50 Years Since the Marr, Ito, and Albus Models of the Cerebellum

Chaotic recurrent neural networks for brain modelling: A review

Even in the absence of external stimuli, the brain is spontaneously active. Indeed, most cortical activity is internally generated by recurrence. Both theoretical and experimental studies suggest that chaotic dynamics characterize this spontaneous activity. While the precise function of brain chaotic activity is still puzzling, we know that chaos confers many advantages. From a computational perspective, chaos enhances the complexity of network dynamics. From a behavioural point of view, chaotic activity could generate the variability required for exploration. Furthermore, information storage and transfer are maximized at the critical border between order and chaos. Despite these benefits, many computational brain models avoid incorporating spontaneous chaotic activity due to the challenges it poses for learning algorithms. In recent years, however, multiple approaches have been proposed to overcome this limitation. As a result, many different algorithms have been developed, initially within the reservoir computing paradigm. Over time, the field has evolved to increase the biological plausibility and performance of the algorithms, sometimes going beyond the reservoir computing framework. In this review article, we examine the computational benefits of chaos and the unique properties of chaotic recurrent neural networks, with a particular focus on those typically utilized in reservoir computing. We also provide a detailed analysis of the algorithms designed to train chaotic RNNs, tracing their historical evolution and highlighting key milestones in their development. Finally, we explore the applications and limitations of chaotic RNNs for brain modelling, consider their potential broader impacts beyond neuroscience, and outline promising directions for future research.

Predictive reward-prediction errors of climbing fiber inputs integrate modular reinforcement learning with supervised learning

Although the cerebellum is typically associated with supervised learning algorithms, it also exhibits extensive involvement in reward processing. In this study, we investigated the cerebellum's role in executing reinforcement learning algorithms, with a particular emphasis on essential reward-prediction errors. We employed the Q-learning model to accurately reproduce the licking responses of mice in a Go/No-go auditory-discrimination task. This method enabled the calculation of reinforcement learning variables, such as reward, predicted reward, and reward-prediction errors in each learning trial. Through tensor component analysis of two-photon Ca2+ imaging data from more than 6,000 Purkinje cells, we found that climbing fiber inputs of the two distinct components, which were specifically activated during Go and No-go cues in the learning process, showed an inverse relationship with predictive reward-prediction errors. Assuming bidirectional parallel-fiber Purkinje-cell synaptic plasticity, we constructed a cerebellar neural-network model with 5,000 spiking neurons of granule cells, Purkinje cells, cerebellar nuclei neurons, and inferior olive neurons. The network model qualitatively reproduced distinct changes in licking behaviors, climbing-fiber firing rates, and their synchronization during discrimination learning separately for Go/No-go conditions. We found that Purkinje cells in the two components could develop specific motor commands for their respective auditory cues, guided by the predictive reward-prediction errors from their climbing fiber inputs. These results indicate a possible role of context-specific actors in modular reinforcement learning, integrating with cerebellar supervised learning capabilities.

Converging and Diverging Cerebellar Pathways for Motor and Social Behaviors in Mice

Evidence from clinical and preclinical studies has shown that the cerebellum contributes to cognitive functions, including social behaviors. Now that the cerebellum's role in a wider range of behaviors has been confirmed, the question arises whether the cerebellum contributes to social behaviors via the same mechanisms with which it modulates movements. This review seeks to answer whether the cerebellum guides motor and social behaviors through identical pathways. It focuses on studies in which cerebellar cells, synapses, or genes are manipulated in a cell-type specific manner followed by testing of the effects on social and motor behaviors. These studies show that both anatomically restricted and cerebellar cortex-wide manipulations can lead to social impairments without abnormal motor control, and vice versa. These studies suggest that the cerebellum employs different cellular, synaptic, and molecular pathways for social and motor behaviors. Future studies warrant a focus on the diverging mechanisms by which the cerebellum contributes to a wide range of neural functions.

