Theta band zero-lag long-range cortical synchronization via hippocampal dynamical relaying

Abstract

Growing evidence suggests that synchronization among distributed neuronal networks underlie functional integration in the brain. Neural synchronization is typically revealed by a consistent phase delay between neural responses generated in two separated sources. But the influence of a third neuronal assembly in that synchrony pattern remains largely unexplored. We investigate here the potential role of the hippocampus in determining cortico-cortical theta synchronization in different behavioral states during motor quiescent and while animals actively explore the environment. To achieve this goal, the two states were modeled with a recurrent network involving the hippocampus, as a relay element, and two distant neocortical sites. We found that cortico-cortical neural coupling accompanied higher hippocampal theta oscillations in both behavioral states, although the highest level of synchronization between cortical regions emerged during motor exploration. Local field potentials recorded from the same brain regions qualitatively confirm these findings in the two behavioral states. These results suggest that zero-lag long-range cortico-cortical synchronization is likely mediated by hippocampal theta oscillations in lower mammals as a function of cognitive demands and motor acts.

Figure 1. Dynamical characterization of the hippocampus and cortical regions during the generation of theta oscillations.

Panels A, B and C show the voltage v time traces of 10 randomly chosen neurons (8 excitatory in black and 2 inhibitory in grey) of each population in the absence of local and long-range connections. Panels D, E and F show the same time traces of neurons locally connected within each population. Panel G shows the ensemble average voltage v of each area: Frontal cortex (F), Visual cortex (V) and the Hippocampus (H). Panel H shows raster plots. Panel I shows an average number of coincident spikes of neuron pairs of the same population, obtained from the auto-correlation function and subtracted from the mean number of coincidences over the delay window. The upper figure in panel I displays cortical groups while the bottom figure stands for the hippocampus. External driving to each neuron is given by 100 independent excitatory neurons spiking according to a Poisson distribution with average rate r = 16.3 Hz.

Figure 2. Modeling neuronal dynamics underlying passive and active behavioral states.

Panel A represents the simple motif connecting the brain regions F, V and H. Each neuron is driven by an independent Poisson process of rate r = 16.3 Hz (r = 15.4 Hz) for the active (passive) state. In panels B and C, the ensemble average voltage for the passive and active states are plot respectively. Panels D and E include the corresponding raster plots.

Figure 3. Zero-lag cortico-cortical synchronization for different motifs.

Simulation results for the ensemble average voltage of the cortical regions are shown for two external drives corresponding to the active (r = 16.3 Hz) and passive (r = 15.4 Hz) states. Regardless of the behavioral state, we found that the two cortical areas (frontal and visual) do not synchronize at zero-lag when mutually connected without the hippocampal relay (panels A–C). Neither we observed zero-lag synchronization when only the hippocampus drives them (panels D–F). The cortical feedback to the hippocampus (as depicted in panel G) is critical to promote zero-lag cortico-cortical synchronization, as depicted in panels H and I.

Figure 4. Spatio-temporal synchronization obtained from the experimental and numerical data.

We plot here the density of spikes in the sliding window of filtered time series cross-correlation (see Materials and Methods section). The window has 300 ms length and is shifted by 50 ms steps and analyzed over the 60-s of continuous artifact-free LFP recordings for each behavioral state and animal (n = 4), separately. Results are normalized in a frame of −110 to 110 ms. Experimental data correspond, in this example, to an individual mouse, although other mice presented qualitatively similar results. Simulations show high agreement with experimental results for both active and passive behavioral states.

Figure 5. Effects of an asymmetric delay time in the inter-population couplings.

If the delay time between the hippocampus and the visual area (T in the figure) is slightly different from that between the hippocampus and the frontal area (20 ms), the maxima of the cross-correlations between the hippocampus and the cortical areas become different, as shown in the experiments (Fig. 4, upper panels).