1. Introduction
Dopamine has been extensively studied for its critical role in reward learning [
1,
2,
3,
4]. Phasic dopamine signals are robustly engaged by rewarding stimuli and their predictors, tracking a prediction error-based signal and supporting the encoding of predictive cue–outcome associations [
5,
6,
7,
8]. Moreover, emerging evidence suggests that dopamine is modulated by the saliency and novelty of external events and stimuli [
9,
10,
11,
12,
13,
14], dynamically integrating this information over time to signal environmental change and uncertainty [
15], suggesting a broader role in shaping learning and memory beyond the classical reward prediction error framework. Beyond reward learning, these processes are increasingly recognized as important for adaptive responses to aversive experiences, and disruption in them contributes to maladaptive learning observed in psychiatric conditions such as substance use disorders, post-traumatic stress disorder (PTSD), and other anxiety or stress disorders [
16,
17]. Consistent with this, fear extinction and recovery depend on coordinated activity across the amygdala, medial prefrontal cortex (mPFC), hippocampus, and nucleus accumbens (NAc), with the NAc increasingly recognized as an integrative hub for motivational and salience-related signals that influence fear recovery [
11,
17,
18,
19,
20].
Recent work has highlighted that dopamine is not limited to reward contexts but is also critically engaged during aversive learning. For example, dopamine neurons in the midbrain and their terminals in the nucleus accumbens (NAc) respond robustly to aversive stimuli and their predictors [
11,
12,
21,
22,
23,
24,
25], and dopamine release is involved in active avoidance and conditioned punishment [
11,
26,
27,
28]. Studies from our group and others demonstrated that NAc core dopamine release is evoked by aversive stimuli, such as footshocks, as well as when those footshocks are predicted but absent [
11,
12,
22,
24,
25,
29]. Moreover, our group and others showed that this dopamine response during the unexpected omission of aversive outcomes causally regulates fear extinction and safety learning [
11,
18,
25,
29]. Specifically, we showed that optogenetic amplification of omission-induced dopamine release in the NAc core impaired fear extinction and safety learning. These findings suggest that dopamine may function as a teaching signal, signaling the unexpected presence or absence of important outcomes and thereby dynamically shaping how animals attend to, encode, and adapt their behavior to both threat- and safety-related experiences.
Previous work has shown that dopamine signaling contributes to both the acquisition and recall of extinction [
11,
18,
27,
30], and that dopamine transients during the omission of expected aversive outcomes promote fear extinction and safety learning [
11,
18,
27,
30]. However, its role in the re-emergence of fear after extinction remains unclear. Fear can return following extinction through several processes, including spontaneous recovery, contextual renewal, and reinstatement, which involve partially overlapping but distinct mechanisms [
17,
31]. Unlike spontaneous recovery or contextual renewal, reinstatement is triggered by renewed exposure to an aversive event following extinction, making it particularly relevant for understanding how unexpected threat experiences reactivate extinguished fear memories [
31,
32,
33]. However, despite evidence implicating dopamine signaling in extinction learning and recall, the involvement of NAc core dopamine in the reinstatement of previously extinguished fear responses remains largely unknown. Therefore, in the present study, we specifically examined dopamine dynamics in the NAc core during shock-induced reinstatement of a previously extinguished fear memory. Clarifying this mechanism is essential for understanding how maladaptive fear responses re-emerge after extinction and may offer critical insight into the neural basis of relapse and recurrence in anxiety- and stress-related disorders.
2. Materials and Methods
Adult male and female C57BL/6J mice (6–8 weeks old; Jackson Laboratories, Bar Harbor, ME, USA; stock no. 000664) were housed five per cage and maintained on a 12-h reverse light/dark cycle with ad libitum access to food and water. All experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at Temple University and conducted in accordance with institutional guidelines.
2.1. Surgical Procedure
At least one hour before surgery, mice received ketoprofen (5 mg/kg, s.c.) for analgesia. Anesthesia was induced with 5% isoflurane and maintained at 2% while mice were secured in a stereotaxic frame (David Kopf Instruments, Tujunga, CA, USA). Using a 0.10-mL NanoFil syringe (WPI, Sarasota, FL, USA) fitted with a 34-gauge needle, AAV5.CAG.dLight1.1 (UC Irvine [
34]) was unilaterally injected into the NAc (AP: +1.4 mm, ML: +1.5 mm, DV: −4.3 mm from bregma; 10° angle) at 50 nL/min for a total volume of 500 nL. The needle was left in place for 7 min before being slowly withdrawn. A fiber-optic cannula (400 μm core, 0.48 NA; Doric Lenses, Quebec, Canada) was then implanted immediately dorsal to the injection site and secured to the skull with adhesive cement (C&B Metabond). Mice were given at least 6 weeks to recover, allowing for robust viral expression prior to behavioral testing.