A computational study of how an α- to γ-motoneurone collateral can mitigate velocity-dependent stretch reflexes during voluntary movement

The primary motor cortex does not uniquely or directly produce alpha motoneurone (α-MN) drive to muscles during voluntary movement. Rather, α-MN drive emerges from the synthesis and competition among excitatory and inhibitory inputs from multiple descending tracts, spinal interneurons, sensory inputs, and proprioceptive afferents. One such fundamental input is velocity-dependent stretch reflexes in lengthening muscles, which should be inhibited to enable voluntary movement. It remains an open question, however, the extent to which unmodulated stretch reflexes disrupt voluntary movement, and whether and how they are inhibited in limbs with numerous multiarticular muscles. We used a computational model of a Rhesus Macaque arm to simulate movements with feedforward α-MN commands only, and with added velocity-dependent stretch reflex feedback. We found that velocity-dependent stretch reflex caused movement-specific, typically large and variable disruptions to arm movements. These disruptions were greatly reduced when modulating velocity-dependent stretch reflex feedback (i) as per the commonly proposed (but yet to be clarified) idealized alpha-gamma (α-γ) coactivation or (ii) an alternative α-MN collateral projection to homonymous γ-MNs. We conclude that such α-MN collaterals are a physiologically tenable propriospinal circuit in the mammalian fusimotor system. These collaterals could still collaborate with α-γ coactivation, and the few skeletofusimotor fibers (β-MNs) in mammals, to create a flexible fusimotor ecosystem to enable voluntary movement. By locally and automatically regulating the highly nonlinear neuro-musculo-skeletal mechanics of the limb, these collaterals could be a critical low-level enabler of learning, adaptation, and performance via higher-level brainstem, cerebellar, and cortical mechanisms.

Cerebellar Functions Beyond Movement and Learning

The cerebellum has a well-established role in controlling motor functions, including coordination, posture, and the learning of skilled movements. The mechanisms for how it carries out motor behavior remain under intense investigation. Interestingly though, in recent years the mechanisms of cerebellar function have faced additional scrutiny since nonmotor behaviors may also be controlled by the cerebellum. With such complexity arising, there is now a pressing need to better understand how cerebellar structure, function, and behavior intersect to influence behaviors that are dynamically called upon as an animal experiences its environment. Here, we discuss recent experimental work that frames possible neural mechanisms for how the cerebellum shapes disparate behaviors and why its dysfunction is catastrophic in hereditary and acquired conditions-both motor and nonmotor. For these reasons, the cerebellum might be the ideal therapeutic target.

An alpha- to gamma-motoneurone collateral can mitigate velocity-dependent stretch reflexes during voluntary movement: A computational study

The primary motor cortex does not uniquely or directly produce alpha motoneurone (α-MN) drive to muscles during voluntary movement. Rather, α-MN drive emerges from the synthesis and competition among excitatory and inhibitory inputs from multiple descending tracts, spinal interneurons, sensory inputs, and proprioceptive afferents. One such fundamental input is velocity-dependent stretch reflexes in lengthening muscles, which should be inhibited to enable voluntary movement. It remains an open question, however, the extent to which unmodulated stretch reflexes disrupt voluntary movement, and whether and how they are inhibited in limbs with numerous multi-articular muscles. We used a computational model of a Rhesus Macaque arm to simulate movements with feedforward α-MN commands only, and with added velocity-dependent stretch reflex feedback. We found that velocity-dependent stretch reflex caused movement-specific, typically large and variable disruptions to arm movements. These disruptions were greatly reduced when modulating velocity-dependent stretch reflex feedback (i) as per the commonly proposed (but yet to be clarified) idealized alpha-gamma (α-γ) co-activation or (ii) an alternative α-MN collateral projection to homonymous γ-MNs. We conclude that such α-MN collaterals are a physiologically tenable, but previously unrecognized, propriospinal circuit in the mammalian fusimotor system. These collaterals could still collaborate with α-γ co-activation, and the few skeletofusimotor fibers (β-MNs) in mammals, to create a flexible fusimotor ecosystem to enable voluntary movement. By locally and automatically regulating the highly nonlinear neuro-musculo-skeletal mechanics of the limb, these collaterals could be a critical low-level enabler of learning, adaptation, and performance via higher-level brainstem, cerebellar and cortical mechanisms.

