A virtual rodent predicts the structure of neural activity across behaviors.

Animals have exquisite control of their bodies, allowing them to perform a diverse range of behaviors. How such control is implemented by the brain, however, remains unclear. Advancing our understanding requires models that can relate principles of control to the structure of neural activity in behaving animals. To facilitate this, we built a 'virtual rodent', in which an artificial neural network actuates a biomechanically realistic model of the rat 1 in a physics simulator 2. We used deep reinforcement learning 3-5 to train the virtual agent to imitate the behavior of freely-moving rats, thus allowing us to compare neural activity recorded in real rats to the network activity of a virtual rodent mimicking their behavior. We found that neural activity in the sensorimotor striatum and motor cortex was better predicted by the virtual rodent's network activity than by any features of the real rat's movements, consistent with both regions implementing inverse dynamics 6. Furthermore, the network's latent variability predicted the structure of neural variability across behaviors and afforded robustness in a way consistent with the minimal intervention principle of optimal feedback control 7. These results demonstrate how physical simulation of biomechanically realistic virtual animals can help interpret the structure of neural activity across behavior and relate it to theoretical principles of motor control.

Dissociating the contributions of sensorimotor striatum to automatic and visually guided motor sequences.

The ability to sequence movements in response to new task demands enables rich and adaptive behavior. However, such flexibility is computationally costly and can result in halting performances. Practicing the same motor sequence repeatedly can render its execution precise, fast and effortless, that is, 'automatic'. The basal ganglia are thought to underlie both types of sequence execution, yet whether and how their contributions differ is unclear. We parse this in rats trained to perform the same motor sequence instructed by cues and in a self-initiated overtrained, or 'automatic,' condition. Neural recordings in the sensorimotor striatum revealed a kinematic code independent of the execution mode. Although lesions reduced the movement speed and affected detailed kinematics similarly, they disrupted high-level sequence structure for automatic, but not visually guided, behaviors. These results suggest that the basal ganglia are essential for 'automatic' motor skills that are defined in terms of continuous kinematics, but can be dispensable for discrete motor sequences guided by sensory cues.

Motor cortex is required for flexible but not automatic motor sequences.

How motor cortex contributes to motor sequence execution is much debated, with studies supporting disparate views. Here we probe the degree to which motor cortex's engagement depends on task demands, specifically whether its role differs for highly practiced, or 'automatic', sequences versus flexible sequences informed by external events. To test this, we trained rats to generate three-element motor sequences either by overtraining them on a single sequence or by having them follow instructive visual cues. Lesioning motor cortex revealed that it is necessary for flexible cue-driven motor sequences but dispensable for single automatic behaviors trained in isolation. However, when an automatic motor sequence was practiced alongside the flexible task, it became motor cortex-dependent, suggesting that subcortical consolidation of an automatic motor sequence is delayed or prevented when the same sequence is produced also in a flexible context. A simple neural network model recapitulated these results and explained the underlying circuit mechanisms. Our results critically delineate the role of motor cortex in motor sequence execution, describing the condition under which it is engaged and the functions it fulfills, thus reconciling seemingly conflicting views about motor cortex's role in motor sequence generation.

Catalyzing next-generation Artificial Intelligence through NeuroAI.

Neuroscience has long been an essential driver of progress in artificial intelligence (AI). We propose that to accelerate progress in AI, we must invest in fundamental research in NeuroAI. A core component of this is the embodied Turing test, which challenges AI animal models to interact with the sensorimotor world at skill levels akin to their living counterparts. The embodied Turing test shifts the focus from those capabilities like game playing and language that are especially well-developed or uniquely human to those capabilities - inherited from over 500 million years of evolution - that are shared with all animals. Building models that can pass the embodied Turing test will provide a roadmap for the next generation of AI.

Long-term stability of single neuron activity in the motor system.

How an established behavior is retained and consistently produced by a nervous system in constant flux remains a mystery. One possible solution to ensure long-term stability in motor output is to fix the activity patterns of single neurons in the relevant circuits. Alternatively, activity in single cells could drift over time provided that the population dynamics are constrained to produce the same behavior. To arbitrate between these possibilities, we recorded single-unit activity in motor cortex and striatum continuously for several weeks as rats performed stereotyped motor behaviors-both learned and innate. We found long-term stability in single neuron activity patterns across both brain regions. A small amount of drift in neural activity, observed over weeks of recording, could be explained by concomitant changes in task-irrelevant aspects of the behavior. These results suggest that long-term stable behaviors are generated by single neuron activity patterns that are themselves highly stable.

Learning-induced changes in the neural circuits underlying motor sequence execution.

