Nociceptin/orphanin FQ receptor agonism attenuates behavioral and neural responses to conditioned aversive stimuli

Abstract

The nociceptin/orphanin FQ peptide (NOP) receptor has emerged as a promising anxiolytic target, as its activation has been shown to reduce anxiety-related behaviors in rodents. However, the mechanisms underlying these effects are not well understood. Here, we investigated the effects of the selective NOP receptor agonist SCH-221510 (0.01–0.1 mg/kg, IM) on behavioral and neural responses to aversive stimuli in squirrel monkeys (n = 3). Subjects underwent Pavlovian fear conditioning, wherein a visual conditioned stimulus (CS) was paired with the presentation of an aversive stimulus. Event-related functional magnetic resonance imaging (fMRI) was conducted in awake subjects to evaluate CS-evoked neural responses. Behavioral and neural responses to the CS were assessed across three experimental phases: pre-conditioning (Pre-C), post-conditioning (Post-C), and Post-C with SCH-221510 administration. In behavioral assessments, CS presentation during Post-C elicited a robust suppression of ongoing operant responding, which was absent during Pre-C and significantly attenuated by SCH-221510 treatment (0.1 mg/kg). fMRI results revealed that, relative to Pre-C, CS presentation during Post-C was associated with increased BOLD activity in brain regions previously implicated in fear processing (e.g., amygdala), expression and regulation (e.g., prefrontal cortex; PFC), as well as sensory integration (e.g., visual cortex). Critically, SCH-221510 (0.1 mg/kg) administration significantly attenuated CS-induced neural activation in these regions. Furthermore, resting-state functional connectivity analysis revealed that SCH-221510 administration decreased connectivity between PFC and amygdala, while enhancing connectivity among PFC subregions. Collectively, these findings suggest that NOP receptor agonism may attenuate conditioned responses to aversive stimuli by modulating functional interactions within a PFC-amygdala circuit.

Introduction

Anxiety disorders are among the most prevalent psychiatric conditions, affecting approximately 300 million individuals worldwide [1]. Dysregulated or exaggerated fear responses have been recognized as significant risk factors for the development and maintenance of these disorders [2]. The neurobiological mechanisms underlying heightened fear-related anxiety have generally been characterized by maladaptive functional dynamics within prefrontal-limbic circuitry, including prefrontal cortex (PFC), amygdala, hippocampus and hypothalamus, which have roles in fear processing and regulation [3, 4]. While the available pharmacological and psychotherapeutic interventions targeting these pathophysiological mechanisms provide symptomatic relief for some individuals, a substantial proportion of patients continue to experience treatment resistance or incomplete symptom remission [5]. These limitations underscore the critical need to identify alternative neuromodulatory systems involved in fear regulation, which may inform the development of more effective and mechanistically targeted interventions for anxiety disorders.

The nociceptin/orphanin FQ (N/OFQ) peptide (NOP) and its receptor system have emerged as promising therapeutic targets for anxiety disorders [6], given their high expression in prefrontal-limbic circuits [7]. Preclinical studies have generally shown that activation of the NOP receptor attenuates anxiety-related behaviors, including fear-potentiated startle, in rodent models [8–11]. Despite these findings, the neurobiological mechanisms underlying these effects, particularly whether NOP system activation modulates the prefrontal-limbic network, including both the activity of key brain regions and their functional connectivity, has not been fully characterized. Addressing these gaps in knowledge may provide new insights into the neurobiological underpinnings of fear-related anxiety disorders and facilitate the development of innovative therapeutic strategies targeting the NOP system.

Event-related functional magnetic resonance imaging (fMRI) has been widely employed in clinical research to investigate the neural correlates of anxiety, particularly in relation to maladaptive fear responses [12, 13]. This technique allows for high temporal resolution in capturing brain activity in response to discrete emotional or conditioned stimuli, making it especially suitable for probing the dynamic engagement of neural circuits involved in fear processing and regulation. Importantly, event-related fMRI has been used as a translational tool to evaluate the efficacy of various anxiolytic compounds by quantifying their modulatory effects on neural responses to aversive or threat-related stimuli [14]. Accordingly, this approach provides a rigorous framework for elucidating the neurobiological mechanisms by which candidate therapeutics, such as NOP receptor agonism, modulate fear-related circuit dynamics to exert anxiolytic effects.

In this study, we investigated the role of the NOP system in modulating fear-related behaviors and associated neural signatures using nonhuman primates (NHPs), given the substantial neuroanatomical homology between NHPs and humans [15], including the CNS distribution of NOP receptors [7, 16]. A Pavlovian fear conditioning procedure, integrated with a food-maintained operant responding task, was implemented to develop behavioral and neural responses to fear-related aversive stimuli [17]. This approach provides a translationally relevant measure of anxiety-related behavioral inhibition, capturing the extent to which conditioned aversive stimuli disrupt goal-directed behavior, which is frequently impaired in individuals with anxiety disorders [18]. Event-related fMRI was employed to investigate the neuronal mechanisms underlying NOP-mediated modulation of fear responses. By integrating behavioral and neuroimaging approaches, this study aims to delineate the circuit-level mechanisms by which NOP signaling regulates fear responses to aversive stimuli. Extrapolating from rodent findings, we hypothesized that pharmacological activation of NOP receptors would attenuate fear-related behavioral inhibition and modulate activity within key prefrontal-limbic circuits, including amygdala, hippocampus, and prefrontal cortex, as well as alter functional connectivity among these regions.

