Multispectral photoacoustic tomography provides a mapping of the tissue chromophore distribution by using a broad range of tunable laser wavelengths. While a large number of employed wavelengths may improve the separation of exogenous and endogenous chromophores, it also decreases the temporal resolution and increases the total energy deposition, which can cause bleaching of fluorescent dyes. With the overall goal of studying the dynamics of cerebrospinal fluid in mice in vivo, the work in this thesis aims to minimize the number of wavelengths in order to reduce the scanning time, improve the temporal resolution, reduce the energy deposition into the tissue and avoid photobleaching of the tracers while maintaining high image quality. Two methods were used to select small sets of wavelengths: the first method of wavelength selection was based on the minimization of the condition number of the extinction matrix, and the second method was based on the peak signal-to-noise ratio optimization of the chromophore images. 3 wavelengths set was selected based on minimum condition number proved to be enough to maintain good image quality and accurate mapping for chromophores and tracer.