A burst-based "Hebbian" learning rule at retinogeniculate synapses links retinal waves to activity-dependent refinement

Patterned spontaneous activity in the developing retina is necessary to drive synaptic refinement in the lateral geniculate nucleus (LGN). Using perforated patch recordings from neurons in LGN slices during the period of eye segregation, we examine how such burst-based activity can instruct this refinement. Retinogeniculate synapses have a novel learning rule that depends on the latencies between pre- and postsynaptic bursts on the order of one second: coincident bursts produce long-lasting synaptic enhancement, whereas non-overlapping bursts produce mild synaptic weakening. It is consistent with “Hebbian” development thought to exist at this synapse, and we demonstrate computationally that such a rule can robustly use retinal waves to drive eye segregation and retinotopic refinement. Thus, by measuring plasticity induced by natural activity patterns, synaptic learning rules can be linked directly to their larger role in instructing the patterning of neural connectivity.     

Synaptic Potentiation Facilitates Memory-like Attractor Dynamics in Cultured In Vitro Hippocampal Networks

Collective rhythmic dynamics from neurons is vital for cognitive functions such as memory formation but how neurons self-organize to produce such activity is not well understood. Attractor-based computational models have been successfully implemented as a theoretical framework for memory storage in networks of neurons. Additionally, activity-dependent modification of synaptic transmission is thought to be the physiological basis of learning and memory. The goal of this study is to demonstrate that using a pharmacological treatment that has been shown to increase synaptic strength within in vitro networks of hippocampal neurons follows the dynamical postulates theorized by attractor models. We use a grid of extracellular electrodes to study changes in network activity after this perturbation and show that there is a persistent increase in overall spiking and bursting activity after treatment. This increase in activity appears to recruit more “errant” spikes into bursts. Phase plots indicate a conserved activity pattern suggesting that a synaptic potentiation perturbation to the attractor leaves it unchanged. Lastly, we construct a computational model to demonstrate that these synaptic perturbations can account for the dynamical changes seen within the network.     

Tinnitus-like “hallucinations” elicited by sensory deprivation in an entropy maximization recurrent neural network

Sensory deprivation has long been known to cause hallucinations or “phantom” sensations, the most common of which is tinnitus induced by hearing loss, affecting 10–20% of the population. An observable hearing loss, causing auditory sensory deprivation over a band of frequencies, is present in over 90% of people with tinnitus. Existing plasticity-based computational models for tinnitus are usually driven by homeostatic mechanisms, modeled to fit phenomenological findings. Here, we use an objective-driven learning algorithm to model an early auditory processing neuronal network, e.g., in the dorsal cochlear nucleus. The learning algorithm maximizes the network’s output entropy by learning the feed-forward and recurrent interactions in the model. We show that the connectivity patterns and responses learned by the model display several hallmarks of early auditory neuronal networks. We further demonstrate that attenuation of peripheral inputs drives the recurrent network towards its critical point and transition into a tinnitus-like state. In this state, the network activity resembles responses to genuine inputs even in the absence of external stimulation, namely, it “hallucinates” auditory responses. These findings demonstrate how objective-driven plasticity mechanisms that normally act to optimize the network’s input representation can also elicit pathologies such as tinnitus as a result of sensory deprivation.     

The gradient clusteron: A model neuron that learns to solve classification tasks via dendritic nonlinearities,structural plasticity,and gradient descent

Synaptic clustering on neuronal dendrites has been hypothesized to play an important role in implementing pattern recognition. Neighboring synapses on a dendritic branch can interact in a synergistic, cooperative manner via nonlinear voltage-dependent mechanisms, such as NMDA receptors. Inspired by the NMDA receptor, the single-branch clusteron learning algorithm takes advantage of location-dependent multiplicative nonlinearities to solve classification tasks by randomly shuffling the locations of “under-performing” synapses on a model dendrite during learning (“structural plasticity”), eventually resulting in synapses with correlated activity being placed next to each other on the dendrite. We propose an alternative model, the gradient clusteron, or G-clusteron, which uses an analytically-derived gradient descent rule where synapses are "attracted to" or "repelled from" each other in an input- and location-dependent manner. We demonstrate the classification ability of this algorithm by testing it on the MNIST handwritten digit dataset and show that, when using a softmax activation function, the accuracy of the G-clusteron on the all-versus-all MNIST task (~85%) approaches that of logistic regression (~93%). In addition to the location update rule, we also derive a learning rule for the synaptic weights of the G-clusteron (“functional plasticity”) and show that a G-clusteron that utilizes the weight update rule can achieve ~89% accuracy on the MNIST task. We also show that a G-clusteron with both the weight and location update rules can learn to solve the XOR problem from arbitrary initial conditions.     