Methods and materials

Subjects

Three adult squirrel monkeys (Saimiri sciureus; one male, two females) were used (see Supplementary Table S1 for additional details). Previous studies from our lab and the literature confirmed that a sample size of three was sufficient to provide adequate statistical power to detect large effect sizes in both behavioral and neuroimaging measures [19–23]. A within-subject, repeated measures design was employed, with each monkey serving as its own control to ensure statistical reliability [24], precluding between-subject randomization and experimenter blinding.

Ethics approval

Animal care and research were conducted according to the guidelines provided by the Institute of Laboratory Animal Resources and the National Institutes of Health Office of Laboratory Animal Welfare. The study was approved by the Institutional Animal Care and Use Committee (IACUC) of McLean Hospital [2023N000171]. Detailed information regarding animal care is provided in the Supplementary Information.

Drugs

SCH-221510 (Tocris, Ellisville, MO), a potent and selective NOP receptor agonist (EC₅₀ ≈ 12 nM), was prepared in a vehicle solution consisting of 10% DMSO, 10% Tween 80, and 80% saline. SCH-221510 or vehicle was administered via intramuscular injection (≤0.3 mL/kg) into the thigh muscle 30 min prior to behavioral assessments or MRI scans. The selected dose range (0.01–0.1 mg/kg) and pretreatment interval were based on prior NHP studies demonstrating CNS activity of the compound [25–27].

Behavioral procedures

A multi-phase procedure was implemented, consisting of operant behavioral training to establish task performance followed by Pavlovian fear conditioning. This protocol integrated operant conditioning principles within a fear-related paradigm [17, 28], enabling measurement of conditioned suppression of operant behavior in response to presentation of visual stimuli associated with aversive stimuli. All experimental procedures were controlled and recorded using MedPC software (v4.2; Med Associates, St. Albans, VT).

Operant behavioral training

Subjects were positioned in a custom-designed MR-compatible chair [29] inside a sound-attenuated, ventilated chamber. A panel with three distinct LEDs (green:left, yellow:center, red:right) was placed 20-cm in front of the subject’s face. A customized lickometer, which also functioned as a milk delivery tube, was positioned within tongue-reach proximity. Each session (300-s) began with illumination of the center LED on the panel to indicate the availability of milk; subjects had 30-s to complete a fixed ratio (FR 1–5) of operant responses on the lickometer. Completion of the required FR response resulted in delivery of 0.3-mL of 30% condensed milk, followed by a 25-s timeout (center LED off). If the subject failed to meet the FR response requirement within the 30-s window, a 25-s timeout ensued without milk delivery. The conclusion of each 300-s session was marked by deactivation of the house-light. Following a 90-s long-timeout period, the house-light reactivated to signal the start of a new session, with a total of four sessions conducted per day. Operant behavioral training continued until subjects achieved a consistent rate of ≥ 90% in completing the FR requirement for three consecutive days.

Pavlovian fear conditioning

Following operant training, subjects underwent Pavlovian fear conditioning to associate a visual conditioned stimulus (CS) with an aversive unconditioned stimulus (US; electric stimulation). During the 300-s session, subjects were allowed to engage in operant responding (i.e., licking the lickometer) to obtain milk reinforcement, consistent with the training phase. The CS, operationalized as illumination of either the left or right LED (counterbalanced across subjects), was presented for 60-s at a randomized timepoint within the session to mitigate temporal predictability, except during test sessions, where it was presented within a predetermined 120–180-s time window. During CS⁺ trials, an electrical pulse (US; 0.5 mA, 0.2 s) was delivered to the subject’s tail at a random time within the 60-s CS⁺ period to establish an aversive association. Conversely, CS⁻ trials (illumination of the alternate LED) involved no stimulation, serving as a control condition. Each conditioning day consisted of four sessions: two baseline (no CS), one CS⁺, one CS⁻. Session order was counterbalanced across days and subjects to mitigate potential order effects.

Conditioned fear was indexed by suppression of operant responses during CS⁺ trials, defined as a reduction in response rate to less than 25% of baseline, despite the absence of US delivery. The 120-s pre-CS period served as the baseline to avoid potential confounding effects of the CS presentation. Notably, in CS⁺ sessions following fear conditioning, responses during the post-CS period were lower than during the pre-CS period (Figure S1A). Response rate was calculated using the following equation: (number of responses during the CS period / 60-s) / (number of responses during the pre-CS period / 120-s) × 100. After conditioning, subjects underwent test sessions to evaluate the effects of the NOP agonist SCH-221510 (0, 0.01, 0.03, and 0.1 mg/kg) pretreatment on conditioned behavioral responses to each CS presentation, in the absence of US delivery. Each subject received all dose conditions once in a within-subject experimental design, with SCH-221510 0 mg/kg (vehicle) serving as the control condition. The order of drug administration was counterbalanced across subjects to minimize potential order effects. Drug test sessions were conducted at intervals of at least 3 days, and CS⁺-related suppression of operant responding was confirmed prior to each test session.

Magnetic resonance imaging (MRI)

