Explore the vast range of titles available.

The brain must maintain a stable world model while rapidly adapting to the environment, but the underlying mechanisms are not known. Here, we posit that cortico-cerebellar loops play a key role in this process. We introduce a computational model of cerebellar networks that learn to drive cortical networks with task-outcome predictions. First, using sensorimotor tasks, we show that cerebellar feedback in the presence of stable cortical networks is sufficient for rapid task acquisition and switching. Next, we demonstrate that, when trained in working memory tasks, the cerebellum can also underlie the maintenance of cognitive-specific dynamics in the cortex, explaining a range of optogenetic and behavioural observations. Finally, using our model, we introduce a systems consolidation theory in which task information is gradually transferred from the cerebellum to the cortex. In summary, our findings suggest that cortico-cerebellar loops are an important component of task acquisition, switching, and consolidation in the brain. How the brain maintains a stable world model while swiftly adapting to environmental changes remains unclear. Here, the authors propose that the cerebellum drives cortical dynamics, enabling rapid task acquisition, switching, and consolidation.

Systems neuroscience seeks explanations for how the brain implements a wide variety of perceptual, cognitive and motor tasks. Conversely, artificial intelligence attempts to design computational systems based on the tasks they will have to solve. In artificial neural networks, the three components specified by design are the objective functions, the learning rules and the architectures. With the growing success of deep learning, which utilizes brain-inspired architectures, these three designed components have increasingly become central to how we model, engineer and optimize complex artificial learning systems. Here we argue that a greater focus on these components would also benefit systems neuroscience. We give examples of how this optimization-based framework can drive theoretical and experimental progress in neuroscience. We contend that this principled perspective on systems neuroscience will help to generate more rapid progress.

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. Behavioral feedback is critical for learning, but it is often not available. Here, the authors introduce a deep learning model in which the cerebellum provides the cerebrum with feedback predictions, thereby facilitating learning, reducing dysmetria, and making several experimental predictions.

The neocortex constructs an internal representation of the world, but the underlying circuitry and computational principles remain unclear. Inspired by self-supervised learning algorithms, we propose a computational theory in which layer 2/3 (L2/3) integrates past sensory input, relayed via layer 4, with top-down context to predict incoming sensory stimuli. Learning is self-supervised by comparing L2/3 predictions with the latent representations of actual sensory input arriving at L5. We demonstrate that our model accurately predicts sensory information in context-dependent temporal tasks, and that its predictions are robust to noisy and occluded sensory input. Additionally, our model generates layer-specific sparsity, consistent with experimental observations. Next, using a sensorimotor task, we show that the model’s L2/3 and L5 prediction errors mirror mismatch responses observed in awake, behaving mice. Finally, through manipulations, we offer testable predictions to unveil the computational roles of various cortical features. In summary, our findings suggest that the multi-layered neocortex empowers the brain with self-supervised predictive learning. How the cortex learns an internal representation of the world and why the cortex relies on a layered structure remain poorly understood. Here, the authors propose a computational theory in which the neocortex engages in self-supervised predictive learning by integrating past sensory input and top-down context in layer 2/3 to anticipate future stimuli, suggesting that a fundamental role of cortical layering is to support self-supervised learning.

Adaptive behavior relies on associating sensory cues with rewarding or aversive outcomes. In mammals, the primary sensory cortex processes stimuli and distributes information to cortical and subcortical targets. Layer 5 (L5) contains two major projection neuron classes, intratelencephalic (IT) and extratelencephalic (ET); however, their roles in associative learning remain unclear. Using transgenic mice, we identified IT and ET neurons in primary somatosensory cortex and tracked their activity with longitudinal two-photon imaging during Pavlovian conditioning with whisker stimulation. IT neurons stably encoded stimulus identity across training, whereas ET neurons showed dynamic changes that paralleled the emergence of anticipatory licking. Chemogenetic silencing of each subtype impaired learning in distinct, phase-specific ways. A reinforcement-learning model reproduced these dynamics, suggesting that IT neurons provide stable sensory representations needed to form cue-reward associations, while ET neurons encode reward expectation to refine behavior. These findings reveal complementary, cell-type-specific contributions of L5 neurons to associative learning. How the brain learns to link a sensory signal to a reward is not fully understood. Here authors show that two types of layer 5 neurons in sensory cortex contribute in different ways, helping the brain recognize relevant sensory cues and refine behavior accordingly.

