Validation of near infrared spectroscopic (NIRS) imaging using programmable phantoms

Abstract

For much of the past decade, we have developed most of the essential hardware and software components needed for practical implementation of dynamic NIRS imaging. Until recently, however, these efforts have been hampered by the lack of calibrating phantoms whose dynamics substantially mimic those seen in tissue. Here we present findings that document the performance of a dynamic phantom based on use of twisted nematic liquid crystal (LC) technology. Programmable time courses of applied voltage cause the opacity of the LC devices, which are embedded in a background matrix consisting of polysiloxane (silicone) admixed with scattering and absorbing materials, to vary in a manner that mimics the spatiotemporal hemodynamic pattern of interest. Methods for producing phantoms with selected absorption and scattering, internal heterogeneity, external geometry, hardness, and number and locations of embedded LCs are described. Also described is a method for overcoming the apparent limitation that arises from LCs being mainly independent of the illumination wavelength. The results presented demonstrate that: the opacity vs. voltage response of LCs are highly stable and repeatable; the dynamic phantom can be driven at physiologically relevant speeds, and will produce time-varying absorption that follows the programmed behavior with high fidelity; image time series recovered from measurements on the phantom have high temporal and spatial location accuracy. Thus the dynamic phantom can fill the need for test media that practitioners may use to confirm the accuracy of computed imaging results, assure the correct operation of imaging hardware, and compare performance of different data analysis algorithms.