Event-related fMRI, incorporating a cue-reactivity paradigm [29] adapted from clinical protocols [12, 13], was utilized to investigate neural responses to CS in awake subjects. During each fMRI session, stimulation-associated (CS⁺) and control (CS⁻) light stimuli were presented in a pseudorandomized sequence to prevent anticipatory responses and ensure sustained task engagement. No electrical stimulations were delivered during fMRI, allowing for the isolated assessment of neural responses elicited by the CS. Each CS was presented for 6-s, with jittered interstimulus intervals of 15–30-s, totaling 23 presentations for each CS. These light presentation parameters were chosen based on previous clinical cue-reactivity studies [30] and the canonical hemodynamic response function, which peaks at 5–8 s and returns to baseline within 15–30 s [31, 32]. Subjects underwent MRI scans across three experimental phases: pre-conditioning (Pre-C), post-conditioning (Post-C), post-conditioning with NOP agonist administration (Post-C + SCH). The effective dose (0.1 mg/kg) of SCH-221510 used for fMRI was determined on the basis on its significant behavioral effects (see Results). In the Pre-C phase, baseline neural response to each CS was assessed to establish a reference for subsequent analyses. During the Post-C phase, conditioning-induced alterations in neural responses to each CS were evaluated. In the subsequent Post-C + SCH phase, the modulatory effects of NOP receptor activation on CS-related neural activities were evaluated. To minimize potential carryover effects of SCH-221510 on BOLD responses, Post-C scans were conducted prior to Post-C + SCH scans, ensuring that baseline post-conditioning neural activity could be measured without pharmacological interference. The interval between Post-C and Post-C + SCH MRI sessions was at least one week, during which daily conditioning sessions were conducted to verify the maintenance of CS⁺-induced behavioral suppression. Details regarding MRI data acquisition and preprocessing are provided in the Supplementary Information.

Analysis of neural responses to CS

Whole-brain analysis was conducted (FSL’s FMRI Expert Analysis Tool) to identify brain regions showing blood oxygen level-dependent (BOLD) signal changes in response to CS presentation across the three experimental phases: Pre-C, Post-C, Post-C + SCH. At the subject-level, event-related general linear models (GLMs) were constructed for each session, incorporating separate regressors for CS⁺ and CS⁻ trials. Each stimulus condition initially consisted of 23 trials; however, to enhance the reliability of cue-reactivity analyses, only trials exhibiting stimulus-evoked BOLD responses in the primary visual cortex during CS presentation were retained. This criterion ensured the inclusion of trials reflecting adequate visual engagement with the stimuli; each subject averaged 13.44 ± 1.303 (SEM) trials per condition in the final analysis. Explanatory variables (EVs) were convolved with a gamma function (phase = 0, SD = 3, mean lag = 6) to model the expected hemodynamic response. Six z-transformed rigid-body motion parameters (three translations, three rotations) and their temporal derivatives (12 regressors total) were included as nuisance regressors to account for residual motion effects. Contrast images were computed for four conditions: CS⁺ activation, CS⁺ deactivation, CS⁻ activation, CS⁻ deactivation. At the group-level, mixed-effects analysis (FLAME 1 + 2) was implemented to account for within- and between-subject variance. A GLM-based ANOVA compared contrast images across the three experimental phases, evaluating phase-specific differences in neural responses to CS. Statistical inference employed a voxel-wise threshold of Z > 3.1 with cluster-level FWE correction (p < 0.05), using Gaussian random field theory for multiple comparisons [33]. Brain regions involved in the identified significant clusters were delineated using a standardized squirrel monkey brain atlas [34] and independently verified by two experienced neuroscientists.

To examine temporal dynamics of neural responses to the CS, time-series BOLD signals were extracted from regions of interest (ROIs; 3D spheres, 0.5 search radius) that exhibited significant CS⁺-evoked activation differences in both the Post-C > Pre-C and Post-C > Post-C + SCH contrasts. BOLD signal intensities were standardized across sessions using mean-based scaling to mitigate inter-session variability. Each CS trial was segmented into 18-s epochs: a 6-s pre-CS baseline, a 6-s CS presentation, and a 6-s post-CS interval. To account for inter-trial variability, normalized BOLD response was calculated as: ((BOLD signal / mean BOLD signal during pre-CS baseline) × 100) - 100, yielding the percentage signal change relative to the baseline. Normalized responses were averaged within subjects and across the group, and plotted for each experimental phase (Pre-C, Post-C, Post-C + SCH). To quantify stimulus-evoked activation, area under the curve (AUC) within the 6–18-s window was subjected to statistical analysis.

Analysis of functional connectivity between ROIs

Resting-state functional connectivity (rsFC) analysis was performed to evaluate the effects of SCH-221510 administration on brain network dynamics, focusing on the 480-s baseline period prior to CS exposure to isolate drug effects from any confounding influence from CS exposure. ROI-based analysis focused on key prefrontal subregions and the amygdala whose functional connectivity has been shown to critically mediate fear-related anxiety [35, 36]. Based on Squirrel Monkey Brain Atlas [37], 10 ROIs were anatomically delineated and defined as 3D spherical masks (0.5 mm radius) centered on atlas coordinates: dmPFC, ventromedial prefrontal cortex (vmPFC), orbitofrontal cortex (OFC), left and right dorsolateral prefrontal cortex (dlPFC_L, dlPFC_R), left and right ventrolateral prefrontal cortex (vlPFC_L, vlPFC_R), left and right lateral orbitofrontal cortex (lOFC_L, lOFC_R), and amygdala (Amyg). For each ROI, BOLD time-series during the baseline period were extracted. Pairwise Pearson correlation coefficients were calculated across all ROI pairs using custom Python scripts, yielding a 10 × 10 rsFC matrix for each subject. To ensure normality and facilitate group-level comparisons, individual correlation coefficients were Fisher z-transformed prior to statistical analysis [38].

Correlation between neural and behavioral responses to CS⁺

To assess whether CS⁺-evoked neural activation in specific brain regions was associated with behavioral performance, brain-behavior correlation analyses were conducted. For each subject and session, mean β-values were extracted from regions showing significant activation differences in both the Post-C > Pre-C and Post-C > Post-C + SCH contrasts. Behavioral performance was quantified as the response rate during CS⁺ presentations across experimental phases. Pearson correlation analyses were performed to examine the relationship between region-specific neural activity and CS⁺-elicited behavioral responses.

