Diffusion modulation of the fMRI signal: Early investigations on the origin of the BOLD signal

The early 1990s was a very special period for functional MRI (fMRI). Many original concepts were formed during that period which helped set up the foundations for modern neuroimaging development. I was fortunate to be in graduate school at the time. I was even more fortunate to be enrolled in one of the pioneer groups in fMRI at the Medical College of Wisconsin, and witnessed some of the early fMRI experiments taking place in the lab. Under the daily influence and steady guidance by the extraordinarily talented researchers there, I also began my own work on the contrast mechanisms of fMRI. In particular, I was developing diffusion weighted strategies to investigate the origin of the functional signal using blood oxygenation level dependent (BOLD) contrast. Our results, that there was significant BOLD signal in large veins and their vicinities at low field strengths (e.g. 1.5 T), played an immediate role in moving fMRI applications to higher fields (3 T and above) where small vessels (e.g. capillaries) contribute more significantly to improve the neuronal specificity of the BOLD signal. This manuscript gathers some of my own recollections concerning this particular development.