Constrained MRI Impedance Imaging
Abstract
An approach for measuring local electric tissue properties in-vivo is proposed. The technique relies upon the integration of MRI data with electrical potential measurements. Through a technique referred to as constrained electric impedance tomography (CEIT), the conductivity of specific regions identified in MRI data can be calculated. A two-dimensional simulation of a phantom containing three-regions of varying electrical conductivity was generated and a finite-element model of electrical current propagation created. The geometry encompassed a total of 24 electrodes arranged around the object with voltages applied between all possible electrode pairs. The resulting potentials for the remaining electrodes were calculated. In order to constrain the reconstruction, co-registered MRI data was used to define the boundaries of the internal structures. Conductivity values were assigned to these regions and the resulting field patterns calculated form a finite element model of the problem. The assumed conductivity was adjusted with an iterative nonlinear optimization technique that minimized the difference between the calculated and measured electrode potentials. When used in the absence of noise, the calculations converged quickly toward the correct conductivity values. However, with the addition of noise (SNR=10), we noted conductivity errors of -20% for high conductivity structures. These encouraging results demonstrate the electric impedance tomography reconstruction is possible by integrating MRI derived spatial constraints. Further improvement would be expected in a full 3D geometry to provide a greater number of independent measures for the iteration process. Further research in this topic is warranted.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 2003
- Accession Number
- ADA423254
Entities
People
- Donald B. Plewes