Structure learning in action.

'Learning to learn' phenomena have been widely investigated in cognition, perception and more recently also in action. During concept learning tasks, for example, it has been suggested that characteristic features are abstracted from a set of examples with the consequence that learning of similar tasks is facilitated-a process termed 'learning to learn'. From a computational point of view such an extraction of invariants can be regarded as learning of an underlying structure. Here we review the evidence for structure learning as a 'learning to learn' mechanism, especially in sensorimotor control where the motor system has to adapt to variable environments. We review studies demonstrating that common features of variable environments are extracted during sensorimotor learning and exploited for efficient adaptation in novel tasks. We conclude that structure learning plays a fundamental role in skill learning and may underlie the unsurpassed flexibility and adaptability of the motor system.

Human muscle spindles act as forward sensory models.

Modern theories of motor control incorporate forward models that combine sensory information and motor commands to predict future sensory states. Such models circumvent unavoidable neural delays associated with on-line feedback control. Here we show that signals in human muscle spindle afferents during unconstrained wrist and finger movements predict future kinematic states of their parent muscle. Specifically, we show that the discharges of type Ia afferents are best correlated with the velocity of length changes in their parent muscles approximately 100-160 ms in the future and that their discharges vary depending on motor sequences in a way that cannot be explained by the state of their parent muscle alone. We therefore conclude that muscle spindles can act as "forward sensory models": they are affected both by the current state of their parent muscle and by efferent (fusimotor) control, and their discharges represent future kinematic states. If this conjecture is correct, then sensorimotor learning implies learning how to control not only the skeletal muscles but also the fusimotor system.

Motor coordination: when two have to act as one.

Trying to pass someone walking toward you in a narrow corridor is a familiar example of a two-person motor game that requires coordination. In this study, we investigate coordination in sensorimotor tasks that correspond to classic coordination games with multiple Nash equilibria, such as "choosing sides," "stag hunt," "chicken," and "battle of sexes". In these tasks, subjects made reaching movements reflecting their continuously evolving "decisions" while they received a continuous payoff in the form of a resistive force counteracting their movements. Successful coordination required two subjects to "choose" the same Nash equilibrium in this force-payoff landscape within a single reach. We found that on the majority of trials coordination was achieved. Compared to the proportion of trials in which miscoordination occurred, successful coordination was characterized by several distinct features: an increased mutual information between the players' movement endpoints, an increased joint entropy during the movements, and by differences in the timing of the players' responses. Moreover, we found that the probability of successful coordination depends on the players' initial distance from the Nash equilibria. Our results suggest that two-person coordination arises naturally in motor interactions and is facilitated by favorable initial positions, stereotypical motor pattern, and differences in response times.

Maintaining internal representations: The role of the human superior parietal lobe

In sensorimotor integration, sensory input and motor output signals are combined to provide an internal estimate of the state of both the world and one's own body. Although a single perceptual and motor snapshot can provide information about the current state, computational models show that the state can be optimally estimated by a recursive process in which an internal estimate is maintained and updated by the current sensory and motor signals. These models predict that an internal state estimate is maintained or stored in the brain. Here we report a patient with a lesion of the superior parietal lobe who shows both sensory and motor deficits consistent with an inability to maintain such an internal representation between updates. Our findings suggest that the superior parietal lobe is critical for sensorimotor integration, by maintaining an internal representation of the body's state.

Deliberation in the motor system: reflex gains track evolving evidence leading to a decision.

Both decision making and sensorimotor control require real-time processing of noisy information streams. Historically these processes were thought to operate sequentially: cognitive processing leads to a decision, and the outcome is passed to the motor system to be converted into action. Recently, it has been suggested that the decision process may provide a continuous flow of information to the motor system, allowing it to prepare in a graded fashion for the probable outcome. Such continuous flow is supported by electrophysiology in nonhuman primates. Here we provide direct evidence for the continuous flow of an evolving decision variable to the motor system in humans. Subjects viewed a dynamic random dot display and were asked to indicate their decision about direction by moving a handle to one of two targets. We probed the state of the motor system by perturbing the arm at random times during decision formation. Reflex gains were modulated by the strength and duration of motion, reflecting the accumulated evidence in support of the evolving decision. The magnitude and variance of these gains tracked a decision variable that explained the subject's decision accuracy. The findings support a continuous process linking the evolving computations associated with decision making and sensorimotor control.

Visuomotor feedback gains upregulate during the learning of novel dynamics.

