A Squeak Is Not Enough: Female Presence and Vocal Playback Have Contrasting Effects on c-Fos Expression by Dorsal Raphé Neurons in Lab Mice

Abstract

The regulation of sensory processing by centralized neuromodulatory systems can alter behavioral responses to social cues. Neuromodulatory systems such as the serotonergic neurons in the dorsal raphé nucleus (DRN) show heterogenous responses to different types of sensory stimuli or to stimulus qualities such as reward, valence, or salience. Sensory neuromodulation could therefore be related to a broader quality of the behavioral context or to specific types of social cues. We assessed this issue by presenting male mice with either playback of female vocal signals associated with defensive aggression (squeaks) or silence, and the presence or absence of a female. Activity in regions of the DRN that project to the auditory midbrain was assessed through co-labeling with antibodies to the serotonin synthetic enzyme tryptophan hydroxylase (TPH) and the immediate early gene product c-Fos. Female presence or absence had the largest effect, decreasing the co-localization of TPH and c-Fos, while the playback of squeaks had effects that were condition-dependent, increasing co-label only when females were absent. Squeak playback further decreased the correlation in the numbers of co-labeled neurons between two dorsal subdivisions of the DRN, the DRD and DRL. These results are inconsistent with an auditory-exclusive feedback loop. Instead, cues associated with female presence heavily influence raphé activity, with squeaks playing a modifying and context-dependent role. Because the elevation of serotonin in the IC causes males to become more responsive to female squeaks, these findings suggest that a nuanced interaction of positive and negative cues during social interaction may fine-tune male responses to the vocalization of social partners, in part through the serotonergic system.

1. Introduction

The regulation of sensory processing by centralized neuromodulatory systems can alter behavioral responses to social cues. In female zebra finches, dopamine agonists in the secondary auditory cortex change the behavioral preferences of females for male song [1]. In mice, acetylcholine and norepinephrine influence the activity of neurons in the olfactory bulb, altering responses to odorants, including those of familiar conspecifics [2,3]. In black ghost knife fish, serotonin alters both the electrosensory and behavioral responses to higher-frequency stimuli [4]. Neuromodulatory neurons or neuromodulatory release are, in turn, sensitive to social signals like song [1,5], electrical signals [6], or rejection from a social partner [7].

The mutual influence of sensory and neuromodulatory systems underlies models in which modulatory neurons integrate information on social context and project this information back to sensory regions through the release of neuromodulators [8]. However, the specificity of this process has not been well-characterized. The responses of neuromodulatory neurons to stimuli with different structures, or of different sensory modalities, are highly heterogeneous, even within well-defined anatomical regions [9,10,11]. This creates the potential for parallel modality-specific loops in which modulatory neurons responding to specific kinds of social cues (for example, vocalizations versus odors) project to separate sensory regions [11]. This could result in sensory processing of social stimuli fine-tuned to the unimodal sensory environment. Alternatively, neuromodulatory systems may be sensitive to characteristics of events such as their intrinsic or associated rewarding or aversive qualities, valence, or salience [11,12,13,14,15].

We explored this issue in the social context of opposite-sex social interaction in mice. In this context, females produce broadband human-audible vocal signals, or squeaks, and calls vary over the course of an interaction [16,17]. During early/appetitive stages of an interaction, squeaks are paired with defensive aggression by females, in the form of kicks or lunges directed at males [16,18,19,20]. A higher number of squeaks during this stage of an interaction is associated with less mounting by males at later stages [16], and the playback of squeaks causes males to reduce the number of ultrasonic vocalizations (USVs) they produce [21]. These events are related to serotonergic signaling in the auditory midbrain nucleus, the inferior colliculus (IC). The number of squeaks produced by females inversely correlates with serotonin levels in the IC of their male partners [7]. In the sources of serotonergic projections to the IC (i.e., specific subregions of the dorsal raphé nucleus (DRN)), the number of serotonergic neurons expressing c-Fos in males likewise corresponds to whether females reject them or not [22,23]. What is unknown is whether serotonergic signaling is responsive to squeaks alone, or to a more complex set of cues related to female presence.

To assess how different types of stimuli drive serotonergic activity, we exposed males to squeak playback or silence in the presence and absence of a female on the opposite side of a divider, and immunohistochemically measured neurons co-labeling for tryptophan hydroxylase (TPH), a synthetic enzyme for serotonin, and c-Fos, an immediate early gene product. As in previous studies [22,23], we found that two dorsal subdivisions of the DRN, the dorsal midline division (DRD) and the dorsal lateral division (DRL), were sensitive to the social stimuli we presented. Female presence or absence had the largest effect on the co-localization of TPH and c-Fos, while the playback of squeaks had effects that were more condition-dependent. This finding is more consistent with sensory neuromodulation tied to a quality of social interaction, like salience or valence, rather than to a modality-specific feedback loop.