Sensory encoding and memory in the mushroom body: signals, noise, and variability

To survive in changing environments, animals need to learn to associate specific sensory stimuli with positive or negative valence. How do they form stimulus-specific memories to distinguish between positively/negatively associated stimuli and other irrelevant stimuli? Solving this task is one of the functions of the mushroom body, the associative memory center in insect brains. Here we summarize recent work on sensory encoding and memory in the Drosophila mushroom body, highlighting general principles such as pattern separation, sparse coding, noise and variability, coincidence detection, and spatially localized neuromodulation, and placing the mushroom body in comparative perspective with mammalian memory systems.

The cognitive reality monitoring network and theories of consciousness

Theories of consciousness abound. However, it is difficult to arbitrate reliably among competing theories because they target different levels of neural and cognitive processing or anatomical loci, and only some were developed with computational models in mind. In particular, theories of consciousness need to fully address the three levels of understanding of the brain proposed by David Marr: computational theory, algorithms and hardware. Most major theories refer to only one or two levels, often indirectly. The cognitive reality monitoring network (CRMN) model is derived from computational theories of mixture-of-experts architecture, hierarchical reinforcement learning and generative/inference computing modules, addressing all three levels of understanding. A central feature of the CRMN is the mapping of a gating network onto the prefrontal cortex, making it a prime coding circuit involved in monitoring the accuracy of one's mental states and distinguishing them from external reality. Because the CRMN builds on the hierarchical and layer structure of the cerebral cortex, it may connect research and findings across species, further enabling concrete computational models of consciousness with new, explicitly testable hypotheses. In sum, we discuss how the CRMN model can help further our understanding of the nature and function of consciousness.

The Cerebellum as an Embodying Machine

Whereas emotion theorists often keep their distance from the embodied approach, theorists of embodiment tend to treat emotion as a mainly physiologic process. However, intimate links between emotions and the body suggest that emotions are privileged phenomena to attempt to reintegrate mind and body and that the body helps the mind in shaping emotional responses. To date, research has favored the cerebrum over other parts of the brain as a substrate of embodied emotions. However, given the widely demonstrated contribution of the cerebellum to emotional processing, research in affective neuroscience should consider embodiment theory as a useful approach for evaluating the cerebellar role in emotion and affect. The aim of this review is to insert the cerebellum among the structures needed to embody emotions, providing illustrative examples of cerebellar involvement in embodied emotions (as occurring in empathic abilities) and in impaired identification and expression of embodied emotions (as occurring in alexithymia).

Emergence of syntax and word prediction in an artificial neural circuit of the cerebellum

The cerebellum, interconnected with the cerebral neocortex, plays a vital role in human-characteristic cognition such as language processing, however, knowledge about the underlying circuit computation of the cerebellum remains very limited. To gain a better understanding of the computation underlying cerebellar language processing, we developed a biologically constrained cerebellar artificial neural network (cANN) model, which implements the recently identified cerebello-cerebellar recurrent pathway. We found that while cANN acquires prediction of future words, another function of syntactic recognition emerges in the middle layer of the prediction circuit. The recurrent pathway of the cANN was essential for the two language functions, whereas cANN variants with further biological constraints preserved these functions. Considering the uniform structure of cerebellar circuitry across all functional domains, the single-circuit computation, which is the common basis of the two language functions, can be generalized to fundamental cerebellar functions of prediction and grammar-like rule extraction from sequences, that underpin a wide range of cerebellar motor and cognitive functions. This is a pioneering study to understand the circuit computation of human-characteristic cognition using biologically-constrained ANNs.

Human Purkinje cells outperform mouse Purkinje cells in dendritic complexity and computational capacity

Purkinje cells in the cerebellum are among the largest neurons in the brain and have been extensively investigated in rodents. However, their morphological and physiological properties remain poorly understood in humans. In this study, we utilized high-resolution morphological reconstructions and unique electrophysiological recordings of human Purkinje cells ex vivo to generate computational models and estimate computational capacity. An inter-species comparison showed that human Purkinje cell had similar fractal structures but were larger than those of mouse Purkinje cells. Consequently, given a similar spine density (2/μm), human Purkinje cell hosted approximately 7.5 times more dendritic spines than those of mice. Moreover, human Purkinje cells had a higher dendritic complexity than mouse Purkinje cells and usually emitted 2-3 main dendritic trunks instead of one. Intrinsic electro-responsiveness was similar between the two species, but model simulations revealed that the dendrites could process ~6.5 times (n = 51 vs. n = 8) more input patterns in human Purkinje cells than in mouse Purkinje cells. Thus, while human Purkinje cells maintained spike discharge properties similar to those of rodents during evolution, they developed more complex dendrites, enhancing computational capacity.

