A novel machine learning-based approach for the detection and analysis of spontaneous synaptic currents

Spontaneous synaptic activity is a hallmark of neural networks. A thorough description of these synaptic signals is essential for understanding neurotransmitter release and the generation of a postsynaptic response. However, the complexity of synaptic current trajectories has either precluded an in-depth analysis or it has forced human observers to resort to manual or semi-automated approaches based on subjective amplitude and area threshold settings. Both procedures are time-consuming, error-prone and likely affected by human bias. Here, we present three complimentary methods for a fully automated analysis of spontaneous excitatory postsynaptic currents measured in major cell types of the mouse retina and in a primary culture of mouse auditory cortex. Two approaches rely on classical threshold methods, while the third represents a novel machine learning-based algorithm. Comparison with frequently used existing methods demonstrates the suitability of our algorithms for an unbiased and efficient analysis of synaptic signals in the central nervous system.

Computational modelling of trans-synaptic nanocolumns, a modulator of synaptic transmission

Nanocolumns are trans-synaptic structures which align presynaptic vesicles release sites and postsynaptic receptors. However, how these nano structures shape synaptic signaling remains little understood. Given the difficulty to probe submicroscopic structures experimentally, computer modelling is a usefull approach to investigate the possible functional impacts of nanocolumns. In our in silico model, as has been experimentally observed, a nanocolumn is characterized by a tight distribution of postsynaptic receptors aligned with the presynaptic vesicle release site and by the presence of trans-synaptic molecules which can modulate neurotransmitter diffusion. We found that nanocolumns can play an important role in reinforcing synaptic current mostly when the presynaptic vesicle contains a small number of neurotransmitters. We also show that synapses with and without nanocolumns could have differentiated responses to spontaneous or evoked events. Our work provides a new methodology to investigate in silico the role of the submicroscopic organization of the synapse.Author summaryNeurotransmitter release, diffusion, and binding to postsynaptic receptors are key steps in synaptic transmission. However, the submicroscopic arrangement of receptors and presynaptic sites of neurotransmitter release remains little investigated. Experimental observations revealed the presence of trans-synaptic nanocolumns which span both the pre and post synaptic sites and fine tune the position of the post synaptic receptors. The functional impact of these nanocolumns (i.e. their influence on synaptic current) is both little understood and difficult to investigate experimentally. Here we construct a novel in silico model to investigate the functional impact of nanocolumns and show that they could play a functional role in reinforcing weak synapses.

Kv4.2-Positive Domains on Dendrites in the Mouse Medial Geniculate Body Receive Ascending Excitatory and Inhibitory Inputs Preferentially From the Inferior Colliculus

The medial geniculate body (MGB) is the thalamic center of the auditory lemniscal pathway. The ventral division of MGB (MGV) receives excitatory and inhibitory inputs from the inferior colliculus (IC). MGV is involved in auditory attention by processing descending excitatory and inhibitory inputs from the auditory cortex (AC) and reticular thalamic nucleus (RTN), respectively. However, detailed mechanisms of the integration of different inputs in a single MGV neuron remain unclear. Kv4.2 is one of the isoforms of the Shal-related subfamily of potassium voltage-gated channels that are expressed in MGB. Since potassium channel is important for shaping synaptic current and spike waveforms, subcellular distribution of Kv4.2 is likely important for integration of various inputs. Here, we aimed to examine the detailed distribution of Kv4.2, in MGV neurons to understand its specific role in auditory attention. We found that Kv4.2 mRNA was expressed in most MGV neurons. At the protein level, Kv4.2-immunopositive patches were sparsely distributed in both the dendrites and the soma of neurons. The postsynaptic distribution of Kv4.2 protein was confirmed using electron microscopy (EM). The frequency of contact with Kv4.2-immunopositive puncta was higher in vesicular glutamate transporter 2 (VGluT2)-positive excitatory axon terminals, which are supposed to be extending from the IC, than in VGluT1-immunopositive terminals, which are expected to be originating from the AC. VGluT2-immunopositive terminals were significantly larger than VGluT1-immunopositive terminals. Furthermore, EM showed that the terminals forming asymmetric synapses with Kv4.2-immunopositive MGV dendritic domains were significantly larger than those forming synapses with Kv4.2-negative MGV dendritic domains. In inhibitory axons either from the IC or from the RTN, the frequency of terminals that were in contact with Kv4.2-positive puncta was higher in IC than in RTN. In summary, our study demonstrated that the Kv4.2-immunopositive domains of the MGV dendrites received excitatory and inhibitory ascending auditory inputs preferentially from the IC, and not from the RTN or cortex. Our findings imply that time course of synaptic current and spike waveforms elicited by IC inputs is modified in the Kv4.2 domains.

