Learning leaves a memory trace in motor cortex.

How are we able to learn new behaviors without disrupting previously learned ones? To understand how the brain achieves this, we used a brain-computer interface (BCI) learning paradigm, which enables us to detect the presence of a memory of one behavior while performing another. We found that learning to use a new BCI map altered the neural activity that monkeys produced when they returned to using a familiar BCI map in a way that was specific to the learning experience. That is, learning left a "memory trace" in the primary motor cortex. This memory trace coexisted with proficient performance under the familiar map, primarily by altering neural activity in dimensions that did not impact behavior. Forming memory traces might be how the brain is able to provide for the joint learning of multiple behaviors without interference.

The role of prospective contingency in the control of behavior and dopamine signals during associative learning.

Associative learning depends on contingency, the degree to which a stimulus predicts an outcome. Despite its importance, the neural mechanisms linking contingency to behavior remain elusive. Here we examined the dopamine activity in the ventral striatum - a signal implicated in associative learning - in a Pavlovian contingency degradation task in mice. We show that both anticipatory licking and dopamine responses to a conditioned stimulus decreased when additional rewards were delivered uncued, but remained unchanged if additional rewards were cued. These results conflict with contingency-based accounts using a traditional definition of contingency or a novel causal learning model (ANCCR), but can be explained by temporal difference (TD) learning models equipped with an appropriate inter-trial-interval (ITI) state representation. Recurrent neural networks trained within a TD framework develop state representations like our best 'handcrafted' model. Our findings suggest that the TD error can be a measure that describes both contingency and dopaminergic activity.

Emergence of belief-like representations through reinforcement learning.

To behave adaptively, animals must learn to predict future reward, or value. To do this, animals are thought to learn reward predictions using reinforcement learning. However, in contrast to classical models, animals must learn to estimate value using only incomplete state information. Previous work suggests that animals estimate value in partially observable tasks by first forming "beliefs"-optimal Bayesian estimates of the hidden states in the task. Although this is one way to solve the problem of partial observability, it is not the only way, nor is it the most computationally scalable solution in complex, real-world environments. Here we show that a recurrent neural network (RNN) can learn to estimate value directly from observations, generating reward prediction errors that resemble those observed experimentally, without any explicit objective of estimating beliefs. We integrate statistical, functional, and dynamical systems perspectives on beliefs to show that the RNN's learned representation encodes belief information, but only when the RNN's capacity is sufficiently large. These results illustrate how animals can estimate value in tasks without explicitly estimating beliefs, yielding a representation useful for systems with limited capacity.

Emergence of belief-like representations through reinforcement learning.

To behave adaptively, animals must learn to predict future reward, or value. To do this, animals are thought to learn reward predictions using reinforcement learning. However, in contrast to classical models, animals must learn to estimate value using only incomplete state information. Previous work suggests that animals estimate value in partially observable tasks by first forming "beliefs"-optimal Bayesian estimates of the hidden states in the task. Although this is one way to solve the problem of partial observability, it is not the only way, nor is it the most computationally scalable solution in complex, real-world environments. Here we show that a recurrent neural network (RNN) can learn to estimate value directly from observations, generating reward prediction errors that resemble those observed experimentally, without any explicit objective of estimating beliefs. We integrate statistical, functional, and dynamical systems perspectives on beliefs to show that the RNN's learned representation encodes belief information, but only when the RNN's capacity is sufficiently large. These results illustrate how animals can estimate value in tasks without explicitly estimating beliefs, yielding a representation useful for systems with limited capacity. Author Summary: Natural environments are full of uncertainty. For example, just because my fridge had food in it yesterday does not mean it will have food today. Despite such uncertainty, animals can estimate which states and actions are the most valuable. Previous work suggests that animals estimate value using a brain area called the basal ganglia, using a process resembling a reinforcement learning algorithm called TD learning. However, traditional reinforcement learning algorithms cannot accurately estimate value in environments with state uncertainty (e.g., when my fridge's contents are unknown). One way around this problem is if agents form "beliefs," a probabilistic estimate of how likely each state is, given any observations so far. However, estimating beliefs is a demanding process that may not be possible for animals in more complex environments. Here we show that an artificial recurrent neural network (RNN) trained with TD learning can estimate value from observations, without explicitly estimating beliefs. The trained RNN's error signals resembled the neural activity of dopamine neurons measured during the same task. Importantly, the RNN's activity resembled beliefs, but only when the RNN had enough capacity. This work illustrates how animals could estimate value in uncertain environments without needing to first form beliefs, which may be useful in environments where computing the true beliefs is too costly.

How learning unfolds in the brain: toward an optimization view.

How do changes in the brain lead to learning? To answer this question, consider an artificial neural network (ANN), where learning proceeds by optimizing a given objective or cost function. This "optimization framework" may provide new insights into how the brain learns, as many idiosyncratic features of neural activity can be recapitulated by an ANN trained to perform the same task. Nevertheless, there are key features of how neural population activity changes throughout learning that cannot be readily explained in terms of optimization and are not typically features of ANNs. Here we detail three of these features: (1) the inflexibility of neural variability throughout learning, (2) the use of multiple learning processes even during simple tasks, and (3) the presence of large task-nonspecific activity changes. We propose that understanding the role of these features in the brain will be key to describing biological learning using an optimization framework.

