Puberty, sex, and fear extinction retention: A neuroimaging study in youth

Abstract

Anxiety disorders affect 31.1 % of U.S. adults, with females exhibiting twice the prevalence of males. While sex differences are well-documented, underlying mechanisms remain unclear. Advanced pubertal status is linked to increased anxiety symptoms in females but not males, suggesting puberty contributes to sex differences in fear-based disorders. Deficits in fear extinction and retention are implicated in anxiety, and prior research suggests sex hormones influence extinction retention. This study examined sex assigned at birth (parent-reported) and pubertal status (self-reported) on extinction retention in 101 youth (47.5 % female) using a Pavlovian fear extinction paradigm. Measures included self-reported anxiety symptoms, extinction retention, and neural activation in the amygdala, hippocampus, and anterior cingulate cortex (ACC).

Graphical Abstract

Highlights

• •Examined puberty and sex effects on fear extinction retention.
• •Mid/Late pubertal females had higher anxiety and poorer retention than males.
• •Better extinction retention linked to lower anxiety in Mid/Late puberty.
• •Mid/Late pubertal females had lower right amygdala activation during retention.
• •ACC activation was higher in Pre/Early puberty than in Mid/Late puberty.

1. Introduction

Anxiety and other fear-based disorders are a significant public health concern, affecting approximately 31.1 % of U.S. adults in their lifetime (National Institute of Mental Health, 2017). Notably, females are twice as likely as males to develop a fear-based disorder (Kessler et al., 2005, Kilpatrick et al., 2013). While this sex difference is well established, the underlying mechanisms are poorly understood. A hallmark feature of fear-based disorders is impaired fear regulation, with fear extinction emerging as a critical paradigm for studying both the mechanisms underlying these disorders and their treatments (Graham and Milad, 2011a).

In a standard Pavlovian fear extinction paradigm, a neutral conditioned stimulus (CS+) is paired with an aversive unconditioned stimulus (US) during fear conditioning. After repeated pairings, the CS+ elicits a conditioned fear response, commonly measured by increased skin conductance responses (SCRs), a measure of automonic arousal, or other physiological outcomes. Fear extinction occurs when the CS+ is repeatedly presented without the US, leading to a gradual reduction in the conditioned fear response. Studies indicate that individuals with anxiety and other fear-based disorders, including posttraumatic stress disorder (PTSD), exhibit heightened SCRs during fear extinction and its later retention, compared to those without anxiety (Treanor et al., 2021). Furthermore, impairments in extinction retention are thought to contribute to persistent fear and anxiety symptoms and may also inform more effective treatments for fear-based disorders, as the concept of extinction forms the basis of exposure-based therapy (Graham and Milad, 2011a, Milad et al., 2009).

Neuroimaging studies have identified key brain regions involved in fear extinction that are conserved across species, including the amygdala, hippocampus, and frontal regions such as the anterior cingulate cortex (ACC) and adjacent medial prefrontal cortex (mPFC; Raber et al., 2019). The amygdala plays a crucial role in fear expression (Davis and Whalen, 2001, LeDoux, 2000), while the frontal brain regions facilitate extinction memory retention through inhibitory connections to the amygdala (Milad and Quirk, 2002, Ochsner and Gross, 2005, Phelps, 2004, Quirk et al., 2003, Quirk and Beer, 2006, Quirk and Mueller, 2008). The hippocampus contributes to context-dependent fear modulation and positively correlates with frontal brain activation during extinction retention (Kalisch et al., 2006, Milad et al., 2007). Individuals with fear-based disorders often show hyperactivity in the amygdala and hypoactivity in frontal regions during emotion processing tasks (Etkin and Wager, 2007).

To explore this in child and adolescents populations, we have previously validated a Pavlovian fear extinction paradigm in youth (Marusak et al., 2017). Our findings indicate that children ages 6–11 years—primarily prepubertal (77 %)—exhibit poor extinction retention (retention index: 13.56 %; Marusak et al., 2018) compared to typical adult performance (74–85 %; Holt et al., 2009; Milad et al., 2007; Milad, Pitman, et al., 2009; Rabinak et al., 2014). Poor retention in late childhood is likely due to underdeveloped frontal regions, such as the ACC and mPFC (Marusak et al., 2018), and their connections with the hippocampus, which are critical for successful extinction retention and continue to mature throughout adolescence (Kalisch et al., 2006).

Given the importance of fear extinction in understanding fear-based disorders, previous studies have explored sex differences in extinction processes. Findings from adult animal models are mixed, with some studies reporting impaired extinction in females compared to males while others report no sex differences (Baker-Andresen et al., 2013, Baran et al., 2009, Baran et al., 2010, Fenton et al., 2016, Gruene et al., 2015, Voulo and Parsons, 2017). Similar inconsistencies exist in human studies (Velasco et al., 2019) suggesting that additional variables, such as sex hormones, may contribute to differences in fear extinction and its neural underpinnings.