At an early stage of learning novel dynamics, changes in muscle activity are mainly due to corrective feedback responses. These feedback contributions to the overall motor command are gradually reduced as feedforward control is learned. The temporary increased use of feedback could arise simply from the large errors in early learning with either unaltered gains or even slightly downregulated gains, or from an upregulation of the feedback gains when feedforward prediction is insufficient. We therefore investigated whether the sensorimotor control system alters feedback gains during adaptation to a novel force field generated by a robotic manipulandum. To probe the feedback gains throughout learning, we measured the magnitude of involuntary rapid visuomotor responses to rapid shifts in the visual location of the hand during reaching movements. We found large increases in the magnitude of the rapid visuomotor response whenever the dynamics changed: both when the force field was first presented, and when it was removed. We confirmed that these changes in feedback gain are not simply a byproduct of the change in background load, by demonstrating that this rapid visuomotor response is not load sensitive. Our results suggest that when the sensorimotor control system experiences errors, it increases the gain of the visuomotor feedback pathways to deal with the unexpected disturbances until the feedforward controller learns the appropriate dynamics. We suggest that these feedback gains are upregulated with increased uncertainty in the knowledge of the dynamics to counteract any errors or disturbances and ensure accurate and skillful movements.

Generalization in Adaptation to Stable and Unstable Dynamics

Humans skillfully manipulate objects and tools despite the inherent instability. In order to succeed at these tasks, the sensorimotor control system must build an internal representation of both the force and mechanical impedance. As it is not practical to either learn or store motor commands for every possible future action, the sensorimotor control system generalizes a control strategy for a range of movements based on learning performed over a set of movements. Here, we introduce a computational model for this learning and generalization, which specifies how to learn feedforward muscle activity in a function of the state space. Specifically, by incorporating co-activation as a function of error into the feedback command, we are able to derive an algorithm from a gradient descent minimization of motion error and effort, subject to maintaining a stability margin. This algorithm can be used to learn to coordinate any of a variety of motor primitives such as force fields, muscle synergies, physical models or artificial neural networks. This model for human learning and generalization is able to adapt to both stable and unstable dynamics, and provides a controller for generating efficient adaptive motor behavior in robots. Simulation results exhibit predictions consistent with all experiments on learning of novel dynamics requiring adaptation of force and impedance, and enable us to re-examine some of the previous interpretations of experiments on generalization. © 2012 Kadiallah et al.

The Effect of Motivation on Movement: A Study of Bradykinesia in Parkinson's Disease

Background: Bradykinesia is a cardinal feature of Parkinson's disease (PD). Despite its disabling impact, the precise cause of this symptom remains elusive. Recent thinking suggests that bradykinesia may be more than simply a manifestation of motor slowness, and may in part reflect a specific deficit in the operation of motivational vigour in the striatum. In this paper we test the hypothesis that movement time in PD can be modulated by the specific nature of the motivational salience of possible action-outcomes. Methodology/Principal Findings: We developed a novel movement time paradigm involving winnable rewards and avoidable painful electrical stimuli. The faster the subjects performed an action the more likely they were to win money (in appetitive blocks) or to avoid a painful shock (in aversive blocks). We compared PD patients when OFF dopaminergic medication with controls. Our key finding is that PD patients OFF dopaminergic medication move faster to avoid aversive outcomes (painful electric shocks) than to reap rewarding outcomes (winning money) and, unlike controls, do not speed up in the current trial having failed to win money in the previous one. We also demonstrate that sensitivity to distracting stimuli is valence specific. Conclusions/Significance: We suggest this pattern of results can be explained in terms of low dopamine levels in the Parkinsonian state leading to an insensitivity to appetitive outcomes, and thus an inability to modulate movement speed in the face of rewards. By comparison, sensitivity to aversive stimuli is relatively spared. Our findings point to a rarely described property of bradykinesia in PD, namely its selective regulation by everyday outcomes. © 2012 Shiner et al.

Model-based influences on humans' choices and striatal prediction errors.

The mesostriatal dopamine system is prominently implicated in model-free reinforcement learning, with fMRI BOLD signals in ventral striatum notably covarying with model-free prediction errors. However, latent learning and devaluation studies show that behavior also shows hallmarks of model-based planning, and the interaction between model-based and model-free values, prediction errors, and preferences is underexplored. We designed a multistep decision task in which model-based and model-free influences on human choice behavior could be distinguished. By showing that choices reflected both influences we could then test the purity of the ventral striatal BOLD signal as a model-free report. Contrary to expectations, the signal reflected both model-free and model-based predictions in proportions matching those that best explained choice behavior. These results challenge the notion of a separate model-free learner and suggest a more integrated computational architecture for high-level human decision-making.

Encoding of marginal utility across time in the human brain.

Marginal utility theory prescribes the relationship between the objective property of the magnitude of rewards and their subjective value. Despite its pervasive influence, however, there is remarkably little direct empirical evidence for such a theory of value, let alone of its neurobiological basis. We show that human preferences in an intertemporal choice task are best described by a model that integrates marginally diminishing utility with temporal discounting. Using functional magnetic resonance imaging, we show that activity in the dorsal striatum encodes both the marginal utility of rewards, over and above that which can be described by their magnitude alone, and the discounting associated with increasing time. In addition, our data show that dorsal striatum may be involved in integrating subjective valuation systems inherent to time and magnitude, thereby providing an overall metric of value used to guide choice behavior. Furthermore, during choice, we show that anterior cingulate activity correlates with the degree of difficulty associated with dissonance between value and time. Our data support an integrative architecture for decision making, revealing the neural representation of distinct subcomponents of value that may contribute to impulsivity and decisiveness.

