We present a study attempting to validate the corticocortical connection information obtainable from diffusion-weighted MR data. We have implemented probabilistic tractography in data acquired in a macaque model and compared this with connection information in a database of invasive tracer studies in the same model. The nature of the corticocortical interconnection information gained from probabilistic tractography is different to that gained from invasive studies, the latter also being sparse. Our results using the LVE00a parcellation scheme indicate that probabilistic tractography is able to give statistically comparable information on corticocortical interconnections to invasive tracer studies.