How Serotonin Circuits Use Movement to Shape Behavior

Summary: Researchers have revealed how serotonergic neurons in larval zebrafish use visual feedback tied to movement to decide when and how much serotonin to release. This neuromodulatory signal then adjusts the fish’s swimming vigor, allowing behavior to be tuned based on the success of previous actions.

The study shows that neurons in the dorsal raphe implement a temporal gate: they only accept visual motion information immediately after a swim burst ends. By admitting sensory input at that specific time, the system isolates relevant feedback that informs whether prior actions were effective, a process analogous to assigning credit for successful behavior.

Key Facts:

  • Effort-based modulation: Serotonergic neurons scale the fish’s future swimming effort according to visual feedback received after each swim.
  • Temporal gating: A post-swim “gate” opens to let sensory motion signals into raphe neurons while filtering out irrelevant visual inputs.
  • Post-inhibitory rebound: Swimming suppresses raphe activity via inhibition, and when swimming stops a rebound excitation amplifies serotonin release.

Source: HHMI

Janelia neuroscientists are mapping the computations that allow neuromodulatory systems to link actions with outcomes, enabling flexible behavior and adaptive learning.

This shows neurons.
The new results illuminate neuromodulation in the raphe, a brain region present in humans as well as zebrafish, and may inform understanding of other neuromodulatory systems. Credit: Neuroscience News

Neuromodulators such as serotonin differ from fast neurotransmitters by acting over slower timescales and tuning the responsiveness of large neuronal populations. These chemical signals influence mood, decision-making, and motor control by regulating circuit dynamics rather than by mediating millisecond synaptic transmission alone.

In larval zebrafish, serotonin released from the dorsal raphe changes the animal’s motor vigor—how strongly it swims—based on how effective recent efforts were at producing motion through the environment. Prior work from the Ahrens Lab at Janelia demonstrated that raphe neurons read out visual motion cues to infer distance traveled and thereby evaluate action outcomes.

The new study, led by the Ahrens Lab, focused on how serotonergic neurons decide when to integrate visual information and how they compute the amount of serotonin to release. While many studies examine how neuromodulators affect circuits, fewer have asked how the neuromodulatory centers themselves perform computations during behavior.

Using a virtual-reality setup for freely swimming larval zebrafish, the team recorded activity in the raphe with high-speed voltage imaging and neurotransmitter sensors developed at Janelia. This combination allowed the researchers to follow input-output transformations in the raphe on millisecond timescales while the fish alternated between short swim bursts and passive coasting.

Zebrafish swim in discrete bouts: short, active strokes followed by coasting intervals. The researchers found that raphe neurons largely ignore sensory motion during the swim itself; instead, a neural gate opens right after a swim ends. This temporal gating ensures that visual signals associated with the animal’s own actions are selectively processed during the immediately subsequent coasting period.

Mechanistically, the gate is produced by inhibition: motor commands initiate inhibitory input to serotonergic cells, reducing their activity during the swim. When the swim stops, the inhibition is released, producing a post-inhibitory rebound in membrane voltage and neuronal firing. During this rebound window, visual motion input can drive excitation through glutamatergic pathways, causing serotonin release that scales with the perceived visual speed.

This arrangement allows the raphe to perform a form of coincidence detection—combining the temporal context of a recent action with the sensory outcome—to encode action effectiveness and adjust future motor vigor. In behavioral terms, the system supports credit assignment by associating a particular movement with the sensory consequence that follows it, thereby guiding learning about which actions are beneficial.

Supporting this circuit model, the team found that removing local GABAergic neurons that mediate the swim-locked inhibition disrupted raphe coding and impaired the animal’s ability to adjust motor vigor, indicating that the inhibitory gate and subsequent rebound are essential for this form of motor learning.

About this serotonin and visual neuroscience research news

Author: Nanci Bompey
Source: HHMI
Contact: Nanci Bompey – HHMI
Image: Credit to Neuroscience News

Original Research: Open access. “Voltage imaging reveals circuit computations in the raphe underlying serotonin-mediated motor vigor learning” by Misha B. Ahrens et al., Neuron. DOI: 10.1016/j.neuron.2025.05.017


Abstract

Voltage imaging reveals circuit computations in the raphe underlying serotonin-mediated motor vigor learning

Adaptive behavior often relies on neuromodulation to change how circuits respond to experience, yet the computations performed by neuromodulatory nuclei during behavior are not well understood. In larval zebrafish, the serotonergic raphe supports motor learning by visually detecting distance traveled during swims, encoding action effectiveness, and adjusting motor vigor accordingly.

Using millisecond-resolution voltage and neurotransmitter imaging, the authors tracked input-output computations in the raphe and found that swimming opens a temporal gate for visual input to influence serotonergic spiking. Swim commands produce GABAergic inhibition of raphe neurons; when inhibition ends, a post-inhibitory rebound in membrane voltage permits visual motion to evoke firing via glutamatergic excitation, triggering serotonin release that modulates future motor vigor. Ablation of local GABAergic neurons impaired raphe encoding and motor learning, indicating that serotonergic neuromodulation in this context emerges from action-outcome coincidence detection within the raphe.