Contributions in computational intelligence with results in functional neuroimaging

This thesis applies computational intelligence methodologies to study functional brain images. It is a state-of-the-art application relative to unsupervised learning domain to functional neuroimaging. There are also contributions related to computational intelligence on topics relative to clustering validation and spatio-temporal clustering analysis. Speci_cally, there are the presentation of a new separation measure based on fuzzy sets theory to establish the validity of the fuzzy clustering outcomes and the presentation of a framework to approach the parcellation of functional neuroimages taking in account both spatial and temporal patterns. These contributions have been applied to neuroimages obtained with functional Magnetic Resonance Imaging, using both active and passive paradigm and using both in-house data and fMRI repository. The results obtained shown, globally, an improvement on the quality of the neuroimaging analysis using the methodological contributions proposed.