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.

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.

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.

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.

Density Encoding Enables Resource-Efficient Randomly Connected Neural Networks.

The deployment of machine learning algorithms on resource-constrained edge devices is an important challenge from both theoretical and applied points of view. In this brief, we focus on resource-efficient randomly connected neural networks known as random vector functional link (RVFL) networks since their simple design and extremely fast training time make them very attractive for solving many applied classification tasks. We propose to represent input features via the density-based encoding known in the area of stochastic computing and use the operations of binding and bundling from the area of hyperdimensional computing for obtaining the activations of the hidden neurons. Using a collection of 121 real-world data sets from the UCI machine learning repository, we empirically show that the proposed approach demonstrates higher average accuracy than the conventional RVFL. We also demonstrate that it is possible to represent the readout matrix using only integers in a limited range with minimal loss in the accuracy. In this case, the proposed approach operates only on small n -bits integers, which results in a computationally efficient architecture. Finally, through hardware field-programmable gate array (FPGA) implementations, we show that such an approach consumes approximately 11 times less energy than that of the conventional RVFL.

Resonator Networks, 2: Factorization Performance and Capacity Compared to Optimization-Based Methods.

We develop theoretical foundations of resonator networks, a new type of recurrent neural network introduced in Frady, Kent, Olshausen, and Sommer (2020), a companion article in this issue, to solve a high-dimensional vector factorization problem arising in Vector Symbolic Architectures. Given a composite vector formed by the Hadamard product between a discrete set of high-dimensional vectors, a resonator network can efficiently decompose the composite into these factors. We compare the performance of resonator networks against optimization-based methods, including Alternating Least Squares and several gradient-based algorithms, showing that resonator networks are superior in several important ways. This advantage is achieved by leveraging a combination of nonlinear dynamics and searching in superposition, by which estimates of the correct solution are formed from a weighted superposition of all possible solutions. While the alternative methods also search in superposition, the dynamics of resonator networks allow them to strike a more effective balance between exploring the solution space and exploiting local information to drive the network toward probable solutions. Resonator networks are not guaranteed to converge, but within a particular regime they almost always do. In exchange for relaxing the guarantee of global convergence, resonator networks are dramatically more effective at finding factorizations than all alternative approaches considered.

Resonator Networks, 1: An Efficient Solution for Factoring High-Dimensional, Distributed Representations of Data Structures.

The ability to encode and manipulate data structures with distributed neural representations could qualitatively enhance the capabilities of traditional neural networks by supporting rule-based symbolic reasoning, a central property of cognition. Here we show how this may be accomplished within the framework of Vector Symbolic Architectures (VSAs) (Plate, 1991; Gayler, 1998; Kanerva, 1996), whereby data structures are encoded by combining high-dimensional vectors with operations that together form an algebra on the space of distributed representations. In particular, we propose an efficient solution to a hard combinatorial search problem that arises when decoding elements of a VSA data structure: the factorization of products of multiple codevectors. Our proposed algorithm, called a resonator network, is a new type of recurrent neural network that interleaves VSA multiplication operations and pattern completion. We show in two examples-parsing of a tree-like data structure and parsing of a visual scene-how the factorization problem arises and how the resonator network can solve it. More broadly, resonator networks open the possibility of applying VSAs to myriad artificial intelligence problems in real-world domains. The companion article in this issue (Kent, Frady, Sommer, & Olshausen, 2020) presents a rigorous analysis and evaluation of the performance of resonator networks, showing it outperforms alternative approaches.

Robust computation with rhythmic spike patterns.

Information coding by precise timing of spikes can be faster and more energy efficient than traditional rate coding. However, spike-timing codes are often brittle, which has limited their use in theoretical neuroscience and computing applications. Here, we propose a type of attractor neural network in complex state space and show how it can be leveraged to construct spiking neural networks with robust computational properties through a phase-to-timing mapping. Building on Hebbian neural associative memories, like Hopfield networks, we first propose threshold phasor associative memory (TPAM) networks. Complex phasor patterns whose components can assume continuous-valued phase angles and binary magnitudes can be stored and retrieved as stable fixed points in the network dynamics. TPAM achieves high memory capacity when storing sparse phasor patterns, and we derive the energy function that governs its fixed-point attractor dynamics. Second, we construct 2 spiking neural networks to approximate the complex algebraic computations in TPAM, a reductionist model with resonate-and-fire neurons and a biologically plausible network of integrate-and-fire neurons with synaptic delays and recurrently connected inhibitory interneurons. The fixed points of TPAM correspond to stable periodic states of precisely timed spiking activity that are robust to perturbation. The link established between rhythmic firing patterns and complex attractor dynamics has implications for the interpretation of spike patterns seen in neuroscience and can serve as a framework for computation in emerging neuromorphic devices.

A Theory of Sequence Indexing and Working Memory in Recurrent Neural Networks.

