Interrogating Roles for VTA Dopamine Neurons in Locomotion and Ultrasonic Vocalizations Elicited by Different Classes of Rewarding Drugs

Abstract

Rationale:

In humans, researchers can readily query the subjective states elicited by drugs of abuse, but no such introspective readout is currently available in rodents.

Objectives:

Rats can convey information relevant to their internal states via ultrasonic vocalizations (USVs). We sought to measure USVs and locomotor behavior as parallel behavioral indices of the effects of drugs and ventral tegmental area dopamine neuron (VTA DA) manipulations.

Methods:

In male and female TH:cre rats (n=17) and wildtype littermates (n=9), effects of IP amphetamine (2 mg/kg), heroin (0.25 mg/kg), ketamine (30 mg/kg), and saline (1 ml/kg) on USVs and locomotion were recorded in a series of 1 hr tests. We also asked how chemogenetically inhibiting or exciting VTA DA impacted these behaviors.

Results:

Each drug had distinct effects on USVs and locomotion, as did chemogenetic manipulations. Amphetamine elicited high-frequency (HF) USVs and increased locomotion, heroin elicited low-frequency USVs in some rats without altering HF-USVs or locomotion, and ketamine suppressed HF-USVs and transiently increased locomotion. VTA DA neuron inhibition suppressed HF-USVs and locomotion seen after saline, and DA stimulation increased locomotion but did not induce USV production. Surprisingly, both inhibiting and stimulating DA neurons appeared to similarly suppress amphetamine- or ketamine-induced locomotion, and amphetamine-induced USVs.

Conclusions:

These results confirm that USVs comprise a readout of certain effects of drugs and brain manipulations that cannot be attained by measuring locomotion alone. Potentially, USVs might therefore be useful for interrogating the subjective states of rats, and thus help bridge human and rodent neuroscience.

Introduction

The subjective states elicited by drugs of abuse are key to the reasons why humans use them, abuse them, and become addicted to them. Since rodents also self-administer most of these drugs (Collins et al., 1983; Ettenberg et al., 1982; Koob and Weiss, 1990), presumably they also experience subjective changes that are relevant to voluntary use. In humans, drug-induced subjective states can be readily measured and compared to other concurrently-measured behavioral and physiological variables (Cooper and Haney, 2009; Hahn et al., 2025; Mayo et al., 2019; Rodriguez Santos et al., 2025; Volkow et al., 1997). In contrast, rodent researchers have powerful and often invasive means of interrogating brain functions, but limited means to interrogate subjective states in their subjects. Therefore, developing a way of understanding how rodents feel after drugs or brain manipulations could help bridge the gap between rodent and human neuroscience research, and thus facilitate translation of basic science research.

Rodents will never provide the rich linguistic descriptions of drug effects humans can provide (Huxley, 1968; Shulgin and Shulgin, 2010), but there may still be a way to interrogate their drug-induced affective states—ultrasonic vocalizations (USVs; Simola et al. 2012). Rodents use USVs to coordinate social interactions with conspecifics (Wöhr, 2017), and to communicate positive or negative affect (Brudzynski, 2015; Knutson et al., 2002), or potentially more complex subjective states (Barker et al., 2015; Hofer, 1996; Okabe et al., 2023; Panksepp and Burgdorf, 2003). Therefore, it may be possible to leverage USVs to learn more about how rats are feeling during behavioral neuroscience experiments, complementing other collected data streams (Brudzynski, 2009; Burgdorf et al., 2008; Coffey et al., 2023; Maier et al., 2010; Simola and Granon, 2019).

Adult rat USVs are commonly dichotomized based on their average frequency (pitch) (Brudzynski, 2015, 2001; Knutson et al., 2002; Litvin et al., 2007; Panksepp and Burgdorf, 2003)—high frequency (HF-USVs) are commonly called “50 kHz”, but occur between ~30–100 kHz, and low frequency (LF-USVs) are commonly called “22 kHz”, but are here defined as 18–30kHz. HF-USVs have been associated with positive affect, social coordination and play, and desire/anticipation for rewards (Burke et al., 2022; Himmler et al., 2014; Kisko et al., 2015; Ma et al., 2010; Scardochio et al., 2015), but numerous subtypes of HF-USVs exist, suggesting more complex information may be present in these communications (Burgdorf et al., 2011, 2020; Lawson et al., 2021; Saltzman et al., 2026; Wada et al., 2026; Wright et al., 2010). LF-USVs are more homogeneous in their characteristics, and long or short duration LF-USVs are elicited in a variety of aversive or dangerous situations to convey negative affect, or warn conspecifics (Blanchard et al., 1991; Borta et al., 2006; Jelen et al., 2003; Kroes et al., 2007; Litvin et al., 2007). Notably, drugs can elicit both HF- and LF-USVs (Barker et al., 2010; Coffey et al., 2023; Mahler et al., 2013; Simola and Costa, 2018). Although ventral tegmental area (VTA) dopamine (DA) neuron activity is relevant to the effects of many drugs (Koob and Volkow, 2010; Volkow et al., 2017), and chemogenetic inhibition of nucleus accumbens (NAc) neurons suppresses amphetamine-induced USVs (Lawson et al., 2021), the roles of VTA DA neurons themselves in drug-induced USVs has not yet been investigated.

Here, we examined how acute injections of different classes of drugs, and VTA DA neuron chemogenetic manipulations, impact USV production and locomotor behavior. Prior reports established a dissociation of drug effects on locomotion and USV production (Banjac et al., 2025; Costa et al., 2015; Simola and Morelli, 2015; Taracha et al., 2014; Wendler et al., 2016; Wright et al., 2012), and we here replicate and extend this finding by examining the role of VTA DA neurons in these effects. We show that drugs and designer receptors exclusively activated by designer drugs (DREADDs) have distinct patterns of effects on these behaviors, suggesting USVs might be conceived of as a window into the otherwise inaccessible internal states of rats in behavioral neuroscience experiments.

Methods

All procedures were approved by the Institutional Animal Care and Use Committee at UC Irvine, and are in accordance with the NIH Guide for the Care and Use of Animals.

Subjects:

TH:cre Long Evans rats (Lawson et al., 2023; Mahler et al., 2019, 2014; Witten et al., 2011) (n=17; female n=8) and their wildtype (WT) littermates (n=9; 4 female) were bred in-house, weaned at postnatal day (PD) 21, and experiments started at PD74+. They were pair-housed in ventilated tub cages with corncob bedding and ad libitum chow/water, in reverse 12:12 hr lighting (tests in dark phase). Experiments were run in three separate cohorts of rats.

Drugs:

D-amphetamine hemisulfate salt (2 mg/kg; Sigma), diacetylmorphine hydrochloride (heroin; 0.25 mg/kg; NIDA Drug Supply Program), ketamine hydrochloride (30 mg/kg; Patterson Veterinary), and saline (0.9%; 1 ml/kg) were prepared daily. Clozapine-n-oxide (CNO; 5 mg/kg; NIDA Drug Supply Program) was stored at 4 °C in powder aliquots in opaque bottles with desiccant, and it was prepared daily for IP injection in a 5% DMSO saline solution vehicle (VEH; 1 ml/kg).

