However, these models were not capable of actually achieving the brain's performance on its wide range of complex computations.

The mouse is one of the most studied animal models in the field of systems neuroscience. Understanding the generalized patterns and decoding the neural representations that are evoked by the diverse range of natural scene stimuli in the mouse visual cortex is one of the key quests in computational vision. In recent years, significant parallels have been drawn between the primate visual cortex and hierarchical deep neural networks. However, their generalized efficacy in understanding mouse vision has been limited. In this study, we investigate the functional alignment between the mouse visual cortex and deep learning models for object classification tasks. We first introduce a generalized representational learning strategy that uncovers a striking resemblance between the functional mapping of the mouse visual cortex and high-performing deep learning models on both top-down (population-level) and bottom-up (single cell-level) scenarios. Next, this representational similarity across the two systems is further enhanced by the addition of Neural Response Normalization (NeuRN) layer, inspired by the activation profile of excitatory and inhibitory neurons in the visual cortex. To test the performance effect of NeuRN on real-world tasks, we integrate it into deep learning models and observe significant improvements in their robustness against data shifts in domain generalization tasks. Our work proposes a novel framework for comparing the functional architecture of the mouse visual cortex with deep learning models. Our findings carry broad implications for the development of advanced AI models that draw inspiration from the mouse visual cortex, suggesting that these models serve as valuable tools for studying the neural representations of the mouse visual cortex and, as a result, enhancing their performance on real-world tasks.

The development of artificial intelligence towards real-time interaction with the environment is a key aspect of embodied intelligence and robotics. Inverse dynamics is a fundamental robotics problem, which maps from joint space to torque space of robotic systems. Traditional methods for solving it rely on direct physical modeling of robots which is difficult or even impossible due to nonlinearity and external disturbance. Recently, data-based model-learning algorithms are adopted to address this issue. However, they often require manual parameter tuning and high computational costs. Neuromorphic computing is inherently suitable to process spatiotemporal features in robot motion control at extremely low costs. However, current research is still in its infancy: existing works control only low-degree-of-freedom systems and lack performance quantification and comparison. In this paper, we propose a neuromorphic control framework to control 7 degree-of-freedom robotic manipulators. We use Spiking Neural Network to leverage the spatiotemporal continuity of the motion data to improve control accuracy, and eliminate manual parameters tuning. We validated the algorithm on two robotic platforms, which reduces torque prediction error by at least 60% and performs a target position tracking task successfully. This work advances embodied neuromorphic control by one step forward from proof of concept to applications in complex real-world tasks.

Understanding how biological neural networks are shaped via local plasticity mechanisms can lead to energy-efficient and self-adaptive information processing systems, which promises to mitigate some of the current roadblocks in edge computing systems. While biology makes use of spikes to seamless use both spike timing and mean firing rate to modulate synaptic strength, most models focus on one of the two. In this work, we present a Hebbian local learning rule that models synaptic modification as a function of calcium traces tracking neuronal activity. We show how the rule reproduces results from spike time and spike rate protocols from neuroscientific studies. Moreover, we use the model to train spiking neural networks on MNIST digit recognition to show and explain what sort of mechanisms are needed to learn real-world patterns. We show how our model is sensitive to correlated spiking activity and how this enables it to modulate the learning rate of the network without altering the mean firing rate of the neurons nor the hyparameters of the learning rule. To the best of our knowledge, this is the first work that showcases how spike timing and rate can be complementary in their role of shaping the connectivity of spiking neural networks.

In this brief and speculative commentary, we explore ideas inspired by neural networks in machine learning, proposing that a simple neural XOR motif, involving both excitatory and inhibitory connections, may provide the basis for a relevant mode of plasticity in neural circuits of living organisms, with homeostasis as the sole guiding principle. This XOR motif simply signals the discrepancy between incoming signals and reference signals, thereby providing a basis for a loss function in learning neural circuits, and at the same time regulating homeostasis by halting the propagation of these incoming signals. The core motif uses a 4:1 ratio of excitatory to inhibitory neurons, and supports broader neural patterns such as the well-known 'winner takes all' (WTA) mechanism. We examined the prevalence of the XOR motif in the published connectomes of various organisms with increasing complexity, and found that it ranges from tens (in C. elegans) to millions (in several Drosophila neuropils) and more than tens of millions (in mouse V1 visual cortex). If validated, our hypothesis identifies two of the three key components in analogy to machine learning models: the architecture and the loss function. And we propose that a relevant type of biological neural plasticity is simply driven by a basic control or regulatory system, which has persisted and adapted despite the increasing complexity of organisms throughout evolution.