For optogenetic experiments, AAV5.Ef1a.DIO.hChR2.eYFP (ChR2; UNC vector core) or AAV5-Ef1a-DIO.eNpHR.3.0-eYFP (NpHR; Addgene, Watertown, MA, USA) and AAV9.rTH.PI.Cre.SV40 (Addgene) was injected into the VTA (unilaterally for ChR2 and bilaterally for NpHR; bregma coordinates: A/P: −3.16 mm; M/L: −0.5 mm; D/V: −4.8 mm; no angle) of C57BL/6J mice. Unilateral (for ChR2) or bilateral (for NpHR) 200 μm-core diameter fiber-optic implants were placed into the NAc core (bregma coordinates matching fiber photometry surgeries). This allowed for the photo-stimulation or photo-inhibition of dopamine responses only in dopamine axons projecting from the VTA and synapsing in the NAc core. Control animals received AAV5.Ef1a.DIO.eYFP injections into the VTA instead of ChR2 or NpHR. Bilateral optical inhibition was used to maximize suppression of dopaminergic signaling and reduce the potential for compensatory activity from the contralateral hemisphere, while unilateral stimulation was sufficient to reliably enhance dopamine signaling during reminder shock. We validated this optogenetic approach in our previously published studies [
12,
25].
2.2. Histology
Mice were deeply anesthetized with an intraperitoneal injection of ketamine/xylazine (100 mg/kg and 10 mg/kg, respectively) and transcardially perfused with 10 mL of 1× PBS followed by 10 mL of cold 4% paraformaldehyde (PFA) in PBS. Brains were removed, post-fixed in 4% PFA at 4 °C for at least 48 h and then transferred to 30% sucrose in PBS at 4 °C for cryoprotection. Once fully sunk, tissue was sectioned at 35 μm on a freezing sliding microtome (Leica SM2010R; Leica Biosystems, Deer Park, IL, USA) and stored in cryoprotectant solution (7.5% sucrose and 15% ethylene glycol in 0.1 M phosphate buffer) at −20 °C until processing. Sections were mounted onto glass slides with ProLong Gold (Invitrogen, Carlsbad, CA, USA) antifade reagent and imaged using a BZ-X700 inverted fluorescence microscope (Keyence, Itasca, IL, USA) equipped with a 20× dry objective (Nikon, Tokyo, Japan). Injection sites and optical fiber placements were verified from serial sections in all experimental animals (
Supplementary Figure S2).
For the optogenetic experiments using AAV9.rTH.PI.Cre.SV40, we also validated the targeting of TH+ cells in the VTA via an anti-TH antibody (mouse anti-TH; Millipore #MAB318, 1:100). Sections were then incubated with secondary antibodies [GFP: goat anti-chicken AlexaFluor 488 (Life Technologies, Carlsbad, CA, USA #A-11039), 1:1000 and TH: donkey anti-mouse AlexaFluor 594 (Life Technologies # A-21203), 1:1000] for 2 h at room temperature. After washing, sections were incubated for 5 min with DAPI (NucBlue, Invitrogen, Carlsbad, CA, USA) to counterstain nuclei, then mounted in Prolong Gold. Representative images of viral expression and quantitative analysis of TH co-localization, and an anatomical map of VTA viral expression and NAc core fiber placements are provided in the
Supplementary Figures S1–S4. Animals were prospectively evaluated using predefined exclusion criteria, including off-target viral expression or implant placement, poor optical signal quality, and technical issues during behavioral or neural recording. No animals met these exclusion criteria; therefore, no animals were excluded from the analyses.
