A microfabricated phantom with application in diffusion tensor imaging (DTI) is presented. Using lithography
technique, we have the capability of creating microchannels in the same scale as actual neural fibers (few to tens of
microns in diameter). The method is flexible in generating different geometrical patterns. Neural bundles were simulated
by designing a large number of microchannels, running parallel to each other.
PDMS, the casting material, is not diffusible to water. This applies the restriction to the diffusion of water molecules in
different directions. The dimensions of the channels can be calculated based on the desired fractional anisotropy (FA)
ratios. Many unresolved issues in studying and improving diffusion tensor imaging can be carefully investigated by
implementing an artificial model of neural bundles with well-known geometrical parameters. Problems affiliated with
the crossing fibers in 'tractography' are among these issues. Since the topology of the tracts is known, whenever the full
characterization of water motion is desired, the elements of the diffusion tensor can be calculated to be compared to the
measured values. Optimization of the pulse sequences and calibration of the gradients in DTI are among other
application of such phantom.
The phantom is made of PDMS, a silicon based, MRI compatible material. Once the mold is generated, creating new
phantoms is easy, quick and inexpensive and requires no special equipment.