The role of the nucleus accumbens in reinforcer discounting
MetadataShow full item record
Delay and uncertainty decrease the value of rewards by a process known as discounting. More impulsive individuals discount the value of delayed rewards more rapidly, whereas more risk-prone individuals discount the value of probabilistic rewards less rapidly. In this study we used a psychophysical test procedure (adjusting amount) to determine the effects of excitotoxic lesions of the nucleus accumbens (NAc) on discounting of reward value by delay and probability. Prior to surgery, rats were trained on the delay (n = 24) and probability discounting (n = 24) tasks with either a constant delay of 4 s to reward or a constant probability of 0.4 of getting the reward. Excitotoxic lesions of the NAC were made by intracranial injections of 0.5 uL 0.15 M quinolinic acid (n = 12) or sham (n = 12) aimed at the NAc (AP + 1.6, ML ± 1.5, DV -7.1) in both the delay and probability trained animals. Following recovery from surgery, the animals were retested with a constant delay (4 s) and probability (0.4) of reinforcement. Under these conditions there was no effect of NAc lesion on delay or probability discounting. The animals were then tested under conditions where the delays (0, 1, 2, 4, & 8s) and probabilities (1.0, 0.7, 0.4, 0.2, & 0.1) were varied between test sessions in order to obtain discount functions. When the delay to reinforcement was varied between test sessions, lesioned rats showed flatter discount curves. In contrast, varying the probability of reinforcement had no differential effect. Upon completion of the discount functions, subjects were given a parametric series of challenges that revealed that lesioned rats were less sensitive to changes in delay and the magnitude of delayed rewards. The rats were also tested post-lesion on acquisition of radial arm maze performance. No impairments on this task were observed, indicating that working and reference memory was not affected. One interpretation of these results is that the NAc may play an important role in reward prediction and error signaling.