Joint analysis of gyral folding and fiber shape patterns

Abstract

Cortical gyral folding pattern can be described by the hinge numbers, e.g., 2-hinge and 3-hinge gyri. Our recent studies have shown that joint analysis of 2-hinge gyral folding and its fiber connection patterns can be a powerful tool to elucidate the regularity and variability of gyral structures. However, other more complex gyral shape patterns such as 3-hinge gyri have not been investigated yet. In this paper, we proposed a novel computational framework to characterize 3-hinge gyral folding patterns and jointly analyze the correlation between folding and DTI-derived fiber orientation patterns in these regions. To characterize and differentiate these 3-hinge gyral folding patterns, an effective decision tree model is designed and applied to classify these 3-hinge gyral patches into the combinations of five basic sub-types of `S', `U', `V', `I', and other shapes. Then, the correlations between the shapes of these sub-types of 3-hinge gyri and their DTI-derived fiber connection patterns are quantitatively assessed, and our results demonstrated their close relationships. This work further replicated and strengthened our previous finding obtained in 2-hinge gyri that gyral folding patterns are closely correlated with fiber connection patterns.