2.3. Fiber Photometry
Fiber photometry recordings were obtained using a dual-LED system (490 nm and 405 nm; Thorlabs, Newton, NJ, USA) driven by an LED controller. The 490-nm LED, filtered through a 470-nm bandpass filter to match the excitation peak of dLight1.1, provided the dopamine-sensitive excitation, while the 405-nm LED served as an isosbestic control [
34]. Light was delivered through a 400-µm, 0.48 NA optical fiber (Doric Lenses) coupled to the chronically implanted fiber-optic cannula in each mouse (fiber tip positioned above the NAc core). LED output was controlled by a real-time processor (RZ5P; Tucker-Davis Technologies, Alachua, FL, USA), and fluorescence emission was separated by multiplexing the two LED channels at distinct carrier frequencies, enabling demodulation of the 490-nm and 405-nm signals. Synapse software (Version 3) (Tucker-Davis Technologies, Alachua, FL, USA) regulated LED timing and intensity and acquired fluorescence signals detected by a photoreceiver (Newport Femtowatt Photoreceiver; Doric Lenses). LED power (125 µW) was measured daily at the fiber tip using a power meter (Thorlabs, Newton, NJ, USA) to ensure consistency across sessions. Behavioral events (e.g., cue onset or footshock delivery) were time-stamped using TTL pulses generated by Med-PC V (Med Associates, St. Albans, VT, USA) and recorded in Synapse via the RZ5P. A built-in 10-Hz low-pass filter in Synapse was applied to reduce high-frequency noise in the raw photometry signal prior to offline analysis. All recordings were performed in freely moving mice, and photometry signals were synchronized with behavioral data via TTL alignment.
2.4. Fiber Photometry Analysis
Raw F470 and F405 signals were first smoothed using locally weighted scatterplot smoothing (LOWESS; MATLAB, Natick, MA, USA “smooth” function, span = 0.0004). The smoothed F405 signal was then fitted to the smoothed F470 signal using a first-order polynomial regression across the full analyzed recording period. The fitted F405 signal was used to correct for motion-related artifacts and slow photobleaching-related signal drift. ΔF/F was calculated as (F470 − fitted F405)/fitted F405.
For event-based analyses, fluorescence traces were aligned to behavioral events using TTL timestamps, and data were extracted from −2 s to +17 s surrounding each event based on our previous published studies [
11,
12,
15,
25]. Z-scores were computed using the pre-event baseline period from −2 to 0 s for each event-aligned trace. Specifically, the mean and standard deviation of the −2 to 0 s baseline were calculated for each trial, and each time point in the corresponding peri-event trace was normalized as z = (signal − baseline mean)/baseline standard deviation. This approach enabled comparison of event-evoked dopamine responses across trials and animals.
For full-session trace analyses, ΔF/F signals were normalized using the same session-wide baseline derived from the early stable recording period (1–10 s after recording onset). The mean and standard deviation of this baseline window were used to z-score the entire trace.
To quantify dopamine responses, the area under the curve (AUC) was calculated from z-scored traces using trapezoidal numerical integration. Analysis windows were selected a priori based on the expected temporal dynamics of dopamine responses and previous fiber photometry studies. For baseline recordings, AUC was calculated across the entire recording period. For shock-evoked responses, AUC was calculated within a fixed post-shock time window that was applied consistently across all animals and experimental groups.
2.5. Behavioral Methods
Mice were first trained in a fear conditioning paradigm (
Figure 1a), where they received six pairings of tone (85 dB, 2.5-kHz frequency, 10 s in duration) and footshock (1 mA intensity and 0.5 s in duration; the onset of the footshock coincided with the tone offset) in a standard conditioning chamber (Med Associates) with distinct contextual cues (e.g., lighting, floor texture, and auditory cues) to facilitate context discrimination. Twenty-four hours after fear conditioning training, mice received 2 sessions of fear extinction, during which the tone was presented for 12 trials and footshocks were omitted (inter-trial interval: variable, ~60–120 s). Following a 7-day hiatus, mice underwent one additional extinction session (6 trials instead of 12) to assess extinction recall and spontaneous recovery. Cue-induced response and dopamine were measured throughout fear conditioning, extinction, and extinction recall sessions, with freezing behavior defined as the absence of all movement except respiration and manually scored by a trained observer blinded to experimental conditions. To induce reinstatement, mice were tested 24 h after the last fear extinction session and received a single unpaired footshock presentation (1 mA, 0.5 s) 30 s into the extinction recall session prior to presentation of the fear cue in extinction (no tone present during shock delivery). We recorded NAc mice via fiber photometry, as explained above, during the extinction recall and reinstatement test sessions, with behavioral events time-locked to photometry recordings via TTL signals to ensure precise alignment between neural and behavioral data. Unless stated otherwise, all behavioral and fiber photometry data were obtained from the same cohort of mice. Sample size was based on prior published studies utilizing similar behavioral, fiber photometry, and optogenetic approaches in fear-learning paradigms [
11,
12,
15,
25]. Both male and female mice were included in all experiments and analyzed together. Since the study was not powered to test sex differences, sex-specific analyses were not performed.