As the old adage goes: practice makes perfect. Yet, the neural mechanisms by which rote repetition transforms a halting behavior into a fluid, effortless, and "automatic" action are not well understood. Here we consider the possibility that well-practiced motor sequences, which initially rely on higher-level decision-making circuits, become wholly specified in lower-level control circuits. We review studies informing this idea, discuss the constraints on such shift in control, and suggest approaches to pinpoint circuit-level changes associated with motor sequence learning.

Distinct roles for motor cortical and thalamic inputs to striatum during motor skill learning and execution.

The acquisition and execution of motor skills are mediated by a distributed motor network, spanning cortical and subcortical brain areas. The sensorimotor striatum is an important cog in this network, yet the roles of its two main inputs, from motor cortex and thalamus, remain largely unknown. To address this, we silenced the inputs in rats trained on a task that results in highly stereotyped and idiosyncratic movement patterns. While striatal-projecting motor cortex neurons were critical for learning these skills, silencing this pathway after learning had no effect on performance. In contrast, silencing striatal-projecting thalamus neurons disrupted the execution of the learned skills, causing rats to revert to species-typical pressing behaviors and preventing them from relearning the task. These results show distinct roles for motor cortex and thalamus in the learning and execution of motor skills and suggest that their interaction in the striatum underlies experience-dependent changes in subcortical motor circuits.

The basal ganglia control the detailed kinematics of learned motor skills.

The basal ganglia are known to influence action selection and modulation of movement vigor, but whether and how they contribute to specifying the kinematics of learned motor skills is not understood. Here, we probe this question by recording and manipulating basal ganglia activity in rats trained to generate complex task-specific movement patterns with rich kinematic structure. We find that the sensorimotor arm of the basal ganglia circuit is crucial for generating the detailed movement patterns underlying the acquired motor skills. Furthermore, the neural representations in the striatum, and the control function they subserve, do not depend on inputs from the motor cortex. Taken together, these results extend our understanding of the basal ganglia by showing that they can specify and control the fine-grained details of learned motor skills through their interactions with lower-level motor circuits.

Geometric deep learning enables 3D kinematic profiling across species and environments.

Comprehensive descriptions of animal behavior require precise three-dimensional (3D) measurements of whole-body movements. Although two-dimensional approaches can track visible landmarks in restrictive environments, performance drops in freely moving animals, due to occlusions and appearance changes. Therefore, we designed DANNCE to robustly track anatomical landmarks in 3D across species and behaviors. DANNCE uses projective geometry to construct inputs to a convolutional neural network that leverages learned 3D geometric reasoning. We trained and benchmarked DANNCE using a dataset of nearly seven million frames that relates color videos and rodent 3D poses. In rats and mice, DANNCE robustly tracked dozens of landmarks on the head, trunk, and limbs of freely moving animals in naturalistic settings. We extended DANNCE to datasets from rat pups, marmosets, and chickadees, and demonstrate quantitative profiling of behavioral lineage during development.

Continuous Whole-Body 3D Kinematic Recordings across the Rodent Behavioral Repertoire.

In mammalian animal models, high-resolution kinematic tracking is restricted to brief sessions in constrained environments, limiting our ability to probe naturalistic behaviors and their neural underpinnings. To address this, we developed CAPTURE (Continuous Appendicular and Postural Tracking Using Retroreflector Embedding), a behavioral monitoring system that combines motion capture and deep learning to continuously track the 3D kinematics of a rat's head, trunk, and limbs for week-long timescales in freely behaving animals. CAPTURE realizes 10- to 100-fold gains in precision and robustness compared with existing convolutional network approaches to behavioral tracking. We demonstrate CAPTURE's ability to comprehensively profile the kinematics and sequential organization of natural rodent behavior, its variation across individuals, and its perturbation by drugs and disease, including identifying perseverative grooming states in a rat model of fragile X syndrome. CAPTURE significantly expands the range of behaviors and contexts that can be quantitatively investigated, opening the door to a new understanding of natural behavior and its neural basis.

The neurobiology of deep reinforcement learning.

In this primer, Ölveczky and Gershman review concepts and advances in deep reinforcement learning and discuss how these can inform the implementation of learning processes in biological neural networks.

Discovering Precise Temporal Patterns in Large-Scale Neural Recordings through Robust and Interpretable Time Warping.

Though the temporal precision of neural computation has been studied intensively, a data-driven determination of this precision remains a fundamental challenge. Reproducible spike patterns may be obscured on single trials by uncontrolled temporal variability in behavior and cognition and may not be time locked to measurable signatures in behavior or local field potentials (LFP). To overcome these challenges, we describe a general-purpose time warping framework that reveals precise spike-time patterns in an unsupervised manner, even when these patterns are decoupled from behavior or are temporally stretched across single trials. We demonstrate this method across diverse systems: cued reaching in nonhuman primates, motor sequence production in rats, and olfaction in mice. This approach flexibly uncovers diverse dynamical firing patterns, including pulsatile responses to behavioral events, LFP-aligned oscillatory spiking, and even unanticipated patterns, such as 7 Hz oscillations in rat motor cortex that are not time locked to measured behaviors or LFP.

