|
|
What are the general features of useful computation in a circuit? Apparently, the computation should be reliable. If the output of the circuit is variable from trial to trail, the information about the input will not be conveyed reliably or accurately. At the same time, it is beneficial for the circuit to be able to flexibly change its properties to effectively function under various environments. For example, if the level of some input becomes low from injury of presynaptic neurons or change in the statistics of input, the other components of the circuit could change their properties to compensate for the input deprivation.
These seemingly incompatible features, namely reliability and flexibility, are hallmarks of many brain circuits. However, we still do not understand well how these are achieved in the same circuit. Here, I will discuss the mechanisms underlying these features at synaptic, cellular, and circuit levels using the Drosophila antennal lobe, an analog of the vertebrate olfactory bulb, as a model circuit. I will show that neurons in the antennal lobe are connected to each other with an extraordinarily high accuracy and the properties of specific connection are modified in response to particular perturbation.
|
|