Anatomy Inspired Hardware for Magnetic Resonance Imaging

Karthik Gopalan

EECS Department
University of California, Berkeley
Technical Report No. UCB/EECS-2022-70
May 11, 2022

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2022/EECS-2022-70.pdf

Magnetic Resonance Imaging (MRI) is a powerful and non-invasive imaging modality that is widely used in medicine. MRI uses a static magnetic field, radiofrequency (RF) pulses, and linearly varying gradient fields to create high quality images of internal anatomy. Despite the ability to safely image soft tissue, MRI inherently suffers from low signal to noise ratio (SNR) which causes long scan times. In addition, the time required to perform multiple scans implies that most imaging is qualitative in nature. Instead of extracting the true value of quantitative parameters like T1, T2, and susceptibility, most MR images capture a single snapshot that is only weighted by these parameters. Speeding up and extracting more information from MR scans is an active area of research. In order to test new reconstruction methods and pulse sequences, it is necessary to have high quality phantoms that accurately mimic human anatomy. SNR of the scan can also be improved with form fitting receive coils since the signal drops off as a cubic function of the distance between the subject and the receiver.

To improve the SNR of MRI scans, this dissertation will discuss two methods for patterning receive coil arrays on complex three-dimensional surfaces. The first method involves spray coating conductive inks and dielectric materials onto a 3D printed surface. The second method uses vacuum forming, sandblasting, and electroless copper plating to deposit conductive traces onto curved polycarbonate substrates. To account for the deformation of vacuum forming, a graphical simulation to pre-distort trace designs was developed. Both processes produce coils that have similar performance to coils made with conventional methods.

This dissertation also discusses the development of slice phantoms that provide a more accurate representation of human anatomy compared to phantoms that are commercially available. A method is presented for preparing and 3D printing these phantoms with any segmented imaging data. A reproducible calibration process is described for creating agar gels that mimic a wide range of T1 and T2 values found in the body.

Advisor: Michael Lustig and Ana Claudia Arias


BibTeX citation:

@phdthesis{Gopalan:EECS-2022-70,
    Author = {Gopalan, Karthik},
    Title = {Anatomy Inspired Hardware for Magnetic Resonance Imaging},
    School = {EECS Department, University of California, Berkeley},
    Year = {2022},
    Month = {May},
    URL = {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2022/EECS-2022-70.html},
    Number = {UCB/EECS-2022-70},
    Abstract = {Magnetic Resonance Imaging (MRI) is a powerful and non-invasive imaging modality that is widely used in medicine. MRI uses a static magnetic field, radiofrequency (RF) pulses, and linearly varying gradient fields to create high quality images of internal anatomy. Despite the ability to safely image soft tissue, MRI inherently suffers from low signal to noise ratio (SNR) which causes long scan times. In addition, the time required to perform multiple scans implies that most imaging is qualitative in nature. Instead of extracting the true value of quantitative parameters like T1, T2, and susceptibility, most MR images capture a single snapshot that is only weighted by these parameters. Speeding up and extracting more information from MR scans is an active area of research. In order to test new reconstruction methods and pulse sequences, it is necessary to have high quality phantoms that accurately mimic human anatomy. SNR of the scan can also be improved with form fitting receive coils since the signal drops off as a cubic function of the distance between the subject and the receiver.

To improve the SNR of MRI scans, this dissertation will discuss two methods for patterning receive coil arrays on complex three-dimensional surfaces. The first method involves spray coating conductive inks and dielectric materials onto a 3D printed surface. The second method uses vacuum forming, sandblasting, and electroless copper plating to deposit conductive traces onto curved polycarbonate substrates. To account for the deformation of vacuum forming, a graphical simulation to pre-distort trace designs was developed. Both processes produce coils that have similar performance to coils made with conventional methods. 

This dissertation also discusses the development of slice phantoms that provide a more accurate representation of human anatomy compared to phantoms that are commercially available. A method is presented for preparing and 3D printing these phantoms with any segmented imaging data. A reproducible calibration process is described for creating agar gels that mimic a wide range of T1 and T2 values found in the body.}
}

EndNote citation:

%0 Thesis
%A Gopalan, Karthik
%T Anatomy Inspired Hardware for Magnetic Resonance Imaging
%I EECS Department, University of California, Berkeley
%D 2022
%8 May 11
%@ UCB/EECS-2022-70
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2022/EECS-2022-70.html
%F Gopalan:EECS-2022-70