2. Materials and Methods

Subjects: All experiments were approved by the Bloomington Institutional Animal Care and Use Committee (protocols 21-020 and 24-025, approvals on 18 August 2021 and 25 July 2024, respectively). In total, 42 male CBA/J mice were obtained from the Jackson Laboratory (Bar Harbor, ME, USA) and socially housed with 3 or 4 males per cage. Mice were used in experiments between 7 and 8 weeks of age. Because experience with females affects vocal behaviors, males were given experience with females five to seven days prior to the experimental day. Males were placed in 10 min interactions with females and rotated through females until they showed mounting behavior. All males mounted after being placed with females no more than 4 times.

Experiments: Behavioral experiments were conducted inside a sound attenuation chamber (IAC Acoustics, Naperville, IL, USA). Males were placed inside a standard mouse cage (12 inch × 6 inch × 6 inch), bisected by a Plexiglas barrier with a small hole 1.25 cm in diameter, permitting limited investigation of the other side (Figure 1). Cages were further modified with an array of small holes drilled in the side to promote acoustic transparency to sound played from a speaker outside the array. Males were placed on one side of the divider. Clean bedding plus soiled bedding from a cage holding multiple females were placed with the male. The opposite side of the cage either remained empty or contained a female mouse. Females used during the behavioral trials were all novel to the males, but were used as social stimuli for up to 4 males. Because multiple experimental groups were balanced across trials conducted in a given time period, there were no females that acted as social stimuli for only one experimental group of males. The presence or absence of females created two social conditions: the ‘female’ and ‘no female’ treatments. This cage arrangement (the split-cage assay) has been used in past studies to assess the effects of call playbacks [21,24]. Females do not produce squeaks in the split-cage assay. Females could be producing a small number of USVs [25,26], but USVs decline in response to squeak playback even when females are absent or anesthetized, indicating that this response must be from males [21]. Awake females are used because they induce a high and sustained production of USVs indicative of male engagement.

Figure 1.

(A) Experimental groups for the study factorially combined female presence versus absence with playback of squeaks or silence within a split arena with males on one side. (B) Oscillogram (top) and spectrogram (bottom) of the exemplar squeak used in the playback.

The two social conditions were factorially combined with a playback of female squeaks (Figure 1B). The playback consisted of a 5 min period of silence followed by a 30 min playback of recorded female squeaks (Figure 1B). Squeak playbacks were constructed using an exemplar squeak that was within one standard deviation of the mean for duration and peak frequency for the BBVs in an opposite-sex interaction with high levels of BBV production and rejection by the female [21]. This exemplar was substituted in time for squeaks recorded from a one-minute segment of an opposite-sex interaction in which vigorous rejection of a male occurred [21]. This one-minute segment was looped to create a 30 min stimulus. Playback of a 5 min loop causes males to suppress their USV production [21,24], and we extended this time to maximize acoustic stimulation and possible changes in c-Fos expression. Squeaks were played through an Ultrasonic Dynamic Speaker (Vifa, range from 1 to 120 kHz) powered by an UltraSoundGate Player 116 (Avisoft Bioacoustics, Glienicke/Nordbahn, Germany). Because female squeaks are usually produced in close proximity to male ears as they show defensive aggression [18], squeak playback was calibrated to the previously measured peak intensity of a microphone placed close to a vigorously squeaking female, at approximately 104 dB. Squeaks were calibrated using a PS9200 system with a ¼ inch microphone (ACO Pacific, Belmont, CA, USA). Manipulating female presence and playback created a total of four experimental groups: no female–no playback, female–no playback, no female–playback, and female–playback.

The day before the experimental day, male subjects were habituated to the recording chamber for one hour. On the day of the experiment, males were again placed inside a sound chamber for a minimum of forty-five minutes before being placed in the split-cage behavioral assay. Following the squeak playback, males were placed in a separate sound attenuation chamber to allow optimal time for c-Fos expression [22]. During the playback, male behaviors were recorded with a Canon VIXIA camera (Q-See, Digital Peripheral Solutions Inc., Anaheim, CA, USA) equipped with a Q-See four-channel DVR PCI video capture card positioned above the split cage. Experiments were also monitored in real time through a monitor on the outside of the sound attenuation chamber. Vocal behaviors were recorded with a 16-bit condenser microphone (CM16/CMPA, Avisoft Bioacoustics, Glienicke/Nordbahn, Germany) placed above the cage. This was connected to an Avisoft Ultrasound Gate 116H (#41163; Avisoft Bioacoustics, Berlin, Germany) with a sampling rate of 250 kHz. Avisoft Recorder software (https://avisoft.com/recorder/, accessed on 21 January 2026) was used to collect audio recordings.