2.6. Optogenetic Inhibition/Stimulation of the NAc Core Dopamine Release at the Time of the Reinstatement Footshock
Three groups of mice were trained in fear conditioning and fear extinction and tested in extinction recall (24 h after the extinction session instead of 7 days) and reinstatement as described above, using identical behavioral parameters and contexts as described for the primary experiments. The mice that were injected with the excitatory opsin (channelrhodopsin, ChR2) and eYFP-expressing control mice received a blue laser (470 nm, 1 s, 20 Hz, 8 mW) during the single presentation of the footshock during reinstatement, whereas the group that received the inhibitory opsin halorhodopsin (NpHR) received yellow laser stimulation (590 nm, 1 s, constant, 8 mW) delivered through implanted optical fibers targeting the NAc core. Laser onset was precisely time-locked to the onset of the footshock via TTL-triggered stimulation, ensuring temporal specificity of the manipulation. The freezing response to cue presentation following the reminder shock was measured, with freezing defined as the absence of all movement except respiration, and was manually scored by observers blinded to experimental conditions. Light power was measured at the fiber tip prior to each session to ensure consistent light delivery across animals.
2.7. Statistical Analysis
Data were analyzed using Prism (version 11; GraphPad) and are presented as mean ± S.E.M. Sample Size (
n = 5–7 animals/group) was selected based on prior published studies utilizing similar behavioral, fiber photometry, and optogenetic approaches in fear-learning paradigms [
11,
12,
15,
25]. Normality was assessed for all datasets using the Kolmogorov–Smirnov test. All groups passed normality testing (
p > 0.1000), except for the normalized freezing responses for the optogenetics experiments, for which the normality assumption was not met, and a Mann–Whitney U test (normality not assumed) was used. Statistical comparisons were performed using Repeated Measures or two-way ANOVA, as appropriate, when assumptions of normality were met, followed by Sidak’s or Dunnett’s post hoc multiple-comparison tests.
4. Discussion
Here, we examined how dopamine release in the NAc core shapes the reinstatement of an extinguished fear memory. During extinction recall, NAc core dopamine levels declined below baseline, whereas an unexpected reminder shock prevented this decline, maintaining dopamine levels near baseline and coinciding with renewed freezing. Optogenetic manipulations further supported a role for dopamine signaling in reinstatement: stimulation of dopamine axons at the time of the reminder shock amplified reinstatement, whereas inhibition markedly blunted it. These results identify dopamine release during unexpected aversive events as a critical regulator of fear relapse.
Recent findings from our group have demonstrated that novelty and unexpected sensory changes in the environment, such as an unexpected presentation of a neutral tone, produce robust increases in NAc core dopamine release [
15]. Importantly, this elevation in dopamine baseline persisted well beyond the time window for the presentation of the novel stimulus and altered how familiar cues were subsequently processed. The present results build directly on this framework. Our results suggest that the sustained dopamine reduction during extinction recall may reflect the encoding of reduced saliency of the previously learned aversive state, consistent with prior observations of dopamine decreases during low-threat conditions [
11,
35,
36]. Importantly, an unexpected aversive event prevented the reduction in dopamine observed during extinction recall, thereby enabling the re-expression of a previously extinguished fear response. Taken together, these studies suggest that NAc core dopamine serves as a general detector of unexpected environmental change, whether neutral, novel, or aversive, and that such signals can reorganize downstream processing in ways that either facilitate learning about new information or reinstate dormant threat associations. Although changes in behavioral state, including freezing-related immobility, may influence dopamine signaling during extinction recall, the prevention of dopamine decline following a reminder shock despite renewed freezing suggests that freezing-related immobility is unlikely to fully account for the observed dopamine responses. Future studies will incorporate detailed movement tracking, which will help further dissociate behavioral state from dopamine signaling during fear recovery.
