Serotonin (5-HT) neurons in the dorsal raphe nucleus (DRN) provide extensive output to numerous brain areas and represent the primary source of 5-HT in the forebrain. 5-HT is involved in modulating and organizing a broad array of behaviors and drugs altering 5-HT levels treat depression and several other major psychiatric disorders. However, it has remained very challenging to assign a general computational role of these neurons in behavioral modulation. Particularly, it has been debated whether these neurons encode reward or punishment signals. Here, we will report our recent optogenetic and electrophysiological data and propose that DRN 5-HT neurons respond positively to reward and encode “beneficialness”, which indicates whether the current state is beneficial to an animal. Optogenetic stimulation of DRN Pet-1 neurons strongly reinforce animal behavior and effectively guide animal learning. Electrophysiology, mouse genetics, and behavioral assays reveal that Pet-1 neurons produce reward signals through the release of both 5-HT and glutamate (1,2). Both fiber photometry and single-unit recording from 5-HT neurons in the DRN of behaving mice show that natural rewards including sucrose, food, sex, and social interaction phasically activate 5-HT neurons. Both expected and unexpected rewards activate 5-HT neurons (3). By conditioning the sucrose delivery with a preceding auditory tone, 5-HT neurons gradually increase their response during the tone and reach the peak at the point of receiving sucrose. In contrast, aversive stimuli and stress do not activate 5-HT neurons, but they inhibit the response to sucrose either during expectation phase or after consumption phase. Finally, we show that, after training, the response profiles of 5-HT neurons and dopamine neurons differ substantially during both anticipation and consumption phases. Thus, DRN 5-HT neurons positively encode a wide range of reward signals during anticipatory and consummatory phases of reward responses. These results lead me to hypothesize that DRN 5-HT neurons encode beneficialness (B), which is determined by pR-C, where p indicates reward probability, R the predicted reward value, and C the cost (4,5). The beneficialness model may explain the extremely diverse behavioral roles of 5-HT in modulating different behaviors.
Biography
Dr. Minmin Luo is an Investigator at the National Institute of Biological Sciences, Beijing.