• Immunohistochemistry:

At 90 min after the start of the playback, males were euthanized with isoflurane fumes and transcardially perfused with Krebs–Henseleit buffer, followed by 4% paraformaldehyde in phosphate-buffered saline. Brains were then postfixed in 4% paraformaldehyde solution overnight before being transferred to a 30% sucrose solution. Four brains were not well-fixed and excluded from immunohistochemical measurement, although behavioral data from these individuals were analyzed. The brains were then sliced by being coated in tissue Tek (Sakura Finetek, Torrance, CA, USA) and cut with 50-micron thickness on a sliding freezing microtome (American Optical Company, Buffalo, NY, USA), then placed in cryoprotectant and stored in a –80 freezer until immunohistochemistry could be performed. Slices were thawed and washed in PBS solution for five minutes five times, then placed in a blocking solution of 10% donkey serum and 0.3% Triton X (Dow Chemical Company, Louisville, KY, USA) in PBS for one hour. After blocking, slices were then incubated overnight in TPH and c-Fos primaries. These consisted of a mouse monoclonal purified IgG, tryptophan hydroxylase2 antibody (348 011, Synaptic Systems, Goettingen, Germany) and a rabbit monoclonal recombinant IgG, c-Fos antibody (226 008, Synaptic Systems, Goettingen, Germany) with 5% donkey serum and 0.3% Triton X. Slices were incubated on an orbital shaker at 4 °C. Sections were removed the following day and rinsed again in PBS for five minutes five times, then transferred into fluorescent labeled secondary antibodies, donkey anti-Mouse IgG (H + L) with Alexa Fluor™ 488 fluorophore (A21202, Life Technologies Corp, Eugene, OR, USA) and Donkey anti-Rabbit IgG (H + L) with Alexa Fluor™ 680 fluorophore (A10043, Life Technologies Corp, Eugene, OR, USA), in PBS with 5% donkey serum and 0.3% Triton X. This was incubated for 2 h in the dark, and washed four times for 5 min. Slices were mounted on 1% gelatin-coated glass slides and cover-slipped with ProLong Gold anti-fade mounting medium (Thermo-Fisher Scientific, Waltham, MA, USA).

• Imaging and analysis:

Confocal Microscopy: Sections were imaged at 10× on a Leica SP8 confocal microscope using simultaneous 488 and 680 nm lasers with 12 z planes separated by 2.41 μm.

Image Analysis: Anatomical boundaries for the DRN subregions were defined as in [22]. Briefly, plates 73, 71, 68, and 66 of the Mouse Brain Atlas [27] were used as templates, and regions fitted to match the TPH+ chemoarchitecture of the DRN relative to landmarks including the median longitudinal fasciculi and the cerebral aqueduct. These regions consisted of the dorsomedial region (DRD), the ventromedial region (DRV), the bilateral dorsal lateral region (DRL), and the bilateral ventral lateral region, also called the posterodorsal region (PDR). These regions have shown distinct functional characteristics in previous studies [22,23].

Regions of interest around these subdivisions were drawn and cropped in Photoshop. C-Fos-labeled neurons were then identified using CellProfiler, followed by manually marking the AI-identified Fos neurons as double-labeled neurons using CellProfiler Analyst with an overlapping image of TPH-labeled neurons. The output images from CellProfiler were then exported to Photoshop to count double-labeled neurons. TPH+ neurons were counted manually in Photoshop by superimposing a grid over TPH+ labeled regions of interest. The same experimenter counted all sections. Figure 2 illustrates the single labels (Figure 2A,B) and merged label (Figure 2C) within the ROIs.

Figure 2.

Immunohistochemical labeling of the dorsal raphé nucleus with (A) an antibody to tryptophan hydroxylase (TPH), (B) an antibody for c-Fos, and (C) a combination of the labels. The inset shows a larger image of a region in the DRD subregion. Arrowheads (B,C) indicate cells positive for both labels. Arrows (B,C) indicate nuclei that were only labeled for c-Fos. Symbols indicate the same cells in (B,C). Abbreviations: DRD—dorsal midline region, DRV—ventral midline region, DRL—dorsal lateral regions, PDR—posterodorsal region corresponding to ventral lateral regions.

