Binding in hippocampal-entorhinal circuits enables compositionality in cognitive maps.

We propose a normative model for spatial representation in the hippocampal formation that combines optimality principles, such as maximizing coding range and spatial information per neuron, with an algebraic framework for computing in distributed representation. Spatial position is encoded in a residue number system, with individual residues represented by high-dimensional, complex-valued vectors. These are composed into a single vector representing position by a similarity-preserving, conjunctive vector-binding operation. Self-consistency between the representations of the overall position and of the individual residues is enforced by a modular attractor network whose modules correspond to the grid cell modules in entorhinal cortex. The vector binding operation can also associate different contexts to spatial representations, yielding a model for entorhinal cortex and hippocampus. We show that the model achieves normative desiderata including superlinear scaling of patterns with dimension, robust error correction, and hexagonal, carry-free encoding of spatial position. These properties in turn enable robust path integration and association with sensory inputs. More generally, the model formalizes how compositional computations could occur in the hippocampal formation and leads to testable experimental predictions.

Enhanced RBE of Particle Radiation Depends on Beam Size in the Micrometer Range.

High-linear energy transfer (LET) radiation, such as heavy ions is associated with a higher relative biological effectiveness (RBE) than low-LET radiation, such as photons. Irradiation with low- and high-LET particles differ in the interaction with the cellular matter and therefore in the spatial dose distribution. When a single high-LET particle interacts with matter, it results in doses of up to thousands of gray (Gy) locally concentrated around the ion trajectory, whereas the mean dose averaged over the target, such as a cell nucleus is only in the range of a Gy. DNA damage therefore accumulates in this small volume. In contrast, up to hundreds of low-LET particle hits are required to achieve the same mean dose, resulting in a quasi-homogeneous damage distribution throughout the cell nucleus. In this study, we investigated the dependence of RBE from different spatial dose depositions using different focused beam spot sizes of proton radiation with respect to the induction of chromosome aberrations and clonogenic cell survival. Human-hamster hybrid (AL) as well as Chinese hamster ovary cells (CHO-K1) were irradiated with focused low LET protons of 20 MeV (LET = 2.6 keV/µm) beam energy with a mean dose of 1.7 Gy in a quadratic matrix pattern with point spacing of 5.4 × 5.4 µm2 and 117 protons per matrix point at the ion microbeam SNAKE using different beam spot sizes between 0.8 µm and 2.8 µm (full width at half maximum). The dose-response curves of X-ray reference radiation were used to determine the RBE after a 1.7 Gy dose of radiation. The RBE for the induction of dicentric chromosomes and cell inactivation was increased after irradiation with the smallest beam spot diameter (0.8 µm for chromosome aberration experiments and 1.0 µm for cell survival experiments) compared to homogeneous proton radiation but was still below the RBE of a corresponding high LET single ion hit. By increasing the spot size to 1.6-1.8 µm, the RBE decreased but was still higher than for homogeneously distributed protons. By further increasing the spot size to 2.7-2.8 µm, the RBE was no longer different from the homogeneous radiation. Our experiments demonstrate that varying spot size of low-LET radiation gradually modifies the RBE. This underlines that a substantial fraction of enhanced RBE originates from inhomogeneous energy concentrations on the µm scale (mean intertrack distances of low-LET particles below 0.1 µm) and quantifies the link between such energy concentration and RBE. The missing fraction of RBE enhancement when comparing with high-LET ions is attributed to the high inner track energy deposition on the nanometer scale. The results are compared with model results of PARTRAC and LEM for chromosomal aberration and cell survival, respectively, which suggest mechanistic interpretations of the observed radiation effects.

Spectral composition of secondary electrons based on the Kiefer-Straaten ion track structure model.

