Learning of sensory cues is believed to rely on synchronous pre- and postsynaptic neuronal firing. Evidence is mounting that such synchronicity is not merely caused by properties of the underlying neuronal network but could also depend on the integrity of gap junctions that connect neurons and astrocytes in networks too. In this perspective, we set out to investigate the effect of astrocytic gap junctions on perceptual learning, introducing a model for coupled neuron-astrocyte networks. In particular, we focus on the fact that astrocytes are rich of GABA transporters (GATs) which can either uptake or release GABA depending on the astrocyte membrane potential, which is a function of local neural activity. We show that GABAergic signaling is a crucial component of intracolumnar neuronal synchronization, thereby promoting learning by neurons in the same cell assembly that are activated by a shared sensory cue. At the same time, we show that this effect can critically depend on astrocytic gap junctions insofar as these latter could synchronize extracellular GABA levels around many neurons and throughout entire cell assemblies. These results are supported by extensive computational arguments and predict that astrocytic gap junctions could improve perceptual learning by controlling extracellular GABA.

Variability is a prominent characteristic of cognitive brain function. For instance, different trials of presentation of the same stimulus yield higher variability in its perception: subjects sometimes fail in perceiving the same stimulus. Perceptual variability could be attributable to ongoing-spontaneous fluctuation in neuronal activity prior to sensory stimulation. Simulating a cortical neural network model, we investigated the underlying neuronal mechanism of perceptual variability in relation to variability in ongoing-spontaneous neuronal activity. In the network model, populations of principal cells (cell assemblies) encode information about sensory features. Each cell assembly is sensitive to one particular feature stimulus. Transporters on GABAergic interneurons regulate ambient GABA concentration in a neuronal activity-dependent manner. Ambient GABA molecules activate extrasynaptic GABA receptors on principal cells and interneurons, and provide them with tonic inhibitory currents. We controlled the variability of ongoing-spontaneous neuronal activity by manipulating the basal level of ambient GABA and assessed the perceptual performance of the network: detection of a feature stimulus. In an erroneous response, stimulus-irrelevant but not stimulus-relevant principal cells were activated, generating trains of action potentials. Perceptual variability, reflected in error rate in detecting the same stimulus that was presented repeatedly to the network, was increased as the variability in ongoing-spontaneous membrane potential among cell assemblies increased. Frequent, transient membrane depolarization below firing threshold was the major cause of the increased neuronal variability, for which a decrease in basal ambient GABA concentration was responsible. We suggest that ambient GABA in the brain may have a role in reducing the variability in ongoing-spontaneous neuronal activity, leading to a decrease in perceptual variability and therefore to reliable sensory perception.

Default mode network (DMN) shows intrinsic, high-level activity at rest. We tested a hypothesis proposed for its role in sensory information processing: Intrinsic DMN activity facilitates neural responses to sensory input. A neural network model, consisting of a sensory network (Nsen) and a DMN, was simulated. The Nsen contained cell assemblies. Each cell assembly comprised principal cells, GABAergic interneurons (Ia, Ib), and glial cells. We let the Nsen carry out a perceptual task: detection of sensory stimuli. During DMN activation, glial cells were hyperpolarized by Ia-to-glia circuitry, by which glial membrane transporters imported GABA molecules from the extracellular space and decreased ambient GABA concentration. Acting on extrasynaptic GABA receptors, the decrease in ambient GABA concentration reduced inhibitory current in a tonic manner. This depolarized principal cells below their firing threshold during the ongoing spontaneous time period and accelerated their reaction speed to a sensory stimulus. During the stimulus presentation period, the Nsen inhibited the DMN and caused DMN deactivation. The DMN deactivation made Nsen Ia cells cease firing, thereby stopping the glial membrane hyperpolarization, quitting the GABA import, returning to the basal ambient GABA level, and thus enhancing global inhibition. Notably, the stimulus-relevant P cell firing could be maintained when GABAergic gliotransmission via Ia-glia signaling worked, decreasing ambient GABA concentration around the stimulus-relevant P cells. This enabled the Nsen to reliably detect the stimulus. We suggest that intrinsic default model network activity may accelerate the reaction speed of the sensory network by modulating its ongoing-spontaneous activity in a subthreshold manner. Ambient GABA contributes to achieve an optimal ongoing spontaneous subthreshold neuronal state, in which GABAergic gliotransmission triggered by the intrinsic default model network activity may play an important role.

For sensory cortices to respond reliably to feature stimuli, the balancing of neuronal excitation and inhibition is crucial. A typical example might be the balancing of phasic excitation within cell assemblies and phasic inhibition between cell assemblies. The former controls the gain of and the latter the tuning of neuronal responses. A change in ambient GABA concentration might affect the dynamic behavior of neurons in a tonic manner. For instance, an increase in ambient GABA concentration enhances the activation of extrasynaptic receptors, augments an inhibitory current, and thus inhibits neurons. When a decrease in ambient GABA concentration occurs, the tonic inhibitory current is reduced, and thus the neurons are relatively excited. We simulated a neural network model in order to examine whether and how such a tonic excitatory-inhibitory mechanism could work for sensory information processing. The network consists of cell assemblies. Each cell assembly, comprising principal cells (P), GABAergic interneurons (Ia, Ib), and glial cells (glia), responds to one particular feature stimulus. GABA transporters, embedded in glial plasma membranes, regulate ambient GABA levels. Hypothetical neuron-glia signaling via inhibitory (Ia-to-glia) and excitatory (P-to-glia) synaptic contacts was assumed. The former let transporters import (remove) GABA from the extracellular space and excited stimulus-relevant P cells. The latter let them export GABA into the extracellular space and inhibited stimulus-irrelevant P cells. The main finding was that the glial membrane transporter gave a combinatorial excitatory-inhibitory effect on P cells in a tonic manner, thereby improving the gain and tuning of neuronal responses. Interestingly, it worked cooperatively with the conventional, phasic excitatory-inhibitory mechanism. We suggest that the GABAergic gliotransmission mechanism may provide balanced intracortical excitation and inhibition so that the best perceptual performance of the cortex can be achieved.

In visual information processing, feedforward projection from primary to secondary visual cortex (V1-to-V2) is essential for integrating combinations of oriented bars in order to extract angular information embedded within contours that represent the shape of objects. For feedback (V2-to-V1) projection, two distinct types of pathways have been observed: clustered projection and diffused projection. The former innervates V1 domains with a preferred orientation similar to that of V2 cells of origin. In contrast, the latter innervates without such orientation specificity. V2 cells send their axons to V1 domains with both similar and dissimilar orientation preferences. It is speculated that the clustered feedback projection has a role in contour integration. The role of the diffused feedback projection, however, remains to be seen. We simulated a minimal, functional V1-V2 neural network model. The diffused feedback projection contributed to achieving ongoing-spontaneous subthreshold membrane oscillations in V1 cells, thereby reducing the reaction time of V1 cells to a pair of bars that represents specific angular information. Interestingly, the feedback influence took place even before V2 responses, which might stem largely from ongoing-spontaneous signaling from V2. We suggest that the diffusive feedback influence from V2 could act early in V1 responses and accelerate their reaction speed to sensory stimulation in order to rapidly extract angular information.