• Behavioral measurement:

Non-vocal: Non-vocal behaviors were scored from pre-recorded videos using Behavioral Observation Research Interactive Software (BORIS v. 7.13.6 [28]). Each video was scored by the same individual (JB). For each subject mouse, rearing, digging, grooming, and jumping were scored according to previously published definitions [21,23]. Rearing was defined as any instance when the mouse’s two front feet left the bottom of the cage, whether the mouse put its feet onto the walls of the cage, or the mouse reared in the middle of the cage. Digging was defined as a mouse shuffling the bedding in the cage with its feet. Grooming was defined as a mouse turning its head to clean its body or rapid head movement with its front paws near its mouth. Jumps were recorded whenever all four of the mouse’s feet left the floor of the cage. The rearing, digging, and grooming by the subject mice were measured as the duration of the behavior (in seconds), whereas jumps were recorded as the number of occurrences.

Vocal: Ultrasonic vocalizations (USVs) were manually counted using Avisoft SASLab Pro (512 FFT-length, Hamming-style window with 50% overlap; Avisoft Bioacoustics, https://avisoft.com/sound-analysis/, access on 21 January 2026). USVs were divided into two classes, non-harmonic and harmonic. Harmonic calls have a pronounced acoustic harmonic at lower frequencies in the range of 40–50 kHz. Harmonic calls show functional distinction from nonharmonic calls in that they are enriched prior to mounting and are produced in higher proportions by socially isolated males [16,17,23,29,30].

• Statistical analysis:

Statistical tests were conducted using SPSS 29.0 (IBM, Armonk, NY, USA). Generalized linear mixed models (GLMMs) were used to assess differences among experimental groups. The normality of distributions was assessed through Shapiro–Wilk tests. For non-normal distributions, negative binomial models were used. The Akaike information criterion was further used to confirm that specific models fit the data well relative to alternative distributions. Tukey’s least significant difference (LSD) was used as the basis for post hoc tests. In terms of specific models, numbers of neurons across groups were compared with a GLMM using a negative binomial distribution, with the DRN subregion as a repeated measure, and female presence versus absence and playback versus silence as fixed factors. The proportions of neurons across groups were assessed with a negative binomial distribution using the number of double-labeled neurons as the target, and the natural log of the total number of TPH+ neurons for each individual as offsets. Numbers of calls were compared across groups with a GLMM using a negative binomial distribution, and female presence versus absence and playback versus silence as fixed factors. Proportions of calls that were harmonic were compared between female presence versus absence using independent samples Mann–Whitney U tests. The nonvocal behaviors were assessed across groups with GLMMs with female presence versus absence and playback versus silence as fixed factors; normal distributions were used for times of digging and rearing, a negative binomial distribution was used for the number of jumps, and a gamma distribution was used for grooming time.

A principal components analysis (PCA) with Varimax rotation and Kaiser normalization was used to reduce the dimensionality of behaviors for comparison to proportions of neurons double-labeled with c-Fos and TPH. Pearson’s correlations were used for assessing whether correlations were significant, and the Benjamini–Hochberg correction was used to adjust for multiple comparisons using a conservative Q value of 0.05 [31].

Data were plotted in Excel, figures were created in Photoshop (Figure 2; Adobe, San Jose, CA, USA), and PowerPoint (all other figures; Microsoft, Redmond, WA, USA) was used for the other figures. Data are presented in Supplemental Table S1.

3. Results

Male mice (n = 42) were placed in cages bisected by Plexiglas barriers with small holes serving as windows to the other side, with either a single unfamiliar female or no female in the other half of the cage (the social condition). In each of these conditions, males were also either presented with a playback of recorded female squeaks or no playback (the playback condition). This factorial design created a total of four experimental groups (Figure 1).

The number of neurons that co-labeled for TPH and c-Fos (Figure 3A) differed among subregions of the DRN, with higher numbers in the more densely populated midline subdivisions (DRD and DRV) relative to the lateral subregions (DRL and PDR) (Figure 3A, Table 1, GLMM with subregion as a repeated measure; F_(3.139) = 77.673, p < 0.001). The number of co-labeled neurons did not differ significantly depending on female presence or absence, or on whether or not squeak playback was present (Table 1). However, when the numbers of co-labeled neurons were corrected for total number of TPH-positive neurons to assess the proportional value, there were significant effects of both the DRN subregion and the absence versus presence of females (Table 1, GLMM with subregion as a repeated measure; subregion: F_(3.138) = 29.219, p < 0.001; social condition: F_(3.138) = 12.884, p < 0.001). The proportion of double-labeled neurons was highest in the DRD and DRV, and highest in the absence of females (Figure 3B). There was no significant interaction between the presence or absence of females and the subregion. However, the difference between females present versus absent was largest in the DRD and DRL subregions (8.89% in the DRD and 7.32% in the DRL versus 2.98% in the PDR and 0.52% in the DRV).

Figure 3.