The major part of energy deposition of ionizing radiation is caused by secondary electrons, independent of the primary radiation type. However, their spatial concentration and their spectral properties strongly depend on the primary radiation type and finally determine the pattern of molecular damage e.g. to biological targets as the DNA, and thus the final effect of the radiation exposure. To describe the physical and to predict the biological consequences of charged ion irradiation, amorphous track structure approaches have proven to be pragmatic and helpful. There, the local dose deposition in the ion track is equated by considering the emission and slowing down of the secondary electrons from the primary particle track. In the present work we exploit the model of Kiefer and Straaten and derive the spectral composition of secondary electrons as function of the distance to the track center. The spectral composition indicates differences to spectra of low linear energy transfer (LET) photon radiation, which we confirm by a comparison with Monte Carlo studies. We demonstrate that the amorphous track structure approach provides a simple tool for evaluating the spectral electron properties within the track structure. Predictions of the LET of electrons across the track structure as well as the electronic dose build-up effect are derived. Implications for biological effects and corresponding predicting models based on amorphous track structure are discussed.

Computing with Residue Numbers in High-Dimensional Representation.

We introduce Residue Hyperdimensional Computing, a computing framework that unifies residue number systems with an algebra defined over random, high-dimensional vectors. We show how residue numbers can be represented as high-dimensional vectors in a manner that allows algebraic operations to be performed with component-wise, parallelizable operations on the vector elements. The resulting framework, when combined with an efficient method for factorizing high-dimensional vectors, can represent and operate on numerical values over a large dynamic range using vastly fewer resources than previous methods, and it exhibits impressive robustness to noise. We demonstrate the potential for this framework to solve computationally difficult problems in visual perception and combinatorial optimization, showing improvement over baseline methods. More broadly, the framework provides a possible account for the computational operations of grid cells in the brain, and it suggests new machine learning architectures for representing and manipulating numerical data.

Learning Energy-Based Models in High-Dimensional Spaces with Multiscale Denoising-Score Matching.

Energy-based models (EBMs) assign an unnormalized log probability to data samples. This functionality has a variety of applications, such as sample synthesis, data denoising, sample restoration, outlier detection, Bayesian reasoning and many more. But, the training of EBMs using standard maximum likelihood is extremely slow because it requires sampling from the model distribution. Score matching potentially alleviates this problem. In particular, denoising-score matching has been successfully used to train EBMs. Using noisy data samples with one fixed noise level, these models learn fast and yield good results in data denoising. However, demonstrations of such models in the high-quality sample synthesis of high-dimensional data were lacking. Recently, a paper showed that a generative model trained by denoising-score matching accomplishes excellent sample synthesis when trained with data samples corrupted with multiple levels of noise. Here we provide an analysis and empirical evidence showing that training with multiple noise levels is necessary when the data dimension is high. Leveraging this insight, we propose a novel EBM trained with multiscale denoising-score matching. Our model exhibits a data-generation performance comparable to state-of-the-art techniques such as GANs and sets a new baseline for EBMs. The proposed model also provides density information and performs well on an image-inpainting task.

Efficient optimization with higher-order ising machines.

A prominent approach to solving combinatorial optimization problems on parallel hardware is Ising machines, i.e., hardware implementations of networks of interacting binary spin variables. Most Ising machines leverage second-order interactions although important classes of optimization problems, such as satisfiability problems, map more seamlessly to Ising networks with higher-order interactions. Here, we demonstrate that higher-order Ising machines can solve satisfiability problems more resource-efficiently in terms of the number of spin variables and their connections when compared to traditional second-order Ising machines. Further, our results show on a benchmark dataset of Boolean k-satisfiability problems that higher-order Ising machines implemented with coupled oscillators rapidly find solutions that are better than second-order Ising machines, thus, improving the current state-of-the-art for Ising machines.

Local interneurons in the murine visual thalamus have diverse receptive fields and can provide feature selective inhibition to relay cells.

