Date of Award


Degree Name

MS in Electrical Engineering


Electrical Engineering


Dennis Derickson


As implantable medical devices are being used more often to treat medical problems for which pharmaceuticals don’t suffice, it is important to understand their interactions with commonly used medical modalities. The interactions between medical implants and Magnetic Resonance Imaging machines have proven to be a risk for patients with implants.

Implanted medical devices with elongated metallic components can create harmful levels of local heating in a Magnetic Resonance Imaging (MRI) environment [1]. The heating of a biological medium under MRI is monitored via the Specific Absorption Rate (SAR). SAR, defined as power absorbed per unit mass (W/kg), can be calculated as , where σ is electrical conductivity of the medium in units of , |E| is the magnitude of the applied electric field in units of , and ρ is the density of the medium in units of . For continuous, uniform power deposition this can be measured experimentally as a rise in temperature over time (∆T/t), where c is the specific heat capacity of the medium in units of. To understand the SAR induced in-vivo, a phantom (Figure 2.10) is used to conduct in-vitro experiments, as it provides a controllable and repeatable experimental setup.

In order to experiment in the phantom, an understanding of the background SAR distribution and in turn the exposure field distribution of the phantom is required as per the ASTMF2182-09 standard [2]. In this work, the background SAR distribution of an ASTM standard torso phantom is measured and studied via fiber optic thermometry. The measurements are compared with an electromagnetic model simulated via FDTD, demonstrating agreement between 10-25%. A custom exposure and data collection setup (including oscilloscope, function generator, RF amplifier, directional coupler, and Neoptix Omniflex Fiber Optic Thermometry system) was integrated and automated using NI LabView.

The purpose of this thesis is to map the field distribution in a torso phantom under RF exposure from a 64 MHz MRI RF Birdcage, compare the results to an electromagnetic simulation, and finally conclude the accuracy of this method for field measurements in a standard torso phantom. Understanding the capabilities and accuracy of the fiber optic thermometry method will ultimately allow researchers to successfully apply this method to monitor background fields in their respective experimental setups (related to MRI implant heating) and understand its limitations.