Functional magnetic resonance imaging (fMRI) has become a tool of major importance in neurology. Typical experiments consist of exposing the subject lying in the scanner to different stimuli, in order to detect activation patterns in the brain. However, the acquired data suffers from low signal-to-noise ratios, and therefore its analysis requires the application of solid statistical methods. Statistical parametric mapping (SPM) is probably the most popular tool for this kind of analysis. Recently, a novel framework based on wavelets has been proposed by our lab. This technique provides a solution to a major disadvantage of the standard wavelet-based approach; i.e., it combines powerful processing in the wavelet domain with statistical testing in the spatial domain. Until now, the technique has been tested and validated using separable wavelet transforms (with dyadic subsampling matrices). The purpose of this diploma project is to incorporate and test another type of wavelets, known as the polyharmonic B-spline wavelets, into the Wavelet Toolbox. The polyharmonic wavelets have several attractive properties that could turn out to be useful for fMRI analysis: (1) the subsampling matrix is quincunx, thus providing a slower progression through scale, (2) the wavelet behaves as the isotropic Laplacian operator, which could be used to better exploit the statistical properties of fMRI data.