Tissue orientation causes bias in MRI measurements of brain blood flow

Venogram image of the brain.

Pictured: a venogram showing blood vessels in the brain.

New research by Dr. Alex Rauscher, an MRI physicist and Canada Research Chair in Developmental Neuroimaging, has shown that about half the blood vessels in the brain's white matter run in parallel with nerve fibres.

Dr. Rauscher's team has been asking fundamental questions to understand how the orientation of brain structures can influence MRI methods. In their most recent study, published in the journal Neuroimage, physics graduate student Jon Doucette observed that brain's blood flow measured with a technique called dynamic susceptibility contrast (DSC) MRI appears much larger in nerve fibres that are perpendicular to the scanner's magnetic field compared to those parallel to the magnetic field.

“Blood vessel-related MRI contrast gives us insight into the brain’s metabolism and function,” explains Dr. Rauscher. “In DSC MRI, we inject a magnetic agent into the blood of an individual and observe how the MRI signal changes when the blood arrives in the brain."

To interpret the their findings, Doucette created a computer model of the brain's blood vessel network and simulated the MRI signal changes due to magnetic agent. The simulations that best matched the measured signal revealed that as much as half of the blood is in vessels that run parallel with the nerve fibres.

These findings have important implications for functional MRI (fMRI), which is widely used in research studies to measure brain activity.

“Our results show that the fMRI signal not only depends on how active a certain brain area is, it also depends strongly on the alignment of nerve fibres. The fMRI signal can be more than twice as high in fibres that go left-right than in those that go in foot-head direction,” said Dr. Rauscher. “This effect creates a systemic bias in fMRI studies, and so researchers need to be more careful when interpreting fMRI results.”

This research is part of a larger project to understand the effects of various tissue properties on the images produced by different MRI applications. This research is supported by the Natural Sciences and Engineering Research Council of Canada (NSERC).