Computational neuroscience creates models of diverse aspects of brain function on the basis of experimental findings. In recent years, close collaborations between computational and experimental neuroscientists have resulted in the development of models that are continuously experimentally tested and subsequently refined, contributing to our understanding of brain function and dysfunction.

Reflecting this trend, this month's issue contains two Reviews that discuss computational models. On page 686, Bradley and Goyal outline various theoretical accounts of how the primate visual system detects a moving object's velocity. They discuss how current experimental evidence fits each model and propose additional experiments that will help to further refine velocity-detection models, highlighting the importance of the interplay between experiment and theory for understanding computation in the nervous system.

Computational models are also used to comprehend the behaviour of neuronal networks in health and disease. In a Review on page 696, Rolls and colleagues discuss a number of computational models that have been developed to explain the symptoms of schizophrenia. They outline a model which suggests that altered functioning of NMDA, GABA and dopamine receptors can reduce the stability of the high-firing-rate states of attractor networks in the prefrontal cortex and elsewhere, and show that such diminished stability could explain the divergent symptoms of schizophrenia.

These two Reviews highlight the significance of collaborations between computational neuroscientists and experimentalists and demonstrate the importance of translational approaches for gaining new insights into nervous-system function.