This paper explores the integration of ring attractors, a mathematical model inspired by neural circuit dynamics, into the reinforcement learning (RL) action selection process. Ring attractors, as specialized brain-inspired structures that encode spatial information and uncertainty, offer a biologically plausible mechanism to improve learning speed and predictive performance. They do so by explicitly encoding the action space, facilitating the organization of neural activity, and enabling the distribution of spatial representations across the neural network in the context of deep RL. The application of ring attractors in the RL action selection process involves mapping actions to specific locations on the ring and decoding the selected action based on neural activity. We investigate the application of ring attractors by both building them as exogenous models and integrating them as part of a Deep Learning policy algorithm. Our results show a significant improvement in state-of-the-art models for the Atari 100k benchmark. Notably, our integrated approach improves the performance of state-of-the-art models by half, representing a 53% increase over selected baselines. This paper addresses the challenge of efficient action selection in reinforcement learning (RL), particularly in environments with spatial structures. Our primary contribution is the novel integration of ring attractors (Kim et al., 2017), a neural circuit model from neuroscience, into the RL framework. This approach improves spatial awareness in action selection and provides a mechanism for uncertainty-aware decision making in RL, leading to more accurate and efficient learning in complex environments.

Nowadays, neuromorphic systems based on Spiking Neural Networks (SNNs) attract attentions of many researchers. There are many studies to improve performances of neuromorphic systems. These studies have been showing satisfactory results. To magnify performances of neuromorphic systems, developing actual neuromorphic systems is essential. For developing them, memristors play key role due to their useful characteristics. Although memristors are essential for actual neuromorphic systems, they are vulnerable to faults. However, there are few studies analyzing effects of fault elements in neuromorphic systems using memristors. To solve this problem, we analyze performance of a memristive neuromorphic system with fault elements changing fault ratios, types, and positions. We choose neurons and synapses to inject faults. We inject two types of faults to synapses: SA0 and SA1 faults. The fault synapses appear in random and important positions. Through our analysis, we discover the following four interesting points. First, memristive characteristics increase vulnerability of neuromorphic systems to fault elements. Second, fault neuron ratios reducing performance sharply exist. Third, performance degradation by fault synapses depends on fault types. Finally, SA1 fault synapses improve performance when they appear in important positions.

Remembering and forgetting mechanisms are two sides of the same coin in a human learning-memory system. Inspired by human brain memory mechanisms, modern machine learning systems have been working to endow machine with lifelong learning capability through better remembering while pushing the forgetting as the antagonist to overcome. Nevertheless, this idea might only see the half picture. Up until very recently, increasing researchers argue that a brain is born to forget, i.e., forgetting is a natural and active process for abstract, rich, and flexible representations. This paper presents a learning model by active forgetting mechanism with artificial neural networks. The active forgetting mechanism (AFM) is introduced to a neural network via a "plug-and-play" forgetting layer (P\&PF), consisting of groups of inhibitory neurons with Internal Regulation Strategy (IRS) to adjust the extinction rate of themselves via lateral inhibition mechanism and External Regulation Strategy (ERS) to adjust the extinction rate of excitatory neurons via inhibition mechanism. Experimental studies have shown that the P\&PF offers surprising benefits: self-adaptive structure, strong generalization, long-term learning and memory, and robustness to data and parameter perturbation. This work sheds light on the importance of forgetting in the learning process and offers new perspectives to understand the underlying mechanisms of neural networks.

Spiking Neural Networks (SNNs) bear the potential of efficient unsupervised and continual learning capabilities because of their biological plausibility, but their complexity still poses a serious research challenge to enable their energy-efficient design for resource-constrained scenarios (like embedded systems, IoT-Edge, etc.). We propose SpikeDyn, a comprehensive framework for energy-efficient SNNs with continual and unsupervised learning capabilities in dynamic environments, for both the training and inference phases. It is achieved through the following multiple diverse mechanisms: 1) reduction of neuronal operations, by replacing the inhibitory neurons with direct lateral inhibitions; 2) a memory- and energy-constrained SNN model search algorithm that employs analytical models to estimate the memory footprint and energy consumption of different candidate SNN models and selects a Pareto-optimal SNN model; and 3) a lightweight continual and unsupervised learning algorithm that employs adaptive learning rates, adaptive membrane threshold potential, weight decay, and reduction of spurious updates. Our experimental results show that, for a network with 400 excitatory neurons, our SpikeDyn reduces the energy consumption on average by 51% for training and by 37% for inference, as compared to the state-of-the-art. Due to the improved learning algorithm, SpikeDyn provides on avg. 21% accuracy improvement over the state-of-the-art, for classifying the most recently learned task, and by 8% on average for the previously learned tasks.

