When dissociated embryonic mouse striatal neurons are mixed with neurons from cerebral cortex of the same aged embryos the resulting culture reconstructs a vital part of the brain system known as the basal ganglia. In whole brain, cortical cells make connections with striatal neurons whose properties are influenced by dopamine released from the substantia nigra. The corticostriatal system has been suggested to be the seat of motor learning and habit formation. We hoped to be able to study the properties of this simple two-neuron circuit in these cultures. Because the characteristic neuronal morphology is not well developed in culture we made the cultures from different mouse lines. One of the groups of neurons (usually cortex) came from a mouse genetically modified to express GFP in all cells (the GFP was driven by human UbiquitinC promoter; Tsirigotis et al 2001) and the other from a normal mouse line. In this kind of culture it was relatively easy to find pairs of neurons that were connected together and to study the properties of the synaptic connections. Of the pairs recorded 45% were not connected. In total 32 pairs were connected by the expected excitatory synaptic connections from cortex to striatal neurons. However, 25 pairs were reciprocally connected and a total of 53 pairs were connected in the striatal to cortical direction. This connection never forms in brains. Although we had an interesting preparation – we could not rely on the properties of this system being under the kind of control that was present in brain. So we tried to develop a culture system in which the cortico-striatal system would make the more normal one-way connections. We have developed such a preparation and in our initial experiments we have been able to show that this time the connections are more appropriate. Cortical and striatal neurons are plated in two compartments separated by a 500µm barrier that is removed after the neurons have attached to the substrate. Over the ensuing weeks the cortical neurons grow connections across the gap to reach the striatal neurons. The striatal cell activity recorded on multielectrode arrays are driven by cortical activity and the striatal activity is almost completely silenced by cutting the connections between the cortical and striatal neurons. If we transfect the cortical cells with Channel Rhodopsin2 then the striatal cells are driven by the light induced activity in the cortical neurons. When the striatal cells are similarly transfected we have not seen any response in cortical neurons. This method may also be useful in other situations where the sparse nature of the in vivo connections makes the study of individual pairs of connected neurons difficult, although exactly how similar such simplified systems are to those in brain needs careful study.