By influencing the type and quality of information that relay cells transmit, local interneurons in thalamus have a powerful impact on cortex. To define the sensory features that these inhibitory neurons encode, we mapped receptive fields of optogenetically identified cells in the murine dorsolateral geniculate nucleus. Although few in number, local interneurons had diverse types of receptive fields, like their counterpart relay cells. This result differs markedly from visual cortex, where inhibitory cells are typically less selective than excitatory cells. To explore how thalamic interneurons might converge on relay cells, we took a computational approach. Using an evolutionary algorithm to search through a library of interneuron models generated from our results, we show that aggregated output from different groups of local interneurons can simulate the inhibitory component of the relay cell's receptive field. Thus, our work provides proof-of-concept that groups of diverse interneurons can supply feature-specific inhibition to relay cells.

Efficient Decoding of Compositional Structure in Holistic Representations.

We investigate the task of retrieving information from compositional distributed representations formed by hyperdimensional computing/vector symbolic architectures and present novel techniques that achieve new information rate bounds. First, we provide an overview of the decoding techniques that can be used to approach the retrieval task. The techniques are categorized into four groups. We then evaluate the considered techniques in several settings that involve, for example, inclusion of external noise and storage elements with reduced precision. In particular, we find that the decoding techniques from the sparse coding and compressed sensing literature (rarely used for hyperdimensional computing/vector symbolic architectures) are also well suited for decoding information from the compositional distributed representations. Combining these decoding techniques with interference cancellation ideas from communications improves previously reported bounds (Hersche et al., 2021) of the information rate of the distributed representations from 1.20 to 1.40 bits per dimension for smaller codebooks and from 0.60 to 1.26 bits per dimension for larger codebooks.

Perceptron Theory Can Predict the Accuracy of Neural Networks.

Multilayer neural networks set the current state of the art for many technical classification problems. But, these networks are still, essentially, black boxes in terms of analyzing them and predicting their performance. Here, we develop a statistical theory for the one-layer perceptron and show that it can predict performances of a surprisingly large variety of neural networks with different architectures. A general theory of classification with perceptrons is developed by generalizing an existing theory for analyzing reservoir computing models and connectionist models for symbolic reasoning known as vector symbolic architectures. Our statistical theory offers three formulas leveraging the signal statistics with increasing detail. The formulas are analytically intractable, but can be evaluated numerically. The description level that captures maximum details requires stochastic sampling methods. Depending on the network model, the simpler formulas already yield high prediction accuracy. The quality of the theory predictions is assessed in three experimental settings, a memorization task for echo state networks (ESNs) from reservoir computing literature, a collection of classification datasets for shallow randomly connected networks, and the ImageNet dataset for deep convolutional neural networks. We find that the second description level of the perceptron theory can predict the performance of types of ESNs, which could not be described previously. Furthermore, the theory can predict deep multilayer neural networks by being applied to their output layer. While other methods for prediction of neural networks performance commonly require to train an estimator model, the proposed theory requires only the first two moments of the distribution of the postsynaptic sums in the output neurons. Moreover, the perceptron theory compares favorably to other methods that do not rely on training an estimator model.

Current state of phoenixin-the implications of the pleiotropic peptide in stress and its potential as a therapeutic target.

Phoenixin is a pleiotropic peptide, whose known functions have broadened significantly over the last decade. Initially first described as a reproductive peptide in 2013, phoenixin is now recognized as being implicated in hypertension, neuroinflammation, pruritus, food intake, anxiety as well as stress. Due to its wide field of involvement, an interaction with physiological as well as psychological control loops has been speculated. It has shown to be both able to actively reduce anxiety as well as being influenced by external stressors. Initial rodent models have shown that central administration of phoenixin alters the behavior of the subjects when confronted with stress-inducing situations, proposing an interaction with the perception and processing of stress and anxiety. Although the research on phoenixin is still in its infancy, there are several promising insights into its functionality, which might prove to be of value in the pharmacological treatment of several psychiatric and psychosomatic illnesses such as anorexia nervosa, post-traumatic stress disorder as well as the increasingly prevalent stress-related illnesses of burnout and depression. In this review, we aim to provide an overview of the current state of knowledge of phoenixin, its interactions with physiological processes as well as focus on the recent developments in stress response and the possible novel treatment options this might entail.

Variable Binding for Sparse Distributed Representations: Theory and Applications.

