Data analysis methods for near-infrared spectroscopy of tissue: problems in determining the relative cytochrome aa3 concentration

摘要

Abstract: In the brain of the adult rat, the ratio of the absorption coefficient of hemoglobin to that of the cytochromes is approximately ten and in the newborn rat brain the ratio is even higher. Additionally the absorption spectra of these compounds overlap markedly. Under these circumstances the accurate determination of cytochrome concentration is difficult. There are many possible sources of error: (i) Non linear measuring equipment. (ii) Inaccurate hemoglobin and cytochrome spectra. (iii) A wavelength dependent effective optical pathlength. (iv) An absorption coefficient dependent effective optical pathlength. (v) Oxygenation dependent changes in tissue scattering. The first two sources of error can be solved with careful instrumental and experimental design. The last three are much more problematic, but can be addressed using time resolved measurements. These are the topic of this paper. A wavelength dependence of the optical pathlength leads to a distortion of the optical spectra of the chromophores in brain tissue. A simple method of examining the wavelength dependant effects is discussed. The selection of the correct wavelength range is important in minimizing these problems. Until recently, all near infrared data processing 'algorithms' have assumed a linear Beer Lambert relationship between the measured attenuation spectra and tissue absorption coefficient. However, picosecond optical techniques have shown that at a single wavelength, the optical pathlength in the rat brain can vary by 10% implying that the Beer Lambert law is not strictly valid. A non linear correction of tissue spectra which can be based on results from time of flight measurements is described.!