Cerebellar nuclei, substructures of the cerebellum, transfer information from the cerebellum to other parts of the brain. Using single-cell transcriptomics, Kebschull et al. have now identified a conserved pattern of cerebellar nuclei structure that has been repeated through evolution (see the Perspective by Hatten). Ranging from mice to chickens to humans, cerebellar nuclei are made up of region-specific excitatory neurons and region-invariant inhibitory neurons. In humans, a facet connecting the cerebellum to the frontal cortex is enhanced. Science , this issue p. [eabd5059][1]; see also p. [1411][2] ### INTRODUCTION The brains of extant animals have evolved over hundreds of millions of years from simple circuits. Cell types diversified, connections elaborated, and new brain regions emerged. Models for brain region evolution range from duplication of existing regions to splitting of previously multifunctional regions and de novo assembly from existing cell types. These models, however, have not been demonstrated in vertebrate brains at cell-type resolution. ### RATIONALE We investigated brain region evolution using the cerebellar nuclei as a model system. The cerebellum is a major hindbrain structure in jawed vertebrates, comprising the cerebellar cortex and cerebellar nuclei. It is thought to act as a feedforward model for motor control and cognitive processes. The cerebellar cortex receives and processes inputs and sends outputs to the cerebellar nuclei, which route the results of cerebellar computations to the rest of the brain. Whereas the cerebellar cortex is well conserved across vertebrates, the cerebellar nuclei vary in number, with none in jawless vertebrates, one pair in cartilaginous fishes and amphibians, two pairs in reptiles and birds, and three pairs in mammals. This pattern suggests that extant cerebellar nuclei evolved from a single ancestral nucleus. Cerebellar nuclei thus provide a good model to interrogate brain region evolution. ### RESULTS We characterized the cerebellar nuclei in mice, chickens, and humans using whole-brain and spinal cord projection mapping in cleared samples, single-nucleus RNA sequencing (snRNAseq), and spatially resolved transcript amplicon readout mapping (STARmap) analysis. We first compared the projection patterns of the three cerebellar nuclei of mice. Our data reveal broad projections of all nuclei, which in common target regions are shifted relative to each other. To understand the transcriptomic differences that underlie these shifting projections, we produced a cell-type atlas of the mouse cerebellar nuclei using snRNAseq. We discovered three region-invariant inhibitory cell classes and 15 region-specific excitatory cell types. Excitatory cell types fall into two classes with distinct gene expression and electrophysiological properties. Members of each class are present in every nucleus and are putative sister cell types. STARmap analysis in mice revealed that the organizational unit of the cerebellar nuclei is cytoarchitectonically distinguishable subnuclei, each of which contains the three inhibitory and two excitatory classes. To test whether this archetypal subnucleus is also the evolutionary unit of the cerebellar nuclei, we performed snRNAseq and STARmap on the chicken cerebellar nuclei. We identified four subnuclei, three of which had direct orthologs in mice. Each chicken subnucleus contained the same cell-type set of three inhibitory and two excitatory classes already identified in mice, confirming our hypothesis. Cerebellar nuclei vary in size across vertebrates. In particular, the human lateral nucleus is markedly expanded. To understand this expansion, we performed snRNAseq in humans. We found that the medial and interposed nuclei maintained the archetypal cerebellar nuclei composition. However, the lateral nucleus expanded one excitatory cell class at the expense of the other. Conditional tracing in the mouse lateral nucleus revealed that the cell class expanded in humans preferentially accesses lateral frontal cortices via specific intermediate thalamic nuclei. ### CONCLUSION We identified a conserved cell-type set that forms an archetypal cerebellar nucleus as the unit of cerebellar nuclei organization and evolution. We propose that this archetypal nucleus was repeatedly duplicated during evolution, accompanied primarily by transcriptomic divergence of excitatory neurons and shifts in their projection patterns. Our data support a model of duplication-and-divergence of entire cell-type sets for brain region evolution. ![Figure][3] Evolution of the cerebellar nuclei. Comparative single-cell transcriptomics in mice, chickens, and humans (top left; neurons are color-coded by type), spatial transcriptomic analyses in mice and chickens (top right; neurons are color-coded by type in raw and processed data), and central nervous system (CNS)–wide projection mapping in mice (bottom left; axons in red in a three-dimensional mouse brain) revealed the unit of cerebellar nuclei organization and evolution. This unit (red box) comprises three inhibitory and two excitatory neuron classes (each colored circle indicates a neuron class). Extant cerebellar nuclei likely derived from the duplication and divergence of this unit, with more dynamic gene expression in excitatory neurons (changed color hues), along with projection target shifts. How have complex brains evolved from simple circuits? Here we investigated brain region evolution at cell-type resolution in the cerebellar nuclei, the output structures of the cerebellum. Using single-nucleus RNA sequencing in mice, chickens, and humans, as well as STARmap spatial transcriptomic analysis and whole–central nervous system projection tracing, we identified a conserved cell-type set containing two region-specific excitatory neuron classes and three region-invariant inhibitory neuron classes. This set constitutes an archetypal cerebellar nucleus that was repeatedly duplicated to form new regions. The excitatory cell class that preferentially funnels information to lateral frontal cortices in mice becomes predominant in the massively expanded human lateral nucleus. Our data suggest a model of brain region evolution by duplication and divergence of entire cell-type sets. [1]: /lookup/doi/10.1126/science.abd5059 [2]: /lookup/doi/10.1126/science.abf4483 [3]: pending:yes