Learning is shaped by abrupt changes in neural engagement.

Internal states such as arousal, attention and motivation modulate brain-wide neural activity, but how these processes interact with learning is not well understood. During learning, the brain modifies its neural activity to improve behavior. How do internal states affect this process? Using a brain-computer interface learning paradigm in monkeys, we identified large, abrupt fluctuations in neural population activity in motor cortex indicative of arousal-like internal state changes, which we term 'neural engagement.' In a brain-computer interface, the causal relationship between neural activity and behavior is known, allowing us to understand how neural engagement impacted behavioral performance for different task goals. We observed stereotyped changes in neural engagement that occurred regardless of how they impacted performance. This allowed us to predict how quickly different task goals were learned. These results suggest that changes in internal states, even those seemingly unrelated to goal-seeking behavior, can systematically influence how behavior improves with learning.

New neural activity patterns emerge with long-term learning.

Learning has been associated with changes in the brain at every level of organization. However, it remains difficult to establish a causal link between specific changes in the brain and new behavioral abilities. We establish that new neural activity patterns emerge with learning. We demonstrate that these new neural activity patterns cause the new behavior. Thus, the formation of new patterns of neural population activity can underlie the learning of new skills.

Constraints on neural redundancy.

Millions of neurons drive the activity of hundreds of muscles, meaning many different neural population activity patterns could generate the same movement. Studies have suggested that these redundant (i.e. behaviorally equivalent) activity patterns may be beneficial for neural computation. However, it is unknown what constraints may limit the selection of different redundant activity patterns. We leveraged a brain-computer interface, allowing us to define precisely which neural activity patterns were redundant. Rhesus monkeys made cursor movements by modulating neural activity in primary motor cortex. We attempted to predict the observed distribution of redundant neural activity. Principles inspired by work on muscular redundancy did not accurately predict these distributions. Surprisingly, the distributions of redundant neural activity and task-relevant activity were coupled, which enabled accurate predictions of the distributions of redundant activity. This suggests limits on the extent to which redundancy may be exploited by the brain for computation.

A Distinct Mechanism of Temporal Integration for Motion through Depth.

Temporal integration of visual motion has been studied extensively within the frontoparallel plane (i.e., 2D). However, the majority of motion occurs within a 3D environment, and it is unknown whether the principles from 2D motion processing generalize to more realistic 3D motion. We therefore characterized and compared temporal integration underlying 2D (left/right) and 3D (toward/away) direction discrimination in human observers, varying motion coherence across a range of viewing durations. The resulting discrimination-versus-duration functions followed three stages, as follows: (1) a steep improvement during the first ∼150 ms, likely reflecting early sensory processing; (2) a subsequent, more gradual benefit of increasing duration over several hundreds of milliseconds, consistent with some form of temporal integration underlying decision formation; and (3) a final stage in which performance ceased to improve with duration over ∼1 s, which is consistent with an upper limit on integration. As previously found, improvements in 2D direction discrimination with time were consistent with near-perfect integration. In contrast, 3D motion sensitivity was lower overall and exhibited a substantial departure from perfect integration. These results confirm that there are overall differences in sensitivity for 2D and 3D motion that are consistent with a sensory difference between binocular and dichoptic sensory mechanisms. They also reveal a difference at the integration stage, in which 3D motion is not accumulated as perfectly as in the 2D motion model system.

Signal multiplexing and single-neuron computations in lateral intraparietal area during decision-making.

Previous work has revealed a remarkably direct neural correlate of decisions in the lateral intraparietal area (LIP). Specifically, firing rate has been observed to ramp up or down in a manner resembling the accumulation of evidence for a perceptual decision reported by making a saccade into (or away from) the neuron's response field (RF). However, this link between LIP response and decision formation emerged from studies where a saccadic target was always stimulating the RF during decisions, and where the neural correlate was the averaged activity of a restricted sample of neurons. Because LIP cells are (1) highly responsive to the presence of a visual stimulus in the RF, (2) heterogeneous, and (3) not clearly anatomically segregated from large numbers of neurons that fail selection criteria, the underlying neuronal computations are potentially obscured. To address this, we recorded single neuron spiking activity in LIP during a well-studied moving-dot direction-discrimination task and manipulated whether a saccade target was present in the RF during decision-making. We also recorded from a broad sample of LIP neurons, including ones conventionally excluded in prior studies. Our results show that cells multiplex decision signals with decision-irrelevant visual signals. We also observed disparate, repeating response "motifs" across neurons that, when averaged together, resemble traditional ramping decision signals. In sum, neural responses in LIP simultaneously carry decision signals and decision-irrelevant sensory signals while exhibiting diverse dynamics that reveal a broader range of neural computations than previously entertained.