Geoffrey Schoenbaum - US grants

DESCRIPTION (provided by applicant): The experiments in this proposal are designed to investigate changes in normal brain function caused by cocaine. These experiments will focus on function in a circuit of structures including orbitofrontal cortex, basolateral amygdala, and nucleus accumbens. Experiments in rats and primates have shown that this circuit - largely conserved across species - is critical for goal-directed behavior or behavior that is guided by associations between cues and the incentive value of associated outcomes. It is now known that chronic intermittent exposure to cocaine - the stereotypical drug of abuse - alters molecular and structural properties of neurons in these brain areas. A growing set of reports indicates that these changes may cause fundamental changes in the normal functions that depend on these regions, changes that may be related to the complex behavioral changes that characterize addiction. However, animal models are necessary to determine whether changes observed in humans are caused by cocaine. To investigate this question, we propose to use data regarding the contribution of these areas to goal-directed behavior in normal rats as a background or model in which to investigate the effects of prior cocaine exposure. The proposed experiments have three components. The first component is a continued characterization of the involvement of these areas in goal-directed behavior in normal rats. The second component is an evaluation of the effects of chronic intermittent cocaine exposure on goal-directed behavior that is affected by lesions in these brain regions. The third component is a comparison of neural representations during goal-directed behavior in these brain regions in normal rats and rats that have received intermittent exposure to cocaine. This comparison will identify changes in neural activity or encoding that relate to the behavioral effects of prior cocaine exposure. These results will then serve as a platform for further investigations into whether and how changes that are identified in the normal operation of this brain circuit may be reversed by abstinence and/or treatment regimes. In addition, changes in function related to other human neuropsychiatric diseases involving these areas could also be pursued using this model.

[unreadable] DESCRIPTION (provided by applicant): The experiments in this proposal are designed to investigate the basis of changes in cognitive flexibility caused by normal aging. Subsets of elderly humans and rats appear to have difficulty changing behaviors when the likely outcomes of those behaviors change. These deficits are evident in tasks that test reversal learning and set-shifting. Reversal learning and set-shifting are critically dependent on two different subdivisions of prefrontal cortex in rats and primates. These circuits are implicated in control of different forms of associative learning. Here we will test the associative basis and circuits involved in these declines in normal aging. Findings will show whether declines in these two types of cognitive flexibility occur independently in the two prefrontal systems or whether they reflect a unitary phenomenon. Associated experiments will identify regional changes in neural activity, dendritic structure, and monoaminergic innervation that correlate with performance changes to identify the location and underlying cause of miscoding associated with the inflexible behavior. Lastly we will test the hypothesis that practice on a task that emphasizes learning in one of these associative domains results in a protective effect on later performance, after aging, within the same associative domain. Improved performance may then be correlated with regional changes in neural activity, dendritic structure and innervation. [unreadable] [unreadable] [unreadable]

DESCRIPTION (provided by applicant): Orbitofrontal dysfunction is implicated in a host of human neuropsychiatric disease, including depression, mania, obsessive-compulsive disorder, attention deficit hyperactivity disorder, autism, and addiction. In most of these disease states, the role of orbitofrontal cortex is only beginning to be understood. Existing theories rely on animal research showing that orbitofrontal cortex is critical to decision-making because of the role it plays in promoting associative learning and the inhibition of inappropriate responses. However recent work in animals now suggests that the conceptualization of orbitofrontal contributions to associative learning is incorrect. Instead orbitofrontal cortex may contribute to associative learning and decision-making due to its role in generating dynamic, prospective predictions regarding the likelihood of future outcomes. Such ?outcome-expectancies? would both guide appropriate decisions and also facilitate associative learning in other brain regions in the face of unexpected outcomes. Such a conceptualization would be fundamentally different from current theories that view orbitofrontal cortex as a static, retrospective ?associative look-up table?. The experiments in this proposal will provide a critical test of this hypothesis, manipulating rats'expectations for rewarding outcomes in two different settings, in order to demonstrate orbitofrontal cortex signals these expectations. Further we will show that these signals facilitate flexible behavior in the face of unexpected outcomes due to their ability to their ability to alter associative representations stored in downstream brain regions. These results will advance our understanding of the fundamental function of orbitofrontal cortex, beyond the current common conceptualization of this area. As such, they will clarify the role this area plays in the neuropsychiatric diseases described above and how manipulation of that role may facilitate treatment of these diseases.

PROJECT SUMMARY Conditioned reinforcement is the process whereby cues that have been paired with primary rewards support the acquisition and maintenance of new instrumental responses. Abnormal responding to such cues is a prominent feature of drug addiction, where drug-associated cues motivate drug-taking and relapse in addicts and in laboratory animals in animal models of drug seeking and relapse. The efficacy of drug-associated cues in animal models depends on circuits that take information from the basolateral amygdala (ABL) to orbitofrontal cortex (OFC) or nucleus accumbens (NAc);however, interpreting these data is hampered by our limited understanding of how these circuits (e.g., ABL-OFC or ABL-NAc) normally support conditioned reinforcement. Here we hypothesize that these circuits support conditioned reinforcement due to their differing roles in Pavlovian associative learning. Output from ABL to OFC is critical to allowing cues to become associated with the value of the outcome (e.g., drug or food availability) Pavlovian cues predict, whereas output from ABL to NAc appears to be critical for the formation of associations between cues and the general affect or emotion (e.g., happiness, fear) evoked by the outcome. Since addictive drugs have divergent effects on these two circuits, they may cause an imbalance in what associative information (outcome-specific mediated by ABL- OFC versus general affect mediated by ABL-NAc) is recruited by drug-associated cues to guide behavior. Here, we will test the hypotheses outlined above. We will train rats to associate cues with the appetitive reward (sucrose) using Pavlovian training procedures that emphasize one or the other type of representation (outcome-specific or general affect). We will then test whether these specialized cues support conditioned reinforcement in normal rats and in rats with lesions of the ABL-OFC or the ABL-NAc circuits. In addition, we will identify neural correlates of conditioned reinforcement for these specialized cues in each circuit. Finally, we will test the effects of cocaine self-administration on conditioned reinforcement for different types of cues and neural correlates of conditioned reinforcement. The results will provide a neuroanatomical framework with which to better understand how reward-associated cues exert effects on behavior under normal (drug-na[unreadable]ve) conditions and after exposure to cocaine.

Associative learning is supported by the ability to recognize errors between expected and actual outcomes. Evidence from primate and rat recording studies suggests that dopaminergic neurons in substantia nigra and the ventral tegmental area (VTA) signal such prediction errors. However, generating such prediction errors presumably requires the comparison of actual outcomes to expected outcomes. Perhaps the best candidate for signaling such outcome expectancies is the orbitofrontal cortex (OFC). This proposal will test this hypothesis, using inactivation and single-unit recording to ask how OFC is involved learning and in the calculation of prediction errors in dopaminergic VTA neurons during Pavlovian blocking.