Dynamic reinforcement learning reveals time-dependent shifts in strategy during reward learning.

Different brain systems have been hypothesized to subserve multiple "experts" that compete to generate behavior. In reinforcement learning, two general processes, one model-free (MF) and one model-based (MB), are often modeled as a mixture of agents (MoA) and hypothesized to capture differences between automaticity vs. deliberation. However, shifts in strategy cannot be captured by a static MoA. To investigate such dynamics, we present the mixture-of-agents hidden Markov model (MoA-HMM), which simultaneously learns inferred action values from a set of agents and the temporal dynamics of underlying "hidden" states that capture shifts in agent contributions over time. Applying this model to a multi-step,reward-guided task in rats reveals a progression of within-session strategies: a shift from initial MB exploration to MB exploitation, and finally to reduced engagement. The inferred states predict changes in both response time and OFC neural encoding during the task, suggesting that these states are capturing real shifts in dynamics.

Trial-history biases in evidence accumulation can give rise to apparent lapses in decision-making.

Trial history biases and lapses are two of the most common suboptimalities observed during perceptual decision-making. These suboptimalities are routinely assumed to arise from distinct processes. However, previous work has suggested that they covary in their prevalence and that their proposed neural substrates overlap. Here we demonstrate that during decision-making, history biases and apparent lapses can both arise from a common cognitive process that is optimal under mistaken beliefs that the world is changing i.e. nonstationary. This corresponds to an accumulation-to-bound model with history-dependent updates to the initial state of the accumulator. We test our model's predictions about the relative prevalence of history biases and lapses, and show that they are robustly borne out in two distinct decision-making datasets of male rats, including data from a novel reaction time task. Our model improves the ability to precisely predict decision-making dynamics within and across trials, by positing a process through which agents can generate quasi-stochastic choices.

Chronic brain functional ultrasound imaging in freely moving rodents performing cognitive tasks.

BACKGROUND: Functional ultrasound imaging (fUS) is an emerging imaging technique that indirectly measures neural activity via changes in blood volume. Chronic fUS imaging during cognitive tasks in freely moving animals faces multiple exceptional challenges: performing large durable craniotomies with chronic implants, designing behavioral experiments matching the hemodynamic timescale, stabilizing the ultrasound probe during freely moving behavior, accurately assessing motion artifacts, and validating that the animal can perform cognitive tasks while tethered. NEW METHOD: We provide validated solutions for those technical challenges. In addition, we present standardized step-by-step reproducible protocols, procedures, and data processing pipelines. Finally, we present proof-of-concept analysis of brain dynamics during a decision making task. RESULTS: We obtain stable recordings from which we can robustly decode task variables from fUS data over multiple months. Moreover, we find that brain wide imaging through hemodynamic response is nonlinearly related to cognitive variables, such as task difficulty, as compared to sensory responses previously explored. COMPARISON WITH EXISTING METHODS: Computational pipelines in fUS are nascent and we present an initial development of a full processing pathway to correct and segment fUS data. CONCLUSIONS: Our methods provide stable imaging and analysis of behavior with fUS that will enable new experimental paradigms in understanding brain-wide dynamics in naturalistic behaviors.

Flow-field inference from neural data using deep recurrent networks.

Computations involved in processes such as decision-making, working memory, and motor control are thought to emerge from the dynamics governing the collective activity of neurons in large populations. But the estimation of these dynamics remains a significant challenge. Here we introduce Flow-field Inference from Neural Data using deep Recurrent networks (FINDR), an unsupervised deep learning method that can infer low-dimensional nonlinear stochastic dynamics underlying neural population activity. Using population spike train data from frontal brain regions of rats performing an auditory decision-making task, we demonstrate that FINDR outperforms existing methods in capturing the heterogeneous responses of individual neurons. We further show that FINDR can discover interpretable low-dimensional dynamics when it is trained to disentangle task-relevant and irrelevant components of the neural population activity. Importantly, the low-dimensional nature of the learned dynamics allows for explicit visualization of flow fields and attractor structures. We suggest FINDR as a powerful method for revealing the low-dimensional task-relevant dynamics of neural populations and their associated computations.

Princeton RAtlas: A Common Coordinate Framework for Fully cleared, Whole Rattus norvegicus Brains.

Whole-brain clearing and imaging methods are becoming more common in mice but have yet to become standard in rats, at least partially due to inadequate clearing from most available protocols. Here, we build on recent mouse-tissue clearing and light-sheet imaging methods and develop and adapt them to rats. We first used cleared rat brains to create an open-source, 3D rat atlas at 25 µm resolution. We then registered and imported other existing labeled volumes and made all of the code and data available for the community (https://github.com/emilyjanedennis/PRA) to further enable modern, whole-brain neuroscience in the rat. Key features • This protocol adapts iDISCO (Renier et al., 2014) and uDISCO (Pan et al., 2016) tissue-clearing techniques to consistently clear rat brains. • This protocol also decreases the number of working hours per day to fit in an 8 h workday. Graphical overview.