Emerging evidence highlights the role of sex hormones—estadiol, a form of estrogen, and testosterone—as modulators of extinction learning and retention (Cover et al., 2014, Maeng and Milad, 2015). While both hormones are present in all individuals and there exists substantial variability, estradiol is the predominate sex hormone in females and testosterone is the predominate sex hormone in males (Collaer and Hines, 1995). Research in both humans and rodent models indicates that individuals in the high-estradiol phase of the menstrual cycle exhibit better extinction retention than those in the low-estradiol phase (Graham and Milad, 2011a, Graham and Milad, 2013, Maeng and Milad, 2015, Milad et al., 2009, Milad et al., 2010). Similarly, those using hormonal contraceptives, which lower estradiol levels, demonstrate poorer extinction retention compared to naturally cycling individuals (Graham and Milad, 2013). These findings suggest that estradiol plays a key role in modulating fear extinction retention (Zeidan et al., 2011). A similar relationship has been explored with testosterone. Evidence suggests that higher testosterone levels in adult male rats are associated with enhanced extinction retention (Maeng et al., 2017). Conversely, the removal of testosterone in adolescent rats results in impaired extinction retention compared to controls (Perry et al., 2020). These findings suggest that testosterone may have a facilitating effect on fear extinction retention.

The influence of sex hormones on fear extinction becomes particularly relevant during puberty, when estradiol and testosterone levels rise in pediatric populations. Puberty typically begins between ages 8 and 14 (Jenner et al., 1972) and is associated with an increased incidence of fear-based disorders in individuals assigned female at birth, but not in those assigned male at birth (Khanal et al., 2022). It remains unclear whether the rise in sex hormones contributes to the higher prevalence of fear-based disorders among females. Few studies have examined extinction retention during the transition into adolescence while considering both pubertal development and sex differences.

To address these gaps, the present study evaluated the effects of sex assigned at birth, puberty, and the sex-by-puberty interaction on anxiety symptoms, extinction retention, and neural activity in a cross-sectional sample of children and adolescents (ages 6–17 years) spanning Pre/Early puberty to Mid/Late puberty. Fear extinction retention was assessed using SCRs, given their association with anxiety symptoms across development (Abend et al., 2020) and their frequent use as an autonomic marker of fear expression (Lonsdorf et al., 2017). Functional magnetic resonance imaging (fMRI) was used to examine activity in extinction-related brain regions, including the amygdala, hippocampus, and ACC during extinction retention. We hypothesized an interaction between sex and puberty in anxiety symptoms and extinction retention, such that prepubertal males and females would exhibit similar anxiety levels and extinction retention. However, with puberty onset, we expected that individuals assigned male at birth would show greater extinction retention, whereas those assigned female at birth in Mid/Late puberty would exhibit poorer retention and higher anxiety symptoms. Additionally, we predicted that Mid/Late puberty females would show reduced hippocampal and ACC activity—regions critical for extinction retention—and heightened amygdala activity, relative to their male counterparts, potentially underlying observed deficits in extinction retention.

2. Methods

2.1. Participants

This study reports on a sample of 101 children and adolescents (48 female) aged 6–17 years (see Table 1). Given prior evidence of age-related differences in fear extinction learning, we initially collected two cohorts: one comprising younger children (ages 6–12; Marusak et al., 2017, Marusak et al., 2018) and the other consisting of youth ages 10–17 years. Due to the age overlap between these groups and the use of identical study procedures—including the same scanner, site, and task parameters—they have been combined for analysis. The only methodological difference between the cohorts was the number of extinction trials, with older participants (adolescents) completing 12 trials per stimulus compared to 8 trials for younger participants (children), as detailed below. Participants were recruited from the Metro Detroit, Michigan, USA, area through various channels, including local schools, recreation centers, pediatrician and behavioral health clinics, flyers, and online advertisements. All participants were right-handed, screened for exclusionary psychiatric and neurological disorders (e.g., epilepsy, schizophrenia, obsessive-compulsive disorder, autism spectrum, bipolar disorder), current oral contraceptive or non-stimulant psychiatric medication use, history of traumatic brain injury, and contraindications for MRI. Written informed consent was obtained from parents/guardians, and assent was obtained from minors in accordance with the Institutional Review Board’s guidelines. The Wayne State University Institutional Review Board (IRB) approved all study procedures and methods were carried out with IRB guidelines and regulations.

Table 1.

Sample characteristics, overall, by sex, and by pubertal group.

