Chapter 16 - Validation of Tractography

Abstract

The validation of tractography is fundamental to the implementation of the technique as a useful biomedical tool. The technique has the potential to help in the diagnosis of patients suffering from brain injury and disease, as well as providing insights into basic neuroscience. It discusses the studies of simple biological models, in which the complexity is increased at the expense of a decrease in control of the underlying architecture and knowledge of the ground truth. Comparison of tractography results with those of invasive tract tracing techniques returns some degree of ground-truth knowledge, allowing the methodology to be tested in vivo and in the brain; however, the invasiveness of dissection studies means that these methods can only be used in animal studies. Yet it is humans that are ultimately the subjects on which the fiber tracking techniques must be validated. The use of known anatomy is an integral part of the validation process, as is the use of circumstantial validation in the form of functional imaging and lesion studies. The search for a “gold standard” for tractography is the search for an ideal model that possesses true axonal characteristics, combined with an ideal model of signal generation—as yet no suitable software or physical phantom has been created. However, a range of work has been reported in which the attributes and pitfalls of the assortment of fiber tracking methodology have been both qualitatively and quantitatively assessed, and which provides valuable information regarding the reliability of the tractography process. The potential of tractography to map the three-dimensional network of connections between brain regions in a non-invasive manner is one of the most exciting recent developments in neuroimaging.