Does a genetic predisposition towards excessive alcohol consumption influence the acquisition of habits? The study of alcohol dependency and its effects on learning and behavior continues to be a subject of growing importance. With an animal model of alcoholism, we sought to uncover some of these potential behavioral consequences. We focused on the development of both flexible and habitual spatial behaviors in rats selectively bred for High Alcohol Drinking (HAD) and Low Alcohol Drinking (LAD). Using a standard plus-maze task, rats were trained to run to an unchanging reward arm. We hypothesized that HAD rats would rely more on response learning strategies than spatial strategies, and would develop a habit more quickly than LAD rats. Initially, visual cues were placed around the room, and two probes were conducted to assess the progress of habitual behaviors in each animal. Later, cues were removed and one last probe was conducted. Preliminary data analysis revealed some unexpected findings, the most striking of which was seen in the LAD rats’ behavior. We expected both groups of animals to form a habitual response by the end of testing, however the percentage of LAD rats using spatial strategies increased with each successive probe. HAD rats, on the other hand, did appear to increasingly rely on response strategies over time, suggesting that they may be more prone towards developing a habitual behavior under these conditions. Additionally, a large number of HAD rats consistently omitted trials and demonstrated increased latencies when compared to LAD rats, possibly indicating that receiving a reward was not enough of an incentive to run the task. These preliminary findings may suggest that genetic preferences for alcohol consumption could influence the learning strategies employed in solving simple behavioral tasks.

It is important that animals be able to modulate the rate at which they learn. Without such a capacity animals might form habits based on single events, or fail to respond to changing circumstances. Previous work suggests that two populations of inhibitory striatal neurons may play antagonistic roles in modulating learning rates. Inactivation of D1 receptor expressing striatonigral neurons slows that rate at which animals learn operant tasks, while inactivation of D2 receptor expressing striatopallidal neurons speeds the rate at which animals learn. This suggests that the normal activity of striatonigral neurons supports learning while the normal activity of striatopallidal neurons suppresses learning. We assessed a maze-based operant learning task that may be useful for examining the roles of these neuronal populations in modulating learning rate. In this task, animals are repeatedly faced with the choice of a right or left turn. One of these turns consistently results in the animal receiving a large reward, while the other results in a small reward. The animal performs twenty trials per day. If after the first ten trials in a day the animal selects the large reward for at least nine trials, saline vehicle or a D1 or D2 receptor specific antagonist is infused into the nucleus accumbens of the animal and the rewarded turn is switched. This allows the assessment of learning rates under conditions of D1 and D2 receptor antagonism within individual animals. Preliminary results are not yet sufficiently extreme to establish a difference between D1 antagonism, D2 antagonism, and vehicle administration. This may result from insufficiently distinct reward magnitudes, an insufficient number of subjects, or from an insufficiently difficult task. Alternatively, it may be that while D1 and D2 receptor expressing neurons contribute antagonistically to learning rate, the activity of D1 and D2 receptors on these neurons does not.