Viral Vectors and Surgery:

AAV2 vectors for cre-dependently expressing inhibitory (hSyn-DIO-hM4D(Gi)-mCherry; titer ≥ 5×10¹² vg/mL) or excitatory (hSyn-DIO-hM3D(Gq)-mCherry; titer ≥ 6×10¹² vg/mL) DREADDs were obtained from AddGene. Both vectors were previously shown to yield functional DREADD expression specifically in DA neurons in VTA of TH:cre rats (Lawson et al., 2023; Mahler et al., 2019, 2014).

Rats were anesthetized with 2.5% isoflourane, given SC saline and meloxicam (1.0 mg/kg), then stereotaxically pressure-injected via glass pipette and Picospritzer with 0.75 μL of virus in each hemisphere, aimed at VTA (coordinates relative to bregma (mm): −5.5 AP, +/− 0.8 ML, −8.1 DV). Pipettes were left in place for 5 min prior to removal to reduce spread. WT littermates without TH:cre knock-in underwent identical surgical procedures, including injection of cre-dependent vectors (yielding no DREADD or fluorophore expression). Rats recovered in their homecages for at least 10 days after surgery.

Behavioral Testing Protocol:

Rats were handled daily for 5 min for at least 5 days prior to beginning test chamber habituation training, which occurred for 1 hr/day on 4 consecutive days. They then underwent a series of behavioral tests, each held at least 48 hr apart. On each of these test days, they first received IP injections of either CNO or VEH, then they were placed in the testing chamber for a 30 min baseline period. They then were briefly removed and injected with saline, amphetamine, heroin, or ketamine, then returned to the chamber for another 60 min (Fig. 1). Drugs were tested in the same order in all rats; amphetamine followed by heroin followed by ketamine, and each of these drugs was tested twice consecutively following injections of both CNO and VEH (counterbalanced order, 48+ hrs between injections). We chose to deliver drugs in this fixed order to standardize potential impacts of conditioning, expectancies, cross-drug tolerance or sensitization on drug-induced locomotion or USVs (Doremus-Fitzwater and Spear, 2010; Panksepp et al., 2002; Vezina et al., 1999). Rats also underwent saline tests, which likewise following counterbalanced CNO and VEH injections. These saline tests occurred either before all drugs (n=8; cohort 1), after all drugs (n=8; cohort 2), or both before and after drugs (n=10; cohort 3), but this difference in saline testing order across cohorts did not affect quantified behaviors (see results for details). We therefore combined data from saline tests for analyses of DREADD and drug effects.

Figure 1: Testing Order and Representative VTA Dopamine Neuron DREADD Expression.

(Top) Testing protocol is shown [Vehicle (VEH) & CNO + Saline tests were conducted before VEH/CNO drug tests in some rats (n=8), after drugs in other rats (n=8), and both before and after drugs in other rats (n=10)]. TH:cre rats and WT Controls were injected with cre-dependent (DIO) vectors. On each test day, rats were injected with 5 mg/kg CNO or VEH, then placed into the recording chamber. 30 min later, they were injected with saline or a drug, then placed back into the recording chamber for 1 hr. (Bottom) Typical expression of DREADD vector (DIO-hM3Dq) in a TH:cre+ rat is shown at ~5.8 mm caudal of bregma. mCherry (the cre-dependent DREADD reporter; red stain) is expressed in TH+ neurons (green stain) of VTA (colocalization appears yellow). DAPI counterstain visualizes nuclei (blue). Insets at right show magnified views of the region defined by a white box (TH: green, mCherry: red, colocalization: yellow).

Behavioral Testing Chamber:

Testing was conducted in clear acrylic tub cages (W × L × H; 25 × 46 × 20 cm) with paper fiber bedding (CareFresh small animal bedding) and a wire cage top. These testing chambers were enclosed within a wood sound attenuating box (120 × 60 × 60 cm), which contained four bays in which testing cages were placed, separated from one another by wood walls. These boxes, and the paper bedding material, helped attenuate ultrasonic noise in recordings. The top of the sound attenuating box was comprised of clear acrylic with a 3 cm hole drilled above the center of each bay, allowing placement of a USV microphone 33 cm above each testing chamber. Avisoft Bioacoustics microphones (model CM16/CMPA; frequency range: 10–200 kHz), receiver (model 816H; sampling rate 250 kHz; 16-bit resolution), and PC software (Version 3.4.4; Fast Fourier Transform 256, sampling rate 250 kHz) were employed, and recordings were done on a Windows laptop. Overhead videos were also recorded with four 1080P webcams directly above each tub cage, using recording software (Open Broadcaster Software; OBS Project) on a Windows laptop. Locomotion (distance travelled in cm) was quantified using EthoVision XT (version 15.0) rat center-point detection.

USV Quantification and Analyses:

DeepSqueak software (version 3.0 with MATLAB) was employed to detect USVs, and to identify their characteristics (Coffey et al., 2019; Lawson et al., 2021). Audio files were run though DeepSqueak’s “Rat Call” neural network, with a band-pass filter of 30–100 kHz to accurately detect HF-USVs. DeepSqueak’s post-hoc denoiser, trained on noise characteristics of our specific experimental setup, was used to remove remaining non-USV artifacts from the dataset. Audio files were also run through DeepSqueak’s “Long Rat Call neural network” with a band pass filter of 18–30 kHz to detect long-duration LF-USVs. Detected LF-USVs were also manually verified by a trained observer. Characteristics of each detected USV were calculated using the “spectrotemporal contours” output of the detection network (Glass et al., 2023; Shembel et al., 2021), which quantified for each USV its: 1) Principal Frequency (mean frequency across the duration of the call), 2) Delta Frequency (maximum change in frequency over its duration), 3) Sinuosity (length of the path between the first and last points on the contour, divided by the Euclidean distance between the first and last points), and 4) Duration.

Confirmation of DREADD Expression:

After behavioral tests, rats were transcardially perfused with ice-cold saline followed by 4% paraformaldehyde, and extracted brains were postfixed for 16 hr in 4% paraformaldehyde, then cryoprotected in 20% sucrose for 3+ days. They were sectioned coronally at 40 μm, and immunolabelled for the DREADD tag mCherry, for tyrosine hydroxylase (TH) to define VTA DA neurons, and with DAPI, using previously-described methods (Mahler et al., 2019, 2014; Martinez et al., 2023). Cre-dependent mCherry expression was confirmed in each TH:cre+ rat to be: 1) bilaterally present in VTA, 2) to not encroach into substantia nigra, 3) to be nearly exclusively present in TH+ cells, and 4) to be absent in cre-negative WT control rats (Mahler et al., 2019, 2014; Paxinos and Watson, 2006); Fig. 1). Two rats were excluded from analyses for unilateral VTA mCherry expression (not included in sample size reporting).