Statistical analysis

Statistical analyses were performed using Prism 9.0 software (GraphPad Software, San Diego, CA). One-tailed paired t-tests were used for pairwise comparisons (e.g., Pre-C vs Post-C in Fig. 1A and D; Post-C vs Post-C + SCH in Fig. 5) to test directional hypotheses. Given the small sample size and the reduced statistical power associated with it, corrections for multiple comparisons were not applied to minimize the risk of Type-II errors. For analysis involving all three conditions, repeated-measures ANOVA was employed, followed by Dunnett’s post hoc test for Fig. 1B and E, or Tukey’s post hoc test for Fig. 3 to correct for multiple comparisons. Formal tests for normality were not conducted due to the small sample size; however, the within-subject repeated measures design, with each monkey serving as its own control, allowed for direct estimation of variability within each condition. Variance between conditions was considered to be approximately similar based on these within-subject comparisons, though caution is warranted in interpretation given the limited sample size. Statistical significance was defined as *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 1. Changes in CS-associated behavioral responses across experimental phases.

Subjects (n = 3) underwent a multi-phase conditioning paradigm consisting of operant training, in which the completion of an operant response was reinforced with milk delivery (0.3 mL/trial), followed by Pavlovian fear conditioning, where a CS⁺ was paired with an aversive US (0.5 mA electric stimulation), whereas the CS⁻ remained unpaired. During the test phase, behavioral responses to each CS were assessed by response rates, calculated as the ratio of the number of responses during the CS presentation period (120–180 s; without US) to the number of responses during the baseline pre-CS interval (0–120 s). A Response rate during CS⁺ presentation was significantly reduced in the Post-C phase compared to the Pre-C phase. B SCH-221510 administration (SCH; 0.1 mg/kg) significantly attenuated CS⁺-induced suppression of operant responding. C Representative response patterns during CS⁺ sessions across experimental phases. D Response rate during CS⁻ presentation did not significantly differ between the Pre-C and Post-C phases. E No significant changes in CS⁻-associated behavioral responses were observed following SCH-221510 treatment at any dose. F Representative response patterns during CS⁻ sessions across experimental phases. Data are presented as mean ± SEM. Statistical significance shown as **p < 0.01.

Fig. 5. NOP activation-induced alterations in rsFC within prefrontal-amygdala circuitry.

ROI-to-ROI connectivity analyses were conducted to assess changes in rsFC within the prefrontal-amygdala circuitry following SCH-221510 administration. Resting-state BOLD signals acquired during a 480-s baseline period were used to estimate rsFC. Pearson’s correlation coefficients between ROI pairs were Fisher z-transformed, and group-level (n = 3) differences in mean rsFC between the Post-C and Post-C + SCH phases were evaluated. A The rsFC matrix across all ROIs shows that SCH-221510 administration generally increased rsFC within PFC subregions, while reducing rsFC between the amygdala and PFC subregions. B Mean rsFC within PFC subregions was significantly enhanced in the Post-C + SCH phase compared to the Post-C phase. C Mean rsFC between PFC subregions and amygdala was significantly decreased in the Post-C + SCH phase relative to the Post-C phase. ROIs included: dorsomedial prefrontal cortex (dmPFC), ventromedial prefrontal cortex (vmPFC), medial orbitofrontal cortex (mOFC), left and right dorsolateral prefrontal cortex (dlPFC_L, dlPFC_R), left and right ventrolateral prefrontal cortex (vlPFC_L, vlPFC_R), left and right lateral orbitofrontal cortex (lOFC_L, lOFC_R), and amygdala (Amyg). Statistical significance was defined as *p < 0.05, ***p < 0.001.

Fig. 3. Changes in CS⁺-associated BOLD response within key ROIs across experimental phases.

Time-series BOLD data were extracted from the dorsomedial prefrontal cortex (dmPFC) and amygdala (Amyg) to visualize CS⁺-evoked neural responses across experimental phases. Each CS trial was segmented into an 18-s epoch comprising a pre-CS baseline period (0–6 s), a CS presentation period (6–12 s), and a post-CS interval (12–18 s). Group-averaged BOLD response time courses (n = 3) were plotted for the three experimental phases: Pre-C, Post-C, and Post-C + SCH. To quantify CS⁺-evoked neural activation, the AUC of the BOLD response (%) from 6 to 18 s was calculated for each phase. In the (A) dmPFC and (B) amygdala, CS⁺-evoked BOLD responses increased in the Post-C phase compared to the Pre-C phase. This conditioning-induced increase in BOLD responses was significantly attenuated following SCH-221510 administration. AUC analyses revealed statistically significant differences across phases, consistent with these observations. Data are presented as mean ± SEM. Statistical significance was defined as *p < 0.05, **p < 0.01, ***p < 0.001.

Results

Effects of NOP activation on conditioned behavioral responses to CS

During the Pre-C phase, operant responding was stably maintained across the session regardless of CS⁺ or CS⁻ presentation. However, in the Post-C phase, the response rate during CS⁺ presentations (Mean ± SEM, 17.03 ± 0.37) was significantly reduced compared to the Pre-C phase (104.70 ± 4.13), indicating CS⁺-induced suppression of operant responding in the absence of the US delivery (Fig. 1A; t = 19.55, p = 0.001, df = 2, Cohen’s d = 11.28). In contrast, responses during the Pre-CS and Post-CS periods did not significantly differ across experimental phases (Figure S1B), suggesting that the suppression of operant responding was both stimulus-specific and temporally constrained to the CS⁺ presentation period. Regarding the suppression of operant responding during CS⁺ presentation, administration of SCH-221510 produced a significant main effect of Dose (Fig. 1B; F (3,6) = 8.85, p = 0.01, η² = 0.82), with the 0.1 mg/kg dose significantly attenuating conditioned suppression (97.43 ± 22.75; p = 0.01). In contrast, CS⁻ presentation did not induce significant changes in operant responding, regardless of the experimental phase, with no significant difference observed in the response rate between the Pre-C (102.8 ± 4.99) and Post-C (92.64 ± 7.56) phases during CS⁻ presentation (Fig. 1D; t = 1.41, p = 0.15, df = 2, Cohen’s d = 0.81). Furthermore, SCH-221510 treatment did not significantly alter the behavioral responses associated with CS⁻ exposure (Fig. 1E; F (3,6) = 0.58, p = 0.65, η² = 0.23). Figure 1C and F illustrate the response patterns during CS⁺ and CS⁻ sessions, respectively, across the three experimental phases (Pre-C, Post-C, and Post-C + SCH) in a representative subject, with the response patterns for all subjects provided in Figure S2.