Effects of female presence or absence, and of playback, on neurons in the DRN co-expressing TPH and c-Fos. (A) Number of neurons labeled for both TPH and c-Fos in each of the examined regions of the dorsal raphé nucleus and each of the 4 experimental treatments. There were significant differences in the number among regions (see text), but no significant differences among the experimental treatments for any region. (B) Proportions of neurons double-labeled for TPH and c-Fos relative to all neurons positive for c-Fos in the same groups as in (A). There were significant differences in the number among regions (see text). Proportions were significantly lower in the presence of females (gray bars) than in the absence of females (white bars). Values represent raw means, with error bars indicating standard error of the mean.

Table 1.

Output of GLMM assessing the effects of squeak playback and female presence on double-labeled neurons.

	Number				Proportion
Source	F	df1	df2	p	F	df1	df2	p
Corrected Model	20.366	12	139	<0.001	9.313	12	139	<0.001
social	0.747	1	139	0.389	11.503	1	139	<0.001
sound	0.611	1	139	0.436	2.981	1	139	0.086
region	77.673	3	139	<0.001	30.314	3	139	<0.001
social × region	2.588	3	139	0.055	1.561	3	139	0.202
sound × region	0.242	3	139	0.867	0.082	3	139	0.97
social × sound	0.616	1	139	0.434	0.982	1	139	0.323

The social and playback conditions also affected the USVs produced during the behavioral trials. There were fewer USVs produced during no-female trials compared to trials with females, with 12.81 ± 4.97 calls in the absence of females versus 2129.67 ± 422.62 when females were present (Figure 4, Table 2, F_(1.38) = 169.769, p < 0.001). As previously reported [17,30], many fewer harmonic than nonharmonic calls were produced across all conditions, with an average of 122.49 ± 33.44 harmonic and 909.34 ± 217.89 nonharmonic calls. The percentage of harmonic calls relative to all USVs was also increased by female presence; 2.71 ± 1.14% of calls were harmonic in the absence of females, while 12.75 ± 2.51% were harmonic in the presence of females (independent samples Mann–Whitney U test, U = 54, p < 0.001).

Figure 4.

Effects of female presence and playback on numbers of ultrasonic vocalizations (USVs). Top: Female presence results in significantly greater numbers of USVs. The inset shows spectrograms of nonharmonic and harmonic call types. Bottom: Numbers of USVs for the no-female condition only replotted on an expanded scale. Values represent raw means, with error bars indicate standard error of the mean. Dashed lines indicate averages presented on a different scale below.

Table 2.

Output of GLMM assessing the effects of squeak playback and female presence on USVs.

Source	F	df1	df2	Sig.
Corrected Model	57.35	3	38	<0.001
sound	1.707	1	38	0.199
social	169.769	1	38	<0.001
sound × social	0.504	1	38	0.482

Several other behaviors were also measured in male mice on their side of the arena: jumping, grooming, digging, and rearing. Of these, jumping and grooming both showed significant decreases with female presence (GLMM). Jumps decreased from 5.71 ± 1.57 to 1.86 ± 1.28 with female presence, and grooming decreased from 93.39 ± 18.87 to 34.63 ± 4.51 s (Figure 5, Table 3).

Figure 5.

Box-and-whisker plots showing that some nonvocal behaviors are significantly affected by female presence versus absence. (A) Jumping and (B) grooming both decreased when females were present. Horizontal lines indicate quartiles with the central line indicating the median, while error bars extend to maximum and minimum values. A single asterisk represents p < 0.05. Double asterisks indicate p < 0.01.

Table 3.

Results of GLMMs assessing the effects of squeak playback and female presence on nonvocal behaviors.

Jumping	F	df1	df2	p
social	5.277	1	38	0.027
playback	1.209	1	38	0.279
social × playback	1.58	1	38	0.216
Grooming
social	25.997	1	38	<0.001
playback	2.611	1	38	0.114
social × playback	4.068	1	38	0.051
Digging
social	3.518	1	38	0.068
playback	0.013	1	38	0.908
social × playback	0.334	1	38	0.567
Rearing
social	0.003	1	38	0.954
playback	0.017	1	38	0.897
social × playback	0.047	1	38	0.83

