Drug addicts display an inability to control drug-seeking behavior despite adverse consequences. Such an inability may be mediated in part by dysfunction of the orbitofrontal cortex (OFC), which has long been implicated in cognitive flexibility. In accord with this hypothesis, chronic cocaine users have demonstrated behavioral deficits in reversal learning tasks, which are exquisitely sensitive to manipulations that disrupt OFC function. However, the precise role that the OFC plays in supporting adaptive behavior in reversal learning remains controversial. As a result, it is difficult to do more than speculate on the neural changes affecting OFC-dependent reversal learning in drug addicts. In addition, while it has been postulated that the ability of drugs of abuse to increase dopamine in limbic regions is crucial for their reinforcing effects, the role of dopamine in drug addiction is much less clear. The reinforcing effects of drugs of abuse may explain the initial drug-taking behavior, but the reinforcing effects per se seem insufficient to explain the compulsive drug intake and the loss of control in the addicts. Here, we propose to test the hypothesis that the OFC contributes to behavioral control in part via support of negative prediction error signaling by the mesolimbic dopaminergic system. We hypothesize that the OFC provides critical information about expected outcomes that leads to the generation of negative prediction error signals in the ventral tegmental area (VTA) to facilitate associative learning. The breakdown of this normal control of the OFC over the VTA in drug addicts may cause abnormal phasic dopamine release in other brain regions, leading to an inability to rapidly engage new learning processes and resulting in abnormal, stimulus-driven behaviors such as compulsive drug-seeking and relapse. Using a Pavlovian over-expectation task, which enables dissociation of learning driven by negative prediction errors from the use of the resultant information to guide behavior, we have recently shown that a pharmacological inactivation of the OFC or VTA prevents learning from an outcome that is worse than expected. This result is consistent with the hypothesis outlined above, but it is still controversial. First, there is no direct evidence that phasic decreases in the firing of dopamine neurons actually drive associative learning as negative reward prediction errors are proposed to do. Second, there is no direct evidence that outcome expectancies signaled by the OFC are necessary for phasic declines in dopamine release in response to negative prediction errors. This proposal will address these specific questions using a combination of cutting-edge behavioral, electrochemical, and genetic techniques. These results will advance our understanding of the dynamic neural circuits underlying normal learning from negative prediction errors, and thus may lead to better understanding of abnormal behaviors such as the compulsive drug seeking of drug addicts.
Changes in how the brain encodes associative information may play a causative role in many neuropsychiatric diseases that impact the health of individuals and society as a whole. For example, addiction is characterized by relapse even after prolonged treatment. Relapse is often driven by cues or contexts that the addict has learned to associate with drug taking. Understanding the processes and brain circuits governing such associative learning will help us understand why these memories are so resistant to change in addicts. This will help us devise novel therapeutic approaches. Experiments in this proposal will further our understanding of the brain circuits mediating associative learning.
| Ogawa, Masaaki; van der Meer, Matthijs A A; Esber, Guillem R et al. (2013) Risk-responsive orbitofrontal neurons track acquired salience. Neuron 77:251-8 |