Effects of NOP activation on neural responses to CS after conditioning

Whole-brain voxel-wise analyses revealed significant differences in CS-evoked BOLD responses across three experimental phases: Pre-C, Post-C, and Post-C + SCH (Table 1).

Table 1.

Brain regions exhibiting significant differences in CS-evoked BOLD responses across the three experimental phases (Pre-C, Post-C, and Post-C + SCH).

CS+ Activation					Atlas Coordinates (mm)			Brain regions
	Cluster	# Voxels	- log10 P	Z-Max	X	Y	Z
Post > Pre	8	825	12.1	6.23	8.25	12.2	7.11	Dorsomedial prefrontal cortex, Ventromedial prefrontal cortex, Ventrolateral prefrontal cortex, Anterior cingulate cortex, Caudate, Putamen, Nucleus accumbens, Hypothalamus, Amygdala_L,
	7	708	10.8	5.99	−4.67	−12.6	−8.79	Cerebellum_posterior lobe, Visual cortex
	6	202	4.3	6.26	−10.6	−5.66	−5.05	Occipito temporal area_L
	5	200	4.27	5.27	0.3	−9.64	1.5	Cerebellum_anterior lobe
	4	93	2.29	5.53	6.26	−2.68	−1.31	Hippocampus_R
	3	73	1.84	4.64	−4.67	−8.64	8.98	Posterior parietal cortex_L
	2	63	1.61	4.51	12.2	−8.64	6.18	Middle temporal cortex_R
	1	55	1.41	5.35	11.2	−1.69	−1.31	Superior temporal association cortex_R
Post < Pre	N/A
Post > +SCH	6	1019	14	6.29	9.24	14.2	3.37	Dorsomedial prefrontal cortex, Ventromedial prefrontal cortex, Ventrolateral prefrontal cortex, Anterior cingulate cortex, Caudate, Putamen, Nucleus accumbens, Hypothalamus, Middle cingulate cortex
	5	931	13.1	5.7	6.26	−11.6	−5.05	Cerebellum_anterior lobe, Cerebellum_posterior lobe, Posterior parietal cortex_L, Visual cortex
	4	266	5.31	5.13	−13.6	−3.68	−0.371	Occipito temporal area_L
	3	74	1.87	5.37	12.2	−1.69	−2.24	Hippocampus_R, Superior temporal association cortex_R
	2	73	1.84	4.83	12.2	−7.65	5.24	Middle temporal cortex_R
	1	57	1.46	5.19	−8.64	7.26	−3.18	Amygdala_L
Post < +SCH	N/A
Pre > +SCH	2	91	2.24	5.07	−1.69	4.27	8.98	Middle cingulate cortex
	1	70	1.78	4.53	−0.694	−22.6	0.565	Visual cortex
Pre < +SCH	N/A

CS- Activation					Atlas Coordinates (mm)			Brain regions
	Cluster	# Voxels	- log10 P	Z-Max	X	Y	Z
Post > Pre	2	113	2.79	5.22	15.2	9.24	4.31	Anterior parietal cortex_R, Premotor cortex_R
	1	97	2.45	6.23	11.2	−2.68	−2.24	Superior temporal association cortex_R
Post < Pre	3	314	6.22	5.21	−8.64	4.27	4.31	Caudate_L, Insula_L, Primary motor cortex_L
	2	125	3.03	4.8	−10.6	19.2	5.24	Ventrolateral prefrontal cortex_L
	1	64	1.69	4.66	3.28	14.2	−0.371	Putamen_R, Nucleus accumbens_R
Post > +SCH	1	71	1.86	4.54	11.2	−1.69	−2.24	Superior temporal association cortex_R
Post < +SCH	2	406	7.22	5.75	−8.64	4.27	4.31	Caudate_L, Insula_L, Primary motor cortex_L
	1	55	1.46	5.08	−9.64	21.2	5.24	Ventrolateral prefrontal cortex_L
Pre > +SCH	N/A
Pre < +SCH	1	72	1.88	4.23	2.29	2.29	−0.371	Thalamus_R

Whole-brain voxel-wise analyses were performed using a GLM-based repeated-measures ANOVA to identify phase-dependent alterations in neural responses to each CS (CS⁺ or CS⁻). Statistical significance was determined using a voxel-level Z-threshold of 3.1 and cluster-level correction for multiple comparisons based on FWE correction (p < 0.05). The brain regions encompassed within significant clusters were anatomically delineated using a standardized squirrel monkey brain atlas.

R right, L left.

