Significance: We present a Monte Carlo (MC) computational framework that simulates near-infrared (NIR) hyperspectral imaging (HSI) aimed at assisting quantification of the in vivo hemodynamic and metabolic states of the exposed cerebral cortex in small animal experiments. This can be done by targeting the NIR spectral signatures of oxygenated (HbO2) and deoxygenated (HHb) hemoglobin for hemodynamics as well as the oxidative state of cytochrome-c-oxidase (oxCCO) for measuring tissue metabolism.
Aim: The aim of this work is to investigate the performances of HSI for this specific application as well as to assess key factors for the future design and operation of a benchtop system.
Approach: The MC framework, based on Mesh-based Monte Carlo (MMC), reproduces a section of the exposed cortex of a mouse from an in vivo image and replicates hyperspectral illumination and detection at multiple NIR wavelengths (up to 121).
Results: The results demonstrate: (1) the fitness of the MC framework to correctly simulate hyperspectral data acquisition; (2) the capability of HSI to reconstruct spatial changes in the concentrations of HbO2, HHb, and oxCCO during a simulated hypoxic condition; (3) that eight optimally selected wavelengths between 780 and 900 nm provide minimal differences in the accuracy of the hyperspectral results, compared to the "gold standard" of 121 wavelengths; and (4) the possibility to mitigate partial pathlength effects in the reconstructed data and to enhance quantification of the hemodynamic and metabolic responses.
Conclusions: The MC framework is proved to be a flexible and useful tool for simulating HSI also for different applications and targets.
Keywords: Monte Carlo methods; biomedical optics; brain hemodynamics and oxygenation; brain metabolism; cytochrome-c-oxidase; hyperspectral imaging.