Dual separable feedback systems govern firing rate homeostasis

Motor neurons within a network use cell-type specific feedback mechanisms to constrain relationships among ion channel mRNAs

Recently, activity has been proposed as a primary feedback mechanism used by continuously bursting neurons to coordinate ion channel mRNA relationships that underlie stable output. However, some neuron types only have intermittent periods of activity and so may require alternative mechanisms that induce and constrain the appropriate ion channel profile in different states of activity. To address this, we used the pyloric dilator (PD; constitutively active) and the lateral gastric (LG; periodically active) neurons of the stomatogastric ganglion (STG) of the crustacean Cancer borealis. We experimentally stimulated descending inputs to the STG to cause release of neuromodulators known to elicit the active state of LG neurons and quantified the mRNA abundances and pairwise relationships of 11 voltage-gated ion channels in active and silent LG neurons. The same stimulus does not significantly alter PD activity. Activation of LG upregulated ion channel mRNAs and lead to a greater number of positively correlated pairwise channel mRNA relationships. Conversely, this stimulus did not induce major changes in ion channel mRNA abundances and relationships of PD cells, suggesting their ongoing activity is sufficient to maintain channel mRNA relationships even under changing modulatory conditions. In addition, we found that ion channel mRNA correlations induced by the active state of LG are influenced by a combination of activity- and neuromodulator-dependent feedback mechanisms. Interestingly, some of these same correlations are maintained by distinct mechanisms in PD, suggesting that these motor networks use distinct feedback mechanisms to coordinate the same mRNA relationships across neuron types.NEW & NOTEWORTHY Neurons use various feedback mechanisms to adjust and maintain their output. Here, we demonstrate that different neurons within the same network can use distinct signaling mechanisms to regulate the same ion channel mRNA relationships.

Neuronal morphology enhances robustness to perturbations of channel densities

Biological neurons show significant cell-to-cell variability but have the striking ability to maintain their key firing properties in the face of unpredictable perturbations and stochastic noise. Using a population of multi-compartment models consisting of soma, neurites, and axon for the lateral pyloric neuron in the crab stomatogastric ganglion, we explore how rebound bursting is preserved when the 14 channel conductances in each model are all randomly varied. The coupling between the axon and other compartments is critical for the ability of the axon to spike during bursts and consequently determines the set of successful solutions. When the coupling deviates from a biologically realistic range, the neuronal tolerance of conductance variations is lessened. Thus, the gross morphological features of these neurons enhance their robustness to perturbations of channel densities and expand the space of individual variability that can maintain a desired output pattern.

Neuron populations use variable combinations of short-term feedback mechanisms to stabilize firing rate

Neurons tightly regulate firing rate and a failure to do so leads to multiple neurological disorders. Therefore, a fundamental question in neuroscience is how neurons produce reliable activity patterns for decades to generate behavior. Neurons have built-in feedback mechanisms that allow them to monitor their output and rapidly stabilize firing rate. Most work emphasizes the role of a dominant feedback system within a neuronal population for the control of moment-to-moment firing. In contrast, we find that respiratory motoneurons use 2 activity-dependent controllers in unique combinations across cells, dynamic activation of an Na+ pump subtype, and rapid potentiation of Kv7 channels. Both systems constrain firing rate by reducing excitability for up to a minute after a burst of action potentials but are recruited by different cellular signals associated with activity, increased intracellular Na+ (the Na+ pump), and membrane depolarization (Kv7 channels). Individual neurons do not simply contain equal amounts of each system. Rather, neurons under strong control of the Na+ pump are weakly regulated by Kv7 enhancement and vice versa along a continuum. Thus, each motoneuron maintains its characteristic firing rate through a unique combination of the Na+ pump and Kv7 channels, which are dynamically regulated by distinct feedback signals. These results reveal a new organizing strategy for stable circuit output involving multiple fast activity sensors scaled inversely across a neuronal population.

Molecular encoding of stimulus features in a single sensory neuron type enables neuronal and behavioral plasticity

Neurons modify their transcriptomes in response to an animal’s experience. How specific experiences are transduced to modulate gene expression and precisely tune neuronal functions are not fully defined. Here, we describe the molecular profile of a thermosensory neuron pair in C. elegans experiencing different temperature stimuli. We find that distinct salient features of the temperature stimulus including its duration, magnitude of change, and absolute value are encoded in the gene expression program in this single neuron, and identify a novel transmembrane protein and a transcription factor whose specific transcriptional dynamics are essential to drive neuronal, behavioral, and developmental plasticity. Expression changes are driven by broadly expressed activity-dependent transcription factors and corresponding cis -regulatory elements that nevertheless direct neuron- and stimulus-specific gene expression programs. Our results indicate that coupling of defined stimulus characteristics to the gene regulatory logic in individual specialized neuron types can customize neuronal properties to drive precise behavioral adaptation.