Figure 2A shows that CS⁺-evoked activation was significantly increased during the Post-C phase compared to the Pre-C phase (Post-C > Pre-C), particularly in prefrontal-limbic regions, including the dmPFC, vmPFC, bilateral vlPFC, anterior cingulate cortex (ACC), putamen, caudate, hypothalamus, amygdala, and hippocampus (see complete list of regions in Table 1). Following SCH-221510 administration, a significant decrease in CS⁺-evoked activation was observed during the Post-C + SCH phase relative to the Post-C phase (Post-C > Post-C + SCH; Fig. 2B), primarily in regions overlapping with the Post-C > Pre-C contrast, as shown in Fig. 2D, with additional reduction in the midcingulate cortex (MCC). Figure 2C demonstrates that SCH-221510 administration attenuated CS⁺-evoked activation in the MCC, as evidenced by significantly greater activation during the Pre-C phase compared to the Post-C + SCH phase (Pre-C > Post-C + SCH).

Fig. 2. Whole-brain mapping of CS⁺-evoked BOLD response across experimental phases.

Whole-brain voxel-wise analyses were conducted to assess differences in CS⁺-evoked BOLD responses across three experimental phases: Pre-C, Post-C, and Post-C + SCH. CS⁺-evoked activation is shown (A) during the Post-C phase compared to the Pre-C phase (Post-C > Pre-C). B following SCH-221510 administration in the Post-C + SCH phase relative to the Post-C phase (Post-C > Post-C + SCH) and, (C) between the Pre-C and Post-C + SCH phases (Pre-C > Post-C + SCH). D The sagittal image shows the locations of the displayed coronal slices along the anterior-posterior axis at the following coordinates: 18, 15.5, 14, 10.5, 7, 3, −3, −8, and -17. Overlapping spatial maps from contrasts (A–C) are visualized with semi-transparent overlays (opacity = 80%), revealing substantial spatial convergence between the Post-C > Pre-C and Post-C > Post-C + SCH clusters. Anatomical localizations of significant clusters were determined using a standardized squirrel monkey brain atlas [34]. Linear interpolation was applied using FSLeyes. Brain regions involved in the identified significant clusters included: dorsomedial prefrontal cortex (dmPFC), ventromedial prefrontal cortex (vmPFC), ventrolateral prefrontal cortex (vlPFC), anterior cingulate cortex (ACC), putamen (Put), caudate (Cau), hypothalamus (Hyp), amygdala (Amyg), middle cingulate cortex (MCC), hippocampus (Hip), superior temporal association cortex (STS), middle temporal cortex (MT), posterior parietal cortex (PPC), occipito-temporal cortex (OT), visual cortex (VC), and cerebellum (Cb).

Time-series analyses were performed on ROIs identified in Table 1, which exhibited significant increase in CS⁺-evoked activation following fear conditioning. Across all ROIs, BOLD response to CS⁺ increased during the Post-C phase relative to Pre-C and was subsequently attenuated following SCH-221510 administration. Representative time-series data from two ROIs, including the dmPFC and the amygdala, are shown in Fig. 3. In the dmPFC, a significant main effect of experimental phase was observed (Fig. 3A; F (2,4) = 37.10, p < 0.001, η² = 0.95), characterized by a robust increase in the area under the curve (AUC) from Pre-C to Post-C (p < 0.001), followed by a significant reduction from Post-C to Post-C + SCH (p < 0.001). Likewise, the amygdala exhibited a significant main effect of phase (Fig. 3B; F (2,4) = 31.75, p = 0.004, η² = 0.94), with post hoc comparisons indicating significant differences across all phase contrasts: increased AUC from Pre-C to Post-C (p = 0.04), and decreased AUC from Post-C to Post-C + SCH (p = 0.003) as well as from Pre-C to Post-C + SCH (p = 0.03).

Relation between CS⁺-evoked BOLD activation and conditioned behavioral suppression

A significant negative correlation was observed between mean β-values from the defined ROIs and response rates during CS⁺ presentation (Fig. 4; r = −0.88, p = 0.001). Specifically, the Post-C phase exhibited significantly elevated β-values concomitant with reduced response rates relative to the Pre-C and Post-C + SCH phases, indicating that increased neural activation within these ROIs was associated with conditioned behavioral suppression.

Fig. 4. Negative correlation between CS⁺-evoked BOLD activation and behavioral response rates.

For each subject (n = 3) and session, neural and behavioral responses to CS⁺ were quantified and included in a correlation analysis. Mean β-values were extracted from ROIs that showed significant activation differences in both the Post-C > Pre-C and Post-C > Post-C + SCH contrasts. Operant response rates during CS⁺ presentations were used as the behavioral measure. Pearson correlation analysis revealed that higher CS⁺-evoked neural activation in the ROIs was associated with lower behavioral response rates. Correlation coefficients (r) and statistical significance (p) are reported.

Effects of NOP agonism on rsFC of prefrontal-amygdala circuitry

Administration of SCH-221510 induced significant alterations in prefrontal-amygdala network dynamics (Fig. 5A). Notably, increased rsFC was observed among several PFC subregions, including vmPFC-vlPFC_R (t = 5.89, p = 0.01, df = 2, Cohen’s d = 3.40), mOFC-dlPFC_L (t = 6.09, p = 0.01, df = 2, Cohen’s d = 3.52), mOFC-vlPFC_L (t = 3.25, p = 0.04, df = 2, Cohen’s d = 1.88), and lOFC_L-vlPFC_L (t = 4.09, p = 0.03, df = 2, Cohen’s d = 2.36). In contrast, a significant reduction in rsFC was detected between vlPFC_R and amygdala (t = 4.12, p = 0.03, df = 2, Cohen’s d = 2.38). Group-level analyses further corroborated these findings, demonstrating a robust enhancement of intra-PFC connectivity (Fig. 5B; t = 9.93, df = 35, p < 0.001, Cohen’s d = 1.66), along with a significant decrease in connectivity between PFC and amygdala (Fig. 5C; t = 5.61, df = 8, p < 0.001, Cohen’s d = 1.87).