An instructive role for retinal waves in the development of retinogeniculate connectivity

A central hypothesis of neural development is that patterned activity drives the refinement of initially imprecise connections. We have examined this hypothesis directly by altering the frequency of spontaneous waves of activity that sweep across the mammalian retina prior to vision. Activity levels were increased in vivo using agents that elevate cAMP. When one eye is made more active, its layer within the LGN is larger despite the other eye having normal levels of activity. Remarkably, when the frequency of retinal waves is increased in both eyes, normally sized layers form. Because relative, rather than absolute, levels of activity between the eyes regulate the amount of LGN territory devoted to each eye, we conclude that activity acts instructively to guide binocular segregation during development.     

“Liking” as an early and editable draft of long-run affective value

Psychological and neural distinctions between the technical concepts of “liking” and “wanting” pose important problems for motivated choice for goods. Why could we “want” something that we do not “like,” or “like” something but be unwilling to exert effort to acquire it? Here, we suggest a framework for answering these questions through the medium of reinforcement learning. We consider “liking” to provide immediate, but preliminary and ultimately cancellable, information about the true, long-run worth of a good. Such initial estimates, viewed through the lens of what is known as potential-based shaping, help solve the temporally complex learning problems faced by animals.

What is the distinction between ’liking’ and ’wanting’? Why could we ’want’ something that we do not ’like,’ or ’like’ something but be unwilling to exert effort to acquire it? This Essay argues that the primary hedonic phenomenon called ’liking’ might solve the temporal credit assignment problem for learning that arises when true reinforcement values are available slowly or late.     

Nob mice wave goodbye to eye-specific segregation

Spontaneous retinal activity is necessary to establish and maintain eye-specific projections to the LGN, but whether the spatial and temporal structure of this activity is important remains unclear. A new study by Demas et al. in the current issue of Neuron shows that when the frequency of spontaneous retinal waves is increased and waves abnormally persist past eye opening, eye-specific projections to the LGN desegregate. These results provide important new insight into the mechanisms that drive eye-specific refinement and stabilization.     

The role of retinal waves and synaptic normalization in retinogeniculate development

The prenatal development of the cat retinogeniculate pathway is thought to involve activity-dependent mechanisms driven by spontaneous waves of retinal activity. The role of these waves upon the segregation of the dorsal lateral geniculate nucleus (LGN) into two eye-specific layers and the development of retinotopic mappings have previously been investigated in a computer model. Using this model, we examine three aspects of retinogeniculate development. First, the mapping of visual space across the whole network into projection columns is shown to be similar to the mapping found in the cat. Second, the simplicity of the model allows us to explore how different forms of synaptic normalization affect development. In comparison to most previous models of ocular dominance, we find that subtractive postsynaptic normalization is redundant and divisive presynaptic normalization is sufficient for normal development. Third, the model predicts that the more often one eye generates waves relative to the other eye, the more LGN units will monocularly respond to the more active eye. In the limit when one eye does not generate any waves, that eye totally disconnects from the LGN allowing the non-deprived eye to innervate all of the LGN. Thus, as well as accounting for normal retinogeniculate development, the model also predicts development under abnormal conditions which can be experimentally tested.     

The Convallis Rule for Unsupervised Learning in Cortical Networks

The phenomenology and cellular mechanisms of cortical synaptic plasticity are becoming known in increasing detail, but the computational principles by which cortical plasticity enables the development of sensory representations are unclear. Here we describe a framework for cortical synaptic plasticity termed the “Convallis rule”, mathematically derived from a principle of unsupervised learning via constrained optimization. Implementation of the rule caused a recurrent cortex-like network of simulated spiking neurons to develop rate representations of real-world speech stimuli, enabling classification by a downstream linear decoder. Applied to spike patterns used in in vitro plasticity experiments, the rule reproduced multiple results including and beyond STDP. However STDP alone produced poorer learning performance. The mathematical form of the rule is consistent with a dual coincidence detector mechanism that has been suggested by experiments in several synaptic classes of juvenile neocortex. Based on this confluence of normative, phenomenological, and mechanistic evidence, we suggest that the rule may approximate a fundamental computational principle of the neocortex.     