To accommodate structured approaches of neural computation, we propose a class of recurrent neural networks for indexing and storing sequences of symbols or analog data vectors. These networks with randomized input weights and orthogonal recurrent weights implement coding principles previously described in vector symbolic architectures (VSA) and leverage properties of reservoir computing. In general, the storage in reservoir computing is lossy, and crosstalk noise limits the retrieval accuracy and information capacity. A novel theory to optimize memory performance in such networks is presented and compared with simulation experiments. The theory describes linear readout of analog data and readout with winner-take-all error correction of symbolic data as proposed in VSA models. We find that diverse VSA models from the literature have universal performance properties, which are superior to what previous analyses predicted. Further, we propose novel VSA models with the statistically optimal Wiener filter in the readout that exhibit much higher information capacity, in particular for storing analog data. The theory we present also applies to memory buffers, networks with gradual forgetting, which can operate on infinite data streams without memory overflow. Interestingly, we find that different forgetting mechanisms, such as attenuating recurrent weights or neural nonlinearities, produce very similar behavior if the forgetting time constants are matched. Such models exhibit extensive capacity when their forgetting time constant is optimized for given noise conditions and network size. These results enable the design of new types of VSA models for the online processing of data streams.

MicroRNA-101 Regulates Multiple Developmental Programs to Constrain Excitation in Adult Neural Networks.

A critical feature of neural networks is that they balance excitation and inhibition to prevent pathological dysfunction. How this is achieved is largely unknown, although deficits in the balance contribute to many neurological disorders. We show here that a microRNA (miR-101) is a key orchestrator of this essential feature, shaping the developing network to constrain excitation in the adult. Transient early blockade of miR-101 induces long-lasting hyper-excitability and persistent memory deficits. Using target site blockers in vivo, we identify multiple developmental programs regulated in parallel by miR-101 to achieve balanced networks. Repression of one target, NKCC1, initiates the switch in γ-aminobutyric acid (GABA) signaling, limits early spontaneous activity, and constrains dendritic growth. Kif1a and Ank2 are targeted to prevent excessive synapse formation. Simultaneous de-repression of these three targets completely phenocopies major dysfunctions produced by miR-101 blockade. Our results provide new mechanistic insight into brain development and suggest novel candidates for therapeutic intervention.

Scalable Semisupervised Functional Neurocartography Reveals Canonical Neurons in Behavioral Networks.

Large-scale data collection efforts to map the brain are underway at multiple spatial and temporal scales, but all face fundamental problems posed by high-dimensional data and intersubject variability. Even seemingly simple problems, such as identifying a neuron/brain region across animals/subjects, become exponentially more difficult in high dimensions, such as recognizing dozens of neurons/brain regions simultaneously. We present a framework and tools for functional neurocartography-the large-scale mapping of neural activity during behavioral states. Using a voltage-sensitive dye (VSD), we imaged the multifunctional responses of hundreds of leech neurons during several behaviors to identify and functionally map homologous neurons. We extracted simple features from each of these behaviors and combined them with anatomical features to create a rich medium-dimensional feature space. This enabled us to use machine learning techniques and visualizations to characterize and account for intersubject variability, piece together a canonical atlas of neural activity, and identify two behavioral networks. We identified 39 neurons (18 pairs, 3 unpaired) as part of a canonical swim network and 17 neurons (8 pairs, 1 unpaired) involved in a partially overlapping preparatory network. All neurons in the preparatory network rapidly depolarized at the onsets of each behavior, suggesting that it is part of a dedicated rapid-response network. This network is likely mediated by the S cell, and we referenced VSD recordings to an activity atlas to identify multiple cells of interest simultaneously in real time for further experiments. We targeted and electrophysiologically verified several neurons in the swim network and further showed that the S cell is presynaptic to multiple neurons in the preparatory network. This study illustrates the basic framework to map neural activity in high dimensions with large-scale recordings and how to extract the rich information necessary to perform analyses in light of intersubject variability.

Direct Imaging of Hippocampal Epileptiform Calcium Motifs Following Kainic Acid Administration in Freely Behaving Mice.

Prolonged exposure to abnormally high calcium concentrations is thought to be a core mechanism underlying hippocampal damage in epileptic patients; however, no prior study has characterized calcium activity during seizures in the live, intact hippocampus. We have directly investigated this possibility by combining whole-brain electroencephalographic (EEG) measurements with microendoscopic calcium imaging of pyramidal cells in the CA1 hippocampal region of freely behaving mice treated with the pro-convulsant kainic acid (KA). We observed that KA administration led to systematic patterns of epileptiform calcium activity: a series of large-scale, intensifying flashes of increased calcium fluorescence concurrent with a cluster of low-amplitude EEG waveforms. This was accompanied by a steady increase in cellular calcium levels (>5 fold increase relative to the baseline), followed by an intense spreading calcium wave characterized by a 218% increase in global mean intensity of calcium fluorescence (n = 8, range [114-349%], p < 10(-4); t-test). The wave had no consistent EEG phenotype and occurred before the onset of motor convulsions. Similar changes in calcium activity were also observed in animals treated with 2 different proconvulsant agents, N-methyl-D-aspartate (NMDA) and pentylenetetrazol (PTZ), suggesting the measured changes in calcium dynamics are a signature of seizure activity rather than a KA-specific pathology. Additionally, despite reducing the behavioral severity of KA-induced seizures, the anticonvulsant drug valproate (VA, 300 mg/kg) did not modify the observed abnormalities in calcium dynamics. These results confirm the presence of pathological calcium activity preceding convulsive motor seizures and support calcium as a candidate signaling molecule in a pathway connecting seizures to subsequent cellular damage. Integrating in vivo calcium imaging with traditional assessment of seizures could potentially increase translatability of pharmacological intervention, leading to novel drug screening paradigms and therapeutics designed to target and abolish abnormal patterns of both electrical and calcium excitation.