Statistical Analyses:

General Linear Model ANOVAs were used to analyze effects of drugs and chemogenetic manipulations on USVs and locomotion, with multiple comparison-corrected Holm or post-hoc t-tests to determine the nature of main effects and interactions, as appropriate. Repeated measures factors employed were Drug (amphetamine, heroin, ketamine, and saline), Treatment (TX; VEH and CNO), and Time (three 20 min bins across the 60 min test session). Between subjects factors employed were Sex (Female or Male), and DREADD Group (Control, hM4Di, or hM3Dq). Video recordings of locomotion for one test session with ketamine (n=2 VEH + ketamine sessions, n=2 CNO + ketamine sessions) were lost due to equipment failure, so for analyses involving those test sessions, these rats were excluded. When rats emitted no vocalizations on a session (n=3 sessions), USV characteristic analysis were not conducted, as there were no vocalizations to analyze. Behavioral data of all types were excluded as statistical outliers only if they were >3 SD from the mean of their group / test condition. Only one datum was excluded for this reason—HF-USV data from one male Control rat that received Saline. Two-tailed tests with significance thresholds of p≤0.05 were used for all analyses. Statistical analyses were conducted in R and SPSS.

Results

Histology:

TH:cre rats showed reliable mCherry/DREADD expression that was restricted to VTA, almost exclusively present in TH+ cells, and which extended throughout the rostral-caudal extent of VTA, as in prior reports (Mahler et al., 2019, 2014); Fig. 1]. No expression was observed in WT Control rats, as expected in rats without the cre gene needed to flip the vector-delivered genes into sense orientation. A total of 9 Control, 8 hM4Di, and 9 hM3Dq rats were included in these experiments. We note that studies were not primarily intended, nor fully powered to uncover sex differences in behaviors.

Saline Testing Order:

We did not find differences in behaviors emitted by rats based on whether saline tests were conducted before or after tests with drugs. Saline testing order did influence the number of HF-USVs produced, or locomotion [TX (VEH, CNO) × saline test Order ANOVA: no main effect of Order on HF-USVs: F_1,34=1.48, p=0.23; or locomotion: F_1,34=1.19, p=0.28; no difference between saline + VEH tests conducted before or after drugs on HF-USVs: t_30.5= −1.70, p=0.10, or locomotion: t_31.9= −0.07, p=0.95]. We also compared behaviors on saline-first and saline-last tests in rats that received saline at both timepoints, and again found that HF-USV production on saline days were statistically equivalent [TX (VEH, CNO) × Order ANOVA: no main effect of Order on HF-USVs: F_1,9=3.26e⁻³, p=0.96; or locomotion: F_1,9=1.09, p=0.32; no difference between saline + VEH tests conducted before or after drugs on HF-USVs: (t₉= −0.30, p=0.77) and locomotion (t₉=1.21, p=0.25)]. Therefore, saline tests were combined for subsequent analyses of drug and CNO effects, with the average data from both tests used in rats tested with saline at both timepoints.

Effects of Drugs on USVs:

We divided calls into HF-USVs (often referred to in the literature as “50 kHz” USVs) and LF-USVs (often referred to as “22 kHz USVs”) for analysis, as these broad categories are thought to convey positive and negative affective states, respectively (Brudzynski, 2015; Panksepp and Burgdorf, 2003). In the absence of VTA DA manipulation (i.e. tests on which rats received VEH, not CNO), we recorded 77,229 HF-USVs, of which 83% (63,978) were emitted on amphetamine tests. About 5% (4,202) of HF-USVs were emitted after heroin, 2% (1,553) after ketamine, and 10% (7,496) after saline.

Accordingly, the different tested drugs elicited statistically different numbers of HF-USVs (Fig. 2a; Sex × Drug ANOVA; main effect of Drug: F_3,72=24.8, p=4.03e⁻¹¹). This was driven primarily by differences between amphetamine and the other drugs (Holm-adjusted paired t-test between amphetamine and saline: t₂₅=4.89, p=4.89e⁻⁵; heroin: t₂₅=4.84, p=5.61e⁻⁵; and ketamine: t₂₅=4.92, p=4.51e⁻⁵). In addition, the number of HF-USVs was also decreased, relative to saline, by ketamine (Holm-adjusted paired t-test: t₂₅= −3.27, p=3.12e⁻³), but not by heroin (t₂₅=1.13, p=0.26).

Figure 2: Effect of Drugs on HF-USV Production and Locomotion.

Effect of drugs on HF-USV production (a) and locomotion (b) on VEH + Drug tests, and the timecourse of USV emission (c) and locomotion (d) on these tests. Animals produce more USVs following amphetamine administration than saline, and saline-relative locomotion was increased by both amphetamine and ketamine USVs were suppressed relative to saline by heroin and ketamine. *Indicates p < 0.05 relative to saline. Each animal’s average shown with symbols: males as Xs and females as Os. Error bars represent SEM.

We also quantified the spectral characteristics of emitted HF-USVs (Fig. 3). The duration of calls differed based on drug (main effect of Drug in Sex × Drug ANOVA: F_3,63=3.15, p=0.03), in that heroin-induced HF-USVs were longer duration than ketamine HF-USVs (Holm-adjusted paired t-test: t₂₂= −2.26, p=0.03), though no drugs significantly differed in call duration from saline (amphetamine: t₂₂=0.32, p=0.76; heroin: t₂₂=1.96, p=0.06; ketamine: t₂₂= −0.60, p=0.56). Sinuosity also differed based on drug (F_3,63=4.11, p=9.88e⁻³), in that amphetamine-induced HF-USV calls had greater sinuosity than saline (t₂₂=2.51, p=0.02), heroin (t₂₂=2.04, p=0.05), or ketamine (t₂₂=3.59, p=1.62e⁻³). Delta frequency also trended toward significant differences across drugs (F_3,63=2.50, p=0.07), with amphetamine causing calls with greater delta frequency than other drugs (amphetamine ve rsus saline: t₂₂=2.23, p=0.04; heroin: t₂₂=2.16, p=0.04; ketamine: t₂₂=2.36, p=0.03). No difference between drugs was seen in principal frequency (F_3,63=1.51, p=0.22).

Figure 3: Effects of Drugs on HF-USV Characteristics.

(a) Mean ± SEM acoustic characteristics of HF-USVs emitted after each of the 4 tested drugs is shown. Each animal’s average value for all the HF-USVs it emitted is shown as points within bars. Rat sex is represented by symbol shape; O=female, X=male. *indicates p<0.05 for the indicated comparison. Error bars represent SEM. (b) Histograms representing the distribution of HF-USV characteristics is shown for each parameter, including all tested drugs, in the absence of DA manipulations (VEH test days).

Amphetamine-induced HF-USVs were consistently increased across the entire session, relative to saline [main effect of Drug (amphetamine, saline): F_1,25=23.7, p=5.16e⁻⁵, no Drug × Time interaction: F_2,50=1.67, p=0.19]. Heroin suppression of HF-USVs relative to saline was specific to the first 20 min of the session (Drug × Time interaction: F_2,50, p=4.23, p=0.02; 0–20 min: t₂₅=2.14, p=0.04), while ketamine suppression was significant for the first 40 min (Drug × Time interaction: F_2,50=13.9, p=1.53e⁻⁵, 0–20 min: t₂₅=4.16, p=3.27e⁻⁴; 20–40 min: t₂₅=2.99, p=6.04e⁻³; Fig. 2c). No main effects of Sex, or Sex × Drug interactions were observed in the number of HF-USVs (Sex: F_1,24=1.61, p=0.21; Sex × Drug: F_3,72=2.11, p=0.11), or in HF-USV characteristics (Sex: Fs<1.75, ps>0.20; Sex × Drug: Fs<1.09, ps>0.36).

