Decoding the brain: From neural representations to mechanistic models.

A central principle in neuroscience is that neurons within the brain act in concert to produce perception, cognition, and adaptive behavior. Neurons are organized into specialized brain areas, dedicated to different functions to varying extents, and their function relies on distributed circuits to continuously encode relevant environmental and body-state features, enabling other areas to decode (interpret) these representations for computing meaningful decisions and executing precise movements. Thus, the distributed brain can be thought of as a series of computations that act to encode and decode information. In this perspective, we detail important concepts of neural encoding and decoding and highlight the mathematical tools used to measure them, including deep learning methods. We provide case studies where decoding concepts enable foundational and translational science in motor, visual, and language processing.

Hippocampal neurons construct a map of an abstract value space.

The hippocampus is thought to encode a "cognitive map," a structural organization of knowledge about relationships in the world. Place cells, spatially selective hippocampal neurons that have been extensively studied in rodents, are one component of this map, describing the relative position of environmental features. However, whether this map extends to abstract, cognitive information remains unknown. Using the relative reward value of cues to define continuous "paths" through an abstract value space, we show that single neurons in primate hippocampus encode this space through value place fields, much like a rodent's place neurons encode paths through physical space. Value place fields remapped when cues changed but also became increasingly correlated across contexts, allowing maps to become generalized. Our findings help explain the critical contribution of the hippocampus to value-based decision-making, providing a mechanism by which knowledge of relationships in the world can be incorporated into reward predictions for guiding decisions.

Modularity and robustness of frontal cortical networks.

Neural activity underlying short-term memory is maintained by interconnected networks of brain regions. It remains unknown how brain regions interact to maintain persistent activity while exhibiting robustness to corrupt information in parts of the network. We simultaneously measured activity in large neuronal populations across mouse frontal hemispheres to probe interactions between brain regions. Activity across hemispheres was coordinated to maintain coherent short-term memory. Across mice, we uncovered individual variability in the organization of frontal cortical networks. A modular organization was required for the robustness of persistent activity to perturbations: each hemisphere retained persistent activity during perturbations of the other hemisphere, thus preventing local perturbations from spreading. A dynamic gating mechanism allowed hemispheres to coordinate coherent information while gating out corrupt information. Our results show that robust short-term memory is mediated by redundant modular representations across brain regions. Redundant modular representations naturally emerge in neural network models that learned robust dynamics.

A neural circuit state change underlying skilled movements.

In motor neuroscience, state changes are hypothesized to time-lock neural assemblies coordinating complex movements, but evidence for this remains slender. We tested whether a discrete change from more autonomous to coherent spiking underlies skilled movement by imaging cerebellar Purkinje neuron complex spikes in mice making targeted forelimb-reaches. As mice learned the task, millimeter-scale spatiotemporally coherent spiking emerged ipsilateral to the reaching forelimb, and consistent neural synchronization became predictive of kinematic stereotypy. Before reach onset, spiking switched from more disordered to internally time-locked concerted spiking and silence. Optogenetic manipulations of cerebellar feedback to the inferior olive bi-directionally modulated neural synchronization and reaching direction. A simple model explained the reorganization of spiking during reaching as reflecting a discrete bifurcation in olivary network dynamics. These findings argue that to prepare learned movements, olivo-cerebellar circuits enter a self-regulated, synchronized state promoting motor coordination. State changes facilitating behavioral transitions may generalize across neural systems.

Representational geometry of perceptual decisions in the monkey parietal cortex.

Lateral intraparietal (LIP) neurons represent formation of perceptual decisions involving eye movements. In circuit models for these decisions, neural ensembles that encode actions compete to form decisions. Consequently, representation and readout of the decision variables (DVs) are implemented similarly for decisions with identical competing actions, irrespective of input and task context differences. Further, DVs are encoded as partially potentiated action plans through balance of activity of action-selective ensembles. Here, we test those core principles. We show that in a novel face-discrimination task, LIP firing rates decrease with supporting evidence, contrary to conventional motion-discrimination tasks. These opposite response patterns arise from similar mechanisms in which decisions form along curved population-response manifolds misaligned with action representations. These manifolds rotate in state space based on context, indicating distinct optimal readouts for different tasks. We show similar manifolds in lateral and medial prefrontal cortices, suggesting similar representational geometry across decision-making circuits.

The Tolman-Eichenbaum Machine: Unifying Space and Relational Memory through Generalization in the Hippocampal Formation.

