Ph.D. Biomedical Engineering In many biological systems, information about the environment is detected by a large array of sensory receptors; it reaches more central regions of the nervous system as parallel streams of spike trains. How this flow of information is processed and which features are most salient to the organism is a central problem in neuroscience. A comprehensive and systematic approach to the analysis of the relationships between stimuli and their neural representations is illustrated. This approach is complementary to the hypothesis driven research paradigm, where the investigator states a hypothesis and then performs experiments in order to validate or invalidate the stated hypothesis. This alternative methodology enlarges the possible relationships between stimuli and their representations beyond one specific hypothesis and aims to evaluate all possible relationships between neural stimuli and their neural representations. For a specific set of data, both the set of stimuli and the set of neural response are characterized either by membership in discrete categories or as a continuous space with a similarity measure. In order to quantify these input-output relationships, existing methods of analysis are adapted or new ones developed where necessary. In the specific data sets analyzed here, timing emerges as a critical parameter for the description of neural stimuli and their representations.