COHERENCY AND CONNECTIVITY IN OSCILLATING NEURAL NETWORKS - LINEAR PARTIALIZATION ANALYSIS

This paper studies the relation between the functional synaptic connec tions between two artificial neural networks and the correlation of th eir spiking activities. The model neurons had realistic non-oscillator y dynamic properties and the networks showed oscillatory behavior as a result of their internal synaptic connectivity. We found that both ex citation and inhibition cause phase locking of the oscillating activit ies. When the two networks excite each other the oscillations synchron ize with zero phase lag, whereas mutual inhibition between the network s resulted in an anti-phase (half period phase difference) synchroniza tion. Correlations between the activities of the two networks can also be caused by correlated external inputs driving the systems (common i nput). Our analysis shows that when the networks exhibit oscillatory b ehavior and the rate of the common input is smaller than a characteris tic network oscillator frequency, the cross-correlation functions betw een the activities of two systems still carry information about the mu tual synaptic connectivity. This information can be retrieved with lin ear partialization, removing the influence of the common input. We fur ther explored the network responses to periodic external input. We fou nd that when the input is of a frequency smaller than a certain thresh old, the network responds with bursts at the same frequency as the inp ut. Above the threshold, the network responds with a fraction of the i nput frequency. This frequency threshold, characterizing the oscillato ry properties of the network, is also found to determine the limit to which linear partialization works.