Visual Receptive Field Properties of Neurons in the Mouse Lateral Geniculate Nucleus

The lateral geniculate nucleus (LGN) is increasingly regarded as a “smart-gating” operator for processing visual information. Therefore, characterizing the response properties of LGN neurons will enable us to better understand how neurons encode and transfer visual signals. Efforts have been devoted to study its anatomical and functional features, and recent advances have highlighted the existence in rodents of complex features such as direction/orientation selectivity. However, unlike well-researched higher-order mammals such as primates, the full array of response characteristics vis-à-vis its morphological features have remained relatively unexplored in the mouse LGN. To address the issue, we recorded from mouse LGN neurons using multisite-electrode-arrays (MEAs) and analysed their discharge patterns in relation to their location under a series of visual stimulation paradigms. Several response properties paralleled results from earlier studies in the field and these include centre-surround organization, size of receptive field, spontaneous firing rate and linearity of spatial summation. However, our results also revealed “high-pass” and “low-pass” features in the temporal frequency tuning of some cells, and greater average contrast gain than reported by earlier studies. In addition, a small proportion of cells had direction/orientation selectivity. Both “high-pass” and “low-pass” cells, as well as direction and orientation selective cells, were found only in small numbers, supporting the notion that these properties emerge in the cortex. ON- and OFF-cells showed distinct contrast sensitivity and temporal frequency tuning properties, suggesting parallel projections from the retina. Incorporating a novel histological technique, we created a 3-D LGN volume model explicitly capturing the morphological features of mouse LGN and localising individual cells into anterior/middle/posterior LGN. Based on this categorization, we show that the ON/OFF, DS/OS and linear response properties are not regionally restricted. Our study confirms earlier findings of spatial pattern selectivity in the LGN, and builds on it to demonstrate that relatively elaborate features are computed early in the visual pathway.     

Short Peptide Induces an “Uncultivable” Microorganism To Grow In Vitro

Microorganisms comprise the bulk of biodiversity, but only a small fraction of this diversity grows on artificial media. This phenomenon was noticed almost a century ago, repeatedly confirmed, and termed the “great plate count anomaly.” Advances in microbial cultivation improved microbial recovery but failed to explain why most microbial species do not grow in vitro. Here we show that at least some of such species can form domesticated variants capable of growth on artificial media. We also present evidence that small signaling molecules, such as short peptides, may be essential factors in initiating growth of nongrowing cells. We identified one 5-amino-acid peptide, LQPEV, that at 3.5 nM induces the otherwise “uncultivable” strain Psychrobacter sp. strain MSC33 to grow on standard media. This demonstrates that the restriction preventing microbial in vitro growth may be different from those offered to date to explain the “great plate count anomaly,” such as deficiencies in nutrient composition and concentrations in standard media, medium toxicity, and inappropriate incubation time. Growth induction of MSC33 illustrates that some microorganisms do not grow in vitro because they are removed from their native communities and the signals produced therein. “Uncultivable” species represent the largest source of unexplored biodiversity, and provide remarkable opportunities for both basic and applied research. Access to cultures of some of these species should be possible through identification of the signaling compounds necessary for growth, their addition to standard medium formulations, and eventual domestication.     

Slowness: An Objective for Spike-Timing–Dependent Plasticity?

Our nervous system can efficiently recognize objects in spite of changes in contextual variables such as perspective or lighting conditions. Several lines of research have proposed that this ability for invariant recognition is learned by exploiting the fact that object identities typically vary more slowly in time than contextual variables or noise. Here, we study the question of how this “temporal stability” or “slowness” approach can be implemented within the limits of biologically realistic spike-based learning rules. We first show that slow feature analysis, an algorithm that is based on slowness, can be implemented in linear continuous model neurons by means of a modified Hebbian learning rule. This approach provides a link to the trace rule, which is another implementation of slowness learning. Then, we show analytically that for linear Poisson neurons, slowness learning can be implemented by spike-timing–dependent plasticity (STDP) with a specific learning window. By studying the learning dynamics of STDP, we show that for functional interpretations of STDP, it is not the learning window alone that is relevant but rather the convolution of the learning window with the postsynaptic potential. We then derive STDP learning windows that implement slow feature analysis and the “trace rule.” The resulting learning windows are compatible with physiological data both in shape and timescale. Moreover, our analysis shows that the learning window can be split into two functionally different components that are sensitive to reversible and irreversible aspects of the input statistics, respectively. The theory indicates that irreversible input statistics are not in favor of stable weight distributions but may generate oscillatory weight dynamics. Our analysis offers a novel interpretation for the functional role of STDP in physiological neurons.     

