Kruskal's Minimum Spanning Tree Approach to Brain Fiber Tractography Computation

Abstract

Kruskal's minimum spanning tree algorithm is a graph algorithm which constructs the tree with the minimum total edge length. The starting node of the tree generated by Kruskal's algorithm is not important. This is an advantage in diffusion tensor fiber analysis. In this study, a tractography approach, which does not depend on user selection, independent from a starting node, covers the whole region of interest, and calculates all the connectivity, is proposed. In this diffusion tensor tractography study, the advantages and limitations of the Kruskal's minimum spanning tree method are discussed, and a tractography algorithm based on this approach is developed.