The hippocampal-entorhinal system is important for spatial and relational memory tasks. We formally link these domains, provide a mechanistic understanding of the hippocampal role in generalization, and offer unifying principles underlying many entorhinal and hippocampal cell types. We propose medial entorhinal cells form a basis describing structural knowledge, and hippocampal cells link this basis with sensory representations. Adopting these principles, we introduce the Tolman-Eichenbaum machine (TEM). After learning, TEM entorhinal cells display diverse properties resembling apparently bespoke spatial responses, such as grid, band, border, and object-vector cells. TEM hippocampal cells include place and landmark cells that remap between environments. Crucially, TEM also aligns with empirically recorded representations in complex non-spatial tasks. TEM also generates predictions that hippocampal remapping is not random as previously believed; rather, structural knowledge is preserved across environments. We confirm this structural transfer over remapping in simultaneously recorded place and grid cells.

The Statistical Structure of the Hippocampal Code for Space as a Function of Time, Context, and Value.

Hippocampal activity represents many behaviorally important variables, including context, an animal's location within a given environmental context, time, and reward. Using longitudinal calcium imaging in mice, multiple large virtual environments, and differing reward contingencies, we derived a unified probabilistic model of CA1 representations centered on a single feature-the field propensity. Each cell's propensity governs how many place fields it has per unit space, predicts its reward-related activity, and is preserved across distinct environments and over months. Propensity is broadly distributed-with many low, and some very high, propensity cells-and thus strongly shapes hippocampal representations. This results in a range of spatial codes, from sparse to dense. Propensity varied ∼10-fold between adjacent cells in salt-and-pepper fashion, indicating substantial functional differences within a presumed cell type. Intracellular recordings linked propensity to cell excitability. The stability of each cell's propensity across conditions suggests this fundamental property has anatomical, transcriptional, and/or developmental origins.

Neuronal Computation Underlying Inferential Reasoning in Humans and Mice.

Every day we make decisions critical for adaptation and survival. We repeat actions with known consequences. But we also draw on loosely related events to infer and imagine the outcome of entirely novel choices. These inferential decisions are thought to engage a number of brain regions; however, the underlying neuronal computation remains unknown. Here, we use a multi-day cross-species approach in humans and mice to report the functional anatomy and neuronal computation underlying inferential decisions. We show that during successful inference, the mammalian brain uses a hippocampal prospective code to forecast temporally structured learned associations. Moreover, during resting behavior, coactivation of hippocampal cells in sharp-wave/ripples represent inferred relationships that include reward, thereby "joining-the-dots" between events that have not been observed together but lead to profitable outcomes. Computing mnemonic links in this manner may provide an important mechanism to build a cognitive map that stretches beyond direct experience, thus supporting flexible behavior.

Saltatory Conduction along Myelinated Axons Involves a Periaxonal Nanocircuit.

The propagation of electrical impulses along axons is highly accelerated by the myelin sheath and produces saltating or "jumping" action potentials across internodes, from one node of Ranvier to the next. The underlying electrical circuit, as well as the existence and role of submyelin conduction in saltatory conduction remain, however, elusive. Here, we made patch-clamp and high-speed voltage-calibrated optical recordings of potentials across the nodal and internodal axolemma of myelinated neocortical pyramidal axons combined with electron microscopy and experimentally constrained cable modeling. Our results reveal a nanoscale yet conductive periaxonal space, incompletely sealed at the paranodes, which separates the potentials across the low-capacitance myelin sheath and internodal axolemma. The emerging double-cable model reproduces the recorded evolution of voltage waveforms across nodes and internodes, including rapid nodal potentials traveling in advance of attenuated waves in the internodal axolemma, revealing a mechanism for saltation across time and space.

Striosomes Mediate Value-Based Learning Vulnerable in Age and a Huntington's Disease Model.

Learning valence-based responses to favorable and unfavorable options requires judgments of the relative value of the options, a process necessary for species survival. We found, using engineered mice, that circuit connectivity and function of the striosome compartment of the striatum are critical for this type of learning. Calcium imaging during valence-based learning exhibited a selective correlation between learning and striosomal but not matrix signals. This striosomal activity encoded discrimination learning and was correlated with task engagement, which, in turn, could be regulated by chemogenetic excitation and inhibition. Striosomal function during discrimination learning was disturbed with aging and severely so in a mouse model of Huntington's disease. Anatomical and functional connectivity of parvalbumin-positive, putative fast-spiking interneurons (FSIs) to striatal projection neurons was enhanced in striosomes compared with matrix in mice that learned. Computational modeling of these findings suggests that FSIs can modulate the striosomal signal-to-noise ratio, crucial for discrimination and learning.