The spinal cord facilitates cerebellar upper limb motor learning and control; inputs from neuromusculoskeletal simulation

Complex interactions between brain regions and the spinal cord (SC) govern body motion, which is ultimately driven by muscle activation. Motor planning or learning are mainly conducted at higher brain regions, whilst the SC acts as a brain-muscle gateway and as a motor control centre providing fast reflexes and muscle activity regulation. Thus, higher brain areas need to cope with the SC as an inherent and evolutionary older part of the body dynamics. Here, we address the question of how SC dynamics affects motor learning within the cerebellum; in particular, does the SC facilitate cerebellar motor learning or constitute a biological constraint? We provide an exploratory framework by integrating biologically plausible cerebellar and SC computational models in a musculoskeletal upper limb control loop. The cerebellar model, equipped with the main form of cerebellar plasticity, provides motor adaptation; whilst the SC model implements stretch reflex and reciprocal inhibition between antagonist muscles. The resulting spino-cerebellar model is tested performing a set of upper limb motor tasks, including external perturbation studies. A cerebellar model, lacking the implemented SC model and directly controlling the simulated muscles, was also tested in the same. The performances of the spino-cerebellar and cerebellar models were then compared, thus allowing directly addressing the SC influence on cerebellar motor adaptation and learning, and on handling external motor perturbations. Performance was assessed in both joint and muscle space, and compared with kinematic and EMG recordings from healthy participants. The differences in cerebellar synaptic adaptation between both models were also studied. We conclude that the SC facilitates cerebellar motor learning; when the SC circuits are in the loop, faster convergence in motor learning is achieved with simpler cerebellar synaptic weight distributions. The SC is also found to improve robustness against external perturbations, by better reproducing and modulating muscle cocontraction patterns.

A Comparative Perspective on the Cerebello-Cerebral System and Its Link to Cognition

The longstanding idea that the cerebral cortex is the main neural correlate of human cognition can be elaborated by comparative analyses along the vertebrate phylogenetic tree that support the view that the cerebello-cerebral system is suited to support non-motor functions more generally. In humans, diverse accounts have illustrated cerebellar involvement in cognitive functions. Although the neocortex, and its transmodal association cortices such as the prefrontal cortex, have become disproportionately large over primate evolution specifically, human neocortical volume does not appear to be exceptional relative to the variability within primates. Rather, several lines of evidence indicate that the exceptional volumetric increase of the lateral cerebellum in conjunction with its connectivity with the cerebral cortical system may be linked to non-motor functions and mental operation in primates. This idea is supported by diverging cerebello-cerebral adaptations that potentially coevolve with cognitive abilities across other vertebrates such as dolphins, parrots, and elephants. Modular adaptations upon the vertebrate cerebello-cerebral system may thus help better understand the neuroevolutionary trajectory of the primate brain and its relation to cognition in humans. Lateral cerebellar lobules crura I-II and their reciprocal connections to the cerebral cortical association areas appear to have substantially expanded in great apes, and humans. This, along with the notable increase in the ventral portions of the dentate nucleus and a shift to increased relative prefrontal-cerebellar connectivity, suggests that modular cerebellar adaptations support cognitive functions in humans. In sum, we show how comparative neuroscience provides new avenues to broaden our understanding of cerebellar and cerebello-cerebral functions in the context of cognition.