LF-USVs were much less commonly observed than HF-USVs on VEH test days, with only 3,064 emitted in total (less than 4% of all calls). Interestingly, the large majority of all LF-USVs (2,692; nearly 88%) were observed on sessions on which rats received heroin. However, all these heroin-induced LF-USVs were emitted by only about one third of the tested rats (n=7/26 rats; 4 of them males). These heroin-induced LF-USVs looked similar to canonical “22 kHz” USVs (Brudzynski, 2001; Litvin et al., 2007; Supplemental Fig. 1), and they had the following characteristics: Duration m ± SEM = 1.02 ± 0.01 sec; Principal Frequency = 22.0 ± 0.03 kHz; Delta Frequency = 0.90 ± 0.02 kHz, Sinuosity = 1.00 ±0.001). About 12% (372) of all observed LF-USVs were emitted on amphetamine sessions, but all of these were emitted by a single male rat (LF-USVs were 16% of the total USVs emitted by that rat), and they did not appear to differ from typical LF-USVs. No LF-USVs were emitted by any rats following ketamine or saline.

Effects of Drugs on Locomotor Behavior:

We also examined distance travelled during the same sessions on which USV responses to drugs were measured (VEH treatment days), and as expected, drugs differentially affected locomotion (Drug × Sex ANOVA: main effect of Drug: F_3,66=22.7, p=3.21e⁻¹⁰; Fig. 2b). Relative to saline, both amphetamine (Holm-corrected paired t-test: t₂₅=5.42, p=1.24e⁻⁵) and ketamine (t₂₃=7.21, p=2.40e⁻⁷) increased locomotion, but heroin did not (t₂₅= −0.39, p=0.70).

Amphetamine-induced locomotion was consistently high across the session [main effect of Drug (amphetamine, saline), no Drug × Time interaction: F_2,50=1.68, p=0.19; Fig. 2d]. Heroin did not suppress total locomotion relative to saline [Main effect of Drug (heroin, saline): F_1,25=0.15, p=0.70), and though a time-dependent effect was observed (Drug × Time interaction: F_2,50=7.96, p=9.92e⁻⁴), the lowest, initial time bin did not reach significance in posthoc comparison (0–20 min: t₂₅=1.52, p=0.13). Ketamine also increased locomotion [main effect of Drug (ketamine, saline): F_1,23=52.0, p=2.40e⁻⁷), and the timecourse of this effect was quite distinct from saline (Drug × Time interaction: F_2,46=12.7, p=3.90e⁻⁵). Specifically, ketamine elicited a strong locomotor response from approximately 20–40 min after injection (20–40 min: t₂₃= −7.83, p=6.08e⁻⁸; Supplemental Fig. 2), during which time rats stumbled around the tub cage, and appeared visibly impaired or intoxicated.

Sex also impacted locomotion, in that females travelled farther than males after all drug/saline treatments (Supplemental Fig. 2; main effect of Sex: F_1,22=45.2, p=9.29e⁻⁷; Sex difference on saline: t_12.8= −3.31, p=5.67e⁻³; amphetamine: t_11.9= −4.24, p=1.14e⁻³; heroin: t_12.4= −2.54, p=0.03; and ketamine: t_16.6= −2.85, p=0.01). The locomotor response of females to amphetamine was particularly strong (Sex × Drug interaction: F_3,66=4.03, p=0.01).

In summary, each drug tested here had a distinct pattern of effects on USVs and locomotion. Amphetamine increased both HF-USVs and locomotion throughout the session. Heroin suppressed HF-USVs, and in some rats elicited LF-USVs, without altering locomotion. Ketamine suppressed HF-USVs, and transiently increased locomotion.

Effects of Dopamine Neuron Stimulation or Inhibition on USVs:

To determine the impacts of VTA DA neuron chemogenetic manipulations on USV production, we examined effects of Treatment (TX; counterbalanced VEH and CNO) in each DREADD group (Control, hM4Di, or hM3Dq) on days in which rats received saline injections (Fig. 4a). As expected, no effect of CNO was seen on HF-USVs in Control rats without DREADDs (VEH versus CNO: t₇= −1.74, p=0.12). In contrast, VTA DA neuron inhibition with CNO suppressed HF-USV production in hM4Di rats (t₇= −4.61, p=2.46e⁻³), but surprisingly, DA neuron stimulation in hM3Dq rats failed to reliably elicit HF-USVs (t₈= −0.50, p=0.63). When we compared effects of TX across all three DREADD groups (TX × DREADD ANOVA), we saw that CNO effects depended on DREADD group (F_2,22=4.04, p=0.03).

Figure 4: Effect of VTA DA manipulation on HF-USVs and locomotion.

VTA DA inhibition suppressed both HF-USV production (a, c) and locomotion (b, d), while VTA DA excitation was sufficient to increase locomotion (b, d) but not USV emission (a, c). Whole session data is shown in (a, b), and timecourse in (c, d). Each animal’s average shown as with a symbol, males as Xs and females as Os. Error bars represent SEM. *p < 0.05, CNO versus VEH.

With regard to the timecourse of VTA DA manipulation effects on HF-USVs, we found no evidence that time modulated CNO effects in Control rats (no TX × Time interaction: F_2,14=1.14, p=0.34) or in hM3Dq rats (TX × Time interaction: F_2,16=0.25, p=0.78), but we did observe that hM4Di inhibition of HF-USVs was strongest in the first 20 min of the session (TX × Time interaction: F_2,14=15.4, p=2.91e⁻⁴; 0–20 min: t₇= −4.20, p=4.01e⁻³; Fig. 4c). There were no sex differences in the effects of VTA DA manipulations on HF-USV production (no main effects of Sex, or Sex × TX interactions in Control rats: Fs<0.571, ps>0.48; hM4Di rats: Fs<0.445, ps>0.52; or hM3Dq rats: Fs<0.3.61, ps>0.10).

Very few LF-USVs were seen after either VEH or CNO in any DREADD group on saline test days. 33 total LF-USVs were observed on these sessions, of which 94% (n=31) were emitted by one male rat after VTA DA stimulation, and 6% (n=2) by one female rat after inhibition of VTA DA neurons.

Effects of Dopamine Neuron Stimulation or Inhibition on Locomotion:

We next determined the impacts of VTA DA manipulations on locomotion during saline tests (Fig. 4b). As expected, no effects of CNO were seen on locomotion in Control rats without DREADDs (VEH versus CNO TX: t₈=0.48, p=0.64). In contrast, VTA DA inhibition with CNO in hM4Di rats suppressed locomotion relative to VEH (t₇= −5.11, p=1.37e⁻³), and VTA DA stimulation with CNO in hM3Dq rats enhanced it (t₈=6.49, p=1.88e⁻⁴). Accordingly, when effects of CNO across all three DREADD groups were statistically compared, locomotion markedly differed (interaction of TX × DREADD group: F_2,23=25.10, p=1.68e⁻⁶).