Towards the Neuroevolution of Low-level artificial general intelligence

In this work, we argue that the search for Artificial General Intelligence should start from a much lower level than human-level intelligence. The circumstances of intelligent behavior in nature resulted from an organism interacting with its surrounding environment, which could change over time and exert pressure on the organism to allow for learning of new behaviors or environment models. Our hypothesis is that learning occurs through interpreting sensory feedback when an agent acts in an environment. For that to happen, a body and a reactive environment are needed. We evaluate a method to evolve a biologically-inspired artificial neural network that learns from environment reactions named Neuroevolution of Artificial General Intelligence, a framework for low-level artificial general intelligence. This method allows the evolutionary complexification of a randomly-initialized spiking neural network with adaptive synapses, which controls agents instantiated in mutable environments. Such a configuration allows us to benchmark the adaptivity and generality of the controllers. The chosen tasks in the mutable environments are food foraging, emulation of logic gates, and cart-pole balancing. The three tasks are successfully solved with rather small network topologies and therefore it opens up the possibility of experimenting with more complex tasks and scenarios where curriculum learning is beneficial.

Minimal requirements for a neuron to co-regulate many properties and the implications for ion channel correlations and robustness

Neurons regulate their excitability by adjusting their ion channel levels. Degeneracy – achieving equivalent outcomes (excitability) using different solutions (channel combinations) – facilitates this regulation by enabling a disruptive change in one channel to be offset by compensatory changes in other channels. But neurons must coregulate many properties. Pleiotropy – the impact of one channel on more than one property – complicates regulation because a compensatory ion channel change that restores one property to its target value often disrupts other properties. How then does a neuron simultaneously regulate multiple properties? Here, we demonstrate that of the many channel combinations producing the target value for one property (the single-output solution set), few combinations produce the target value for other properties. Combinations producing the target value for two or more properties (the multioutput solution set) correspond to the intersection between single-output solution sets. Properties can be effectively coregulated only if the number of adjustable channels (n_in) exceeds the number of regulated properties (n_out). Ion channel correlations emerge during homeostatic regulation when the dimensionality of solution space (n_in − n_out) is low. Even if each property can be regulated to its target value when considered in isolation, regulation as a whole fails if single-output solution sets do not intersect. Our results also highlight that ion channels must be coadjusted with different ratios to regulate different properties, which suggests that each error signal drives modulatory changes independently, despite those changes ultimately affecting the same ion channels.

Age-related changes in Kv4/Shal and Kv1/Shaker expression in Drosophila and a role for reactive oxygen species

Age-related changes in ion channel expression are likely to affect neuronal signaling. Here, we examine how age affects Kv4/Shal and Kv1/Shaker K+ channel protein levels in Drosophila. We show that Kv4/Shal protein levels decline sharply from 3 days to 10 days, then more gradually from 10 to 40 days after eclosion. In contrast, Kv1/Shaker protein exhibits a transient increase at 10 days that then stabilizes and eventually declines at 40 days. We present data that begin to show a relationship between reactive oxygen species (ROS), Kv4/Shal, and locomotor performance. We show that Kv4/Shal levels are negatively affected by ROS, and that over-expression of Catalase or RNAi knock-down of the ROS-generating enzyme, Nicotinamide Adenine Dinucleotide Phosphate (NADPH) Oxidase (NOX), can attenuate the loss of Kv4/Shal protein. Finally, we compare levels of Kv4.2 and Kv4.3 in the hippocampus, olfactory bulb, cerebellum, and motor cortex of mice aged 6 weeks and 1 year. While there was no global decline in Kv4.2/4.3 that parallels what we report in Drosophila, we did find that Kv4.2/4.3 are differentially affected in various brain regions; this survey of changes may help inform mammalian studies that examine neuronal function with age.