Cerebellar circuits for disinhibition and synchronous inhibition

The cerebellar cortex contributes to diverse behaviors by transforming mossy fiber inputs into predictions in the form of Purkinje cell (PC) outputs, and then refining those predictions1. Molecular layer interneurons (MLIs) account for approximately 80% of the inhibitory interneurons in the cerebellar cortex2, and are vital to cerebellar processing1,3. MLIs are thought to primarily inhibit PCs and suppress the plasticity of excitatory synapses onto PCs. MLIs also inhibit, and are electrically coupled to, other MLIs4-7, but the functional significance of these connections is not known1,3. Behavioral studies suggest that cerebellar-dependent learning is gated by disinhibition of PCs, but the source of such disinhibition has not been identified8. Here we find that two recently recognized MLI subtypes2, MLI1 and MLI2, have highly specialized connectivity that allows them to serve very different functional roles. MLI1s primarily inhibit PCs, are electrically coupled to each other, fire synchronously with other MLI1s on the millisecond time scale in vivo, and synchronously pause PC firing. MLI2s are not electrically coupled, they primarily inhibit MLI1s and disinhibit PCs, and are well suited to gating cerebellar-dependent learning8. These findings require a major reevaluation of processing within the cerebellum in which disinhibition, a powerful circuit motif present in the cerebral cortex and elsewhere9-17, greatly increases the computational power and flexibility of the cerebellum. They also suggest that millisecond time scale synchronous firing of electrically-coupled MLI1s helps regulate the output of the cerebellar cortex by synchronously pausing PC firing, which has been shown to evoke precisely-timed firing in PC targets18.

Dynamic organization of cerebellar climbing fiber response and synchrony in multiple functional components reduces dimensions for reinforcement learning

Cerebellar climbing fibers convey diverse signals, but how they are organized in the compartmental structure of the cerebellar cortex during learning remains largely unclear. We analyzed a large amount of coordinate-localized two-photon imaging data from cerebellar Crus II in mice undergoing 'Go/No-go' reinforcement learning. Tensor component analysis revealed that a majority of climbing fiber inputs to Purkinje cells were reduced to only four functional components, corresponding to accurate timing control of motor initiation related to a Go cue, cognitive error-based learning, reward processing, and inhibition of erroneous behaviors after a No-go cue. Changes in neural activities during learning of the first two components were correlated with corresponding changes in timing control and error learning across animals, indirectly suggesting causal relationships. Spatial distribution of these components coincided well with boundaries of Aldolase-C/zebrin II expression in Purkinje cells, whereas several components are mixed in single neurons. Synchronization within individual components was bidirectionally regulated according to specific task contexts and learning stages. These findings suggest that, in close collaborations with other brain regions including the inferior olive nucleus, the cerebellum, based on anatomical compartments, reduces dimensions of the learning space by dynamically organizing multiple functional components, a feature that may inspire new-generation AI designs.

Low Vision Impairs Implicit Sensorimotor Adaptation in Response to Small Errors, But Not Large Errors

Successful goal-directed actions require constant fine-tuning of the motor system. This fine-tuning is thought to rely on an implicit adaptation process that is driven by sensory prediction errors (e.g., where you see your hand after reaching vs. where you expected it to be). Individuals with low vision experience challenges with visuomotor control, but whether low vision disrupts motor adaptation is unknown. To explore this question, we assessed individuals with low vision and matched controls with normal vision on a visuomotor task designed to isolate implicit adaptation. We found that low vision was associated with attenuated implicit adaptation only for small visual errors, but not for large visual errors. This result highlights important constraints underlying how low-fidelity visual information is processed by the sensorimotor system to enable successful implicit adaptation.

Electrophysiological activity from over the cerebellum and cerebrum during eye blink conditioning in human subjects

We report the results of an experiment in which electrophysiological activity was recorded from the human cerebellum and cerebrum in a sample of 14 healthy subjects before, during and after a classical eye blink conditioning procedure with an auditory tone as conditional stimulus and a maxillary nerve unconditional stimulus. The primary aim was to show changes in the cerebellum and cerebrum correlated with behavioral ocular responses. Electrodes recorded EMG and EOG at peri-ocular sites, EEG from over the frontal eye-fields and the electrocerebellogram (ECeG) from over the posterior fossa. Of the 14 subjects half strongly conditioned while the other half were resistant. We confirmed that conditionability was linked under our conditions to the personality dimension of extraversion-introversion. Inhibition of cerebellar activity was shown prior to the conditioned response, as predicted by Albus (1971). However, pausing in high frequency ECeG and the appearance of a contingent negative variation (CNV) in both central leads occurred in all subjects. These led us to conclude that while conditioned cerebellar pausing may be necessary, it is not sufficient alone to produce overt behavioral conditioning, implying the existence of another central mechanism. The outcomes of this experiment indicate the potential value of the noninvasive electrophysiology of the cerebellum.