We also examined the timecourse of VTA DA manipulation effects on locomotion (Fig. 4d). CNO treatment was again without effect in Control rats (no main effect of TX: F_1,8=0.24, p=0.64). VTA DA neuron inhibition of locomotion in hM4Di rats was strongest early in the session (TX × Time interaction: F_2,14=9.29, p=2.69e⁻³), though there was significant VEH-relative suppression by CNO in this group at all three time bins (0–20 min: t₇= −6.73, p=2.69e⁻⁴; 20–40 min: t₇= −3.35, p=0.01; 40–60 min: t₇= −4.21, p=3.99e⁻³). VTA DA stimulation robustly increased locomotion to a similar extent across the entirety of the session (main effect of TX in hM3Dq rats: F_1,8=42.20, p=1.88e⁻⁴; no TX × Time interaction: F_2,16=0.70, p=0.51). There were no sex differences in the effects of VTA DA manipulations on locomotion in Control and hM3Dq rats (no main effects of Sex, or Sex × TX interactions in Control rats: Fs<4.24, ps>0.07; hM3Dq rats: Fs<2.18, ps>0.18), though we found slightly greater locomotor suppression in hM4Di females than males (TX × Sex interaction: F_1,6=8.43, p=0.03).

Effects of VTA DA Neuron Manipulations on Drug-Induced USVs:

We next asked how VTA DA manipulations affected USV production that was elicited by drugs. The drugs we tested markedly differed in their propensity to elicit USVs, with amphetamine eliciting by far the most, and both heroin and ketamine suppressing USV production relative to saline. Therefore, primary analyses examined effects of VTA DA manipulations on USVs separately for each drug. Furthermore, the prior sections demonstrated that CNO had no off-target effects in Control rats without DREADDs, and we confirmed this result here as well [TX (VEH, CNO) × Drug (saline, amphetamine, heroin, ketamine) ANOVA in Control rats; HF-USV: no main effect of TX: F_1,8=0.34, p=0.58; or TX × Drug interaction: F_3,24=0.03, p=0.99; Locomotion: no main effect of TX: F_1,6=0.003, p=0.98; or TX × Drug interaction: F_3,18=0.31, p=0.81; Control rat data is shown in Supplemental Fig. 3]. Therefore, to maximize interpretability of the interactive effects of excitatory or inhibitory DREADD manipulations on USVs and locomotion elicited by drugs, we directly compared VTA DA neuron stimulation to VTA DA neuron inhibition [TX (VEH, CNO) × DREADD (hM3Dq, hM4Di)] in the following analyses, and show these data as change from VEH after CNO in these DREADD groups in Fig. 5.

Figure 5: Effects of VTA DA Manipulations on Drug-Induced HF-USVs and Locomotion.

(a) Effects of chemogenetic manipulations on drug-induced HF-USVs are shown by plotting CNO day minus VEH day data for the same drug. For amphetamine (Amph.) tests, CNO in both inhibitory and excitatory DREADD rats suppressed HF-USVs. DREADDs did not significantly alter HF-USVs elicited by heroin or ketamine. (b) Effects of chemogenetic manipulations on drug-induced locomotion is shown by plotting CNO day minus VEH day data for same drug. For amphetamine and ketamine tests, CNO in both inhibitory and excitatory DREADD rats suppressed locomotion. For heroin tests, CNO in inhibitory but not excitatory DREADD rats suppressed locomotion. Each animal’s change from VEH day shown as a point, males as Xs and females as Os. Error bars represent SEM. *indicates p<0.05 significant difference, and #indicates p≤0.1 trend toward CNO effect (CNO-VEH, compared to zero in t-test) for the associated bar.

For amphetamine, both inhibition and excitation of VTA DA suppressed HF-USV production, and both manipulations did so to a similar extent [TX (VEH, CNO) × Group (hM4Di, hM3Dq) ANOVA; main effect of TX: F_1,15=5.96, p=0.03; no TX × DREADD interaction: F_1,15=0.04, p=0.85; Fig. 5a], though statistically significant CNO effects were not noted in posthoc tests of either hM4Di (t₇= −1.46, p=0.18) or hM3Dq (t₈= −1.99, p=0.08) groups. No effect of CNO was seen in Control rats without DREADDs (t₈= −0.11, p=0.91; Supplemental Fig. 3a). VTA DA manipulations did not alter the timecourse of amphetamine-induced HF-USV production (TX × DREADD × Time ANOVA; no main effect of Time, or TX × Time interactions: Fs<2.35, ps>0.11). Females emitted more HF-USVs following amphetamine than males, but this effect did not differ with VTA DA manipulation (Sex × TX ANOVA: main effect of Sex: F_1,15=5.54, p=0.03; no Sex × TX interaction: F_1,15=2.00, p=0.18).

Amphetamine again induced very few LF-USVs, and this was not altered by VTA DA manipulations. 372 LF-USVs were emitted by 1 male hM3Dq rat after VEH + amphetamine. On CNO + amphetamine tests, a total of 137 LF-USVs were observed, 45% (n=62) of which were emitted by a male hM4Di rat, another 45% (n=62) of which were emitted by the same male hM3Dq rat which emitted LF-USVs after VEH. The remaining 10% (n=13) of LF-USVs seen after CNO were emitted by a male Control rat.

For heroin and ketamine, VTA DA manipulations did not measurably alter production of HF-USVs (heroin: no main effect of TX: F_1,15=2.96, p=0.11; no TX × DREADD interaction F_1,15=0.28, p=0.61; ketamine: no main effect of TX: F_1,15=4.05, p=0.06; or TX × DREADD interaction F_1,15=2.13, p=0.17), though we note that this analysis was likely impacted by a floor effect due to the low number of USVs emitted after both drugs. No time-dependent effects (TX × DREADD × Time ANOVA; no TX × Time interaction following heroin: Fs<2.39, ps>0.11), nor sex differences were found in VTA DA manipulation effects on HF-USVs elicited by heroin (no main effect of Sex, or Sex × TX interaction in any DREADD group: Fs<0.541, ps>0.47). CNO altered the timecourse of ketamine-induced HF-USV production [TX × Time × DREADD (hM3Dq, hM4Di) ANOVA: Time × TX interaction: F_2,30=4.64, p=0.02] with the most potent suppression in the beginning of the session (0–20 min: t₁₆= −2.12, p=0.05). There were no sex differences in effects of VTA DA manipulations on HF-USVs elicited by ketamine (no main effect of Sex, or Sex × TX interaction: Fs<0.00177, ps>0.97).