A Novel Machine Learning-based Approach for the Detection and Analysis of Spontaneous Synaptic Currents

Spontaneous synaptic activity is a hallmark of neural networks. A thorough description of these synaptic signals is essential for understanding neurotransmitter release and the generation of a postsynaptic response. However, the complexity of synaptic current trajectories has either precluded an in-depth analysis or it has forced human observers to resort to manual or semi-automated approaches based on subjective amplitude and area threshold settings. Both procedures are time-consuming, error-prone and likely affected by human bias. Here, we present three complimentary methods for a fully automated analysis of spontaneous excitatory postsynaptic currents measured in major cell types of the mouse retina and in a primary culture of mouse auditory cortex. Two approaches rely on classical threshold methods, while the third represents a novel machine learning-based algorithm. Comparison with frequently used existing methods demonstrates the suitability of our algorithms for an unbiased and efficient analysis of synaptic signals in the central nervous system.

Intracellular NASPM allows an unambiguous functional measure of GluA2-lacking calcium-permeable AMPA receptor prevalence

Calcium-permeable AMPA-type glutamate receptors (CP-AMPARs) contribute to many forms of synaptic plasticity and pathology. They can be distinguished from GluA2-containing calcium-impermeable AMPARs by the inward rectification of their currents, which reflects voltage-dependent block by intracellular spermine. However, the efficacy of this weakly permeant blocker is differentially altered by the presence of AMPAR auxiliary subunits – including transmembrane AMPAR regulatory proteins, cornichons and GSG1L – that are widely expressed in neurons and glia. This complicates the interpretation of rectification as a measure of CP-AMPAR expression. Here we show that inclusion of the spider toxin analogue 1-naphthylacetyl spermine (NASPM) in the intracellular recording solution results in complete block of GluA1-mediated outward currents irrespective of the type of associated auxiliary subunit. In neurons from GluA2-knockout mice expressing only CP-AMPARs, intracellular NASPM, unlike spermine, blocks all outward synaptic current. Thus, our results identify an unambiguous functional measure, sensitive solely to changes in CP-AMPAR prevalence.

Energetics of stochastic BCM type synaptic plasticity and storing of accurate information

Excitatory synaptic signaling in cortical circuits is thought to be metabolically expensive. Two fundamental brain functions, learning and memory, are associated with long-term synaptic plasticity, but we know very little about energetics of these slow biophysical processes. This study investigates the energy requirement of information storing in plastic synapses for an extended version of BCM plasticity with a decay term, stochastic noise, and nonlinear dependence of neuron’s firing rate on synaptic current (adaptation). It is shown that synaptic weights in this model exhibit bistability. In order to analyze the system analytically, it is reduced to a simple dynamic mean-field for a population averaged plastic synaptic current. Next, using the concepts of nonequilibrium thermodynamics, we derive the energy rate (entropy production rate) for plastic synapses and a corresponding Fisher information for coding presynaptic input. That energy, which is of chemical origin, is primarily used for battling fluctuations in the synaptic weights and presynaptic firing rates, and it increases steeply with synaptic weights, and more uniformly though nonlinearly with presynaptic firing. At the onset of synaptic bistability, Fisher information and memory lifetime both increase sharply, by a few orders of magnitude, but the plasticity energy rate changes only mildly. This implies that a huge gain in the precision of stored information does not have to cost large amounts of metabolic energy, which suggests that synaptic information is not directly limited by energy consumption. Interestingly, for very weak synaptic noise, such a limit on synaptic coding accuracy is imposed instead by a derivative of the plasticity energy rate with respect to the mean presynaptic firing, and this relationship has a general character that is independent of the plasticity type. An estimate for primate neocortex reveals that a relative metabolic cost of BCM type synaptic plasticity, as a fraction of neuronal cost related to fast synaptic transmission and spiking, can vary from negligible to substantial, depending on the synaptic noise level and presynaptic firing.