Distinct molecular pathways govern presynaptic homeostatic plasticity

Presynaptic homeostatic plasticity (PHP) stabilizes synaptic transmission by counteracting impaired neurotransmitter receptor function through neurotransmitter release potentiation. PHP is thought to be triggered by impaired receptor function and to involve a stereotypic signaling pathway. However, here we demonstrate that different receptor perturbations that similarly reduce synaptic transmission result in different responses at the Drosophila neuromuscular junction. While receptor inhibition by the glutamate receptor (GluR) antagonist γ-D-glutamylglycine (γDGG) is not compensated by PHP, the GluR inhibitors Philanthotoxin-433 (PhTx) and Gyki-53655 (Gyki) induce compensatory PHP. Intriguingly, PHP triggered by PhTx and Gyki involve separable signaling pathways, including inhibition of distinct GluR subtypes, differential modulation of the active-zone scaffold Bruchpilot, and short-term plasticity. Moreover, while PHP upon Gyki treatment does not require genes promoting PhTx-induced PHP, it involves presynaptic protein kinase D. Thus, synapses not only respond differentially to similar activity impairments, but achieve homeostatic compensation via distinct mechanisms, highlighting the diversity of homeostatic signaling.

Biophysical models of intrinsic homeostasis: Firing rates and beyond

In view of ever-changing conditions both in the external world and in intrinsic brain states, maintaining the robustness of computations poses a challenge, adequate solutions to which we are only beginning to understand. At the level of cell-intrinsic properties, biophysical models of neurons permit one to identify relevant physiological substrates that can serve as regulators of neuronal excitability and to test how feedback loops can stabilize crucial variables such as long-term calcium levels and firing rates. Mathematical theory has also revealed a rich set of complementary computational properties arising from distinct cellular dynamics and even shaping processing at the network level. Here, we provide an overview over recently explored homeostatic mechanisms derived from biophysical models and hypothesize how multiple dynamical characteristics of cells, including their intrinsic neuronal excitability classes, can be stably controlled.

Activity-dependent compensation of cell size is vulnerable to targeted deletion of ion channels

In many species, excitable cells preserve their physiological properties despite significant variation in physical size across time and in a population. For example, neurons in crustacean central pattern generators generate similar firing patterns despite several-fold increases in size between juveniles and adults. This presents a biophysical problem because the electrical properties of cells are highly sensitive to membrane area and channel density. It is not known whether specific mechanisms exist to sense membrane area and adjust channel expression to keep a consistent channel density, or whether regulation mechanisms that sense activity alone are capable of compensating cell size. We show that destabilising effects of growth can be specifically compensated by feedback mechanism that senses average calcium influx and jointly regulate multiple conductances. However, we further show that this class of growth-compensating regulation schemes is necessarily sensitive to perturbations that alter the expression of subsets of ion channel types. Targeted perturbations of specific ion channels can trigger a pathological response of the regulation mechanism and a failure of homeostasis. Our findings suggest that physiological regulation mechanisms that confer robustness to growth may be specifically vulnerable to deletions or mutations that affect subsets of ion channels.

Homeostatic plasticity fails at the intersection of autism-gene mutations and a novel class of common genetic modifiers

We identify a set of common phenotypic modifiers that interact with five independent autism gene orthologs (RIMS1, CHD8, CHD2, WDFY3, ASH1L) causing a common failure of presynaptic homeostatic plasticity (PHP) in Drosophila. Heterozygous null mutations in each autism gene are demonstrated to have normal baseline neurotransmission and PHP. However, PHP is sensitized and rendered prone to failure. A subsequent electrophysiology-based genetic screen identifies the first known heterozygous mutations that commonly genetically interact with multiple ASD gene orthologs, causing PHP to fail. Two phenotypic modifiers identified in the screen, PDPK1 and PPP2R5D, are characterized. Finally, transcriptomic, ultrastructural and electrophysiological analyses define one mechanism by which PHP fails; an unexpected, maladaptive up-regulation of CREG, a conserved, neuronally expressed, stress response gene and a novel repressor of PHP. Thus, we define a novel genetic landscape by which diverse, unrelated autism risk genes may converge to commonly affect the robustness of synaptic transmission.

Neuronal Homeostasis: Voltage Brings It All Together

Neurons generate consistent patterns of activity combining a large set of ionic currents and variable conductance levels. How they regulate this variability so that it does not run out of control seems to depend most prominently on neuronal activity itself.