Heroin caused some rats to emit LF-USVs, but VTA DA manipulations did not appear to affect these. On CNO + heroin test days, a total of 2,579 LF-USVs were recorded, which were emitted by 8 of 26 tested rats—three of these also emitted LF-USVs on VEH + heroin test days (1 hM4Di rat emitted 1 LF-USV on VEH and 35 on CNO, 1 hM3Dq rat emitted 693 following VEH and 192 on CNO, and 1 Control rat emitted 408 following VEH and 39 on CNO). In the 5 rats that emitted LF-USVs only on CNO + heroin tests, 3% (78 total) of them were emitted by 3 hM4Di rats (3 of them female), 66% (1,706 total) were emitted by 2 hM3Dq rats (1 female), and the remaining 31% (795 total) were emitted by 2 male Control rats. We did not observe any LF-USVs following ketamine administration.

Effects of VTA DA Neuron Manipulations on Drug-Induced Locomotion:

We next asked how VTA DA manipulations affected locomotion elicited by drugs, using the same analysis strategy described above for USVs.

On amphetamine tests, inhibition and excitation of VTA DA similarly suppressed locomotion [TX (VEH, CNO) × DREADD (hM3Dq, hM4Di) ANOVA: main effect of TX: F_1,15=9.04, p=8.84e⁻³; no TX × DREADD interaction: F_1,15=0.32, p=0.58; Fig. 5b], with posthoc comparison for hM3Dq reaching statistical significance (hM3Dq: t₈= −2.39, p=0.04) and hM4Di showing a trend (hM4Di: t₇= −2.02, p=0.08). No effect of CNO was seen in Control rats without DREADDs (No effect of TX in Control rats: t₈=0.06, p=0.95; Supplemental Fig. 3b). CNO did not alter the timecourse of amphetamine-induced locomotion [TX × Time × DREADD (hM3Dq, hM4Di) ANOVA: no TX × Time interaction: F_2,32=0.99, p=0.38]. On amphetamine tests, females again travelled farther than males (main effect of Sex: F_1,13=7.13, p=0.02), and VTA DA manipulation did not alter this (no Sex × TX interactions in any DREADD group: Fs<2.12, ps>0.17).

On heroin tests, VTA DA inhibition suppressed locomotion in hM4Di rats, but hM3Dq stimulation did not affect locomotion [TX (VEH, CNO) × DREADD (hM3Dq, hM4Di) ANOVA: main effect of TX: F_1,15=4.75, p=0.05; TX × DREADD interaction: F_1,15=6.12, p=0.03; Fig. 5b]. No effect of CNO was seen in Control rats without DREADDs (No effect of TX in Control rats: t₈=0.28, p=0.78). In hM4Di rats, VTA DA inhibition suppressed locomotion after heroin (t₇= −3.32, p=0.01), which was equivalent to the DA inhibition-induced locomotor suppression seen on saline test days in these rats [no TX (VEH, CNO) × Drug (heroin, saline) interaction: F_1,7=0.06, p=0.82]. Stimulating DA with CNO in hM3Dq rats did not increase locomotion on heroin tests, as it did in the absence of heroin on saline days [TX (VEH, CNO) × Drug (heroin, saline) interaction: F_1,8=5.89, p=0.04; no TX effect in heroin tests of hM4Di rats: t₈=0.11, p=0.91; but a TX effect on saline tests of these rats: t₈=6.49, p=1.88e⁻⁴]. CNO did not alter the timecourse of heroin-induced suppression of locomotion in hM4Di rats (no TX × Time interaction: F_2,14=0.40, p=0.68) or in hM3Dq rats (no TX × Time interaction: F_2,16=2.26, p=0.14). Males and females travelled the same distance after heroin (no main effect of Sex in hM3Dq: F_1,7=0.88, p=0.38 or hM4Di: F_1,6=0.85, p=0.39 rats), and VTA DA manipulation did not alter this (no Sex × TX interactions in hM3Dq: F_1,7=0.02, p=0.89 or hM4Di: F_1,6=3.32, p=0.12) rats.

On ketamine tests, locomotion was similarly suppressed after both inhibition and excitation of VTA DA neurons [TX (VEH, CNO) × DREADD (hM3Dq, hM4Di) ANOVA: main effect of TX: F_1,13=9.48, p=8.77e⁻³; no TX × DREADD interaction: F_1,13=9.87, p=0.98; Fig. 5b], though in posthoc tests the effect of CNO was significant in hM4Di rats (t₇= −2.44, p=0.04), but not hM3Dq rats (t₆= −1.95, p=0.10). No effect of CNO was seen in Control rats without DREADDs (t₆=0.25, p=0.81). CNO altered the timecourse of ketamine-induced locomotion [TX × Time × DREADD (hM3Dq, hM4Di) ANOVA: Time × TX interaction: F_2,26=4.37, p=0.02] with the most potent suppression of locomotion seen in the middle of the session (20–40: t₁₄= −4.23, p=8.37e⁻⁴), when ketamine induced the most locomotor enhancement. Females travelled farther than males overall after ketamine (main effect of Sex: F_1,11=14.10, p=3.11e⁻³) but this effect did not differ with VTA DA manipulation (no interactions between Sex and other variables: Fs<0.39, ps>0.55).

Discussion

Here we confirmed that USVs provide information about the effects of drugs of abuse, and of VTA DA neuron manipulations, that cannot be captured by measuring locomotion alone. We examined USVs and locomotor behavior elicited by acute administration of amphetamine, heroin, ketamine, and saline, and also by chemogenetic stimulation or inhibition of VTA dopamine neurons. We demonstrate that USVs and locomotion yield complementary, largely non-overlapping behavioral readouts, and we report several surprising findings that may lend insight into the effects of drugs and VTA DA neuron manipulations on rat subjective states.

Effects of Drugs on USVs and Locomotion:

Drugs of abuse induce rewarding and pleasurable effects in most humans who use them, and preclinical scientists generally presume this is also the case in rodents. Many such drugs also impact activity levels, and both arousing and sedative drugs can be rewarding and addictive, so we sought to examine the relationship of drug-induced rat USVs to locomotion. We found that amphetamine increased both reward-related HF-USVs and locomotion, ketamine instead decreased HF-USVs while transiently increasing locomotion, and heroin did not significantly alter HF-USVs or locomotion, replicating previous findings with these drugs (Banjac et al., 2025; Costa et al., 2015; Simola and Morelli, 2015; Taracha et al., 2014; Wendler et al., 2016; Wright et al., 2012). Clearly these results indicate that USVs and locomotion reflect distinct drug effects that are not inherently linked, and suggest that USVs do not merely reflect behavioral activation or arousal.

One possibility is that instead of reflecting generalized arousal, USVs may be linked instead to activation of linguistic or communicative arousal in particular that is elicited by psychostimulants like amphetamine. Consistent with this speculation, amphetamines induce vocalization in rats (Ahrens et al., 2009; Rippberger et al., 2015; Schwarting, 2023; Wöhr, 2021), prairie voles (Ma et al., 2014), monkeys (Bellarosa et al., 1980; Miczek and Gold, 1983), and birds (de Lanerolle, 2008; Koc and Marley, 1982; Matheson and Sakata, 2015), and they robustly induce speech in humans (Heishman and Stitzer, 1989; Higgins et al., 1989; Higgins and Stitzer, 1989; Marrone et al., 2010; Stitzer et al., 1978; Strakowski et al., 1996; Wardle et al., 2012).