Mutagenesis of the “Leucine Gate” To Explore the Basis of Catalytic Versatility in Soluble Methane Monooxygenase

Soluble methane monooxygenase (sMMO) from methane-oxidizing bacteria is a multicomponent nonheme oxygenase that naturally oxidizes methane to methanol and can also cooxidize a wide range of adventitious substrates, including mono- and diaromatic hydrocarbons. Leucine 110, at the mouth of the active site in the α subunit of the hydroxylase component of sMMO, has been suggested to act as a gate to control the access of substrates to the active site. Previous crystallography of the wild-type sMMO has indicated at least two conformations of the enzyme that have the “leucine gate” open to different extents, and mutagenesis of homologous enzymes has indicated a role for this residue in the control of substrate range and regioselectivity with aromatic substrates. By further refinement of the system for homologous expression of sMMO that we developed previously, we have been able to prepare a range of site-directed mutations at position 110 in the α subunit of sMMO. All the mutants (with Gly, Cys, Arg, and Tyr, respectively, at this position) showed relaxations of regioselectivity compared to the wild type with monoaromatic substrates and biphenyl, including the appearance of new products arising from hydroxylation at the 2- and 3- positions on the benzene ring. Mutants with the larger Arg and Trp residues at position 110 also showed shifts in regioselectivity during naphthalene hydroxylation from the 2- to the 1- position. No evidence that mutagenesis of Leu 110 could allow very large substrates to enter the active site was found, however, since the mutants (like the wild type) were inactive toward the triaromatic hydrocarbons anthracene and phenanthrene. Thus, our results indicate that the “leucine gate” in sMMO is more important in controlling the precision of regioselectivity than the sizes of substrates that can enter the active site.     

Par1b Induces Asymmetric Inheritance of Plasma Membrane Domains via LGN-Dependent Mitotic Spindle Orientation in Proliferating Hepatocytes

The development and maintenance of polarized epithelial tissue requires a tightly controlled orientation of mitotic cell division relative to the apical polarity axis. Hepatocytes display a unique polarized architecture. We demonstrate that mitotic hepatocytes asymmetrically segregate their apical plasma membrane domain to the nascent daughter cells. The non-polarized nascent daughter cell can form a de novo apical domain with its new neighbor. This asymmetric segregation of apical domains is facilitated by a geometrically distinct “apicolateral” subdomain of the lateral surface present in hepatocytes. The polarity protein partitioning-defective 1/microtubule-affinity regulating kinase 2 (Par1b/MARK2) translates this positional landmark to cortical polarity by promoting the apicolateral accumulation of Leu-Gly-Asn repeat-enriched protein (LGN) and the capture of nuclear mitotic apparatus protein (NuMA)–positive astral microtubules to orientate the mitotic spindle. Proliferating hepatocytes thus display an asymmetric inheritance of their apical domains via a mechanism that involves Par1b and LGN, which we postulate serves the unique tissue architecture of the developing liver parenchyma.     

Differential Sensitivity of “Old” versus “New” APOBEC3G to Human Immunodeficiency Virus Type 1 Vif

HIV-1 Vif counteracts the antiviral activity of APOBEC3G by inhibiting its encapsidation into virions. Here, we compared the relative sensitivity to Vif of APOBEC3G in stable HeLa cells containing APOBEC3G (HeLa-A3G cells) versus that of newly synthesized APOBEC3G. We observed that newly synthesized APOBEC3G was more sensitive to degradation than preexisting APOBEC3G. Nevertheless, preexisting and transiently expressed APOBEC3G were packaged with similar efficiencies into vif-deficient human immunodeficiency virus type 1 (HIV-1) virions, and Vif inhibited the encapsidation of both forms of APOBEC3G into HIV particles equally well. Our results suggest that HIV-1 Vif preferentially induces degradation of newly synthesized APOBEC3G but indiscriminately inhibits encapsidation of “old” and “new” APOBEC3G.     

Time-Warp–Invariant Neuronal Processing

Fluctuations in the temporal durations of sensory signals constitute a major source of variability within natural stimulus ensembles. The neuronal mechanisms through which sensory systems can stabilize perception against such fluctuations are largely unknown. An intriguing instantiation of such robustness occurs in human speech perception, which relies critically on temporal acoustic cues that are embedded in signals with highly variable duration. Across different instances of natural speech, auditory cues can undergo temporal warping that ranges from 2-fold compression to 2-fold dilation without significant perceptual impairment. Here, we report that time-warp–invariant neuronal processing can be subserved by the shunting action of synaptic conductances that automatically rescales the effective integration time of postsynaptic neurons. We propose a novel spike-based learning rule for synaptic conductances that adjusts the degree of synaptic shunting to the temporal processing requirements of a given task. Applying this general biophysical mechanism to the example of speech processing, we propose a neuronal network model for time-warp–invariant word discrimination and demonstrate its excellent performance on a standard benchmark speech-recognition task. Our results demonstrate the important functional role of synaptic conductances in spike-based neuronal information processing and learning. The biophysics of temporal integration at neuronal membranes can endow sensory pathways with powerful time-warp–invariant computational capabilities.     

