Material below summarizes the article Learning and Stress Shape the Reward Response Patterns of Serotonin Neurons, published on September 13, 2017, in JNeurosci and authored by Weixin Zhong, Yi Li, Qiru Feng, and Minmin Luo.
The ability to predict future events is critical for the survival of an organism. Prediction via associative learning can prepare animals to gain rewards while avoiding disadvantages.
Classical Pavlovian conditioning is an important means to create predictive associations. During conditioning, a previously neutral item — unconditioned stimulus (US), which can be either rewarding or aversive, is repeatedly paired with a biologically salient stimulus — conditioned stimulus (CS). The CS acquires importance after conditioning if it consistently predicts the occurrence of the US.
The reward association process is believed to involve dopamine neurons in the ventral tegmental area (VTA). Dopamine neurons are activated by unexpected rewards. When animals have learned that a cue will follow a reward, the VTA dopamine neurons no longer respond to the expected reward. Instead, they are activated by the reward-predicting cue.
It has therefore been proposed that dopamine neurons encode the discrepancy between the predicted reward and the actual reward received (reward prediction error). It has also been proposed that the shift in phasic activation of dopamine neurons from the US to the CS occurs gradually across trials during the process of associative learning.
This proposal has been captured in the temporal difference (TD) model. Data simulated with this model matches very well with the response pattern of dopamine neurons before and after associative learning. However, it has been difficult to confirm this model by tracking the gradual shift of dopamine neuronal activity pattern throughout the entire process of associative learning.
In addition to VTA dopamine neurons, serotonin neurons in the dorsal raphe nucleus (DRN) participate in reward-related information processing.
Recent studies reported that DRN serotonin neurons were activated by primary rewards and reward-predicting cues, but there has been no data demonstrating the dynamic change in DRN serotonin neuron activity during reward associative learning. Particularly, we do not know whether the activity pattern of DRN serotonin neurons evolves similarly to that of VTA dopamine neurons.
Fiber photometry can overcome the limitation of in vivo electrophysiological recording and allow us to track neurons’ activity during a long-term associative learning process. We used fiber photometry to monitor the activities of genetically identified serotonin neurons in the DRN and dopamine neurons in the VTA during the entire process of associative learning.
First, we trained mice with Pavlovian conditioning by repetitively presenting an auditory cue (CS) with delayed delivery of a sucrose solution (US).
Initially, DRN serotonin neurons and VTA dopamine neurons showed similar activation when sucrose was received. As learning proceeded, however, the response pattern of serotonin neurons become clearly different from that of dopamine neurons.
Serotonin neurons gradually developed a slow ramp-up response to the reward-predicting cue, and ultimately remained activated when the US was received. In contrast, dopamine neurons gradually increased their response to the CS but reduced their response to the US.
The initial and final activity pattern of dopamine neurons was consistent with previous studies, but we did not observe a gradual propagation of neuronal activity from the US back to the CS.
Dopamine and serotonin neurons showed different responses when the reward value or delivery schedule was altered.
In a reward omission task, serotonin neurons gradually recovered the previously established ramp up response pattern when the reward reinstated. For dopamine neurons, reinstating reward produced a dramatically stronger activation compared to the response before extinction. In a reward value discrimination task, serotonin neurons show stronger CS and US responses to the larger reward size than that of the smaller reward size. In contrast, dopamine neurons responded only to the CS and the US of the larger reward value.
These results lead us to suggest that DRN serotonin neurons track the absolute reward value, whereas dopamine neurons track the relative reward value.
As is known, limited ability to adapt to stressful contexts cause defects in emotion control and may lead to psychological disorders such as depression, anxiety, and post-traumatic stress disorder. However, very little is known about how stress can directly influence the reward responses of DRN serotonin neurons and VTA dopamine neurons.
We therefore tested how stressors influence the reward response patterns of DRN serotonin neurons and VTA dopamine neurons.
We used two separate stressor treatments to challenge the mice: head restraint and fearful context.
For both DRN serotonin neurons and VTA dopamine neurons, we observed that acute stressors have a dramatic effect on reward responses: the head restraint and fearful context stressors substantially reduced the response strength of both neuron types. For mice that underwent Pavlovian conditioning, this reduction occurred with the reward-predicting cue and the reward itself.
Our study thus revealed that learning differentially shapes the response patterns of both neuron types, suggesting that DRN serotonin neurons and VTA dopamine neurons play complementary roles in reward processing.
Moreover, we revealed that acute stress reduces the extent of activation associated with rewards and reward-predicting cues, implying that acute stress may negatively affect reward processing by suppressing serotonin and dopamine signals.
Learning and Stress Shape the Reward Response Patterns of Serotonin Neurons. Weixin Zhong, Yi Li, Qiru Feng, and Minmin Luo. JNeurosci Sep 2017, 37 (37) 8863-8875; DOI: https://doi.org/10.1523/JNEUROSCI.1181-17.2017