Amphetamines also altered the characteristics of HF-USVs, increasing their sinuosity and frequency range (Fig.3), and we note that psychostimulants also alter the linguistic structure of speech in humans (Bedi et al., 2014; Marrone et al., 2010). Amphetamine effects here are consistent with prior reports showing amphetamine preferentially induces frequency modulated and “trill” HF-USVs (Ahrens et al., 2009; Lawson et al., 2021; Mulvihill and Brudzynski, 2018; Schwarting, 2023), which some have suggested may reflect strongly arousing motivational or affective internal states, as opposed to shorter, less frequency modulated calls which may preferentially serve social coordination functions (Brudzynski, 2015). We are intrigued by HF-USV heterogeneity, including the possibility that complexity in these vocalizations may even represent additional “content” that could be useful for interpreting subjective states in rats (Lawson et al., 2021; Takahashi et al., 2010; Wright et al., 2010). However, since amphetamine induced the bulk of all observed calls here, we were unable to ascertain how other drugs and DA neuron manipulations impacted USV characteristics/subtypes. Additional work studying other doses and drugs, other co-occurring behaviors in more complex and social situations, and using larger sample sizes will be required to further explore how USV characteristics may inform interpretations of what USVs heterogeneity “means.” Such investigations could lead to even richer understanding of drug-induced subjective states in rats, and how they relate to the subjective experiences of humans experiencing effects of the same drugs.

This said, the present results suggest that HF-USVs in general are very unlikely to merely reflect pleasure or euphoria that is elicited by the intoxicating drugs examined here. The doses of amphetamine (2 mg/kg), heroin (0.25 mg/kg) and ketamine (30 mg/kg) tested are generally behaviorally rewarding, as measured using assays like conditioned place preference (Häggkvist et al., 2009; Tzschentke et al., 2006; van der Kam et al., 2009), and humans taking recreational doses of each of these drugs also usually report pleasurable feelings (Dahan and Niesters, 2026; Johanson and Uhlenhuth, 1980; Martinez et al., 2022; Seecof and Tennant, 1986). For example, many humans who use opioids like heroin consider their euphoric effects to be very salient and desirable, especially during early experiences with them (Bornstein and Pickard, 2020; Zacny and Gutierrez, 2003). However, we here replicated prior findings that experimenter-administered opioids do not elicit HF-USVs (Best et al., 2017; Hamed and Boguszewski, 2018; Shepherd et al., 1992; van der Poel et al., 1989; Vivian and Miczek, 1993), though prior studies showed that rats do emit them during heroin and fentanyl self-administration (Adamatzky et al., 2025; Avvisati et al., 2016; Coffey et al., 2023). Further work is required to clarify the difference between the drugs and doses which do or do not tend to elicit HF-USVs in order to understand the factors which govern this distinction.

It is also the case drugs can have negative, aversive effects in some individuals. This may be especially true for opioids, which are initially experienced not as rewarding, but as noxious, nauseating, and aversive in a significant number of people (Porreca and Ossipov, 2009; Smith and Laufer, 2014). Here, a minority of rats emitted a large number of aversion-linked LF-USVs after heroin, and we suspect this may reflect an analogous individual difference. We did not test whether rats that emitted LF-USV differed from others in terms of their preference for, or willingness to self-administer heroin or other drugs, but it would be of interest to test this possibility directly in a future experiment. Of course individuals also differ in their subjective enjoyment of amphetamine and ketamine drug effects (Aguado et al., 1997; Castells et al., 2018; Stahl, 2017), but we did not see notable LF-USV production after these drugs. More work is required to determine how drug-induced LF-USVs may shed light on individual differences in drug effects that may have implications for risk or resilience to their addictive potential.

Therefore, the present findings suggest, as does most of the prior literature, that rat USVs are not a straightforward readout of mere drug-induced arousal, reward, or dysphoria. Their modulation by social factors and their communicative functions should not be ignored, and individual differences in their production may be useful for parsing both normal and psychopathology-relevant brain states in rat behavioral neuroscience experiments (Barker et al., 2015, 2010; Brudzynski, 2015; Burgdorf et al., 2020; Schwarting, 2023; Simola, 2015; Taracha et al., 2012; Wendler et al., 2016; Wöhr, 2021).

Effects of VTA DA Neuron Manipulations:

Both USVs and locomotion elicited by drugs, and especially psychostimulants, are tightly linked to mesolimbic dopamine (Burgdorf et al., 2001; Serra et al., 2024; Simola and Morelli, 2015; Thompson et al., 2006; Williams and Undieh, 2016; Willuhn et al., 2014), and the present study points to an interesting new layer of complexity in this relationship. We used a well-established chemogenetic method to selectively activate or inhibit VTA DA neurons and their release of dopamine in transgenic TH:cre rats, a technique that is only possible in animals. The subjective effects of these manipulations are of interest but unknown, and therefore we sought to interrogate them by examining USVs.

We previously showed that chemogenetically activating VTA dopamine neurons elicits robust locomotion for at least 2 hrs (Bonaventura et al., 2019; Mahler et al., 2019). In this manner it is superficially similar to amphetamine, a drug that also elicits robust locomotion on a similar timescale, as we also replicated here. Since amphetamine is thought to increase locomotion largely via its ability to increase synaptic DA levels in nucleus accumbens and other striatal regions (Ikemoto, 2002; Steinkellner et al., 2014), and since amphetamine also robustly induces HF-USVs, we hypothesized that stimulating VTA DA neuron activity would also lead to USV production. However, this was not the case—we saw no evidence of chemogenetic stimulation consistently eliciting either HF- or LF-USVs, despite the fact that locomotion was markedly increased by the same manipulation. This further supports the notion that HF-USVs do not reflect mere psychomotor activation, and suggests that fundamentally distinct neural substrates underlie these behavioral effects. Prior studies support such a dissociation as well, since for example several pharmacological manipulations in NAc are capable of blocking USVs, but not locomotion induced by amphetamine (de Oliveira Guaita et al., 2018; Rippberger et al., 2015).

Our findings are also in line with prior reports (Brudzynski, 2015; Simola et al., 2016; Wintink and Brudzynski, 2001; Wöhr et al., 2015a) suggesting that amphetamine’s ability to induce HF-USVs does not rely solely upon the drug’s actions on VTA DA projections to NAc. Prior work using pharmacology showed that amphetamine injections induce USVs in a manner dependent upon both D1 and D2 dopamine receptors, including in NAc (Serra et al., 2024; Thompson et al., 2006; Wright et al., 2013), and we previously showed using chemogenetics that inhibiting NAc neurons themselves suppresses amphetamine-induced USVs (Lawson et al., 2021). Since VTA is the main source of dopaminergic inputs to NAc, it is unclear why chemogenetically stimulating DA cells there fails to induce USVs. The answer is unlikely to involve lack of accumbens DA release caused by this DREADD approach, since we have previously shown that this occurs in this preparation (Mahler et al., 2019). It is possible that non-VTA dopaminergic cells preferentially promote amphetamine-induced USVs, but it also seems likely that non-DA neurotransmitters which modulate amphetamine-induced USVs, like serotonin and adenosine (Rippberger et al., 2015; Simola et al., 2016; Wöhr et al., 2015b, 2015a), also contribute to this difference between amphetamine and direct VTA DA neuron stimulation. In addition, it is possible that DREADD stimulation simply yields a different level or pattern of increased DA signaling than amphetamine. Dedicated research will be required to further explore these possibilities.

