Recent experiments show that the change of a synaptic weight from a neuron i to a neuron j depends on the relative timing of the pulses of neurons i and j. What are the computational consequences of such asymmetric learning rules?
Some animal brains contain a spatial representation of their environmont other’s don’t. What could be the use of spatial maps for navigation? How much can be achieved with pure reflexes without navigation? How can learning be incorporated in spatial representations?