Reinforcement learning, one of the most active research areas in artificial intelligence, is a computational approach to learning whereby an agent tries to maximize the total amount of reward it receives when interacting with a complex, uncertain environment. In Reinforcement Learning, Richard Sutton and Andrew Barto provide a clear and simple account of the key ideas and algorithms of reinforcement learning. Their discussion ranges from the history of the field's intellectual foundations to the most recent developments and applications. The only necessary mathematical background is familiarity with elementary concepts of probability. The book is divided into three parts. Part I defines the reinforcement learning problem in terms of Markov decision processes. Part II provides basic solution methods: dynamic programming, Monte Carlo methods, and temporal-difference learning. Part III presents a unified view of the solution methods and incorporates artificial neural networks, eligibility traces, and planning; the two final chapters present case studies and consider the future of reinforcement learning.

/pdf/reinforcement-learning-an-introduction-rzxgej9p17.pdf

Reinforcement Learning: An Introduction

I have developed "tennis elbow" from lugging this book around the past four weeks, but it is worth the pain, the effort, and the aspirin. It is also worth the (relatively speaking) bargain price. Including appendixes, this book contains 894 pages of text. The entire panorama of the neural sciences is surveyed and examined, and it is comprehensive in its scope, from genomes to social behaviors. The editors explicitly state that the book is designed as "an introductory text for students of biology, behavior, and medicine," but it is hard to imagine any audience, interested in any fragment of neuroscience at any level of sophistication, that would not enjoy this book. The editors have done a masterful job of weaving together the biologic, the behavioral, and the clinical sciences into a single tapestry in which everyone from the molecular biologist to the practicing psychiatrist can find and appreciate his or

Principles of Neural Science

http://umh1946.edu.umh.es/wp-content/uploads/sites/172/2015/04/Neurocircuitry-of-addiction1.pdf

Neurocircuitry of Addiction

Drug addiction is increasingly viewed as the endpoint of a series of transitions from initial drug use--when a drug is voluntarily taken because it has reinforcing, often hedonic, effects--through loss of control over this behavior, such that it becomes habitual and ultimately compulsive. Here we discuss evidence that these transitions depend on interactions between pavlovian and instrumental learning processes. We hypothesize that the change from voluntary drug use to more habitual and compulsive drug use represents a transition at the neural level from prefrontal cortical to striatal control over drug seeking and drug taking behavior as well as a progression from ventral to more dorsal domains of the striatum, involving its dopaminergic innervation. These neural transitions may themselves depend on the neuroplasticity in both cortical and striatal structures that is induced by chronic self-administration of drugs.

/pdf/neural-systems-of-reinforcement-for-drug-addiction-from-3s2f74bu84.pdf

Neural systems of reinforcement for drug addiction: from actions to habits to compulsion

A broad range of neural and behavioral data suggests that the brain contains multiple systems for behavioral choice, including one associated with prefrontal cortex and another with dorsolateral striatum. However, such a surfeit of control raises an additional choice problem: how to arbitrate between the systems when they disagree. Here, we consider dual-action choice systems from a normative perspective, using the computational theory of reinforcement learning. We identify a key trade-off pitting computational simplicity against the flexible and statistically efficient use of experience. The trade-off is realized in a competition between the dorsolateral striatal and prefrontal systems. We suggest a Bayesian principle of arbitration between them according to uncertainty, so each controller is deployed when it should be most accurate. This provides a unifying account of a wealth of experimental evidence about the factors favoring dominance by either system.

/pdf/uncertainty-based-competition-between-prefrontal-and-s5b6rfvmil.pdf

Uncertainty-based competition between prefrontal and dorsolateral striatal systems for behavioral control

Many organisms, especially humans, are characterized by their capacity for intentional, goal-directed actions. However, similar behaviours often proceed automatically, as habitual responses to antecedent stimuli. How are goal-directed actions transformed into habitual responses? Recent work combining modern behavioural assays and neurobiological analysis of the basal ganglia has begun to yield insights into the neural basis of habit formation.

