropy approaches with respect to 2 metrics: modularity and average efficiency.