The neural network was trained to simulate the olfactory system of a fruit fly

Scientists have created a three-layer neural network that classifies odors. After training, the model of communication between neurons accurately reproduced the structure of the olfactory system of a fruit fly.

In fruit flies, the organism in which the olfactory circuit of the brain is best studied, the processing of information about smell begins in the antennae. Sensory neurons equipped with odor receptors convert the chemical bond of molecules with a certain odor into electrical activity. These neurons, which make up the first layer of the olfactory network, send a signal to the second layer, in which sensory neurons having the same receptor converge to the same neuron of the second layer. Since there are fewer neurons in the second layer than in the first, this part is considered a compression layer. The neurons of the second layer, in turn, transmit a signal to a larger set of neurons in the third layer – the expansion layer.

Scientists have created a network of artificial neurons consisting of an input layer, a compression layer and an expansion layer and with the same number of neurons as in the olfactory system of a fruit fly, but without its inherent structure, forcing connections between neurons to be reprogrammed as the model was trained to classify odors. Self-organization of the neural network took several minutes. The resulting structure turned out to be similar to that found in the brain of a fruit fly.

The results of the study suggest that biological neural networks that interpret olfactory information are optimally organized to perform their task. The work also demonstrates the importance of artificial neural networks for research in neuroscience.

Subscribe
Notify of
guest

0 Comments
Inline Feedbacks
View all comments

aischool