The non-invasive research of information inside biological tissues can be made by means of setups using continuous, time-dependent or frequency modulated light sources, which emit in the visible or near-infrared range. Moreover, the biological structures such as brain, breast or fruits, can be regarded as closer to a spherical shape than a slab. This paper focus on the retrieval of tissue optical parameters in a spherical geometry using fittings with analytical solutions adapted for semi-infinite geometry. The data were generated using three different optical spectroscopy methods: frequency-resolved, spatially-resolved, and time-resolved modes. Simulations based on a Monte Carlo code were performed on homogeneous spheres, with 18 spaced detectors located on their boundary. First, data are examined in the frequency domain. Second, they are treated with optimization algorithms to assess the optical coefficients. The computations show that the spatially-resolved measurements are often more robust than those related to the frequency-resolved mode. In the temporal domain, errors on the estimates are also exhibited with the fitting by the Fourier transform of a solution based on the semi-infinite geometry. Furthermore, when the analytical solution is modified by taking into account the spherical shape, the retrieval of the coefficients is improved.