Cerebro-cerebellar networks facilitate learning through feedback decoupling

Behavioural feedback is critical for learning in the cerebral cortex. However, such feedback is often not readily available. How the cerebral cortex learns efficiently despite the sparse nature of feedback remains unclear. Inspired by recent deep learning algorithms, we introduce a systems-level computational model of cerebro-cerebellar interactions. In this model a cerebral recurrent network receives feedback predictions from a cerebellar network, thereby decoupling learning in cerebral networks from future feedback. When trained in a simple sensorimotor task the model shows faster learning and reduced dysmetria-like behaviours, in line with the widely observed functional impact of the cerebellum. Next, we demonstrate that these results generalise to more complex motor and cognitive tasks. Finally, the model makes several experimentally testable predictions regarding cerebro-cerebellar task-specific representations over learning, task-specific benefits of cerebellar predictions and the differential impact of cerebellar and inferior olive lesions. Overall, our work offers a theoretical framework of cerebro-cerebellar networks as feedback decoupling machines.

The molecular diversity of plasticity mechanisms underlying memory: An evolutionary perspective

Experience triggers molecular cascades in organisms (learning) that lead to alterations (memory) to allow the organism to change its behavior based on experience. Understanding the molecular mechanisms underlying memory, particularly in the nervous system of animals, has been an exciting scientific challenge for neuroscience. We review what is known about forms of neuronal plasticity that underlie memory highlighting important issues in the field: (1) the importance of being able to measure how neurons are activated during learning to identify the form of plasticity that underlies memory, (2) the many distinct forms of plasticity important for memories that naturally decay both within and between organisms, and (3) unifying principles underlying the formation and maintenance of long-term memories. Overall, the diversity of molecular mechanisms underlying memories that naturally decay contrasts with more unified molecular mechanisms implicated in long-lasting changes. Despite many advances, important questions remain as to which mechanisms of neuronal plasticity underlie memory.

A cerebellum inspired spiking neural network as a multi-model for pattern classification and robotic trajectory prediction

Spiking neural networks were introduced to understand spatiotemporal information processing in neurons and have found their application in pattern encoding, data discrimination, and classification. Bioinspired network architectures are considered for event-driven tasks, and scientists have looked at different theories based on the architecture and functioning. Motor tasks, for example, have networks inspired by cerebellar architecture where the granular layer recodes sparse representations of the mossy fiber (MF) inputs and has more roles in motor learning. Using abstractions from cerebellar connections and learning rules of deep learning network (DLN), patterns were discriminated within datasets, and the same algorithm was used for trajectory optimization. In the current work, a cerebellum-inspired spiking neural network with dynamics of cerebellar neurons and learning mechanisms attributed to the granular layer, Purkinje cell (PC) layer, and cerebellar nuclei interconnected by excitatory and inhibitory synapses was implemented. The model’s pattern discrimination capability was tested for two tasks on standard machine learning (ML) datasets and on following a trajectory of a low-cost sensor-free robotic articulator. Tuned for supervised learning, the pattern classification capability of the cerebellum-inspired network algorithm has produced more generalized models than data-specific precision models on smaller training datasets. The model showed an accuracy of 72%, which was comparable to standard ML algorithms, such as MLP (78%), Dl4jMlpClassifier (64%), RBFNetwork (71.4%), and libSVM-linear (85.7%). The cerebellar model increased the network’s capability and decreased storage, augmenting faster computations. Additionally, the network model could also implicitly reconstruct the trajectory of a 6-degree of freedom (DOF) robotic arm with a low error rate by reconstructing the kinematic parameters. The variability between the actual and predicted trajectory points was noted to be ± 3 cm (while moving to a position in a cuboid space of 25 × 30 × 40 cm). Although a few known learning rules were implemented among known types of plasticity in the cerebellum, the network model showed a generalized processing capability for a range of signals, modulating the data through the interconnected neural populations. In addition to potential use on sensor-free or feed-forward based controllers for robotic arms and as a generalized pattern classification algorithm, this model adds implications to motor learning theory.

