Pedunculopontine tegmental nucleus lesions impair probabilistic reversal learning by reducing sensitivity to positive reward feedback

Abstract

Recent findings indicate that pedunculopontine tegmental nucleus (PPTg) neurons encode reward-related information that is context-dependent. This information is critical for behavioral flexibility when reward outcomes change signaling a shift in response patterns should occur. The present experiment investigated whether NMDA lesions of the PPTg affects the acquisition and/or reversal learning of a spatial discrimination using probabilistic reinforcement. Male Long-Evans rats received a bilateral infusion of NMDA (30nmoles/side) or saline into the PPTg. Subsequently, rats were tested in a spatial discrimination test using a probabilistic learning procedure. One spatial location was rewarded with an 80% probability and the other spatial location rewarded with a 20% probability. After reaching acquisition criterion of 10 consecutive correct trials, the spatial location - reward contingencies were reversed in the following test session. Bilateral and unilateral PPTg-lesioned rats acquired the spatial discrimination test comparable to that as sham controls. In contrast, bilateral PPTg lesions, but not unilateral PPTg lesions, impaired reversal learning. The reversal learning deficit occurred because of increased regressions to the previously 'correct' spatial location after initially selecting the new, 'correct' choice. PPTg lesions also reduced the frequency of win-stay behavior early in the reversal learning session, but did not modify the frequency of lose-shift behavior during reversal learning. The present results suggest that the PPTg contributes to behavioral flexibility under conditions in which outcomes are uncertain, e.g. probabilistic reinforcement, by facilitating sensitivity to positive reward outcomes that allows the reliable execution of a new choice pattern.

Keywords: Brainstem; Pedunculopontine tegmental nucleus; Reinforcement; Reversal learning.

Figure 1

Photomicrographs of the PPTg region from a NMDA-lesioned rat (left side) and sham lesion rat (right side), approximately 7.8mm posterior to bregma, processed for NeuN immunoreactivity. Approximate borders of the PPTg are shown with dashed lines.

Figure 2

Representation of lesion damage in rats classified as having bilateral PPTg lesions. Numbers in the bottom left of each section indicate the distance that each section is from bregma (mm). The black shaded areas represents the minimum extent of neuronal loss for all bilaterally lesioned rats. The surrounding grey areas represents the maximum extent of neuronal loss in the bilateral lesion group. Sections were adapted from those in Paxinos & Watson (1998).

Figure 3

PPTg lesion effect on spatial acquisition and reversal learning. Neither unilateral or bilateral PPTg lesions affected trials to criterion in the acquisition phase. Bilateral PPTg lesions significantly increased trials to criterion in reversal compared to that of sham controls and the unilateral PPTg lesion group. There was not a significant difference in reversal learning performance between sham controls and the unilateral PPTg lesion group. * = p < 0.05 vs sham control and unilateral PPTg lesion group.

Figure 4

Errors committed during probabilistic reversal learning. All groups committed a similar number of perseverative errors during reversal learning. Bilateral PPTg lesions significantly increased regressive errors compared to that of sham controls and unilateral PPTg lesion lesions during reversal. There was not a significant difference in regressive errors between sham controls and the unilateral PPTg lesion group. * = p < 0.05 vs sham control and unilateral PPTg lesion group.

Figure 5

Overall win-stay and lose-shift performance during reversal learning. All groups exhibited comparable win-stay performance and lose-shift probabilities.

Figure 6

“Early” win-stay lose-shift analysis. Win-stay and lose-shift probabilities were analyzed on reversal learning trials until a rat made five consecutive correct choices. Bilateral PPTg lesions led to a significantly lower probability of win-stay trials compared to that of sham and the unilateral PPTg lesion group. There was not a significant difference in win-stay probabilities between the sham and unilateral PPTg lesion groups. All groups exhibited comparable lose-shift probabilities early in reversal learning. * = p < 0.05 vs sham control and unilateral PPTg lesion group.