Skip to main content Skip to secondary navigation
Main content start

Numerous Shades of the Raphe: Behavioral Heterogeneity in the Serotonin System

Box 9-1 introduces the neuromodulatory systems that have been discussed in multiple chapters of Principles of Neurobiology. Unlike other fast-acting neurotransmitters–such as glutamate and GABA–that take effect in a few milliseconds, neuromodulators may seem rather sluggish. Yet, neuromodulators induce longer-term and amplified effects via their interaction with G-protein-coupled receptors on postsynaptic neurons, thereby modulating postsynaptic neuronal activities.

Not only for their physiological impact, but the neuromodulatory systems are also renowned for their anatomical characteristics: as illustrated in Figure 9-32, the four neuromodulatory systems—the dopamine, serotonin, norepinephrine, and acetylcholine systems—nicely exemplify the extensive axonal projections that broadcast neuromodulatory signals to the whole brain.

Despite the general physiological and anatomical attributes they share, the detailed architectures of the neuromodulatory systems are quite different from each other. For example, while the norepinephrine neurons in the locus coeruleus share their collateralized projection targets forming a largely homogeneous projectomic cluster, midbrain dopamine and dorsal raphe (DR) serotonin populations can be further divided into multiple subgroups based on various criteria, such as their projection targets, physiological response properties, and behavioral functions.

A recent study by Paquelet and colleagues (2022) epitomizes the heterogeneity in the DR serotonin population. Unlike many of the previous studies that recorded population neuronal activities in the DR, the authors adapted microendoscopy and recorded the neural activities of >2,000 DR serotonin neurons in mice during a series of behavioral tasks—sucrose consumption, social interaction, random foot shock, and open arm exploration.

One take-home message is that the functional composition of the DR neuronal population is indeed very heterogenous: across the behavioral experiments, the authors observed that various portions (~20%–80%) of the recorded neurons were activated depending on the behavioral task the mice performed. This indicates that individual DR serotonin neurons have their own preferred emotional stimuli, as the activated neurons in one behavioral task did not necessarily respond to other behavioral tasks. Interestingly, serotonin neurons that responded to the same stimuli appeared to be loosely clustered in DR, implying that the DR functional subgroups may correspond to anatomical subdivisions in the raphe.

What would make the raphe neurons fire during a particular behavior? The authors answer this question in a clever way. First, they compared DR neuronal responses during sucrose or quinine consumption to that during water drinking. Interestingly, the majority of the neurons that were active during sucrose consumption were also activated during drinking quinine solution. This observation suggests that the DR serotonin neurons respond to emotional saliency, not the positive or negative valence of a given stimulus. In the next experiment, thirsty mice were trained to relate a neutral audiovisual stimulus to a water reward. Before training, only a small fraction of DR neurons was either excited or inhibited by the sensory stimulus; however, after the mice learned that the sensory input is associated with reward, there was a substantial increase in the proportion of raphe cells responding to the sensory cue. In other words, the originally “uninteresting” stimulus became emotionally meaningful. The contrariwise is also hinted in other behavioral experiments: DR neurons often showed decreased responses when mice were repeatedly exposed to the same emotional stimuli, implying that raphe serotonin cells can become blasé about once emotionally salient inputs.

It is noteworthy that, because the activities of single DR neurons were recorded, the neural activities of the same raphe serotonin neurons could be thoroughly examined and compared across the behavior experiments conducted. The authors take full advantage of this rich dataset to answer how individual neurons with mixed selectivity respond to multiple behavioral experiments conducted. From the pairwise behavior analysis, it was shown that the DR neurons were not strictly divided based on the type of behavioral task, as many neurons were activated by multiple combinations of stimuli that didn’t share noticeable values, such as valence and stimulus type. The only exception was the overlap between sugar consumption and foot shock, which sit at either extreme end of the valence axis.

Lastly, the authors examined whether DR serotonin neurons projecting to specific regions of the brain are shown coherent activities during behavioral tasks. To this end, the activities of two projection groups were recorded—one innervating the ventral tegmental area (VTA), a region strongly associated with reward, and the other projecting to the bed nucleus of stria terminalis (BNST), a main target of the DR neurons that play crucial roles in anxiety-related behaviors. Both the VTA- and BNST-projecting serotonin neurons had heterogenous responses just like the DR as a whole; however, interestingly, these populations were functionally biased: VTA-projecting neurons were preferentially activated during sucrose consumption, whereas BNST-projecting ones showed more responses during anxiety-inducing tasks, such as foot shocks and elevated plus maze.

Together, these findings again highlight the heterogeneity of the serotonin system. What are the biological advantages that arise from this high level of complexity in such a small group of neurons? What is the most efficient way to systematically describe such a heterogeneous system? In fact, it is the heterogeneity that makes the serotonin system more attractive to study!

 

Figure 1.  (Top row) The behavioral tests used by Paquelet and colleagues. The combination is designed to induce emotional states different from each other. (Middle row) Activated serotonin neurons in the dorsal raphe during animal behavior. Note that the neurons activated during a particular behavior are forming a loose spatial cluster. (Bottom row) Functionally heterogeneous composition of the raphe neurons.

Reference

Paquelet, G. E., Carrion, K., Lacefield, C. O., Zhou, P., Hen, R., & Miller, B. R. (2022). Single-cell activity and network properties of dorsal raphe nucleus serotonin neurons during emotionally salient behaviors. Neuron, 110(16), 2664-2679. Link