Specific Dystrophins Selectively Associate with Inhibitory and Excitatory Synapses of the Mouse Cerebellum and their Loss Alters Expression of P2X7 Purinoceptors and Pro-Inflammatory Mediators

Duchenne muscular dystrophy (DMD) patients, having mutations of the DMD gene, present with a range of neuropsychiatric disorders, in addition to the quintessential muscle pathology. The neurobiological basis remains poorly understood because the contributions of different DMD gene products (dystrophins) to the different neural networks underlying such symptoms are yet to be fully characterised. While full-length dystrophin clusters in inhibitory synapses, with inhibitory neurotransmitter receptors, the precise subcellular expression of truncated DMD gene products with excitatory synapses remains unresolved. Furthermore, inflammation, involving P2X purinoceptor 7 (P2RX7) accompanies DMD muscle pathology, yet any association with brain dystrophins is yet to be established. The aim of this study was to investigate the comparative expression of different dystrophins, alongside ionotropic glutamate receptors and P2RX7s, within the cerebellar circuitry known to express different dystrophin isoforms. Immunoreactivity for truncated DMD gene products was targeted to Purkinje cell (PC) distal dendrites adjacent to, or overlapping with, signal for GluA1, GluA4, GluN2A, and GluD2 receptor subunits. P2X7R immunoreactivity was located in Bergmann glia profiles adjacent to PC-dystrophin immunoreactivity. Ablation of all DMD gene products coincided with decreased mRNA expression for Gria2, Gria3, and Grin2a and increased GluD2 immunoreactivity. Finally, dystrophin-null mice showed decreased brain mRNA expression of P2rx7 and several inflammatory mediators. The data suggest that PCs target different dystrophin isoforms to molecularly and functionally distinct populations of synapses. In contrast to muscle, dystrophinopathy in brain leads to the dampening of the local immune system.

In search of a naturalistic neuroimaging approach: Exploration of general feasibility through the case of VR-fMRI and application in the domain of episodic memory

Virtual Reality (VR) is an increasingly widespread tool for research as it allows the creation of experiments taking place in multimodal and daily-life-like environments, while keeping a strong experimental control. Adding neuroimaging to VR leads to a better understanding of the underlying brain networks activated during a naturalistic task, whether for research purposes or rehabilitation. The present paper focuses on the specific use of concurrent VR and fMRI and its technical challenges and feasibility, with a brief examination of the general existing solutions. Following the PRISMA guidelines, the review investigates the particular case of how VR-fMRI has explored episodic memory so far, with a comparison of object- and place-based episodic memory. This review confirms the involvement of cerebral regions well-known to be implicated in episodic memory and unravels other regions devoted to bodily and narrative aspects of the self, promoting new avenues of research in the domain of naturalistic episodic memory. Future studies should develop more immersive and interactive virtual neuroimaging features to increase ecological and embodied neurocognition aspects.

The neural mechanisms of manual dexterity

The hand endows us with unparalleled precision and versatility in our interactions with objects, from mundane activities such as grasping to extraordinary ones such as virtuoso pianism. The complex anatomy of the human hand combined with expansive and specialized neuronal control circuits allows a wide range of precise manual behaviours. To support these behaviours, an exquisite sensory apparatus, spanning the modalities of touch and proprioception, conveys detailed and timely information about our interactions with objects and about the objects themselves. The study of manual dexterity provides a unique lens into the sensorimotor mechanisms that endow the nervous system with the ability to flexibly generate complex behaviour.