Transitions in dynamical regime and neural mode underlie perceptual decision-making.

Perceptual decision-making is the process by which an animal uses sensory stimuli to choose an action or mental proposition. This process is thought to be mediated by neurons organized as attractor networks 1,2 . However, whether attractor dynamics underlie decision behavior and the complex neuronal responses remains unclear. Here we use an unsupervised, deep learning-based method to discover decision-related dynamics from the simultaneous activity of neurons in frontal cortex and striatum of rats while they accumulate pulsatile auditory evidence. We show that contrary to prevailing hypotheses, attractors play a role only after a transition from a regime in the dynamics that is strongly driven by inputs to one dominated by the intrinsic dynamics. The initial regime mediates evidence accumulation, and the subsequent intrinsic-dominant regime subserves decision commitment. This regime transition is coupled to a rapid reorganization in the representation of the decision process in the neural population (a change in the "neural mode" along which the process develops). A simplified model approximating the coupled transition in the dynamics and neural mode allows inferring, from each trial's neural activity, the internal decision commitment time in that trial, and captures diverse and complex single-neuron temporal profiles, such as ramping and stepping 3-5 . It also captures trial-averaged curved trajectories 6-8 , and reveals distinctions between brain regions. Our results show that the formation of a perceptual choice involves a rapid, coordinated transition in both the dynamical regime and the neural mode of the decision process, and suggest pairing deep learning and parsimonious models as a promising approach for understanding complex data.

Trial-history biases in evidence accumulation can give rise to apparent lapses.

Trial history biases and lapses are two of the most common suboptimalities observed during perceptual decision-making. These suboptimalities are routinely assumed to arise from distinct processes. However, several hints in the literature suggest that they covary in their prevalence and that their proposed neural substrates overlap - what could underlie these links? Here we demonstrate that history biases and apparent lapses can both arise from a common cognitive process that is normative under misbeliefs about non-stationarity in the world. This corresponds to an accumulation-to-bound model with history-dependent updates to the initial state of the accumulator. We test our model's predictions about the relative prevalence of history biases and lapses, and show that they are robustly borne out in two distinct rat decision-making datasets, including data from a novel reaction time task. Our model improves the ability to precisely predict decision-making dynamics within and across trials, by positing a process through which agents can generate quasi-stochastic choices.

Neural circuit models for evidence accumulation through choice-selective sequences.

Decision making is traditionally thought to be mediated by populations of neurons whose firing rates persistently accumulate evidence across time. However, recent decision-making experiments in rodents have observed neurons across the brain that fire sequentially as a function of spatial position or time, rather than persistently, with the subset of neurons in the sequence depending on the animal's choice. We develop two new candidate circuit models, in which evidence is encoded either in the relative firing rates of two competing chains of neurons or in the network location of a stereotyped pattern ("bump") of neural activity. Encoded evidence is then faithfully transferred between neuronal populations representing different positions or times. Neural recordings from four different brain regions during a decision-making task showed that, during the evidence accumulation period, different brain regions displayed tuning curves consistent with different candidate models for evidence accumulation. This work provides mechanistic models and potential neural substrates for how graded-value information may be precisely accumulated within and transferred between neural populations, a set of computations fundamental to many cognitive operations.

Corrigendum: An accumulation-of-evidence task using visual pulses for mice navigating in virtual reality.

[This corrects the article DOI: 10.3389/fnbeh.2018.00036.].

Value representations in the rodent orbitofrontal cortex drive learning, not choice.

Humans and animals make predictions about the rewards they expect to receive in different situations. In formal models of behavior, these predictions are known as value representations, and they play two very different roles. Firstly, they drive choice: the expected values of available options are compared to one another, and the best option is selected. Secondly, they support learning: expected values are compared to rewards actually received, and future expectations are updated accordingly. Whether these different functions are mediated by different neural representations remains an open question. Here, we employ a recently developed multi-step task for rats that computationally separates learning from choosing. We investigate the role of value representations in the rodent orbitofrontal cortex, a key structure for value-based cognition. Electrophysiological recordings and optogenetic perturbations indicate that these representations do not directly drive choice. Instead, they signal expected reward information to a learning process elsewhere in the brain that updates choice mechanisms.

Multiple timescales of sensory-evidence accumulation across the dorsal cortex.