To assess whether c-Fos activation in the DRN was correlated with behavior, we reduced the behaviors in dimension through a principal components analysis, and compared the resulting factors to the proportions of double-labeled neurons in the DRD and DRL across individual males. The first two principal components accounted for a cumulative 58.8% of the variation in the data. PC1 loaded strongly for the harmonic and nonharmonic USVs, and PC2 loaded strongly for jumping and rearing (Table 4). Figure 6 plots PC1 versus PC2, with groups including females versus no females clearly differentiating along the axis of PC1. We assessed correlations between PC1 and PC2 and the proportions of double-labeled neurons in DRD and DRL. Across all experimental groups, PC2 correlated with the proportion of double-labeled neurons in the DRL, but not the DRD (Pearson’s correlation, DRL: p = 0.018, DRD: p = 0.07), while there were no correlations between the proportion of co-labeled neurons and PC1 (Pearson’s correlation, DRL: p = 0.248, DRD: p = 0.566). However, the value for the correlation between PC2 and the proportion of co-labeled neurons in the DRL was no longer significant when corrected for multiple comparisons. Since the presence versus absence of females had the strongest effect on double-labeled neurons, we further broke these data down into groups that were exposed to females or not. There were correlations between PC2 and the proportion of double-labeled neurons in the DRD and DRL only in the group without female contact (Pearson’s correlation, DRL: p = 0.041, DRD: p = 0.043), but these again were no longer significant after corrections for multiple comparisons.

Table 4.

Eigenvector coefficients for principal components analysis of behaviors.

Behavior	Component
	1	2
Jump	−0.305	0.737
Groom	−0.399	−0.232
Dig	−0.295	−0.244
Rear	0.061	0.836
Harm	0.936	−0.08
Nonharm	0.931	−0.097

Figure 6.

Values of the first two principal components of all measured behaviors, including vocalizations, are differently distributed when females are present (orange symbols) versus absent (blue symbols).

The correlation in the proportion of co-labeled neurons between the DRD and DRL, the two subregions in which double-labeled cell proportions were sensitive to female presence and absence, was generally high, whether females were present or not (Pearson’s correlation, females present: p = 0.004, females absent: p < 0.001). However, correlation coefficients were decreased by the playback of squeaks. Figure 7 shows that the correlation coefficient between the DRL and DRD for the groups in which no females were present decreased from 0.91 to 0.712 following squeak playback as opposed to silence. Likewise, for the groups in which females were present, the coefficient decreased from 0.907 to 0.598 following playback versus silence, and was no longer significant.

Figure 7.

Scatterplots comparing proportions of double-labeled neurons in the DRD and DRL in all four experimental groups. Correlation coefficients are lower during playback of squeaks (lower row) than in the no-playback condition (upper row), regardless of whether females were present (left column) or not (right column). Gray dots represent individual mice. Note that the x-axis scale is different for the no-female–playback condition. Dashed lines indicate significant correlations (p < 0.05).

4. Discussion

In addition to important roles in central functions like the regulation of affect, appetite, cognition, and aggression and other social behaviors [32,33,34,35], the serotonergic system is an important modulator of sensory function [36,37,38]. In the auditory midbrain and brainstem of mice, serotonergic projections arise from two anatomically distinct subregions of the DRN, the DRD and DRL [22,39,40]. Functionally, these connections are relevant to important social behaviors, including aggression and opposite-sex interaction [22,24,40,41]. In males, serotonergic increases in the inferior colliculus negatively correlate with female squeaks [7]. This correlation could result from direct responsiveness of DRN neurons to squeaks themselves, in line with the modality specificity of the responses of some DRN neurons [11]. Alternatively, the correspondence between the activity of DRN neurons and serotonin release in the IC could arise from a broader sensitivity of DRN neurons to information on the social context. For example, some DRN neurons, or serotonin release in specific regions, are sensitive to qualities such as the valence or salience of a stimulus [7,11,42]. Squeaks themselves are associated with defensive aggression [16,18].

To assess which of these possibilities is best supported in serotonin-auditory interactions, we presented male mice with a female partner or no partner, and with playback of squeaks or silence. We measured the expression of c-Fos, a proxy marker for neural activity [43,44], in neurons positive for TPH in different subregions of the DRN. We predicted that fewer TPH+ neurons would show c-Fos expression during squeak playback, reflecting the inverse association between squeaks and serotonin release in the IC [7]. Although this prediction was incorrect, both female presence and squeak playback did influence c-Fos expression in TPH+ neurons in select subregions of the DRN, and female presence additionally affected across-individual correlations between c-Fos expression in different DRN subregions.

• A squeak is not enough:

The social manipulation with the largest effect on c-Fos expression by DRN neurons was the presence or absence of females on the other side of a barrier from males. Female presence decreased the proportion of double-labeled neurons in the DRN. This change was largest in two of the anatomically defined subregions we examined, the DRD and the DRL, both of which contain neurons projecting to the IC [22,40]. Female presence also had a powerful effect on vocal behavior, increasing the number of USVs by over two orders of magnitude. A variety of experimental approaches, including devocalization, anesthetized social partners, localizing calls with microphone arrays, and attaching microphones to individual mice, were all in agreement in finding that male mice make the overwhelming majority of USVs in opposite-sex contexts [25,26,29,45,46,47]. In addition to increasing the overall number of USVs, female presence increased the percentage of USVs showing a pronounced harmonic at frequencies ~50 kHz and below. This type of call is associated in time with mounting by males in direct interactions [16,17,29]. In comparison, the playback of squeaks had smaller effects on both the expression of c-Fos by serotonergic neurons and USV production, which were not significant.

