Chapter 20 – Validation of Tractography

This chapter leads the reader through the scaffold of validation studies that have been carried out with the aim of supporting tractography. We start with software and physical phantoms, which are user-definable and easy to manipulate, but are gross approximations of the in vivo brain. We then discuss simple biological models, in which complexity is increased at the expense of decreased control of underlying architecture and ground truth. Comparing tractography with invasive tract-tracing techniques returns some degree of ground-truth knowledge, allowing testing in the in vivo brain. However, the invasiveness of dissection studies and the potential toxicity of classical histological and MR-visible tracers, means these methods can only be used in animals. It is humans that are ultimately the subjects on which fiber tracking must be validated. The use of known anatomy and circumstantial validation, such as functional imaging and lesion studies, is an integral part of the validation process.