Highly Connected Neurons Influence Spontaneously Synchronized Bursts in Cortex

Figure 1. Effects of removing hub neurons on cortical activity patterns. (A) Spontaneous bursting activity in the intact cortical circuit. The x-axis is labeled time in seconds. The y-axis for the top raster plot is labeled with the cortical layer, with each horizontal set of rasters from an individual neuron. The y-axis in the bottom plot is labeled with the total number of spikes per second. (B) Same as A except 2,977 hub neurons were turned off. (D) Same as B except 2,977 randomly selected neurons were turned off. From figure 2 in “The Role of Hub Neurons in Modulating Cortical Dynamics” (2021).

Brain cells with more than the average number of connections with other neurons are known as hub neurons. Hub neurons have been shown to significantly decrease the average path length of communication from any one neuron to another, which results in what is called a small world network. Recent research investigated hub neurons and their contributions to simulated cortical circuit activity (“The Role of Hub Neurons in Modulating Cortical Dynamics“, 2021).

Synchronous bursts of activity at about 1 Hz (Figure 1A above) was spontaneously generated by a data-driven simulation of a 0.3 cubic millimeters piece of cortex. The cortical microcircuit contained 31,000 neurons, about 37 million synapses, and 55 morphological cell types. Interestingly, 1 Hz is the delta oscillation frequency, which is most prominent in cortex during deep sleep. Unfortunately, spectral analysis was not addressed in this paper.

The number of network bursts was precipitously reduced and periodicity was far less sharply defined when the investigators turned off 2,977 hub neurons selected randomly among the hub neuron sub-population (Figure 1B above). In contrast, network bursting remained unaltered when they turned off 2,977 randomly selected neurons (Figure 1D above). Hub neurons clearly contributed to the 1 Hz spontaneously synchronous bursts of activity.

Next the research team turned off hub neurons one cortical layer at a time (except for layer 1) and then observed network activity. Turning off random hub neurons across all layers had the most robust affect on all network activity measures: reduced bursts, reduced firing rates, reduced coefficient of variation, and decreased correlation. Removing layer 5 hub neurons had the largest effect of the layer specific manipulations. Interestingly, removing a relatively small number of layer 4 hub neurons reduced spontaneously synchronized burst activity well before effects were seen from hub neuron removal from any other layer or from across all layers.

In summary, hub neurons in general and layer 5 hub neurons in particular contributed to the spontaneously synchronized bursts of activity in simulated cortical microcircuit experiments (“The Role of Hub Neurons in Modulating Cortical Dynamics“, 2021). These intriguing results suggest taking closer looks at the effects of specific layer 5 neuron types that have hub connectivity. It would also be informative to see a more detailed look at the cortical circuit’s neural dynamics. In particular, is the pre-manipulated simulated cortical circuit activity equivalent to awake activity in real brains? Deep sleep activity in real brains? Cortical slice activity? And what do the spectral analyses show?

See Also

Cortical Microcircuit States: Follow-up to Yesterday’s Post