Production of mouse ultrasonic vocalizations and distress calls is associated with different patterns of Fos expression in the nucleus retroambiguus

This study demonstrates that the mouse nucleus retroambiguus contains distinct, overlapping, and shared neuronal populations that are differentially recruited to produce ultrasonic vocalizations and distress calls, thereby refining models of the premotor control of vocalization.

The Gist

Imagine the mouse brain as a massive, high-tech control room for making sounds. Just like a human conductor uses different sections of an orchestra to play a soft lullaby versus a loud, frantic drum solo, mice use different parts of their brain to make different types of calls.

This paper investigates a specific "switchboard" in the mouse brain called the Nucleus Retroambiguus (RAm). Scientists wanted to know: Is this switchboard one big, messy pile of wires that does everything, or is it organized into specific teams for specific jobs?

Here is the story of what they found, broken down simply:

1. The Two Types of Mouse "Talk"

Mice have two main ways of screaming or singing:

• The Ultrasonic Song (USVs): These are high-pitched, inaudible-to-humans "songs" mice sing when they are happy, social, or courting a mate. Think of this as a romantic serenade.
• The Squeak: This is a loud, distress call made when a mouse is scared, hurt, or being chased. Think of this as a fire alarm or a panic scream.

2. The Experiment: Lighting Up the Brain

To see which parts of the brain were working during these calls, the scientists used a clever trick. They treated the brain like a security camera system.

• When a neuron (a brain cell) gets really active, it turns on a "light" called Fos.
• The scientists let mice make their calls, waited a couple of hours, and then looked at the brain under a microscope to see where the "lights" were on.

3. The Big Discovery: Different Neighborhoods for Different Jobs

The scientists expected to see the same group of neurons lighting up for both the "serenade" and the "panic scream." They thought maybe the brain just turned up the volume for the scream.

They were wrong. The brain is much more organized than that.

• The "Serenade" Team: When mice sang their ultrasonic songs, a huge crowd of neurons lit up all over the control room, from the front to the back. It was a full-team effort.
• The "Panic" Team: When mice squeaked in distress, a much smaller group of neurons lit up, but they were clustered in a very specific spot at the back of the control room.

It's like if you needed to bake a cake, you'd use the whole kitchen (mixers, ovens, counters). But if you just needed to open a door, you'd only use the doorknob. The brain uses a "whole kitchen" approach for singing, but a "specialized doorknob" approach for screaming.

4. The "Shared" Neighbors

To be sure, the scientists did a second, even cooler experiment. They used a "time-travel" labeling technique (TRAP2) to mark the neurons that were active during the first call, and then checked if those same neurons were active during a second call.

They found that while there is a tiny bit of overlap (a few neurons that help with both), most of the neurons are specialists.

• There are Singing Neurons.
• There are Squeaking Neurons.
• There are a few Multi-tasking Neurons that do both.

The Takeaway

This study changes how we think about animal communication. We used to think the brain might just have one "volume knob" that gets turned up or down to make different sounds.

Instead, the mouse brain is more like a specialized factory.

• To make a happy song, it activates Assembly Line A (which is big and complex).
• To make a distress scream, it activates Assembly Line B (which is smaller and located in a different wing).

Why does this matter?
Understanding how the brain separates these signals helps us understand how animals communicate complex emotions. It also gives scientists a roadmap to figure out how the brain controls other complex behaviors, showing that even in the smallest parts of the brain, there is a sophisticated, organized division of labor.

Technical Summary

1. Problem Statement

While the Nucleus Retroambiguus (RAm) in the hindbrain is established as a critical premotor region for vocal production in mammals, the neural organization underlying the production of distinct vocalization types remains poorly understood. Specifically, it is unclear whether the production of Ultrasonic Vocalizations (USVs) (social/mating calls) and distress calls/squeaks (aversive calls) relies on:

• A single, homogenous population of RAm neurons where activation intensity dictates vocal type.
• Distinct, non-overlapping populations of neurons.
• Partially overlapping populations with specific subsets recruited for each vocal type.

Previous studies suggested some overlap (e.g., ablation of specific RAm neurons blocked both types), but direct comparisons of neuronal recruitment patterns for these specific vocalizations in the same or comparable subjects were lacking.

2. Methodology

The study employed a combination of between-subjects and within-subjects experimental designs using C57BL/6 and TRAP2;Ai14 transgenic mice.

• Behavioral Paradigms:

  • USV Production: Males and females were single-housed for 3 days to induce high rates of USVs during social interactions (male-female or female-female).
  • Squeak Production: Females were exposed to courtship interactions with males (to elicit "courtship squeaks"). Males and females were also subjected to mild footshocks (0.5 mA) to elicit distress squeaks, though courtship interactions yielded higher rates in females.
  • Control: "Low vocalization" groups were created using brief interactions to correlate vocalization time with neural activation.