Mastering navigation in environments with limited visibility is crucial for survival. Although the hippocampus has been associated with goal-oriented navigation, its role in real-world behaviour remains unclear. To investigate this, we combined deep reinforcement learning (RL) modelling with behavioural and neural data analysis. First, we trained RL agents in partially observable environments using egocentric and allocentric tasks. We show that agents equipped with recurrent hippocampal circuitry, but not purely feedforward networks, learned the tasks in line with animal behaviour. Next, we used dimensionality reduction of the agents’ internal representations to extract components reflecting reward, strategy, and temporal representations, which we validated experimentally against hippocampal recordings from rats. Moreover, hippocampal RL agents predicted state-specific trajectories, mirroring empirical findings. In contrast, agents trained in fully observable environments failed to capture experimental observations. Finally, we show that hippocampal-like RL agents demonstrated improved generalisation across novel task conditions. In summary, our findings suggest an important role of hippocampal networks in facilitating reinforcement learning in naturalistic environments. Neural mechanisms underlying reinforcement learning in naturalistic environments are not fully understood. Here authors show that reinforcement learning (RL) agents with hippocampal-like recurrence, unlike feedforward networks, match animal behaviour and neural data in navigation tasks, revealing that hippocampal circuits support RL in naturalistic environments.

A plethora of experimental studies have shown that long-term synaptic plasticity can be expressed pre- or postsynaptically depending on a range of factors such as developmental stage, synapse type, and activity patterns. The functional consequences of this diversity are not clear, although it is understood that whereas postsynaptic expression of plasticity predominantly affects synaptic response amplitude, presynaptic expression alters both synaptic response amplitude and short-term dynamics. In most models of neuronal learning, long-term synaptic plasticity is implemented as changes in connective weights. The consideration of long-term plasticity as a fixed change in amplitude corresponds more closely to post- than to presynaptic expression, which means theoretical outcomes based on this choice of implementation may have a postsynaptic bias. To explore the functional implications of the diversity of expression of long-term synaptic plasticity, we adapted a model of long-term plasticity, more specifically spike-timing-dependent plasticity (STDP), such that it was expressed either independently pre- or postsynaptically, or in a mixture of both ways. We compared pair-based standard STDP models and a biologically tuned triplet STDP model, and investigated the outcomes in a minimal setting, using two different learning schemes: in the first, inputs were triggered at different latencies, and in the second a subset of inputs were temporally correlated. We found that presynaptic changes adjusted the speed of learning, while postsynaptic expression was more efficient at regulating spike timing and frequency. When combining both expression loci, postsynaptic changes amplified the response range, while presynaptic plasticity allowed control over postsynaptic firing rates, potentially providing a form of activity homeostasis. Our findings highlight how the seemingly innocuous choice of implementing synaptic plasticity by single weight modification may unwittingly introduce a postsynaptic bias in modelling outcomes. We conclude that pre- and postsynaptically expressed plasticity are not interchangeable, but enable complimentary functions.

Growing experimental evidence shows that both homeostatic and Hebbian synaptic plasticity can be expressed presynaptically as well as postsynaptically. In this review, we start by discussing this evidence and methods used to determine expression loci. Next, we discuss the functional consequences of this diversity in pre- and postsynaptic expression of both homeostatic and Hebbian synaptic plasticity. In particular, we explore the functional consequences of a biologically tuned model of pre- and postsynaptically expressed spike-timing-dependent plasticity complemented with postsynaptic homeostatic control. The pre- and postsynaptic expression in this model predicts (i) more reliable receptive fields and sensory perception, (ii) rapid recovery of forgotten information (memory savings), and (iii) reduced response latencies, compared with a model with postsynaptic expression only. Finally, we discuss open questions that will require a considerable research effort to better elucidate how the specific locus of expression of homeostatic and Hebbian plasticity alters synaptic and network computations. This article is part of the themed issue ‘Integrating Hebbian and homeostatic plasticity’.