Recovering Tissue Volumes From a Single Bioimpedance Spectrum: An Error Analysis From Phantoms and Simulation
Files
Date
Authors
Supervisor
Item type
Degree name
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Bioelectrical impedance analysis (BIA) estimates body composition through equations calibrated on reference populations, and its error for an individual is typically about ±8 percentage points of body fat. A physics-based estimate from a single spectrum would remove this dependence on calibration, although the errors that limit such an estimate have not yet been quantified against a known truth. This work quantifies these errors with three-layer phantoms of known composition, a finite element (FE) model verified against Ansys Maxwell, and simulations with phantom and tissue properties over 1–251 kHz. Without fitted parameters, the FE model predicted the impedance of two cylindrical phantoms measured with spot electrodes to within 1.1% and 6%. With the measurement scale fixed from the phantom dimensions, the fat fraction of all three phantoms was recovered to within 3 percentage points. The fat-free and water fractions, whose material spectra were very similar, were poorly separated. Without a known scale, recovery failed for the two gelatine phantoms. The fat fraction of an agar phantom with a more distinct fat spectrum, however, was still recovered to within 1 percentage point. In simulations, tissue spectra allowed recovery without a known scale, with typical errors of 3–5 percentage points, but only when the tissue properties were known exactly. A 10–20% spread in these properties raised the typical errors to 7–15 percentage points, and an extension of the band to 2 MHz did not reduce them. An uncorrected setup error, measured as 120 pF with a 14 ns delay, reduced recovery of the tissue set from 97% to 58%. Physics-based recovery therefore requires either a known scale, together with a validated electrode model and a corrected setup, or material spectra distinct enough to fix the scale. These findings define the requirements for phantoms intended to test such methods.