From internal models toward metacognitive AI

In several papers published in Biological Cybernetics in the 1980s and 1990s, Kawato and colleagues proposed computational models explaining how internal models are acquired in the cerebellum. These models were later supported by neurophysiological experiments using monkeys and neuroimaging experiments involving humans. These early studies influenced neuroscience from basic, sensory-motor control to higher cognitive functions. One of the most perplexing enigmas related to internal models is to understand the neural mechanisms that enable animals to learn large-dimensional problems with so few trials. Consciousness and metacognition-the ability to monitor one's own thoughts, may be part of the solution to this enigma. Based on literature reviews of the past 20 years, here we propose a computational neuroscience model of metacognition. The model comprises a modular hierarchical reinforcement-learning architecture of parallel and layered, generative-inverse model pairs. In the prefrontal cortex, a distributed executive network called the "cognitive reality monitoring network" (CRMN) orchestrates conscious involvement of generative-inverse model pairs in perception and action. Based on mismatches between computations by generative and inverse models, as well as reward prediction errors, CRMN computes a "responsibility signal" that gates selection and learning of pairs in perception, action, and reinforcement learning. A high responsibility signal is given to the pairs that best capture the external world, that are competent in movements (small mismatch), and that are capable of reinforcement learning (small reward-prediction error). CRMN selects pairs with higher responsibility signals as objects of metacognition, and consciousness is determined by the entropy of responsibility signals across all pairs. This model could lead to new-generation AI, which exhibits metacognition, consciousness, dimension reduction, selection of modules and corresponding representations, and learning from small samples. It may also lead to the development of a new scientific paradigm that enables the causal study of consciousness by combining CRMN and decoded neurofeedback.

Projection-dependent heterogeneity of cerebellar granule cell calcium responses

Cerebellar granule cells (GCs) relay mossy fiber (MF) inputs to Purkinje cell dendrites via their axons, the parallel fibers (PFs), which are individually located at a given sublayer of the molecular layer (ML). Although a certain degree of heterogeneity among GCs has been recently reported, variability of GC responses to MF inputs has never been associated with their most notable structural variability, location of their projecting PFs in the ML. Here, we utilize an adeno-associated virus (AAV)-mediated labeling technique that enables us to categorize GCs according to the location of their PFs, and compare the Ca²⁺ responses to MF stimulations between three groups of GCs, consisting of either GCs having PFs at the deep (D-GCs), middle (M-GCs), or superficial (S-GCs) sublayer. Our structural analysis revealed that there was no correlation between position of GC soma in the GC layer and location of its PF in the ML, confirming that our AAV-mediated labeling was important to test the projection-dependent variability of the Ca²⁺ responses in GCs. We then found that the Ca²⁺ responses of D-GCs differed from those of M-GCs. Pharmacological experiments implied that the different Ca²⁺ responses were mainly attributable to varied distributions of GABA_A receptors (GABA_ARs) at the synaptic and extrasynaptic regions of GC dendrites. In addition to GABA_AR distributions, amounts of extrasynaptic NMDA receptors appear to be also varied, because Ca²⁺ responses were different between D-GCs and M-GCs when glutamate spillover was enhanced. Whereas the Ca²⁺ responses of S-GCs were mostly equivalent to those of D-GCs and M-GCs, the blockade of GABA uptake resulted in larger Ca²⁺ responses in S-GCs compared with D-GCs and M-GCs, implying existence of mechanisms leading to more excitability in S-GCs with increased GABA release. Thus, this study reveals MF stimulation-mediated non-uniform Ca²⁺ responses in the cerebellar GCs associated with the location of their PFs in the ML, and raises a possibility that combination of inherent functional variability of GCs and their specific axonal projection contributes to the information processing through the GCs.

Human-scale Brain Simulation via Supercomputer: A Case Study on the Cerebellum

Performance of supercomputers has been steadily and exponentially increasing for the past 20 years, and is expected to increase further. This unprecedented computational power enables us to build and simulate large-scale neural network models composed of tens of billions of neurons and tens of trillions of synapses with detailed anatomical connections and realistic physiological parameters. Such "human-scale" brain simulation could be considered a milestone in computational neuroscience and even in general neuroscience. Towards this milestone, it is mandatory to introduce modern high-performance computing technology into neuroscience research. In this article, we provide an introductory landscape about large-scale brain simulation on supercomputers from the viewpoints of computational neuroscience and modern high-performance computing technology for specialists in experimental as well as computational neurosciences. This introduction to modeling and simulation methods is followed by a review of various representative large-scale simulation studies conducted to date. Then, we direct our attention to the cerebellum, with a review of more simulation studies specific to that region. Furthermore, we present recent simulation results of a human-scale cerebellar network model composed of 86 billion neurons on the Japanese flagship supercomputer K (now retired). Finally, we discuss the necessity and importance of human-scale brain simulation, and suggest future directions of such large-scale brain simulation research.