The role of the basal ganglia in habit formation

Habits are controlled by antecedent stimuli rather than by goal expectancy. Interval schedules of feedback have been shown to generate habits, as revealed by the insensitivity of behaviour acquired under this schedule to outcome devaluation treatments. Two experiments were conducted to assess the role of the dorsolateral striatum in habit learning. In Experiment 1, sham operated controls and rats with dorsolateral striatum lesions were trained to press a lever for sucrose under interval schedules. After training, the sucrose was devalued by inducing taste aversion to it using lithium chloride, whereas saline injections were given to the controls. Only rats given the devaluation treatment reduced their consumption of sucrose and this reduction was similar in both the sham and the lesioned groups. All rats were then returned to the instrumental chamber for an extinction test, in which the lever was extended but no sucrose was delivered. In contrast to sham operated controls, rats with dorsolateral striatum lesions refrained from pressing the lever if the outcome was devalued. To assess the specificity of the role of dorsolateral striatum in this effect a second experiment was conducted in which a group with lesions of dorsomedial striatum was added. In relation now to both the sham and the dorsomedial lesioned groups, only rats with lesions of dorsolateral striatum significantly reduced responding after outcome devaluation. In conclusion, this study provides direct evidence that the dorsolateral striatum is necessary for habit formation. Furthermore, it suggests that, when the habit system is disrupted, control over instrumental performance reverts to the system controlling the performance of goal-directed instrumental actions.

Lesions of dorsolateral striatum preserve outcome expectancy but disrupt habit formation in instrumental learning.

Considerable evidence suggests that, in instrumental conditioning, rats learn the relationship between actions and their specific consequences or outcomes. The present study examined the role of the dorsomedial striatum (DMS) in this type of learning after excitotoxic lesions and reversible, muscimol-induced inactivation. In three experiments, rats were first trained to press two levers for distinct outcomes, and then tested after training using a variety of behavioural assays that have been established to detect action-outcome learning. In Experiment 1, pre-training lesions of the posterior DMS abolished the sensitivity of rats' instrumental performance to both outcome devaluation and contingency degradation when tested in extinction, whereas lesions of the anterior DMS had no effect. In Experiment 2, both pre-training and post-training lesions of the posterior DMS were equally effective in reducing the sensitivity of performance both to devaluation and degradation treatments. In Experiment 3, the infusion of muscimol into the posterior DMS selectively abolished sensitivity of performance to devaluation and contingency degradation without impairing the ability of rats to discriminate either the instrumental actions performed or the identity of the earned outcomes. Taken together, these results suggest that the posterior region of the DMS is a crucial neural substrate for the acquisition and expression of action-outcome associations in instrumental conditioning.

/pdf/the-role-of-the-dorsomedial-striatum-in-instrumental-1dntwb2tqs.pdf

The role of the dorsomedial striatum in instrumental conditioning.

The learning of new skills is characterized by an initial phase of rapid improvement in performance and a phase of more gradual improvements as skills are automatized and performance asymptotes. Using in vivo striatal recordings, we observed region-specific changes in neural activity during the different phases of skill learning, with the associative or dorsomedial striatum being preferentially engaged early in training and the sensorimotor or dorsolateral striatum being engaged later in training. Ex vivo recordings from medium spiny striatal neurons in brain slices of trained mice revealed that the changes observed in vivo corresponded to regional- and training-specific changes in excitatory synaptic transmission in the striatum. Furthermore, the potentiation of glutamatergic transmission observed in dorsolateral striatum after extensive training was preferentially expressed in striatopallidal neurons, rather than striatonigral neurons. These findings demonstrate that region- and pathway-specific plasticity sculpts the circuits involved in the performance of the skill as it becomes automatized.

/pdf/dynamic-reorganization-of-striatal-circuits-during-the-3qvn0ygrvl.pdf

Dynamic reorganization of striatal circuits during the acquisition and consolidation of a skill.

https://people.duke.edu/~hy43/omission.pdf

Henry H. Yin

Papers

The role of the basal ganglia in habit formation

Lesions of dorsolateral striatum preserve outcome expectancy but disrupt habit formation in instrumental learning.

The role of the dorsomedial striatum in instrumental conditioning.

Dynamic reorganization of striatal circuits during the acquisition and consolidation of a skill.

Inactivation of dorsolateral striatum enhances sensitivity to changes in the action–outcome contingency in instrumental conditioning