When any animal receives sensory inputs, these inputs go to different parts of its brain where they are processed in order to produce some kind of response from the animal. The processing is done by specialized cells called neurons. They typically react to inputs by producing patterns of electrical activity which are propagated to other neurons in the brain. There are many ways to record this electrical activity including special dyes which react to changes in voltage by producing an optical signal. The advantage of these dyes is that you can `see' the activity of many neurons at the same time. Another way to record the activity is through an electrode placed in the brain. This provides better resolution but you can only see the activity of a few cells at once.

In this project we modeled the response of neurons to an odor stimulus in a region of the turtle brain called the olfactory bulb. This is the first place in the brain where smell is processed after turning on the sensory receptors in the nose. The response in the bulb consists of synchronized oscillations of many neurons. There are three distinct rhythms and our project focussed on the one farthest from the point of input. The rhythm starts out at a high frequency and then spontaneously switches to a frequency which is half the original.

The model describes a network of two kinds of cells: excitatory which turn on other cells and inhibitory which turn off other cells. The negative feedback between the two populations is responsible for the rhythm. We suggest that initially there are two populations of excitatory cells which fire out of phase from each other, like people clapping alternately. However, as the excitatory cells fire, they release a substance which causes them to pull together and fire synchronously. This leads to a single effective population firing together at a frequency half the original. slides>>