The fact that female presence had opposite effects on both the proportions of double-labeled neurons and on USV production is consistent with the suites of behaviors associated with this stimulus in the current and previous studies. Female presence or female odor elicits prosocial behavior from males, including not only increased production of USVs but also the trailing, olfactory investigation, and mounting of females [25,48,49,50,51]. USVs, in turn, are attractive to female mice, although this depends on the female estrous stage and social housing [52,53,54,55]. In other studies, squeaks act as a modifier to these behaviors. Squeaks are associated with physical rejection by females, including kicking and lunging at males [16,18,19]. Playback of squeaks causes the suppression of USVs that is observable on timescales of less than a second to minutes [21]. Similarly, in the absence of females but the presence of female odor, squeak playback causes fewer males to produce USVs [50]. However, it is important to note that the responses of males to squeaks are not always suppressive, but depend on contextual factors. Males avoid the source of squeak playback when paired with the odor of cats but not the odor of females [56]. Males may even increase the proportional production of USVs relative to other call types in response to playbacks of squeaks, when females are not present [57]. The lack of effects of squeak playback in the current study could therefore be related to the details of the experimental paradigm.

• Context matters:

A surprising aspect of our findings was that the expression of c-Fos by TPH+ neurons was decreased by female presence and conditionally increased by squeak playback, because this finding is seemingly in the opposite direction from previous findings. In the IC and the hippocampus, serotonin is elevated in the presence of females or of preferred social stimuli [22,42]. There are several possible explanations for this discrepancy. One is that the electrochemical measurement of serotonin used in past studies in the IC [7] is not comparable to c-Fos expression in DRN neurons, which reflects the integration of activity over longer time periods [43]. A second possibility is that DRN neurons, even within specific DRN subregions, have heterogenous response properties, connectivity, developmental origin, and transcriptomic profiles [11,58,59,60]. It is therefore likely that our measurements of c-Fos expression captured responses from an expanded population of neurons in addition to IC-projecting neurons.

A final possibility is that the context of the split-cage paradigm used in the current study, with males separated from females, is different from direct interactions in which males can approach females, collect additional sensory cues, and evoke female responses. This could change the relationships between c-Fos in the DRN and the behavior that we observed in previous studies [22,23]. In fact, a previous study has suggested that these relationships are plastic, since the correlation between social behavior and c-Fos expression in the DRD is inverted in socially isolated versus socially housed groups of males [23]. These findings all suggest that behavioral and serotonergic responsiveness to squeaks and female presence are related to each other, but that the strength and direction of the relationship depend on the context in which these stimuli occur.

• Correlations:

An unanticipated finding in the current study was that the playback of squeaks influenced the correlation coefficient between the number of double-labeled neurons in the DRD and DRL. These numbers were generally highly positively correlated across males, regardless of whether females were present or not. The playback of squeaks decreased the correlation coefficient for co-labeled neurons between these two regions and caused the relationship between c-Fos activity in the DRN and DRL to become non-significant when females were present. This phenomenon may have been driven by greater variation among individuals in the squeak playback groups, and could result from afferents to the DRN versus DRL that are differentially influenced by squeaks, or through mutual interconnections that exist among groups of neurons in the DRN, some of which are thought to be inhibitory [61,62]. A similar phenomenon was observed in female mice, in which the correlation between neural activity and male sexual behaviors was inverted in the DRL compared to the DRD [22]. Overall, these findings suggest that activity in the DRN and DRL can be correlated or independent depending on the context, including the presence of squeaks, a signal of defensive aggression.

• Functional implications:

Our findings have functional implications for auditory regulation in the brainstem and midbrain during social interactions. Although it was beyond the scope of the current study to directly measure serotonin release in the IC, the activity of neurons in the DRN is predicted to trigger serotonin release in the IC and their other targets. Neurons from the DRD and DRL subregions provide the majority of the serotonergic innervation of the IC [22,40]. Unexplored aspects of this projection could create a mismatch between c-Fos activity in the DRN subregions and serotonin release in the IC. The quantitative relationship between the activity of DRN neurons, c-Fos expression, and serotonin release in the IC has not been explored and may not be linear. Serotonergic axons are also somewhat denser in the IC shell regions than in the central IC [39,63], so that the level or dynamics of serotonin release may depend on the IC subregion. Despite these sources of uncertainty, there are intriguing similarities in c-Fos expression in the DRN and previously measured serotonin release in the IC. Chief among these is a correlation of c-Fos expression with the valence of an opposite-sex interaction from the perspective of a male interacting with a female. Both serotonin release in the IC and the number of neurons expressing c-Fos in the DRD of socially housed males show negative correlations with female rejection [7,23]. This correspondence suggests that broadly measured c-Fos activity in the DRD and DRL is predictive of serotonin release in the IC, at least in some contexts.