Improved PeT molecules for optically sensing voltage in neurons.

VoltageFluor (VF) dyes have the potential to measure voltage optically in excitable membranes with a combination of high spatial and temporal resolution essential to better characterize the voltage dynamics of large groups of excitable cells. VF dyes sense voltage with high speed and sensitivity using photoinduced electron transfer (PeT) through a conjugated molecular wire. We show that tuning the driving force for PeT (ΔGPeT + w) through systematic chemical substitution modulates voltage sensitivity, estimate (ΔGPeT + w) values from experimentally measured redox potentials, and validate the voltage sensitivities in patch-clamped HEK cells for 10 new VF dyes. VF2.1(OMe).H, with a 48% ΔF/F per 100 mV, shows approximately 2-fold improvement over previous dyes in HEK cells, dissociated rat cortical neurons, and medicinal leech ganglia. Additionally, VF2.1(OMe).H faithfully reports pharmacological effects and circuit activity in mouse olfactory bulb slices, thus opening a wide range of previously inaccessible applications for voltage-sensitive dyes.

Sexual attraction: sex-specific wiring of neural circuitry.

Two recent studies describe mechanisms by which sexually dimorphic responses to pheromones in the nematode worm Caenorhabditis elegans are driven by differences in the balance of neural circuits that control attraction and repulsion behaviors.

Optically monitoring voltage in neurons by photo-induced electron transfer through molecular wires.

Fluorescence imaging is an attractive method for monitoring neuronal activity. A key challenge for optically monitoring voltage is development of sensors that can give large and fast responses to changes in transmembrane potential. We now present fluorescent sensors that detect voltage changes in neurons by modulation of photo-induced electron transfer (PeT) from an electron donor through a synthetic molecular wire to a fluorophore. These dyes give bigger responses to voltage than electrochromic dyes, yet have much faster kinetics and much less added capacitance than existing sensors based on hydrophobic anions or voltage-sensitive ion channels. These features enable single-trial detection of synaptic and action potentials in cultured hippocampal neurons and intact leech ganglia. Voltage-dependent PeT should be amenable to much further optimization, but the existing probes are already valuable indicators of neuronal activity.

Faces and text attract gaze independent of the task: Experimental data and computer model.

Previous studies of eye gaze have shown that when looking at images containing human faces, observers tend to rapidly focus on the facial regions. But is this true of other high-level image features as well? We here investigate the extent to which natural scenes containing faces, text elements, and cell phones-as a suitable control-attract attention by tracking the eye movements of subjects in two types of tasks-free viewing and search. We observed that subjects in free-viewing conditions look at faces and text 16.6 and 11.1 times more than similar regions normalized for size and position of the face and text. In terms of attracting gaze, text is almost as effective as faces. Furthermore, it is difficult to avoid looking at faces and text even when doing so imposes a cost. We also found that subjects took longer in making their initial saccade when they were told to avoid faces/text and their saccades landed on a non-face/non-text object. We refine a well-known bottom-up computer model of saliency-driven attention that includes conspicuity maps for color, orientation, and intensity by adding high-level semantic information (i.e., the location of faces or text) and demonstrate that this significantly improves the ability to predict eye fixations in natural images. Our enhanced model's predictions yield an area under the ROC curve over 84% for images that contain faces or text when compared against the actual fixation pattern of subjects. This suggests that the primate visual system allocates attention using such an enhanced saliency map.

The relation of phase noise and luminance contrast to overt attention in complex visual stimuli.

Models of attention are typically based on difference maps in low-level features but neglect higher order stimulus structure. To what extent does higher order statistics affect human attention in natural stimuli? We recorded eye movements while observers viewed unmodified and modified images of natural scenes. Modifications included contrast modulations (resulting in changes to first- and second-order statistics), as well as the addition of noise to the Fourier phase (resulting in changes to higher order statistics). We have the following findings: (1) Subjects' interpretation of a stimulus as a "natural" depiction of an outdoor scene depends on higher order statistics in a highly nonlinear, categorical fashion. (2) Confirming previous findings, contrast is elevated at fixated locations for a variety of stimulus categories. In addition, we find that the size of this elevation depends on higher order statistics and reduces with increasing phase noise. (3) Global modulations of contrast bias eye position toward high contrasts, consistent with a linear effect of contrast on fixation probability. This bias is independent of phase noise. (4) Small patches of locally decreased contrast repel eye position less than large patches of the same aggregate area, irrespective of phase noise. Our findings provide evidence that deviations from surrounding statistics, rather than contrast per se, underlie the well-established relation of contrast to fixation.