Discussion

The present findings suggest that activation of the NOP system via systemic administration of SCH-221510 attenuates behavioral responses to aversive stimuli and modulates neural activity and functional connectivity within fear-associated circuits in NHPs. From a methodological perspective, this study builds upon existing translational approaches by employing a Pavlovian conditioning paradigm in combination with event-related fMRI in awake animals [39], a technique that remains relatively underutilized in preclinical research. Combining these methods enhances cross-species translational validity and enables a more comprehensive characterization of the neural substrates underlying maladaptive fear processing, thereby contributing to the development of mechanism-based therapeutic strategies for anxiety disorders.

Following fear conditioning, subjects exhibited robust suppression of operant responding during CS⁺ presentation despite the absence of US delivery, indicative of behavioral inhibition in response to the conditioned aversive stimulus. Notably, activation of the NOP receptor significantly attenuated this behavioral suppression, leading to a recovery of operant responding during CS⁺ presentation, consistent with prior preclinical findings that the NOP receptor agonist exerts anxiolytic-like effects in rodents [8–10]. Although one potential interpretation of the restoration is that NOP receptor activation enhances general operant behavior, this explanation seems unlikely, as responding during the pre-CS period was not elevated following NOP agonist administration (Figure S1C). These findings suggest that the primary effect of NOP receptor activation is not the enhancement of general motivated behavior, but rather the selective attenuation of conditioned fear-related behavioral responses to aversive stimuli. The specificity of this anxiolytic effect was further supported by the absence of significant behavioral effects of NOP activation during CS⁻ sessions indicating that NOP receptor modulation selectively targets fear-associated behavioral inhibition without producing nonspecific sedative or salience-reducing effects.

Whole-brain voxel-wise analyses revealed robust BOLD activation in response to CS⁺ following fear conditioning across multiple brain networks, including regions implicated in fear processing [40–44] (amygdala, hypothalamus, caudate, putamen, hippocampus), fear expression and regulation [45–47] (dmPFC, vmPFC, vlPFC, ACC), and sensory integration [48–50] (visual cortex, parietal cortex, temporal cortex, cerebellum). These findings align with previous studies demonstrating the engagement of fronto-limbic and sensorimotor networks in visually mediated fear signaling pathways [51]. Notably, SCH-221510 administration significantly attenuated CS⁺-evoked activation across most of these regions, indicating that NOP receptor activation suppresses conditioned neural responses to aversive stimuli. In addition, response to the CS⁺ in MCC, a region implicated in threat detection and appraisal [52], was significantly reduced by SCH-221510 treatment relative to both the Pre-C and Post-C conditions (Fig. 2B and C). A similar dampening effect on response to CS⁻ presentation were also observed. Specifically, during the Post-C phase, CS⁻ presentation evoked reduced neural activation in several fear-related regions (caudate, putamen, vlPFC), while increased activity was observed in sensory processing areas (temporal cortex) (Table 1). Taken together, NOP agonism blunted responses to both CS⁺ and CS⁻, suggesting that NOP signaling may diminish neural discrimination between aversive and non-aversive cues by modulating circuits involved in fear processing and sensory salience. This widespread, yet regionally selective modulatory effect likely reflects the neuroanatomical distribution of NOP receptors, which are densely expressed across key regions of the fronto-limbic circuitry, including PFC, cingulate cortex, amygdala, hypothalamus, dorsal striatum, and hippocampus [7]. Further, considering that visual responses to fearful stimuli are modulated by limbic regions including amygdala [53], the observed reduction in activity across sensory integration areas may suggest that NOP receptor agonism attenuates sensory processing of aversive stimuli, potentially as a downstream consequence of reduced limbic activation.

Time-series analysis was performed to examine how neural responses evolved over the course of CS presentations and to assess whether NOP activation modulated the transient or sustained effects of CS exposure. This analysis provided superior temporal resolution compared to conventional block-averaged methods, enabling the detection of dynamic fluctuations in BOLD signals in response to CS presentation. Furthermore, this approach enabled characterization of phasic changes in BOLD responses to CS⁺ across experimental phases, including the conditioning-induced increase and its subsequent attenuation following SCH-221510 administration, across all ROIs identified in the whole-brain analysis. Notably, representative regions such as the dmPFC and amygdala exhibited robust increases in CS⁺-evoked activation following conditioning, consistent with previous studies showing the activation in these regions during fear processing [54, 55]. The attenuation of this activation through NOP receptor agonism indicates that the NOP system serves as a key neuromodulatory pathway in regulating activity within fear-associated neural circuits.

Correlation analyses further supported the finding that increased activation within fear-associated ROIs was associated with decreased operant responding during CS⁺ presentations, particularly in the Post-C phase. Activation of the NOP receptor attenuated this heightened neural reactivity within fronto-limbic circuits in response to conditioned aversive stimuli, which was associated with a corresponding reduction in behavioral suppression, thereby providing mechanistic insight into the anxiolytic-like effects of NOP receptor agonism.