Molecular mechanisms of homeostatic plasticity in central pattern generator networks

Central pattern generator (CPG) networks rely on a balance of intrinsic and network properties to produce reliable, repeatable activity patterns. This balance is maintained by homeostatic plasticity where alterations in neuronal properties dynamically maintain appropriate neural output in the face of changing environmental conditions and perturbations. However, it remains unclear just how these neurons and networks can both monitor their ongoing activity and use this information to elicit homeostatic physiological responses to ensure robustness of output over time. Evidence exists that CPG networks use a mixed strategy of activity-dependent, activity-independent, modulator-dependent, and synaptically regulated homeostatic plasticity to achieve this critical stability. In this review, we focus on some of the current understanding of the molecular pathways and mechanisms responsible for this homeostatic plasticity in the context of central pattern generator function, with a special emphasis on some of the smaller invertebrate networks that have allowed for extensive cellular-level analyses that have brought recent insights to these questions.

Firing Rate Homeostasis Can Occur in the Absence of Neuronal Activity-Regulated Transcription

Despite dynamic inputs, neuronal circuits maintain relatively stable firing rates over long periods. This maintenance of firing rate, or firing rate homeostasis, is likely mediated by homeostatic mechanisms such as synaptic scaling and regulation of intrinsic excitability. Because some of these homeostatic mechanisms depend on transcription of activity-regulated genes, including Arc and Homer1a, we hypothesized that activity-regulated transcription would be required for firing rate homeostasis. Surprisingly, however, we found that cultured mouse cortical neurons from both sexes grown on multi-electrode arrays homeostatically adapt their firing rates to persistent pharmacological stimulation even when activity-regulated transcription is disrupted. Specifically, we observed firing rate homeostasis in Arc knock-out neurons, as well as knock-out neurons lacking the activity-regulated transcription factors AP1 and SRF. Firing rate homeostasis also occurred normally during acute pharmacological blockade of transcription. Thus, firing rate homeostasis in response to increased neuronal activity can occur in the absence of neuronal-activity-regulated transcription.SIGNIFICANCE STATEMENT Neuronal circuits maintain relatively stable firing rates even in the face of dynamic circuit inputs. Understanding the molecular mechanisms that enable this firing rate homeostasis could potentially provide insight into neuronal diseases that present with an imbalance of excitation and inhibition. It has long been proposed that activity-regulated transcription could underlie firing rate homeostasis because activity-regulated genes turn on when neurons are above their target firing rates and include many genes that could regulate firing rate. Surprisingly, despite this prediction, we found that cortical neurons can undergo firing rate homeostasis in the absence of activity-regulated transcription, indicating that firing rate homeostasis can be controlled by non-transcriptional mechanisms.

Shaw and Shal voltage-gated potassium channels mediate circadian changes in Drosophila clock neuron excitability

As in mammals, Drosophila circadian clock neurons display rhythms of activity with higher action potential firing rates and more positive resting membrane potentials during the day. This rhythmic excitability has been widely observed but, critically, its regulation remains unresolved. We have characterized and modelled the changes underlying these electrical activity rhythms in the lateral ventral clock neurons (LNvs). We show that currents mediated by the voltage-gated potassium channels Shaw (Kv3) and Shal (Kv4) oscillate in a circadian manner. Disruption of these channels, by expression of dominant negative (DN) subunits, leads to changes in circadian locomotor activity and shortens lifespan. LNv whole-cell recordings then show that changes in Shaw and Shal currents drive changes in action potential firing rate and that these rhythms are abolished when the circadian molecular clock is stopped. A whole-cell biophysical model using Hodgkin-Huxley equations can recapitulate these changes in electrical activity. Based on this model and by using dynamic clamp to manipulate clock neurons directly, we can rescue the pharmacological block of Shaw and Shal, restore the firing rhythm, and thus demonstrate the critical importance of Shaw and Shal. Together, these findings point to a key role for Shaw and Shal in controlling circadian firing of clock neurons and show that changes in clock neuron currents can account for this. Moreover, with dynamic clamp we can switch the LNvs between morning-like and evening-like states of electrical activity. We conclude that changes in Shaw and Shal underlie the daily oscillation in LNv firing rate.

Allostasis: A Brain-Centered, Predictive Mode of Physiological Regulation

Although the concept of allostasis was proposed some 30 years ago, doubts persist about its precise meaning and whether it is useful. Here we review the concept in the context of recent studies as a strategy to efficiently regulate physiology and behavior. The brain, sensing the internal and external milieu, and consulting its database, predicts what is likely to be needed; then, it computes the best response. The brain rewards a better-than-predicted result with a pulse of dopamine, thereby encouraging the organism to learn effective regulatory behaviors. The brain, by prioritizing behaviors and dynamically adjusting the flows of energy and nutrients, reduces costly errors and exploits more opportunities. Despite significant costs of computation, allostasis pays off and can now be recognized as a core principle of organismal design.