AgRP neurons trigger long-term potentiation and facilitate food seeking

Sufficient feeding is essential for animals’ survival, which requires a cognitive capability to facilitate food seeking, but the neurobiological processes regulating food seeking are not fully understood. Here we show that stimulation of agouti-related peptide-expressing (AgRP) neurons triggers a long-term depression (LTD) of spontaneous excitatory post-synaptic current (sEPSC) in adjacent pro-opiomelanocortin (POMC) neurons and in most of their distant synaptic targets, including neurons in the paraventricular nucleus of the thalamus (PVT). The AgRP-induced sEPCS LTD can be enhanced by fasting but blunted by satiety signals, e.g. leptin and insulin. Mice subjected to food-seeking tasks develop similar neural plasticity in AgRP-innervated PVT neurons. Further, ablation of the majority of AgRP neurons, or only a subset of AgRP neurons that project to the PVT, impairs animals’ ability to associate spatial and contextual cues with food availability during food seeking. A similar impairment can be also induced by optogenetic inhibition of the AgRP→PVT projections. Together, these results indicate that the AgRP→PVT circuit is necessary for food seeking.

Slow AMPA receptors in hippocampal principal cells

SummaryGlutamate receptor ion channels such as the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor mediate the majority of fast excitatory neurotransmission in the vertebrate CNS. AMPA receptors canonically provide the fast, millisecond component of the synaptic current. However, we found that about two-thirds of principal cells in the mouse hippocampus express AMPA receptors that do not desensitize and stay active for up to half a second. These receptors are expressed at synapses with a sparse but flat spatial distribution. The resulting increase in charge transfer allows single connections to reliably trigger action potentials. Biophysical and pharmacological observations imply that slow AMPA receptors incorporate γ-8 and other auxiliary proteins, and their activation lengthens individual miniature synaptic currents. Synaptic connections harboring slow AMPARs should have unique roles in hippocampal function.

Selectivity to approaching motion in retinal inputs to the dorsal visual pathway

To efficiently navigate through the environment and avoid potential threats, an animal must quickly detect the motion of approaching objects. Current models of primate vision place the origins of this complex computation in the visual cortex. Here, we report that detection of approaching motion begins in the retina. Several ganglion cell types, the retinal output neurons, show selectivity to approaching motion. Synaptic current recordings from these cells further reveal that this preference for approaching motion arises in the interplay between presynaptic excitatory and inhibitory circuit elements. These findings demonstrate how excitatory and inhibitory circuits interact to mediate an ethologically relevant neural function. Moreover, the elementary computations that detect approaching motion begin early in the visual stream of primates.

Quantitative and qualitative analysis of asynchronous neural activity

The activity of a sparse network of leaky integrate-and-fire neurons is carefully revisited with reference to a regime of a bona-fide asynchronous dynamics. The study is preceded by a finite-size scaling analysis, carried out to identify a setup where collective synchronization is negligible. The comparison between quenched and annealed networks reveals the emergence of substantial differences when the coupling strength is increased, via a scenario somehow reminiscent of a phase transition. For sufficiently strong synaptic coupling, quenched networks exhibit a highly bursting neural activity, well reproduced by a self-consistent approach, based on the assumption that the input synaptic current is the superposition of independent renewal processes. The distribution of interspike intervals turns out to be relatively long-tailed; a crucial feature required for the self-sustainment of the bursting activity in a regime where neurons operate on average (much) below threshold. A semi-quantitative analogy with Ornstein-Uhlenbeck processes helps validating this interpretation. Finally, an alternative explanation in terms of Poisson processes is offered under the additional assumption of mutual correlations among excitatory and inhibitory spikes.