Preliminary Fusion of Structural Brain Data and Developmental Scores in Epileptic Children Through Canonical Polyadic Decomposition (CPD)

This study analyzes structural brain connectivity and developmental scores in children with epilepsy using Canonical Polyadic Decomposition (CPD). By leveraging Diffusion Tensor Imaging (DTI) to generate connectivity matrices, we integrate them with MRI volumes and phenotypic scores and compare them to a healthy group. The model demonstrates accuracy and stability in identifying specific features associated with cognitive and behavioral functions in both groups, revealing significant differences. Briefly, connections in the Lingual-Gyrus, Temporal-Occipital-Fusiform-Cortex, Occipital-Fusiform-Gyrus, Right-Pallidum, and the volume of the Left-Putamen are highlighted. These results enhance the understanding of brain connectivity patterns in epilepsy and their relationship to developmental outcomes, potentially improving diagnosis, treatment planning, and symptom management for young epileptic children.