A Unified Framework for Dopamine Signals across Timescales.

Rapid phasic activity of midbrain dopamine neurons is thought to signal reward prediction errors (RPEs), resembling temporal difference errors used in machine learning. However, recent studies describing slowly increasing dopamine signals have instead proposed that they represent state values and arise independent from somatic spiking activity. Here we developed experimental paradigms using virtual reality that disambiguate RPEs from values. We examined dopamine circuit activity at various stages, including somatic spiking, calcium signals at somata and axons, and striatal dopamine concentrations. Our results demonstrate that ramping dopamine signals are consistent with RPEs rather than value, and this ramping is observed at all stages examined. Ramping dopamine signals can be driven by a dynamic stimulus that indicates a gradual approach to a reward. We provide a unified computational understanding of rapid phasic and slowly ramping dopamine signals: dopamine neurons perform a derivative-like computation over values on a moment-by-moment basis.

The Geometry of Abstraction in the Hippocampus and Prefrontal Cortex.

The curse of dimensionality plagues models of reinforcement learning and decision making. The process of abstraction solves this by constructing variables describing features shared by different instances, reducing dimensionality and enabling generalization in novel situations. Here, we characterized neural representations in monkeys performing a task described by different hidden and explicit variables. Abstraction was defined operationally using the generalization performance of neural decoders across task conditions not used for training, which requires a particular geometry of neural representations. Neural ensembles in prefrontal cortex, hippocampus, and simulated neural networks simultaneously represented multiple variables in a geometry reflecting abstraction but that still allowed a linear classifier to decode a large number of other variables (high shattering dimensionality). Furthermore, this geometry changed in relation to task events and performance. These findings elucidate how the brain and artificial systems represent variables in an abstract format while preserving the advantages conferred by high shattering dimensionality.

Plugging in to Human Memory: Advantages, Challenges, and Insights from Human Single-Neuron Recordings.

We describe single-neuron recordings in the human hippocampal formation, performed in epileptic patients for clinical reasons, and highlight their advantages, challenges, and limitations compared with non-invasive recordings in humans and invasive recordings in animals. We propose a unified framework to explain different findings-responses to novel stimuli, spatial locations, and specific concepts-linking the rodent and human literature regarding the function of the hippocampal formation. Moreover, we propose a model of how memories are encoded in this area, suggesting that the context-independent, invariant coding by concept cells may provide a uniquely human neural mechanism underlying memory representations.

Continual Learning in a Multi-Layer Network of an Electric Fish.

Distributing learning across multiple layers has proven extremely powerful in artificial neural networks. However, little is known about how multi-layer learning is implemented in the brain. Here, we provide an account of learning across multiple processing layers in the electrosensory lobe (ELL) of mormyrid fish and report how it solves problems well known from machine learning. Because the ELL operates and learns continuously, it must reconcile learning and signaling functions without switching its mode of operation. We show that this is accomplished through a functional compartmentalization within intermediate layer neurons in which inputs driving learning differentially affect dendritic and axonal spikes. We also find that connectivity based on learning rather than sensory response selectivity assures that plasticity at synapses onto intermediate-layer neurons is matched to the requirements of output neurons. The mechanisms we uncover have relevance to learning in the cerebellum, hippocampus, and cerebral cortex, as well as in artificial systems.

Development of Concurrent Retinotopic Maps in the Fly Motion Detection Circuit.

Understanding how complex brain wiring is produced during development is a daunting challenge. In Drosophila, information from 800 retinal ommatidia is processed in distinct brain neuropiles, each subdivided into 800 matching retinotopic columns. The lobula plate comprises four T4 and four T5 neuronal subtypes. T4 neurons respond to bright edge motion, whereas T5 neurons respond to dark edge motion. Each is tuned to motion in one of the four cardinal directions, effectively establishing eight concurrent retinotopic maps to support wide-field motion. We discovered a mode of neurogenesis where two sequential Notch-dependent divisions of either a horizontal or a vertical progenitor produce matching sets of two T4 and two T5 neurons retinotopically coincident with pairwise opposite direction selectivity. We show that retinotopy is an emergent characteristic of this neurogenic program and derives directly from neuronal birth order. Our work illustrates how simple developmental rules can implement complex neural organization.