• Immunohistochemistry (Between-Subjects):

  • Mice were sacrificed 120 minutes post-behavior.
  • Brains were sectioned, and Fos protein (an immediate early gene marker of neural activity) was stained to map activated neurons in the RAm.
  • Quantification: Fos-positive neurons were manually counted across five coronal sections (approx. -7.2 mm to -8.0 mm caudal to bregma) within defined Regions of Interest (ROIs). Heatmaps were generated to visualize spatial distribution.

• TRAP2 Activity-Dependent Labeling (Within-Subjects):

  • Strategy: TRAP2;Ai14 mice were exposed to a first vocalization context (e.g., USVs), followed by an injection of 4-hydroxytamoxifen (4-OHT) to permanently label active neurons with tdTomato.
  • Second Exposure: Ten days later, mice were exposed to a second context (either the same vocalization type or the alternative type).
  • Analysis: Brains were processed for Fos immunostaining. Overlap was calculated as the proportion of tdTomato-labeled neurons (from the first event) that were also Fos-positive (during the second event).

• Statistical Analysis:

  • Used ANOVA (repeated measures), Tukey's HSD post-hoc tests, and linear regressions to compare Fos counts, spatial distributions, and correlations with vocalization time.

3. Key Results

A. Distinct Spatial Patterns of Fos Expression

• USVs: Production of USVs (in both males and females) drove robust Fos expression distributed throughout the entire rostro-caudal extent of the RAm (all 5 sections). There was a strong positive correlation between total USV time and Fos counts across all sections.
• Squeaks: Production of courtship squeaks in females drove significantly weaker overall Fos expression. Crucially, this activation was biased toward the caudal-most sections (Sections 4 and 5) of the RAm. Activation in rostral sections was minimal.
• Vocalization Time Control: The difference in Fos patterns was not due to differences in total vocalization time. Even when male USV cases were matched to female squeak cases for vocalization duration, or when low-vocalizing males were analyzed, the "USV-like" (broad) pattern persisted, while the "squeak-like" (caudal-biased) pattern remained distinct.

B. Limited Overlap in Neuronal Recruitment (TRAP2 Data)

• Control Groups: In groups exposed to the same vocalization twice (USV-USV or Squeak-Squeak), there was high overlap (70–80%) between the initially labeled (tdTomato) and secondarily activated (Fos) neurons, validating the TRAP2 method.
• Cross-Modal Groups: In groups exposed to different vocalizations (e.g., Squeak $\to$ USV or USV $\to$ Squeak), the overlap between the first event's labeled neurons and the second event's Fos-positive neurons was significantly lower across all sections compared to control groups.
• Conclusion: The population of neurons activated during squeak production is not a simple nested subset of those activated during USV production. Instead, the two vocalization types recruit largely distinct ensembles.

4. Key Contributions

1. Refined Anatomical Model: The study proposes a model where the RAm contains three distinct functional populations:
   • USV-related neurons: Distributed across the entire rostro-caudal axis.
   • Squeak-related neurons: Localized primarily to the caudal RAm.
   • Shared neurons: A smaller subset recruited during both vocalization types.
2. Decoupling Vocal Type from Intensity: The findings refute the hypothesis that vocal type is determined solely by the intensity or number of activated RAm neurons (e.g., that squeaks are just "low intensity" USVs). Instead, the identity of the vocalization is determined by the recruitment of specific, spatially segregated neuronal subpopulations.
3. Sex Differences in Paradigms: The study successfully characterized squeak production in female mice during courtship, providing a robust behavioral model for studying distress calls in a social context, which showed distinct neural signatures compared to footshock-induced squeaks.

5. Significance

• Circuit Organization: This work demonstrates that the hindbrain premotor control of vocalization is more heterogeneous than previously thought. It suggests a "labeled line" or specific population coding mechanism rather than a rate-coding mechanism for different vocal types.
• Evolutionary Context: The findings align with recent work in Drosophila and birds, suggesting that complex vocal repertoires may rely on distinct, intermingled but functionally separate neuronal populations even at the level of the brainstem.
• Future Directions: The study highlights the need to characterize the specific axonal projections (e.g., to nucleus ambiguus vs. spinal expiratory centers) and molecular markers of these distinct RAm populations. It sets the stage for investigating how the Periaqueductal Gray (PAG) gates these distinct hindbrain circuits to produce context-appropriate vocalizations.
• Limitations & Next Steps: The authors acknowledge that Fos expression is a proxy for activity and may miss neurons that are active but do not upregulate Fos. Future work using in vivo calcium imaging and functional manipulation (optogenetics/chemogenetics) of these specific populations is required to confirm causal roles.