Prof. Coleman's work is chiefly in the areas of network information theory and computational neuroscience. In network information theory, his focus is on developing theoretical limits of reliable communication for universal coding, wireless networks, and transmisson of correlated sources across networks, and also on the construction of robust architectures and practical encoding/decoding algorithms that achieve reliable performance near the theoretical boundaries. In computational neuroscience, his work includes the modeling of dynamic representations between stimuli and neural activity in a canonical way, with measures of accuracy that can be quantified statistically; the development of practical mixed-modality information fusion algorithms for neural signal analysis over varying time-scales; and the development of new algorithms and interfaces for neural prosthetics.