Variable binding is a cornerstone of symbolic reasoning and cognition. But how binding can be implemented in connectionist models has puzzled neuroscientists, cognitive psychologists, and neural network researchers for many decades. One type of connectionist model that naturally includes a binding operation is vector symbolic architectures (VSAs). In contrast to other proposals for variable binding, the binding operation in VSAs is dimensionality-preserving, which enables representing complex hierarchical data structures, such as trees, while avoiding a combinatoric expansion of dimensionality. Classical VSAs encode symbols by dense randomized vectors, in which information is distributed throughout the entire neuron population. By contrast, in the brain, features are encoded more locally, by the activity of single neurons or small groups of neurons, often forming sparse vectors of neural activation. Following Laiho et al. (2015), we explore symbolic reasoning with a special case of sparse distributed representations. Using techniques from compressed sensing, we first show that variable binding in classical VSAs is mathematically equivalent to tensor product binding between sparse feature vectors, another well-known binding operation which increases dimensionality. This theoretical result motivates us to study two dimensionality-preserving binding methods that include a reduction of the tensor matrix into a single sparse vector. One binding method for general sparse vectors uses random projections, the other, block-local circular convolution, is defined for sparse vectors with block structure, sparse block-codes. Our experiments reveal that block-local circular convolution binding has ideal properties, whereas random projection based binding also works, but is lossy. We demonstrate in example applications that a VSA with block-local circular convolution and sparse block-codes reaches similar performance as classical VSAs. Finally, we discuss our results in the context of neuroscience and neural networks.

News without the buzz: reading out weak theta rhythms in the hippocampus.

Local field potentials (LFPs) reflect the collective dynamics of neural populations, yet their exact relationship to neural codes remains unknown1. One notable exception is the theta rhythm of the rodent hippocampus, which seems to provide a reference clock to decode the animal's position from spatiotemporal patterns of neuronal spiking2 or LFPs3. But when the animal stops, theta becomes irregular4, potentially indicating the breakdown of temporal coding by neural populations. Here we show that no such breakdown occurs, introducing an artificial neural network that can recover position-tuned rhythmic patterns (pThetas) without relying on the more prominent theta rhythm as a reference clock. pTheta and theta preferentially correlate with place cell and interneuron spiking, respectively. When rats forage in an open field, pTheta is jointly tuned to position and head orientation, a property not seen in individual place cells but expected to emerge from place cell sequences5. Our work demonstrates that weak and intermittent oscillations, as seen in many brain regions and species, can carry behavioral information commensurate with population spike codes.

Optimal experiment design for element specific atom counting using multiple annular dark field scanning transmission electron microscopy detectors.

This paper investigates the possible benefits for counting atoms of different chemical nature when analysing multiple 2D scanning transmission electron microscopy (STEM) images resulting from independent annular dark field (ADF) detector regimes. To reach this goal, the principles of statistical detection theory are used to quantify the probability of error when determining the number of atoms in atomic columns consisting of multiple types of elements. In order to apply this theory, atom-counting is formulated as a statistical hypothesis test, where each hypothesis corresponds to a specific number of atoms of each atom type in an atomic column. The probability of error, which is limited by the unavoidable presence of electron counting noise, can then be computed from scattering-cross sections extracted from multiple ADF STEM images. Minimisation of the probability of error as a function of the inner and outer angles of a specified number of independent ADF collection regimes results in optimal experimental designs. Based on simulations of spherical Au@Ag and Au@Pt core-shell nanoparticles, we investigate how the combination of two non-overlapping detector regimes helps to improve the probability of error when unscrambling two types of atoms. In particular, the combination of a narrow low angle ADF detector with a detector formed by the remaining annular collection regime is found to be optimal. The benefit is more significant if the atomic number Z difference becomes larger. In addition, we show the benefit of subdividing the detector regime into three collection areas for heterogeneous nanostructures based on a structure consisting of three types of elements, e.g., a mixture of Au, Ag and Al atoms. Finally, these results are compared with the probability of error resulting when one would ultimately use a pixelated 4D STEM detector and how this could help to further reduce the incident electron dose.