The role of human orbitofrontal cortex in value comparison for incommensurable objects.

The human orbitofrontal cortex is strongly implicated in appetitive valuation. Whether its role extends to support comparative valuation necessary to explain probabilistic choice patterns for incommensurable goods is unknown. Using a binary choice paradigm, we derived the subjective values of different bundles of goods, under conditions of both gain and loss. We demonstrate that orbitofrontal activation reflects the difference in subjective value between available options, an effect evident across valuation for both gains and losses. In contrast, activation in dorsal striatum and supplementary motor areas reflects subjects' choice probabilities. These findings indicate that orbitofrontal cortex plays a pivotal role in valuation for incommensurable goods, a critical component process in human decision making.

Anchors, scales and the relative coding of value in the brain.

People are alarmingly susceptible to manipulations that change both their expectations and experience of the value of goods. Recent studies in behavioral economics suggest such variability reflects more than mere caprice. People commonly judge options and prices in relative terms, rather than absolutely, and display strong sensitivity to exemplar and price anchors. We propose that these findings elucidate important principles about reward processing in the brain. In particular, relative valuation may be a natural consequence of adaptive coding of neuronal firing to optimise sensitivity across large ranges of value. Furthermore, the initial apparent arbitrariness of value may reflect the brains' attempts to optimally integrate diverse sources of value-relevant information in the face of perceived uncertainty. Recent findings in neuroscience support both accounts, and implicate regions in the orbitofrontal cortex, striatum, and ventromedial prefrontal cortex in the construction of value.

Dopamine-dependent prediction errors underpin reward-seeking behaviour in humans.

Theories of instrumental learning are centred on understanding how success and failure are used to improve future decisions. These theories highlight a central role for reward prediction errors in updating the values associated with available actions. In animals, substantial evidence indicates that the neurotransmitter dopamine might have a key function in this type of learning, through its ability to modulate cortico-striatal synaptic efficacy. However, no direct evidence links dopamine, striatal activity and behavioural choice in humans. Here we show that, during instrumental learning, the magnitude of reward prediction error expressed in the striatum is modulated by the administration of drugs enhancing (3,4-dihydroxy-L-phenylalanine; L-DOPA) or reducing (haloperidol) dopaminergic function. Accordingly, subjects treated with L-DOPA have a greater propensity to choose the most rewarding action relative to subjects treated with haloperidol. Furthermore, incorporating the magnitude of the prediction errors into a standard action-value learning algorithm accurately reproduced subjects' behavioural choices under the different drug conditions. We conclude that dopamine-dependent modulation of striatal activity can account for how the human brain uses reward prediction errors to improve future decisions.

Cortical substrates for exploratory decisions in humans.

Decision making in an uncertain environment poses a conflict between the opposing demands of gathering and exploiting information. In a classic illustration of this 'exploration-exploitation' dilemma, a gambler choosing between multiple slot machines balances the desire to select what seems, on the basis of accumulated experience, the richest option, against the desire to choose a less familiar option that might turn out more advantageous (and thereby provide information for improving future decisions). Far from representing idle curiosity, such exploration is often critical for organisms to discover how best to harvest resources such as food and water. In appetitive choice, substantial experimental evidence, underpinned by computational reinforcement learning (RL) theory, indicates that a dopaminergic, striatal and medial prefrontal network mediates learning to exploit. In contrast, although exploration has been well studied from both theoretical and ethological perspectives, its neural substrates are much less clear. Here we show, in a gambling task, that human subjects' choices can be characterized by a computationally well-regarded strategy for addressing the explore/exploit dilemma. Furthermore, using this characterization to classify decisions as exploratory or exploitative, we employ functional magnetic resonance imaging to show that the frontopolar cortex and intraparietal sulcus are preferentially active during exploratory decisions. In contrast, regions of striatum and ventromedial prefrontal cortex exhibit activity characteristic of an involvement in value-based exploitative decision making. The results suggest a model of action selection under uncertainty that involves switching between exploratory and exploitative behavioural modes, and provide a computationally precise characterization of the contribution of key decision-related brain systems to each of these functions.

Temporal difference models describe higher-order learning in humans.

The ability to use environmental stimuli to predict impending harm is critical for survival. Such predictions should be available as early as they are reliable. In pavlovian conditioning, chains of successively earlier predictors are studied in terms of higher-order relationships, and have inspired computational theories such as temporal difference learning. However, there is at present no adequate neurobiological account of how this learning occurs. Here, in a functional magnetic resonance imaging (fMRI) study of higher-order aversive conditioning, we describe a key computational strategy that humans use to learn predictions about pain. We show that neural activity in the ventral striatum and the anterior insula displays a marked correspondence to the signals for sequential learning predicted by temporal difference models. This result reveals a flexible aversive learning process ideally suited to the changing and uncertain nature of real-world environments. Taken with existing data on reward learning, our results suggest a critical role for the ventral striatum in integrating complex appetitive and aversive predictions to coordinate behaviour.