The medial forebrain bundle: Evidence for multiple systems of reward.

Abstract

In order to determine if multiple reward systems exist within the medial forebrain bundle, lateral hypothalamic (LH; anterior medial forebrain bundle) and ventral tegmental (VTA; posterior medial forebrain bundle) pulse frequencies that sustained the same proportion of the maximum self-stimulation rate (equipotent or threshold frequencies) were compared in a twin-lever psychophysical preference test. In a first experiment, the capacity of rats to discriminate between rewarding pulse frequencies delivered to the same reward site was determined. Results showed that rats equally preferred identical rewarding stimuli within 0.009 logfrequency units. Secondly, they discriminated between competing rewarding stimuli that differed by only 0.016 logfrequency units. Based on these criteria, it was hypothesized that if rats failed to equally prefer LH and VTA threshold frequencies, then each site may convey qualitatively different rewarding signals. Results of Experiment Two showed that LH and VTA threshold frequencies were equally preferred at one third of the pairs of sites tested. These data suggest that the medial forebrain bundle conveys equally rewarding signals at some electrode placements. However, at another third of the pairs of sites tested, rats failed to equally prefer anterior and posterior medial forebrain bundle threshold frequencies. These results suggest that this pathway may convey qualitatively different reward signals at other electrode placements. Mixed results were obtained at the remaining pairs of sites, where rats neither consistently chose one stimulus over the other, nor equally preferred both stimuli. A collision test was conducted in Experiment Three between all pairs of sites tested in the previous experiment in order to determine if the number of common reward fibres between electrodes was correlated with preference behavior. We rationalized that if the LH and VTA electrodes stimulated a substantial proportion of common rewarding fibres (as evidenced by a strong collision effect), then rats should equally prefer threshold frequencies. Unfortunately, a substantial collision effect was not noted in any of the subjects. Nevertheless, no correlation existed between modest collision effects or summation and preference behavior. Experiment Four examined the hypothesis that a reward-irrelevant factor, namely the co-activation of aversive neurons, contributed to the lack of equal preference observed at some pairs of sites. Results showed that concomitant aversive nucleus reticularis gigantocellularis (Gi) pulses did indeed initially increase the VTA self-stimulation threshold frequency. However, the aversive pulses progressively lost their capacity to do so with repeated testing. Furthermore, the preference test conducted in two rats showed that the animals did not prefer VTA over combined VTA+Gi stimulation. In sum, a modest degree of anatomical linkage between electrodes could not predict animals' preference behavior. Secondly, the co-activation of aversive fibres could not account for the lack of equal preference observed between some medial forebrain bundle sites. In conclusion, this series of experiments supports the notion of multiple reward systems within the medial forebrain bundle.