Modeling and tissue parameter extraction challenges for free space broadband fNIR brain imaging systems

Fiber based functional near infra-red (fNIR) spectroscopy has been considered as a cost effective imaging modality. To achieve a better spatial resolution and greater accuracy in extraction of the optical parameters (i.e., μa and μ's), broadband frequency modulated systems covering multi-octave frequencies of 10-1000MHz is considered. A helmet mounted broadband free space fNIR system is considered as significant improvement over bulky commercial fiber fNIR realizations that are inherently uncomfortable and dispersive for broadband operation. Accurate measurements of amplitude and phase of the frequency modulated NIR signals (670nm, 795nm, and 850nm) is reported here using free space optical transmitters and receivers realized in a small size and low cost modules. The tri-wavelength optical transmitter is based on vertical cavity semiconductor lasers (VCSEL), whereas the sensitive optical receiver is based on either PIN or APD photodiodes combined with transimpedance amplifiers. This paper also has considered brain phantoms to perform optical parameter extraction experiments using broadband modulated light for separations of up to 5cm. Analytical models for predicting forward (transmittance) and backward (reflectance) scattering of modulated photons in diffused media has been modeled using Diffusion Equation (DE). The robustness of the DE modeling and parameter extraction algorithm was studied by experimental verification of multi-layer diffused media phantoms. In particular, comparison between analytical and experimental models for narrow band and broadband has been performed to analyze the advantages of our broadband fNIR system.