Research Article: Theta rhythm-like bidirectional cycling dynamics of living neuronal networks in vitro

Date Published: February 7, 2018

Publisher: Public Library of Science

Author(s): Arseniy Gladkov, Oleg Grinchuk, Yana Pigareva, Irina Mukhina, Victor Kazantsev, Alexey Pimashkin, Gennady Cymbalyuk.

http://doi.org/10.1371/journal.pone.0192468

Abstract

The phenomena of synchronization, rhythmogenesis and coherence observed in brain networks are believed to be a dynamic substrate for cognitive functions such as learning and memory. However, researchers are still debating whether the rhythmic activity emerges from the network morphology that developed during neurogenesis or as a result of neuronal dynamics achieved under certain conditions. In the present study, we observed self-organized spiking activity that converged to long, complex and rhythmically repeated superbursts in neural networks formed by mature hippocampal cultures with a high cellular density. The superburst lasted for tens of seconds and consisted of hundreds of short (50–100 ms) small bursts with a high spiking rate of 139.0 ± 78.6 Hz that is associated with high-frequency oscillations in the hippocampus. In turn, the bursting frequency represents a theta rhythm (11.2 ± 1.5 Hz). The distribution of spikes within the bursts was non-random, representing a set of well-defined spatio-temporal base patterns or motifs. The long superburst was classified into two types. Each type was associated with a unique direction of spike propagation and, hence, was encoded by a binary sequence with random switching between the two “functional” states. The precisely structured bidirectional rhythmic activity that developed in self-organizing cultured networks was quite similar to the activity observed in the in vivo experiments.

Partial Text

Synchronization and the interplay between excitation and inhibition in neural networks play crucial roles in the organization of rhythmic activity in the brain [1–5]. Rhythmic oscillatory activity with various frequencies represents a multi-clock substrate for cognitive function, memory and sleep [6, 7]. However, researchers still question whether the rhythmicity emerges from the specific network morphology that develops during neurogenesis [7–11] or it is generated spontaneously due to nonlinear network dynamics mediated by an interplay between excitation and inhibition that is sustained by a homeostatic balance [12–14]. An answer to this fundamental question promises to define the network mechanisms of pathological seizure activity and, hence, to determine treatment approaches. Many brain network functions, normal and pathological states have recently been studied using in vitro models [13–23]. In these models, dissociated neuronal cultures provide researchers a unique opportunity to model network dynamics and rhythmicity in vitro.

First, we analysed the spontaneous activity of the hippocampal cultures. We obtained complex bursting patterns similar to those reported previously in cortical cultures [26]. An example of the spikes recorded from a single electrode within a small burst is shown in Fig 1C. After 3–4 weeks of culture in vitro, we obtained the activity described as a superburst (Fig 1E). A typical superburst consisted of a sequence of 3–20 small bursts of 50–100 ms in duration and a 50–150 ms interburst interval. During the period of 30–40 DIV, the cultures generated long superbursts that were similar to regular superbursts, but that lasted for 10–30 seconds and consisted of hundreds of regular small bursts. In summary, we analysed 11 cultures from 3 plating experiments and observed long superbursts in 8 cultures. Six cultures generated more than 6 long superbursts for at least 20 minutes, which were included in the statistical analysis. In other cultures, we observed no more than 2 long superbursts. The signals from a single electrode during the long superburst on timescales of 15 and 2 seconds are illustrated in Fig 1A and 1B. Raster plots of the spiking activity recorded from all 59 electrodes during the superburst and long superburst are shown in Fig 1G. Each point on the raster plot represents the time at which a spike occurred at a particular electrode. The long superbursts were composed of relatively long initiation bursts (50–100 ms) followed by shorter bursts, i.e., the small bursts. The initiation bursts and the small bursts were easily identified by K-means clustering (Fig 1F) using burst firing rate features (see the Methods).

In the present study, the neuronal networks formed by mature hippocampal cultures (30 DIV and older) generated specific network activity with surprisingly long sequences of bursts, i.e., long superbursts of up to hundreds of constituent bursts, with highly regular spiking patterns. In previous studies, superburst activity was reported to display a much shorter duration (up to ten bursts) [26]. In our experiments, we also observed similar activity, but in addition, more than 70% of the cultures (8 of 11) at DIV 30–35 began to spontaneously generate long superbursts with durations of up to hundreds of seconds. Six of 11 cultures generated more than 6 long superbursts during at least 30 minutes. The other cultures generated less than 2 long superbursts with a regular superburst in the background. The precise biophysical mechanism underlying the long superbursts is still largely unknown. We hypothesize that mature cultures (DIV from 35) spontaneously organize into networks with an optimal excitatory-inhibitory balance in which cycling dynamics represent a homeostatic (“natural mode”) pattern that “sustains” the functional connectivity.

 

Source:

http://doi.org/10.1371/journal.pone.0192468

 

0 0 vote
Article Rating
Subscribe
Notify of
guest
0 Comments
Inline Feedbacks
View all comments