We also found that chemogenetically inhibiting VTA DA neurons suppresses HF-USV production and locomotor activity. Combined with DA neuron stimulation effects discussed above, this may imply that VTA DA is necessary, but not sufficient for USV production. Alternatively, it could reflect a nonspecific suppression of behaviors in general by VTA DA neuron inhibition, rather than a specific decrease in USVs. However, this possibility seems unlikely since we found that VTA DA neuron chemogenetic inhibition suppressed cue driven food seeking (Halbout et al., 2019), cocaine- or yohimbine-primed reinstatement of cocaine seeking (Mahler et al., 2019), effortful operant responding for food (Lawson et al., 2023), and locomotion (Mahler et al., 2019), but did not affect spontaneous rearing or cue-induced cocaine reinstatement [in fact, pressing of a non-reinforced lever during cocaine cue reinstatement was actually increased by DA neuron inhibition (Mahler et al., 2019), arguing against a global inhibition of all behaviors by this manipulation].

One might also suspect that suppressing DA might cause changes in hedonic or affective states that resemble anhedonia, a DA-linked symptom cluster typical in disorders like depression or addiction (Nestler and Carlezon, 2006; Russo and Nestler, 2013; Wise, 2006, 1982), but if so, this was not reflected in USVs. We found that chemogenetically inhibiting VTA DA neurons did not elicit negative affect-related LF-USVs, even in individual animals that sometimes produced these aversion- or threat-related calls when they were given heroin. This is consistent with a disconnection between DA neurotransmission and negative affect (Berridge, 2007; Bromberg-Martin et al., 2010; Matthews et al., 2016; Rutledge et al., 2017). It is also consistent with prior work showing that LF-USVs are linked much more clearly to cholinergic rather than dopaminergic signaling (Brudzynski, 2015).

Effects of VTA DA Neuron Manipulations on Drug-Induced Locomotion and USVs:

A surprising finding of this experiment was that both stimulating and inhibiting DA neurons appeared to decrease the ability of amphetamine and ketamine to increase locomotion, and the ability of amphetamine to elicit USVs (the low levels of ketamine-induced vocalizations were not affected by DA manipulations). This result is somewhat preliminary due to borderline statistical power of posthoc analyses, but the trends are clear, and puzzling. We expected that VTA DA inhibition would suppress amphetamine-induced USVs and locomotor behavior, as has been reported previously (Runegaard et al., 2019), but we were surprised to find that stimulating VTA DA also suppresses amphetamine- and ketamine-induced locomotion, and amphetamine-induced USVs. On its face, this could imply that either excessively high- or excessively-low activity of DA neurons disrupts amphetamine and ketamine-induced locomotion and/or USV production. It is also possible that amphetamine or ketamine alter the ability of DREADDs to modulate DA neuron activity. This possibility is also consistent with our finding that although DA neuron inhibition suppressed locomotion in the presence of heroin, stimulating DA neurons in the presence of heroin failed to increase locomotion, as normally occurs with great consistency (Bonaventura et al., 2019; Brodnik et al., 2020; Mahler et al., 2019). Further mechanistic investigations will be needed to unravel the meaning of these intriguing preliminary findings.

Limitations of These Studies:

This study has a number of limitations that should be considered. First of all, we only tested a single dose of each drug, and it is not clear how other doses of these drugs would affect behavioral outcomes. Follow-up studies would also be useful to examine other routes of volitional or non-contingent drug administration, other classes of drugs, or other drugs within the tested classes of psychostimulants, opioids, and dissociative anesthetics, as this could help elucidate potential impacts of pharmacokinetic, pharmacodynamic, metabolism, and other factors not examined here. Conceivably, other doses of CNO used to actuate DREADDs could also have altered conclusions, for example if a lower dose elicited milder DA stimulation that yielded USV production in addition to locomotion.

Since amphetamine was administered before other drugs in this experiment and it induced the most USVs of any drug measured, we also cannot exclude the possibility that lingering primary or conditioned effects of amphetamine could have influenced subsequently measured behaviors. This said, the lack of difference between saline tests held before versus after the other drugs argues against the likelihood of this possibility. We only tested two behaviors here (USVs and locomotion), and additional behavioral readouts would also likely enrich findings in subsequent experiments of this kind.

It would also be interesting to examine how manipulating other cell populations relevant to drug effects, including neurons in substantia nigra, amygdala, and other mesocorticolimbic regions would impact conclusions about mechanisms of drug-induced USV production. Likewise, DREADDs can also be used to manipulate specific VTA dopamine pathways, and since projections to accumbens, cortex and amygdala are known to have distinct behavioral functions (Halbout et al., 2019; Mahler et al., 2019), analogous studies in these circuits would also be interesting. Potential impacts of species and strain on these findings may also have impacted behavioral effects observed here (Barker et al., 2015; Caruso et al., 2022; Peleh et al., 2019; Premoli et al., 2023). Finally, though most of the key results were statistically robust with the sample sizes tested, larger samples would be desirable, especially when it comes to potential sex-dependent effects, interacting effects of sex, drugs, and DREADDs, or effects where statistical floor effects might have been present (e.g. ketamine and heroin induced few USVs, so it would be difficult to observe a chemogenetic suppression of these behaviors). Despite these limitations, we hope that the present report will serve as a framework for future experiments of this kind which explore the potential of USVs to provide unique, novel insight into the subjective experiences of rats.

Summary:

We intend these studies primarily as a proof of concept, setting the stage for more detailed behavioral and mechanistic research on USVs elicited by drugs and DREADDs, thus leveraging the potential explanatory value of rat USVs for behavioral neuroscience. If USVs provide complementary information relevant to rat subjective states as we and others have hoped, this would be a great advance toward the goal of bridging human and animal neuroscience research. If so, this will likely facilitate translation of basic science results to understanding of the human brain, and may help leverage these insights into new clinical strategies for identifying and treating disorders like addiction.

Supplementary Material

Supplemental Figure 1

Supplemental Figure 1: Heroin-Induced LF-USVs. Example spectrograph of LF-USVs recorded from a rat after heroin administration.

Supplemental Figure 2

Supplemental Figure 2: Effect of Drugs on Locomotion Timecourse. Effect of drugs locomotion in both sexes over the 1 hr session on VEH day. Males and females both travel farther following amphetamine and ketamine than saline, and the effects of these drugs are more pronounced in females. Error bars represent SEM

Supplemental Figure 3

Supplemental Figure 3: Effects of CNO on HF-USVs and Locomotion in Control Rats. There were no effects of CNO on (a) drug-induced HF-USV production or (b) locomotion in Control rats during CNO tests, relative to VEH day tests with the same drugs. Each animal’s average shown as a point, males as Xs and females as Os. Error bars represent SEM.

Data availability statement:

Data will be made available upon request.