Cortical areas seem to form a hierarchy of intrinsic timescales, but the relevance of this organization for cognitive behavior remains unknown. In particular, decisions requiring the gradual accrual of sensory evidence over time recruit widespread areas across this hierarchy. Here, we tested the hypothesis that this recruitment is related to the intrinsic integration timescales of these widespread areas. We trained mice to accumulate evidence over seconds while navigating in virtual reality and optogenetically silenced the activity of many cortical areas during different brief trial epochs. We found that the inactivation of all tested areas affected the evidence-accumulation computation. Specifically, we observed distinct changes in the weighting of sensory evidence occurring during and before silencing, such that frontal inactivations led to stronger deficits on long timescales than posterior cortical ones. Inactivation of a subset of frontal areas also led to moderate effects on behavioral processes beyond evidence accumulation. Moreover, large-scale cortical Ca2+ activity during task performance displayed different temporal integration windows. Our findings suggest that the intrinsic timescale hierarchy of distributed cortical areas is an important component of evidence-accumulation mechanisms.

Stable choice coding in rat frontal orienting fields across model-predicted changes of mind.

During decision making in a changing environment, evidence that may guide the decision accumulates until the point of action. In the rat, provisional choice is thought to be represented in frontal orienting fields (FOF), but this has only been tested in static environments where provisional and final decisions are not easily dissociated. Here, we characterize the representation of accumulated evidence in the FOF of rats performing a recently developed dynamic evidence accumulation task, which induces changes in the provisional decision, referred to as "changes of mind". We find that FOF encodes evidence throughout decision formation with a temporal gain modulation that rises until the period when the animal may need to act. Furthermore, reversals in FOF firing rates can be accounted for by changes of mind predicted using a model of the decision process fit only to behavioral data. Our results suggest that the FOF represents provisional decisions even in dynamic, uncertain environments, allowing for rapid motor execution when it is time to act.

Sequential and efficient neural-population coding of complex task information.

Recent work has highlighted that many types of variables are represented in each neocortical area. How can these many neural representations be organized together without interference and coherently maintained/updated through time? We recorded from excitatory neural populations in posterior cortices as mice performed a complex, dynamic task involving multiple interrelated variables. The neural encoding implied that highly correlated task variables were represented by less-correlated neural population modes, while pairs of neurons exhibited a spectrum of signal correlations. This finding relates to principles of efficient coding, but notably utilizes neural population modes as the encoding unit and suggests partial whitening of task-specific information where different variables are represented with different signal-to-noise levels. Remarkably, this encoding function was multiplexed with sequential neural dynamics yet reliably followed changes in task-variable correlations throughout the trial. We suggest that neural circuits can implement time-dependent encodings in a simple way using random sequential dynamics as a temporal scaffold.

Subpopulations of neurons in lOFC encode previous and current rewards at time of choice.

Studies of neural dynamics in lateral orbitofrontal cortex (lOFC) have shown that subsets of neurons that encode distinct aspects of behavior, such as value, may project to common downstream targets. However, it is unclear whether reward history, which may subserve lOFC's well-documented role in learning, is represented by functional subpopulations in lOFC. Previously, we analyzed neural recordings from rats performing a value-based decision-making task, and we documented trial-by-trial learning that required lOFC (Constantinople et al., 2019). Here, we characterize functional subpopulations of lOFC neurons during behavior, including their encoding of task variables. We found five distinct clusters of lOFC neurons, either based on clustering of their trial-averaged peristimulus time histograms (PSTHs), or a feature space defined by their average conditional firing rates aligned to different task variables. We observed weak encoding of reward attributes, but stronger encoding of reward history, the animal's left or right choice, and reward receipt across all clusters. Only one cluster, however, encoded the animal's reward history at the time shortly preceding the choice, suggesting a possible role in integrating previous and current trial outcomes at the time of choice. This cluster also exhibits qualitatively similar responses to identified corticostriatal projection neurons in a recent study (Hirokawa et al., 2019), and suggests a possible role for subpopulations of lOFC neurons in mediating trial-by-trial learning.

Interrogating theoretical models of neural computation with emergent property inference.

A cornerstone of theoretical neuroscience is the circuit model: a system of equations that captures a hypothesized neural mechanism. Such models are valuable when they give rise to an experimentally observed phenomenon -- whether behavioral or a pattern of neural activity -- and thus can offer insights into neural computation. The operation of these circuits, like all models, critically depends on the choice of model parameters. A key step is then to identify the model parameters consistent with observed phenomena: to solve the inverse problem. In this work, we present a novel technique, emergent property inference (EPI), that brings the modern probabilistic modeling toolkit to theoretical neuroscience. When theorizing circuit models, theoreticians predominantly focus on reproducing computational properties rather than a particular dataset. Our method uses deep neural networks to learn parameter distributions with these computational properties. This methodology is introduced through a motivational example of parameter inference in the stomatogastric ganglion. EPI is then shown to allow precise control over the behavior of inferred parameters and to scale in parameter dimension better than alternative techniques. In the remainder of this work, we present novel theoretical findings in models of primary visual cortex and superior colliculus, which were gained through the examination of complex parametric structure captured by EPI. Beyond its scientific contribution, this work illustrates the variety of analyses possible once deep learning is harnessed towards solving theoretical inverse problems.