Adaptive Regulation of Motor Variability.

Trial-to-trial movement variability can both drive motor learning and interfere with expert performance, suggesting benefits of regulating it in context-specific ways. Here we address whether and how the brain regulates motor variability as a function of performance by training rats to execute ballistic forelimb movements for reward. Behavioral datasets comprising millions of trials revealed that motor variability is regulated by two distinct processes. A fast process modulates variability as a function of recent trial outcomes, increasing it when performance is poor and vice versa. A slower process tunes the gain of the fast process based on the uncertainty in the task's reward landscape. Simulations demonstrated that this regulation strategy optimizes reward accumulation over a wide range of time horizons, while also promoting learning. Our results uncover a sophisticated algorithm implemented by the brain to adaptively regulate motor variability to improve task performance. VIDEO ABSTRACT.

What is cognition?

Eleven authors with disparate relevant backgrounds give their view on what is meant by the word "cognition".

Flexibility in motor timing constrains the topology and dynamics of pattern generator circuits.

Temporally precise movement patterns underlie many motor skills and innate actions, yet the flexibility with which the timing of such stereotyped behaviors can be modified is poorly understood. To probe this, we induce adaptive changes to the temporal structure of birdsong. We find that the duration of specific song segments can be modified without affecting the timing in other parts of the song. We derive formal prescriptions for how neural networks can implement such flexible motor timing. We find that randomly connected recurrent networks, a common approximation for how neocortex is wired, do not generally conform to these, though certain implementations can approximate them. We show that feedforward networks, by virtue of their one-to-one mapping between network activity and time, are better suited. Our study provides general prescriptions for pattern generator networks that implement flexible motor timing, an important aspect of many motor skills, including birdsong and human speech.

The promise and perils of causal circuit manipulations.

The development of increasingly sophisticated methods for recording and manipulating neural activity is revolutionizing neuroscience. By probing how activity patterns in different types of neurons and circuits contribute to behavior, these tools can help inform mechanistic models of brain function and explain the roles of distinct circuit elements. However, in systems where functions are distributed over large networks, interpreting causality experiments can be challenging. Here we review common assumptions underlying circuit manipulations in behaving animals and discuss the strengths and limitations of different approaches.

Automated long-term recording and analysis of neural activity in behaving animals.

Addressing how neural circuits underlie behavior is routinely done by measuring electrical activity from single neurons in experimental sessions. While such recordings yield snapshots of neural dynamics during specified tasks, they are ill-suited for tracking single-unit activity over longer timescales relevant for most developmental and learning processes, or for capturing neural dynamics across different behavioral states. Here we describe an automated platform for continuous long-term recordings of neural activity and behavior in freely moving rodents. An unsupervised algorithm identifies and tracks the activity of single units over weeks of recording, dramatically simplifying the analysis of large datasets. Months-long recordings from motor cortex and striatum made and analyzed with our system revealed remarkable stability in basic neuronal properties, such as firing rates and inter-spike interval distributions. Interneuronal correlations and the representation of different movements and behaviors were similarly stable. This establishes the feasibility of high-throughput long-term extracellular recordings in behaving animals.

The Role of Variability in Motor Learning.

Trial-to-trial variability in the execution of movements and motor skills is ubiquitous and widely considered to be the unwanted consequence of a noisy nervous system. However, recent studies have suggested that motor variability may also be a feature of how sensorimotor systems operate and learn. This view, rooted in reinforcement learning theory, equates motor variability with purposeful exploration of motor space that, when coupled with reinforcement, can drive motor learning. Here we review studies that explore the relationship between motor variability and motor learning in both humans and animal models. We discuss neural circuit mechanisms that underlie the generation and regulation of motor variability and consider the implications that this work has for our understanding of motor learning.

Rules and mechanisms for efficient two-stage learning in neural circuits.

Trial-and-error learning requires evaluating variable actions and reinforcing successful variants. In songbirds, vocal exploration is induced by LMAN, the output of a basal ganglia-related circuit that also contributes a corrective bias to the vocal output. This bias is gradually consolidated in RA, a motor cortex analogue downstream of LMAN. We develop a new model of such two-stage learning. Using stochastic gradient descent, we derive how the activity in 'tutor' circuits (e.g., LMAN) should match plasticity mechanisms in 'student' circuits (e.g., RA) to achieve efficient learning. We further describe a reinforcement learning framework through which the tutor can build its teaching signal. We show that mismatches between the tutor signal and the plasticity mechanism can impair learning. Applied to birdsong, our results predict the temporal structure of the corrective bias from LMAN given a plasticity rule in RA. Our framework can be applied predictively to other paired brain areas showing two-stage learning.