Clock-Generated Temporal Codes Determine Synaptic Plasticity to Control Sleep.

Neurons use two main schemes to encode information: rate coding (frequency of firing) and temporal coding (timing or pattern of firing). While the importance of rate coding is well established, it remains controversial whether temporal codes alone are sufficient for controlling behavior. Moreover, the molecular mechanisms underlying the generation of specific temporal codes are enigmatic. Here, we show in Drosophila clock neurons that distinct temporal spike patterns, dissociated from changes in firing rate, encode time-dependent arousal and regulate sleep. From a large-scale genetic screen, we identify the molecular pathways mediating the circadian-dependent changes in ionic flux and spike morphology that rhythmically modulate spike timing. Remarkably, the daytime spiking pattern alone is sufficient to drive plasticity in downstream arousal neurons, leading to increased firing of these cells. These findings demonstrate a causal role for temporal coding in behavior and define a form of synaptic plasticity triggered solely by temporal spike patterns.

Language in Brains, Minds, and Machines.

It has long been argued that only humans could produce and understand language. But now, for the first time, artificial language models (LMs) achieve this feat. Here we survey the new purchase LMs are providing on the question of how language is implemented in the brain. We discuss why, a priori, LMs might be expected to share similarities with the human language system. We then summarize evidence that LMs represent linguistic information similarly enough to humans to enable relatively accurate brain encoding and decoding during language processing. Finally, we examine which LM properties-their architecture, task performance, or training-are critical for capturing human neural responses to language and review studies using LMs as in silico model organisms for testing hypotheses about language. These ongoing investigations bring us closer to understanding the representations and processes that underlie our ability to comprehend sentences and express thoughts in language.

Grid Cells in Cognition: Mechanisms and Function.

The activity patterns of grid cells form distinctively regular triangular lattices over the explored spatial environment and are largely invariant to visual stimuli, animal movement, and environment geometry. These neurons present numerous fascinating challenges to the curious (neuro)scientist: What are the circuit mechanisms responsible for creating spatially periodic activity patterns from the monotonic input-output responses of single neurons? How and why does the brain encode a local, nonperiodic variable-the allocentric position of the animal-with a periodic, nonlocal code? And, are grid cells truly specialized for spatial computations? Otherwise, what is their role in general cognition more broadly? We review efforts in uncovering the mechanisms and functional properties of grid cells, highlighting recent progress in the experimental validation of mechanistic grid cell models, and discuss the coding properties and functional advantages of the grid code as suggested by continuous attractor network models of grid cells.

Cortical Layer-Dependent Signaling in Cognition: Three Computational Modes of the Canonical Circuit.

The cerebral cortex performs computations via numerous six-layer modules. The operational dynamics of these modules were studied primarily in early sensory cortices using bottom-up computation for response selectivity as a model, which has been recently revolutionized by genetic approaches in mice. However, cognitive processes such as recall and imagery require top-down generative computation. The question of whether the layered module operates similarly in top-down generative processing as in bottom-up sensory processing has become testable by advances in the layer identification of recorded neurons in behaving monkeys. This review examines recent advances in laminar signaling in these two computations, using predictive coding computation as a common reference, and shows that each of these computations recruits distinct laminar circuits, particularly in layer 5, depending on the cognitive demands. These findings highlight many open questions, including how different interareal feedback pathways, originating from and terminating at different layers, convey distinct functional signals.

Predictive Processing: A Circuit Approach to Psychosis.

Predictive processing is a computational framework that aims to explain how the brain processes sensory information by making predictions about the environment and minimizing prediction errors. It can also be used to explain some of the key symptoms of psychotic disorders such as schizophrenia. In recent years, substantial advances have been made in our understanding of the neuronal circuitry that underlies predictive processing in cortex. In this review, we summarize these findings and how they might relate to psychosis and to observed cell type-specific effects of antipsychotic drugs. We argue that quantifying the effects of antipsychotic drugs on specific neuronal circuit elements is a promising approach to understanding not only the mechanism of action of antipsychotic drugs but also psychosis. Finally, we outline some of the key experiments that should be done. The aims of this review are to provide an overview of the current circuit-based approaches to psychosis and to encourage further research in this direction.