Geometry of abstract learned knowledge in the hippocampus.

Hippocampal neurons encode physical variables1-7 such as space1 or auditory frequency6 in cognitive maps8. In addition, functional magnetic resonance imaging studies in humans have shown that the hippocampus can also encode more abstract, learned variables9-11. However, their integration into existing neural representations of physical variables12,13 is unknown. Here, using two-photon calcium imaging, we show that individual neurons in the dorsal hippocampus jointly encode accumulated evidence with spatial position in mice performing a decision-making task in virtual reality14-16. Nonlinear dimensionality reduction13 showed that population activity was well-described by approximately four to six latent variables, which suggests that neural activity is constrained to a low-dimensional manifold. Within this low-dimensional space, both physical and abstract variables were jointly mapped in an orderly manner, creating a geometric representation that we show is similar across mice. The existence of conjoined cognitive maps suggests that the hippocampus performs a general computation-the creation of task-specific low-dimensional manifolds that contain a geometric representation of learned knowledge.

Collicular circuits for flexible sensorimotor routing.

Context-based sensorimotor routing is a hallmark of executive control. Pharmacological inactivations in rats have implicated the midbrain superior colliculus (SC) in this process. But what specific role is this, and what circuit mechanisms support it? Here we report a subset of rat SC neurons that instantiate a specific link between the representations of context and motor choice. Moreover, these neurons encode animals' choice far earlier than other neurons in the SC or in the frontal cortex, suggesting that their neural dynamics lead choice computation. Optogenetic inactivations revealed that SC activity during context encoding is necessary for choice behavior, even while that choice behavior is robust to inactivations during choice formation. Searches for SC circuit models matching our experimental results identified key circuit predictions while revealing some a priori expected features as unnecessary. Our results reveal circuit mechanisms within the SC that implement response inhibition and context-based vector inversion during executive control.

Extracting the dynamics of behavior in sensory decision-making experiments.

Decision-making strategies evolve during training and can continue to vary even in well-trained animals. However, studies of sensory decision-making tend to characterize behavior in terms of a fixed psychometric function that is fit only after training is complete. Here, we present PsyTrack, a flexible method for inferring the trajectory of sensory decision-making strategies from choice data. We apply PsyTrack to training data from mice, rats, and human subjects learning to perform auditory and visual decision-making tasks. We show that it successfully captures trial-to-trial fluctuations in the weighting of sensory stimuli, bias, and task-irrelevant covariates such as choice and stimulus history. This analysis reveals dramatic differences in learning across mice and rapid adaptation to changes in task statistics. PsyTrack scales easily to large datasets and offers a powerful tool for quantifying time-varying behavior in a wide variety of animals and tasks.

Amplitude modulations of cortical sensory responses in pulsatile evidence accumulation.

How does the brain internally represent a sequence of sensory information that jointly drives a decision-making behavior? Studies of perceptual decision-making have often assumed that sensory cortices provide noisy but otherwise veridical sensory inputs to downstream processes that accumulate and drive decisions. However, sensory processing in even the earliest sensory cortices can be systematically modified by various external and internal contexts. We recorded from neuronal populations across posterior cortex as mice performed a navigational decision-making task based on accumulating randomly timed pulses of visual evidence. Even in V1, only a small fraction of active neurons had sensory-like responses time-locked to each pulse. Here, we focus on how these 'cue-locked' neurons exhibited a variety of amplitude modulations from sensory to cognitive, notably by choice and accumulated evidence. These task-related modulations affected a large fraction of cue-locked neurons across posterior cortex, suggesting that future models of behavior should account for such influences.

An approach for long-term, multi-probe Neuropixels recordings in unrestrained rats.

The use of Neuropixels probes for chronic neural recordings is in its infancy and initial studies leave questions about long-term stability and probe reusability unaddressed. Here, we demonstrate a new approach for chronic Neuropixels recordings over a period of months in freely moving rats. Our approach allows multiple probes per rat and multiple cycles of probe reuse. We found that hundreds of units could be recorded for multiple months, but that yields depended systematically on anatomical position. Explanted probes displayed a small increase in noise compared to unimplanted probes, but this was insufficient to impair future single-unit recordings. We conclude that cost-effective, multi-region, and multi-probe Neuropixels recordings can be carried out with high yields over multiple months in rats or other similarly sized animals. Our methods and observations may facilitate the standardization of chronic recording from Neuropixels probes in freely moving animals.