Serotonergic projections to the orbitofrontal and medial prefrontal cortices differentially modulate waiting for future rewards

Abstract

Optogenetic activation of serotonergic neurons in the dorsal raphe nucleus (DRN) enhances patience when waiting for future rewards, and this effect is maximized by both high probability and high timing uncertainty of reward. Here, we explored which serotonin projection areas contribute to these effects using optogenetic axon terminal stimulation. We found that serotonin stimulation in the orbitofrontal cortex (OFC) is nearly as effective as that in the DRN for promoting waiting, while in the nucleus accumbens, it does not promote waiting. We also found that serotonin stimulation in the medial prefrontal cortex (mPFC) promotes waiting only when the timing of future rewards is uncertain. Our Bayesian decision model of waiting assumed that the OFC and mPFC calculate the posterior probability of reward delivery separately. These results suggest that serotonin in the mPFC affects evaluation of time committed, while serotonin in the OFC is responsible for overall valuation of delayed rewards.

Fig. 1. The sequential tone-food waiting task and location of optic fibers.

(A) Experimental apparatus of the tone-food waiting task. (B) Diagram of the tone-food waiting task in which optogenetic stimulation was applied during waiting for delayed food, defined as a reward-delay period. Each trial started with a nose poke into tone site for 0.3 s until an 8-kHz tone was presented. After tone presentation, mice had to continue nose poking at reward site until food presentation. Seventy-five percent of trials were rewarded (i.e., prior probability of the tone-food waiting task was 0.75). Four reward-delay tests in which the timing of reward delivery was changed were introduced (i.e., change of reward timing uncertainty). To examine how serotonergic neuron activation promotes waiting for delayed rewards, we focused on waiting time in the 25% of trials with no reward (i.e., omission). Mice had to nose poke at tone site again for the next trial. (C) Locations of optic fibers in the OFC and DRN and representative fiber trace for the OFC and DRN. Light blue circles in the OFC represent tip positions of optic fibers. Light blue bars in the DRN represent tracks of implanted optic fibers. Coronal drawings were adapted from (54) with permission.

Fig. 2. Optogenetic activation of the OFC enhances waiting in all reward delay conditions.

(A) Average waiting time in no-activation (yellow) and activation (blue) of OFC serotonergic neuron terminals during the four reward-delay tests. Gray lines indicate waiting times for individual ChR2-expressing mice (n = 5). (B) Average waiting time in no-activation (yellow) and activation (blue) of DRN serotonergic neurons during the four reward-delay tests. Gray lines indicate waiting times for individual ChR2-expressing mice (n = 5). (C) Distribution of waiting times during omission trials in no-activation (yellow) and activation (blue) of OFC serotonergic neuron terminals during the four reward-delay tests. (D) Distribution of waiting times during omission trials in no-activation (yellow) and activation (blue) of DRN serotonergic neurons during the four reward-delay tests. ***P < 0.001 by paired t test. Error bars represent the SEM.

Fig. 3. The effectiveness of OFC photostimulation in promoting patience is similar in fixed-delay tests, but less in uncertain timing compared with DRN photostimulation.

(A) Waiting-time ratios of OFC (left) and DRN (right) optogenetic stimulation in the four reward-delay tests. *P < 0.05, **P < 0.01, and ***P < 0.001 by post hoc Bonferroni correction. n.s., not significant. Error bars present the SEM. (B) Comparison of waiting-time ratios between OFC and DRN optogenetic stimulation in the four reward-delay tests. Gray lines indicate waiting-time ratios for individual ChR2-expressing mice (n = 5). *P < 0.05, **P < 0.01, and ***P < 0.001 by paired t test. Error bars represent the SEM.

Fig. 4. Optogenetic activation of the mPFC enhances waiting in high reward timing uncertainty.

(A) Average waiting time in no-activation (yellow) and activation (blue) of mPFC serotonergic neuron terminals during the four reward-delay tests. Gray lines indicate waiting time for individual ChR2-expressing mice (n = 5). (B) Distribution of waiting time during omission trials in no-activation (yellow) and activation (blue) of mPFC serotonergic neuron terminals during the four reward-delay tests. ***P < 0.001 by paired t test. Error bars represent the SEM. (C) Waiting-time ratios of mPFC (left) and DRN (right) optogenetic stimulation in the four reward-delay tests. *P < 0.05, **P < 0.01, and ***P < 0.001 by post hoc Bonferroni correction. Error bars present the SEM. (D) Comparison of waiting-time ratios between DRN and mPFC optogenetic stimulation in the four reward-delay tests. Gray lines indicate waiting-time ratios for individual ChR2-expressing mice (n = 5). *P < 0.05, **P < 0.01, and ***P < 0.001 by paired t test. Error bars represent the SEM.

Fig. 5. Optogenetic activation of NAc serotonergic neuron terminals does not enhance waiting.

(A) Average waiting time during no-activation (yellow) and activation (blue) of NAc serotonin neuron terminals during the four reward-delay tests. Gray lines indicate waiting time for individual ChR2-expressing mice (n = 5). (B) Distribution of waiting times during omission trials in no-activation (yellow) and activation (blue) of NAc serotonin neuron terminals during the four reward-delay tests. ***P < 0.001 by paired t test. Error bars represent the SEM. (C) Waiting-time ratios of NAc (left) and DRN (right) optogenetic stimulation in the four reward-delay tests. *P < 0.05, **P < 0.01, and ***P < 0.001 by post hoc Bonferroni correction. Error bars present the SEM. (D) Comparison of waiting-time ratios of DRN and NAc optogenetic stimulation in the four reward-delay tests. Gray lines indicate waiting-time ratios for individual ChR2-expressing mice (n = 5). *P < 0.05, **P < 0.01, and ***P < 0.001 by paired t test. Error bars represent the SEM.

Fig. 6. A model assuming that the OFC and mPFC independently calculate posterior probabilities reproduces features of OFC, mPFC, and DRN photostimulation.

(A) The model assumes that the OFC and mPFC have individual probabilistic models of reward delivery timing (red lines for OFC and black lines for mPFC), which are assumed to be gamma distributions donated by μ and σ. As the time passes without reward delivery, the likelihood of a reward trial diminishes according to the cumulative density function (green lines for OFC and magenta lines for mPFC). (B) Simulation of waiting distribution change caused by OFC optogenetic activation. OFC photostimulation shifts the prior probability of the OFC [Prior(OFC)] from 0.75 to 0.94 and keeps the prior probability of the mPFC [Prior(mPFC)] 0.75. (C) Simulation of waiting distribution change by mPFC optogenetic activation. mPFC photostimulation shifts Prior(mPFC) from 0.75 to 0.94 and keeps Prior(OFC) 0.75. (D) Simulation of waiting distribution change by DRN optogenetic activation. DRN photostimulation shifts both Prior(OFC) and Prior(mPFC) from 0.75 to 0.94. Blue and orange lines show the time of quitting with and without increased prior probability, respectively. Yellow and blue shaded regions indicate distribution of waiting times during omission trials in no-activation and activation of serotonergic neurons, respectively.