It is known that on the single cell level, neurons retain many of the properties that they have in-vivo, ex-vivo. In this research we investigate whether the same is true for networks of neurons also. We concentrate on one of the most fundamental properties of neuronal networks: learning and memory. To do this we culture neuronal networks on multi-electrode arrays (MEA's). This allows us to measure the action potentials generated by a number of neurons (up to 61 in our current configuration) simultaneously. Besides measuring, we can also provoke action potentials, by applying a voltage to an electrode. This allows us to set up an ‘environment' by making stimulation a function of what we measure. Also, the network can be given a ‘goal' within this environment. This allows us to asses the performance of the network and the effectiveness of a learning strategy. We also want to couple the changes in performance to changes in physiological parameters of the network.
Interactions with a physical environment requires us to interpret (decode) the patterns of activity and translate them into actions (e.g. muscle contractions). The problem is circumvented if the performance measure is based directly on the activity patterns.
Address: 5636 Lemon Ave.
Dallas TX 75209
Phone: +1 214 5203694