Distinct roles for spontaneous and visual activity in remodeling of the retinogeniculate synapse

Sensory experience and spontaneous activity play important roles in development of sensory circuits; however, their relative contributions are unclear. Here, we test the role of different forms of activity on remodeling of the mouse retinogeniculate synapse. We found that the bulk of maturation occurs without patterned sensory activity over 4 days spanning eye opening. During this early developmental period, blockade of spontaneous retinal activity by tetrodotoxin, but not visual deprivation, retarded synaptic strengthening and inhibited pruning of excess retinal afferents. Later in development, synaptic remodeling becomes sensitive to changes in visually evoked activity, but only if there has been previous visual experience. Synaptic strengthening and pruning were disrupted by visual deprivation following 1 week of vision, but not by chronic deprivation from birth. Thus, spontaneous activity is necessary to drive the bulk of synaptic refinement around the time of eye opening, while sensory experience is important for the subsequent maintenance of connections.     

An assessment of terminology for intraspecific diversity in fishes,with a focus on “ecotypes” and “life histories”

Understanding and preserving intraspecific diversity (ISD) is important for species conservation. However, ISD units do not have taxonomic standards and are not universally recognized. The terminology used to describe ISD is varied and often used ambiguously. We compared definitions of terms used to describe ISD with use in recent studies of three fish taxa: sticklebacks (Gasterosteidae), Pacific salmon and trout (Oncorhynchus spp., “PST”), and lampreys (Petromyzontiformes). Life history describes the phenotypic responses of organisms to environments and includes biological parameters that affect population growth or decline. Life‐history pathway(s) are the result of different organismal routes of development that can result in different life histories. These terms can be used to describe recognizable life‐history traits. Life history is generally used in organismal‐ and ecology‐based journals. The terms paired species/species pairs have been used to describe two different phenotypes, whereas in some species and situations a continuum of phenotypes may be expressed. Our review revealed overlapping definitions for race and subspecies, and subspecies and ecotypes. Ecotypes are genotypic adaptations to particular environments, and this term is often used in genetic‐ and evolution‐based journals. “Satellite species” is used for situations in which a parasitic lamprey yields two or more derived, nonparasitic lamprey species. Designatable Units, Evolutionary Significant Units (ESUs), and Distinct Population Segments (DPS) are used by some governments to classify ISD of vertebrate species within distinct and evolutionary significant criteria. In situations where the genetic or life‐history components of ISD are not well understood, a conservative approach would be to call them phenotypes.

The terminology used to describe intraspecific diversity is varied and often used ambiguously. “Ecotype” was originally used to describe patterns in genes and ecology, and recent studies employing this term tend to report a genetic basis in ISD. By contrast, “life history” describes biological parameters that affect demography, and this term tends to be used in organismal‐ and ecology‐based journals.     

Functional Brain Networks Develop from a “Local to Distributed” Organization

The mature human brain is organized into a collection of specialized functional networks that flexibly interact to support various cognitive functions. Studies of development often attempt to identify the organizing principles that guide the maturation of these functional networks. In this report, we combine resting state functional connectivity MRI (rs-fcMRI), graph analysis, community detection, and spring-embedding visualization techniques to analyze four separate networks defined in earlier studies. As we have previously reported, we find, across development, a trend toward ‘segregation’ (a general decrease in correlation strength) between regions close in anatomical space and ‘integration’ (an increased correlation strength) between selected regions distant in space. The generalization of these earlier trends across multiple networks suggests that this is a general developmental principle for changes in functional connectivity that would extend to large-scale graph theoretic analyses of large-scale brain networks. Communities in children are predominantly arranged by anatomical proximity, while communities in adults predominantly reflect functional relationships, as defined from adult fMRI studies. In sum, over development, the organization of multiple functional networks shifts from a local anatomical emphasis in children to a more “distributed” architecture in young adults. We argue that this “local to distributed” developmental characterization has important implications for understanding the development of neural systems underlying cognition. Further, graph metrics (e.g., clustering coefficients and average path lengths) are similar in child and adult graphs, with both showing “small-world”-like properties, while community detection by modularity optimization reveals stable communities within the graphs that are clearly different between young children and young adults. These observations suggest that early school age children and adults both have relatively efficient systems that may solve similar information processing problems in divergent ways.