The neuron groups of the serotonergic system projecting to the auditory midbrain and brainstem play a role in the regulation of at least two classes of social behavior. IC-projecting serotonergic neurons are involved in regulating aggression. Some DRN neurons projecting to the IC have been defined molecularly as PET1+ neurons that also express dopamine 2 (D2) receptors. This neuron class projects not only to the midbrain IC, but to many brainstem auditory nuclei, as well as nonauditory brain regions [40]. D2-expressing serotonergic neurons suppress intermale aggression when they are activated [40]. They also show nuanced sex differences in terms of the correlated patterns of projection to different auditory nuclei, suggesting the possibility of greater coordination of release across auditory regions in males [41].

As we have described, a second behavioral function relates to male responses to the vocal signals of female social partners. Naturally occurring elevation in serotonin in the IC of males correlates inversely with female rejection, including squeaks produced by females [7]. This association of squeaks with rejection of males is especially prominent at early times in an interaction that corresponds to an appetitive phase [16,20]. Playback of squeaks causes males to suppress their production of USVs [21,50]. This responsiveness to squeaks could be advantageous in avoiding injury, but it may also serve the function of responding to female cues and result in pacing an interaction to achieve an optimal reproductive outcome [64]. Male house mice in a semiwild population are able to adjust their reproductive strategies to some extent in response to external cues of the social environment [65]. Our current findings add to the evidence that males can adjust their courtship strategies on a shorter timescale in response to female cues [16,21,57,66]. Serotonin can influence this process; pharmacologically induced elevation of serotonin, either systemically or locally within the IC, significantly enhances the suppression of USVs in response to squeak playback [21,24].

5. Conclusions

These past findings and the current study all support a version of the serotonergic feedback model, in which information on the quality of an ongoing social interaction is fed back into the auditory system through serotonergic projections, altering the processing of acoustic signals and modulating the behavioral responses to those signals [8]. What the current study contributes to this model is evidence that it is not squeaks alone that drive serotonergic feedback, but a set of cues embodied in female presence, which could include fresh odor, tactile cues, mutual responsiveness of males and females, or a combination of these. Squeak playback may play a modifying role by altering the correlation of activity in different DRN subregions. These effects of female presence versus squeaks are strikingly similar to the effects of these stimuli on male vocal behaviors, in which the effects of vocal playback modify rather than drive male vocal responses to the presentation of female odor cues [50]. The serotonergic system, in its sensitivity to female signals and ability to modulate male courtship and aggression, is well-positioned as a mechanism that can contribute to male responsiveness to a dynamic social context.

Abbreviations

The following abbreviations are used in this manuscript:

USV	Ultrasonic vocalization
DRN	Dorsal raphé nucleus
TPH	Tryptophan hydroxylase
DRD	Dorsal midline subdivision of the DRN
DRV	Ventral midline subdivision of the DRN
DRL	Dorsal lateral subdivisions of the DRN
PDR	Posterodorsal subdivision of the DRN

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/brainsci16020148/s1, Table S1: data.

Author Contributions

Conceptualization, L.M.H. and M.W.; methodology, L.M.H. and M.W.; formal analysis, L.M.H., M.W. and J.B.; resources, L.M.H.; data curation, M.W. and L.M.H.; writing—original draft preparation, L.M.H.; writing—review and editing, L.M.H., M.W. and J.B.; visualization, L.M.H.; supervision, L.M.H.; project administration, L.M.H. and M.W.; funding acquisition, L.M.H. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

The study was conducted in accordance with the Guidelines for Care and Use of Laboratory Animals (National Institutes of Health). All experiments were approved by the Bloomington Institutional Animal Care and Use Committee (protocols 21-020 and 24-025, approved on 18 August 2021 and 25 July 2024, respectively) and approved by the Institutional Review Board.

Informed Consent Statement

Not applicable.

Data Availability Statement

The original contributions presented in this study are included in the Supplementary Material Table S1. Further inquiries can be directed to the corresponding author.

Conflicts of Interest

The authors declare no conflicts of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

Funding Statement

This research was funded by the National Science Foundation, grant number 1856436.