Oscillatory hierarchy in a network of leaky integrate-and-fire neurons with short-term plasticity

Simulation in a network composed of 8000 pyramidal cells and 2000 interneurons. (A) Firing rates of pyramidal cells ν_E and interneurons ν_I in red and blue, respectively, smoothen within a window of 25 ms duration. The frequency of slow oscillations is approximately 0.2 Hz. (B1) Continuous wavelet transform of the firing rate ν_I using Morlet wavelets. The color scale represents the energy for each coefficient of the wavelet decomposition (arbitrary scale). (B2) Normalized cumulated energy for frequencies shown in (B1). (C) Firing rate after applying a band pass filter around frequencies corresponding to slow, δ, α, θ, and γ, respectively. (D1-D3) Cross-frequency coupling between two oscillatory modes f₁ and f₂ represented in polar coordinates. Radius and angle are the normalized average amplitude of f₂ and phase of f₁, respectively, obtained from the wavelet decomposition (method adapted from [1]). The blue line is the distribution calculated from the simulation of leaky integrate-and-fire neurons. The red line is the fitted von Mises distribution and red dot the phase of f₁ corresponding to the maximal average amplitude of f₂. (D1) f₁ = δ, f₂ = α. (D2) f₁ = δ, f₂ = θ. (D3) f₁ = θ, f₂ = γ.