A prediction-error framework also offers an interpretation of these findings. In classical models, unexpected outcomes, whether better or worse than predicted, evoke phasic changes in dopamine that update internal representations of the environment [
1,
6]. In our paradigm, the reminder footshock occurred after a period in which the animal had learned that the cue no longer predicted shock, making it an unexpected aversive event that may engage processes consistent with an aversive prediction error, heightened salience, or novelty. Such a violation of expectation may contribute to the transient increase in dopamine, consistent with the photometry signal observed at the time of the reminder shock. The dopamine response could then promote renewed engagement of the previously extinguished association and reinstatement of freezing. From this perspective, the dopamine surge at the moment of an unexpected shock may function as a teaching or updating signal that promotes re-expression of the threat memory, while optogenetic enhancement or suppression of this signal modulates the strength of reinstatement.
Importantly, dopamine responses to aversive events are heterogeneous across mesolimbic circuits, with distinct VTA dopamine subpopulations and NAc subregions showing differential responses to aversive stimuli and fear-related learning [
11,
12,
21,
22,
24]. Thus, our findings specifically implicate NAc core dopamine in reinstatement and may not be interpreted as a uniform dopaminergic response across the mesolimbic circuitry. Nevertheless, these findings suggest that dopamine also gates the re-expression of previously suppressed aversive memories. Traditional frameworks emphasize dopamine’s roles in prediction error, salience detection, and safety learning [
1,
6,
11,
12,
25,
37], but the present findings indicate that dopamine may also influence whether extinguished threat associations remain suppressed or become reactivated following unexpected environmental changes. This expands dopamine’s influence from shaping how new associations are acquired to regulating whether old associations are permitted to re-emerge. Alongside our results showing that novelty increases NAc dopamine and transforms stimulus processing, our findings suggest that dopaminergic responses to unexpected change, whether innocuous or aversive, serve as a shared pathway through which such events restore behavioral control and contribute to the persistence and recurrence of maladaptive fear.
Although caution should be exercised when extrapolating findings from rodent fear reinstatement models to human psychopathology, the present results may provide insight into neural mechanisms that contribute to the recovery of extinguished fear. Fear reinstatement is commonly used as a preclinical model for studying the return of fear following successful extinction and may capture certain features relevant to trauma- and anxiety-related disorders. In this context, our findings suggest that dopaminergic responses to unexpected aversive events may contribute to the re-emergence of previously extinguished fear memories. Specifically, unexpected aversive experiences may engage nucleus accumbens dopamine signaling in a manner that promotes the recovery of latent threat associations. Consistent with our previous work demonstrating that accumbal dopamine neurons are sensitive to salient, surprising, and novel events [
11,
12,
15,
25], the present findings further support a role for dopamine signaling in updating behavioral responses when environmental contingencies change. Future studies will be needed to determine whether similar mechanisms contribute to maladaptive fear recovery in clinical populations and how dopamine signaling interacts with other neural systems implicated in trauma- and anxiety-related disorders.
Despite these findings, a few limitations should be considered. Although sample sizes were consistent with prior studies, the modest group sizes may limit the generalizability of the findings. Moreover, while the reminder shock prevented the decline in dopamine observed during extinction recall and bidirectionally altered reinstatement behavior, supporting a contribution of dopamine signaling to fear recovery. However, because reinstatement was examined only after reminder shock, the extent to which these effects reflect shock-induced reinstatement rather than spontaneous recovery or contextual exposure remains unclear. Future studies incorporating a no-shock control group will be important for addressing this distinction. In addition, while our optogenetic findings are consistent with a role for dopamine signaling, a subset of VTA TH+ neurons co-release glutamate at accumbal terminals, raising the possibility that co-transmission contributed to the observed behavioral effects. Additionally, because the eYFP control group received blue light stimulation, whereas the NpHR group received yellow light, nonspecific effects related to wavelength or optical stimulation parameters cannot be fully excluded. Future studies incorporating wavelength-matched controls, greater cell-type specificity, and temporally controlled manipulations outside the reminder shock window will help determine whether dopamine activity during the shock specifically contributes to reinstatement or reflects a broader mechanism through which unexpected events reactivate previously extinguished fear responses.
Together, these findings show that NAc core dopamine is a key factor in restoring the expression of extinguished threat memories in response to unexpected events. Unpredicted aversive stimuli elevate dopamine, prevent recall-associated decline, and contribute to reinstatement, supporting a mechanism for dormant fear responses’ re-emergence. Alongside recent work showing that novelty also elevates dopamine and reshapes processing, this study demonstrates that dopamine is a signal for unexpected change that governs whether past aversive memories remain suppressed or regain control over behavior. This framework clarifies the mechanistic basis for understanding fear relapse and highlights dopaminergic circuits as a potential target for preventing the return of maladaptive threat.