The gene order in the nuo-operon is not essential for the assembly of E. coli complex I.

Energy-converting NADH: ubiquinone oxidoreductase, respiratory complex I, plays an important role in cellular energy metabolism. Bacterial complex I is generally composed of 14 different subunits, seven of which are membranous and the other seven are globular proteins. They are encoded by the nuo-operon, whose gene order is strictly conserved in bacteria. The operon starts with nuoA encoding a membranous subunit followed by genes encoding globular subunits. To test the idea that NuoA acts as a seed to initiate the assembly of the complex in the membrane, we generated mutants that either lacked nuoA or contain nuoA at a different position within the operon. To enable the detection of putative assembly intermediates, the globular subunit NuoF and the membranous subunit NuoM were individually decorated with the fluorescent protein mCherry. Deletion of nuoA led to the assembly of an inactive complex in the membrane containing NuoF and NuoM. Re-arrangement of nuoA within the nuo-operon led to a slightly diminished amount of complex I in the membrane that was fully active. Thus, nuoA but not its distinct position in the operon is required for the assembly of E. coli complex I. Furthermore, we detected a previously unknown assembly intermediate in the membrane containing NuoM that is present in greater amounts than complex I.

Consciousness is supported by near-critical slow cortical electrodynamics.

Mounting evidence suggests that during conscious states, the electrodynamics of the cortex are poised near a critical point or phase transition and that this near-critical behavior supports the vast flow of information through cortical networks during conscious states. Here, we empirically identify a mathematically specific critical point near which waking cortical oscillatory dynamics operate, which is known as the edge-of-chaos critical point, or the boundary between stability and chaos. We do so by applying the recently developed modified 0-1 chaos test to electrocorticography (ECoG) and magnetoencephalography (MEG) recordings from the cortices of humans and macaques across normal waking, generalized seizure, anesthesia, and psychedelic states. Our evidence suggests that cortical information processing is disrupted during unconscious states because of a transition of low-frequency cortical electric oscillations away from this critical point; conversely, we show that psychedelics may increase the information richness of cortical activity by tuning low-frequency cortical oscillations closer to this critical point. Finally, we analyze clinical electroencephalography (EEG) recordings from patients with disorders of consciousness (DOC) and show that assessing the proximity of slow cortical oscillatory electrodynamics to the edge-of-chaos critical point may be useful as an index of consciousness in the clinical setting.

Is Neuroscience FAIR? A Call for Collaborative Standardisation of Neuroscience Data.

In this perspective article, we consider the critical issue of data and other research object standardisation and, specifically, how international collaboration, and organizations such as the International Neuroinformatics Coordinating Facility (INCF) can encourage that emerging neuroscience data be Findable, Accessible, Interoperable, and Reusable (FAIR). As neuroscientists engaged in the sharing and integration of multi-modal and multiscale data, we see the current insufficiency of standards as a major impediment in the Interoperability and Reusability of research results. We call for increased international collaborative standardisation of neuroscience data to foster integration and efficient reuse of research objects.

Vector Symbolic Architectures as a Computing Framework for Emerging Hardware.

This article reviews recent progress in the development of the computing framework Vector Symbolic Architectures (also known as Hyperdimensional Computing). This framework is well suited for implementation in stochastic, emerging hardware and it naturally expresses the types of cognitive operations required for Artificial Intelligence (AI). We demonstrate in this article that the field-like algebraic structure of Vector Symbolic Architectures offers simple but powerful operations on high-dimensional vectors that can support all data structures and manipulations relevant to modern computing. In addition, we illustrate the distinguishing feature of Vector Symbolic Architectures, "computing in superposition," which sets it apart from conventional computing. It also opens the door to efficient solutions to the difficult combinatorial search problems inherent in AI applications. We sketch ways of demonstrating that Vector Symbolic Architectures are computationally universal. We see them acting as a framework for computing with distributed representations that can play a role of an abstraction layer for emerging computing hardware. This article serves as a reference for computer architects by illustrating the philosophy behind Vector Symbolic Architectures, techniques of distributed computing with them, and their relevance to emerging computing hardware, such as neuromorphic computing.

