To present the argument that the only secure foundation for a theory of behaviour, and ultimately of mind, rests at the level of single neurons, and to assess progress at this level of explanation.

Relevant data were obtained by a search of PubMed, last updated in January 2007, focused on implemented models from single-neuron studies.

Technical limitations on recording neural activity produce trade-offs between temporal and spatial resolution and the ability to track the massively parallel activity of the nervous system. The properties of the single neuron that would need to be measured and the techniques available to obtain the data are described. The concept of a fixed neuronal identity may be impeding progress and should be replaced with the concept of dynamically assigned neuron identity.

Modern data collection techniques make it possible to obtain data at the single-neuron level on the complete nervous systems of simple organisms. Present models based on this data do not provide an integrated explanation of behaviour. However, there do not appear to be insurmountable theoretical or practical obstacles to building such models in the future or of scaling the data collection up to more complex organisms.