Maximizing the coding capacity of neuronal networks
Chowdhary, S.; Assisi, C.
Show abstract
Information in neuronal networks is encoded as spatiotemporal patterns of activity. The capacity of a network may thus be thought of as the number of stable spatiotemporal patterns it can generate. To understand what structural attributes of a network enable it to generate a profusion of stable patterns, we simulated an array of 9 x 9 neurons modelled as pulse-coupled oscillators. The structure of the network was inspired by the popular puzzle Sudoku such that its periodic responses mapped to solutions of the puzzle. Given that there are nearly a 109 possible Sudokus, this networks could possibly generate 109 spatiotemporal patterns. We show that the number of stable patterns were maximized when excitatory and inhibitory inputs to each neuron were balanced. When this balance was disrupted, only a subset of patterns with certain symmetries survived.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.