Development of a rodent probabilistic reward task used to objectively assess human anhedonia: Implications for translational research
Also presented as part of satellite meeting for the Society of Neuroscience.
Patients with major depression display dysfunctional processing of positively reinforcing stimuli when assessed using a probabilistic reward task [Pizzagalli et al. (2008) J Psychiatr Res 43(1)]. In this task, human subjects are exposed to two visual stimuli that are difficult to discriminate. Correct responses to one stimulus are rewarded three times more frequently (“rich”) compared to correct responses to the other stimulus (“lean”). Healthy human subjects modulate their behavior during testing as a function of prior reinforcement by gradually developing a biased response for the “rich” stimulus, whereas depressed subjects fail to develop this response bias. The result is a quantitative task that objectively measures anhedonia, or decreased response to positively reinforcing stimuli. The goal of the present study was to develop a rat analogue of this probabilistic reward task that can be used to conduct translational preclinical studies. Food restricted Wistar and Long-Evans rats were trained in operant boxes to press a lever to receive a food pellet as a reward. Using a discrete-trial procedure, rats were then presented with one of two tones varying in duration, but identical in all other parameters. Each tone was paired with one of two levers so that rats were rewarded with a food pellet upon correctly pressing the lever associated with a particular tone duration. Once the tone-lever associations were learned, the two tone durations were made more ambiguous during a test session to allow differential positive reinforcement of correct responses to influence subsequent behavior. Correct responses on the lever associated with either the short or the long tone (counterbalanced) were reinforced three times more (“rich”) than correct responses to the other tone (“lean”). During the initial one-third of the test session, when subjects are first exposed to the probabilistic nature of the reinforcement schedule, both Wistar and Long Evans rats exhibited a minimal response bias. During the final one-third of the test session, rats developed a significant response bias toward the “rich” stimulus, similar to the response bias quantified in healthy human subjects. Other task measures, such as ability to discriminate between the different tone stimuli and increased accuracy for the “rich” versus “lean” stimuli, were also comparable between rats and healthy human subjects. The results indicate that rats alter their response patterns over time as a function of reinforcement history, similar to humans. These data are the first to provide a rodent analogue of a probabilistic reward task in human subjects that can be used to quantify dysfunctional reward processing in animals.