Cellular Automata Can Reduce Memory Requirements of Collective-State Computing.

Various nonclassical approaches of distributed information processing, such as neural networks, reservoir computing (RC), vector symbolic architectures (VSAs), and others, employ the principle of collective-state computing. In this type of computing, the variables relevant in computation are superimposed into a single high-dimensional state vector, the collective state. The variable encoding uses a fixed set of random patterns, which has to be stored and kept available during the computation. In this article, we show that an elementary cellular automaton with rule 90 (CA90) enables the space-time tradeoff for collective-state computing models that use random dense binary representations, i.e., memory requirements can be traded off with computation running CA90. We investigate the randomization behavior of CA90, in particular, the relation between the length of the randomization period and the size of the grid, and how CA90 preserves similarity in the presence of the initialization noise. Based on these analyses, we discuss how to optimize a collective-state computing model, in which CA90 expands representations on the fly from short seed patterns-rather than storing the full set of random patterns. The CA90 expansion is applied and tested in concrete scenarios using RC and VSAs. Our experimental results show that collective-state computing with CA90 expansion performs similarly compared to traditional collective-state models, in which random patterns are generated initially by a pseudorandom number generator and then stored in a large memory.

Alpha-Particle Exposure Induces Mainly Unstable Complex Chromosome Aberrations which do not Contribute to Radiation-Associated Cytogenetic Risk.

The mechanism underlying the carcinogenic potential of α radiation is not fully understood, considering that cell inactivation (e.g., mitotic cell death) as a main consequence of exposure efficiently counteracts the spreading of heritable DNA damage. The aim of this study is to improve our understanding of the effectiveness of α particles in inducing different types of chromosomal aberrations, to determine the respective values of the relative biological effectiveness (RBE) and to interpret the results with respect to exposure risk. Human peripheral blood lymphocytes (PBLs) from a single donor were exposed ex vivo to doses of 0-6 Gy X rays or 0-2 Gy α particles. Cells were harvested at two different times after irradiation to account for the mitotic delay of heavily damaged cells, which is known to occur after exposure to high-LET radiation (including α particles). Analysis of the kinetics of cells reaching first or second (and higher) mitosis after irradiation and aberration data obtained by the multiplex fluorescence in situ hybridization (mFISH) technique are used to determine of the cytogenetic risk, i.e., the probability for transmissible aberrations in surviving lymphocytes. The analysis shows that the cytogenetic risk after α exposure is lower than after X rays. This indicates that the actually observed higher carcinogenic effect of α radiation is likely to stem from small scale mutations that are induced effectively by high-LET radiation but cannot be resolved by mFISH analysis.

A Neural Network MCMC Sampler That Maximizes Proposal Entropy.

Markov Chain Monte Carlo (MCMC) methods sample from unnormalized probability distributions and offer guarantees of exact sampling. However, in the continuous case, unfavorable geometry of the target distribution can greatly limit the efficiency of MCMC methods. Augmenting samplers with neural networks can potentially improve their efficiency. Previous neural network-based samplers were trained with objectives that either did not explicitly encourage exploration, or contained a term that encouraged exploration but only for well structured distributions. Here we propose to maximize proposal entropy for adapting the proposal to distributions of any shape. To optimize proposal entropy directly, we devised a neural network MCMC sampler that has a flexible and tractable proposal distribution. Specifically, our network architecture utilizes the gradient of the target distribution for generating proposals. Our model achieved significantly higher efficiency than previous neural network MCMC techniques in a variety of sampling tasks, sometimes by more than an order magnitude. Further, the sampler was demonstrated through the training of a convergent energy-based model of natural images. The adaptive sampler achieved unbiased sampling with significantly higher proposal entropy than a Langevin dynamics sample. The trained sampler also achieved better sample quality.