While NOP receptor activation attenuated the response to the CS⁺, we also investigated whether it modulated baseline brain activity independently of stimulus presentation, specifically by altering rsFC within the PFC-amygdala circuitry prior to CS onset. This circuitry was targeted because reduced PFC-amygdala connectivity is widely implicated in anxiety-related pathophysiology in clinical populations [56–58]. Based on clinical studies suggesting that anxiolytics such as selective serotonin reuptake inhibitors often enhance PFC-amygdala rsFC [59–61], we initially hypothesized that NOP receptor activation would similarly increase connectivity within this pathway. Contrary to this expectation, however, NOP receptor activation significantly decreased functional connectivity between PFC subregions and amygdala. This discrepancy may be attributable to the use of healthy subjects in the present study, who underwent short-term conditioning and exhibited anxiety-related responses only in specific contexts, potentially reflecting neural processing mechanisms that differ from those observed in clinical populations [61]. Furthermore, NOP receptor activation represents a functionally distinct mechanism from classical anxiolytic interventions. Given that the PFC-amygdala pathway plays a central role in transmitting aversive stimulus information necessary for the initiation and expression of conditioned fear responses [62], this reduction may suggest that NOP receptor activation disrupts the transmission of fear-related signals. This finding aligns with prior studies demonstrating that reduced PFC-amygdala coupling is associated with attenuated fear expression, as evidenced by reduced behavioral reactivity to aversive stimuli [35]. In contrast, NOP receptor activation enhanced intra-PFC connectivity, indicating increased functional integration among PFC subregions involved in the regulation of affective responses [63]. This finding is in line with existing literature reporting that decreased intra-PFC connectivity is often observed in heightened fear-related anxiety conditions [64]. Taken together, these findings suggest that NOP receptor activation may suppress the propagation of fear-related signals originating from the amygdala while concurrently strengthening intra-PFC coordination that facilitates prefrontal regulatory control. This bidirectional modulation of rsFC within the prefrontal-amygdala circuit represents a novel circuit-level mechanism underlying the observed attenuation of both behavioral and neural responses to conditioned aversive stimuli.

Importantly, this is the first evidence that NOP receptor activation reduces prefrontal-amygdala connectivity, which implies that NOP receptor blockade may conversely enhance connectivity within this circuit. This provides a potential mechanistic explanation for the paradoxical effects of the NOP system [11, 65]. Specifically, receptor activation may produce anxiolytic effects by suppressing the transmission of aversive signals between the prefrontal cortex and amygdala, whereas receptor blockade may elicit antidepressant-like effects [66], potentially by restoring prefrontal-amygdala circuit function involved in stress regulation and mood control [67], which is commonly reduced in patients with depression [68, 69]. Systematic follow-up studies using advanced neuroimaging and pharmacological approaches are necessary to clarify the role of the NOP system in modulating anxiety- and depression-related neural circuits.

Several considerations, particularly the relatively small sample size that is a common constraint in NHP research, warrant careful interpretation of these findings. While correction for multiple comparisons is important to control for Type I error, the limited sample size substantially reduces statistical power, increasing the likelihood of Type II errors following correction [70]. Consequently, uncorrected results are presented with appropriate caution and transparency. Despite this limitation, the convergence of results across multiple methodological approaches, including behavioral paradigms, event-related BOLD responses, and rsFC analyses, supports the internal consistency and construct validity of the observed effects. Future studies that include larger sample sizes will be essential for enhancing statistical sensitivity and confirming the replicability of these findings. Moreover, in the rsFC analyses, the amygdala was defined as a single ROI, as our fMRI resolution was insufficient to reliably distinguish activity in distinct amygdala subregions. We recognize that individual amygdala subregions may have distinct roles in fear-related anxiety conditions [71, 72], and future studies employing higher-resolution imaging will be necessary to enable subregion-specific analyses.

In conclusion, the present study provides compelling evidence that pharmacological activation of the NOP receptor attenuates behavioral and neural responses to conditioned aversive stimuli in NHPs. Specifically, NOP receptor agonism appeared to modulate functional interactions within the PFC-amygdala circuitry and suppress aversive stimulus-evoked neural activation across the fronto-limbic network, thereby resulting in a selective attenuation of fear-related behavioral responses. Taken together, these findings provide the first evidence of anxiolytic-like effects of NOP receptor activation in NHPs, extending previous rodent work. While these results highlight the translational potential of targeting the NOP system for fear-related anxiety disorders, clinical studies in humans are still lacking and will be necessary to determine therapeutic applicability.

Author contributions

Kwang-Hyun Hur: Conceptualization, Investigation, Methodology, Formal analysis, Writing - Original Draft, Visualization, Funding acquisition. Diego A. Pizzagalli: Writing - Review & Editing, Supervision, Funding acquisition. Jessi Stover: Investigation. Kenroy Cayetano: Software, Resources. Stephen J. Kohut: Conceptualization, Methodology, Resources, Writing - Review & Editing, Supervision, Funding acquisition.

Funding statement

Kwang-Hyun Hur and Diego A. Pizzagalli disclose support for the research of this work from the National Institute of Mental Health [P50MH119467]. Stephen J. Kohut discloses support for publication of this work from National Institute on Drug Abuse [K01DA039306]. All other authors declare no relevant funding.

Data availability

All data needed to evaluate the conclusions in the paper are present in the paper. Raw data from this study are available from the corresponding author upon reasonable request.

Competing interests

Over the past three years, Diego A. Pizzagalli has received consulting fees from Arronhead Pharmaceuticals, Boehringer Ingelheim, Circular Genomics, Compass Pathways, Engrail Therapeutics, Neumora Therapeutics, Neurocrine Biosciences, Neuroscience Software, TAP Sciences, and Xenon Pharmaceuticals; he has received honoraria from the American Psychological Association, Psychonomic Society and Springer (for editorial work) and from Alkermes; he has received research funding from the BIRD Foundation, Brain and Behavior Research Foundation, Circular Genomics, Dana Foundation, DARPA, Millennium Pharmaceuticals, NIMH and Wellcome Leap MCPsych; he has received stock options from Ceretype Neuromedicine, Compass Pathways, Engrail Therapeutics, Neumora Therapeutics, and Neuroscience Software. No funding from these entities was used to support the current work, and all views expressed are solely those of the authors. The other authors declare no competing financial interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41398-026-04111-5.