		By Sex			By Pubertal Group
Variable Name	Overall (N = 101)	Females (n = 48)	Males (n = 53)	P-value	Pre/Early Pubertal (n = 38)	Mid/Late Pubertal (n = 63)	P-value
Age, in years, M (SD)	11.73 (2.96)	11.83 (3.12)	11.65 (2.82)	0.746	9.55 (1.655)	13.05 (2.79)	< 0.001
Anxiety symptoms (SCARED), M (SD)	23.36 (13.87)	30.15 (14.65)	17.21 (9.73)	< 0.001	20.39 (11.98)	25.14(14.70)	0.096
Pubertal (Tanner) stage, M (SD)	3.13 (.49)	3.32 (1.27)	2.96 (1.35)	0.164	1.671 (0.54)	4.02 (0.67)	< 0.001
Pubertal (Tanner) group, N (%)				0.095			< 0.001
Pre/Early (Stages 1–2) Mid/Late (Stages 3–5)	38 (37.6 %) 63 (62.4 %)	14 (29.2 %) 34 (70.8 %)	24 (45.3 %) 29 (54.7 %)		38 (100 %) 0 (0 %)	0 (0 %) 63 (100 %)
Sex, N (%)				< 0.001			0.095
Male Female	53 (52.5 %) 48 (47.5 %)	0 (0 %) 48 (100 %)	53 (100 %) 0 (0 %)		24 (63.2 %) 14 (36.8 %)	29 (46.0 %) 34 (54.0 %)
Gender, N (%)				< 0.001			0.159
Male Female Nonbinary	52 (51.5 %) 48 (47.5 %) 1 (1.0 %)	0 (0 %) 48 (100 %) 0 (0 %)	52 (98.1 %) 0 (0 %) 1 (1.9 %)		24 (63.2 %) 14 (36.8 %) 0 (0 %)	28 (44.4 %) 34 (54.0 %) 1 (1.6 %)
Annual Household Income, N (%)				0.837			0.898
< $100,000 ≥ $100,000 Not Reported	65 (64.3 %) 34 (33.7 %) 2 (1.6 %)	32 (66.7 %) 16 (33.3 %) 0 (0 %)	33 (62.3 %) 18 (34.0 %) 2 (3.8 %)		24 (63.2 %) 13 (34.2 %) 1 (2.6 %)	41 (65.1 %) 21 (33.3 %) 1 (1.6 %)
Race/Ethnicity, N (%)				0.328			0.122
White, Non-Hispanic Black, Non-Hispanic White, Hispanic Asian/Pacific Islander Native American Biracial Other or Not Reported	37 (36.6 %) 41 (40.6 %) 7 (6.9 %) 6 (5.9 %) 1 (1.0 %) 8 (8.0 %) 1 (1.0 %)	16 (33.2 %) 24 (50 %) 2 (4.2 %) 3 (6.3 %) 0 (0 %) 2 (4.2 %) 1 (2.1 %)	21 (39.6 %) 17 (32.1 %) 5 (9.4 %) 3 (5.7 %) 1 (1.9 %) 6 (11.3 %) 0 (0 %)		19 (50 %) 11 (28.9 %) 4 (10.6 %) 1 (2.6 %) 1 (2.6 %) 2 (5.3 %) 0 (0 %)	18 (28.6 %) 30 (47.6 %) 3 (4.8 %) 5 (7.9 %) 0 (0 %) 6 (9.5 %) 1 (1.6 %)

2.2. Questionnaires

Youth sex assigned at birth was measured using parent report and youth gender was measured using self-report. Pubertal development was assessed using the self-reported Tanner stages questionnaire (Tanner, 1962). Consistent with prior research, participants were categorized as Pre/Early pubertal (stages 1–2) or Mid/Late pubertal (stages 3–5; Morgan et al., 2013). Anxiety symptoms were measured using the 41-item self-reported Screen for Child Anxiety Related Emotional Disorders (SCARED), validated for ages 8–18 and commonly used to screen for anxiety symptoms in pediatric populations (Birmaher et al., 1997). The anxiety total score was used for anxiety symptoms in analyses.

2.3. Pavlovian fear extinction paradigm

The Pavlovian fear extinction paradigm was based on a novel version initially established in adults (Milad et al., 2007) and later adapted and validated for children and adolescents by our lab (Marusak et al., 2017). Briefly, in this task, the conditioned stimuli (CS) are virtual avatars, chosen due to the high rates of interpersonal violence in our population. The avatars vary in race/ethnicity to reflect the demographic makeup of our sample, and the identity is counterbalanced across participants. The unconditioned stimulus (US) is a 0.5 s 95 dB white noise burst, delivered through headphones at a partial reinforcement rate of 75 %. Avatars are presented in one of two virtual contexts (hallways), counterbalanced across participants. The two-day design includes fear conditioning in one context (CXT+), and, following a 10-minute delay, extinction in a distinct (CXT−) context, as extinction retention has been shown to be context-dependent (Hermann et al., 2017, Ji and Maren, 2007, Kalisch et al., 2006). Between-session testing of extinction retention was assessed the following day in the safety context (CXT−). Although extinction training and extinction retention testing both occurred in the same virtual context (CXT−), they were not always conducted in the same physical environment (i.e., within MRI scanner, or in lab using virtual reality) due to scheduling constraints, funding, and scanner availability. Additional task information can be found in the Supplemental Material.

During the conditioning phase, participants were presented with two conditioned stimuli (CS+), both presented for 4 s and paired with the US. Of these, one conditioned stimulus (CS+E) was subsequently extinguished, while the other (CS+U) remained unextinguished. A third stimulus (CS−) was never paired with the US. The fear conditioning phase consisted of six CS+ –US pairings, two additional non-reinforced CS+ presentations, and eight CS− presentations, resulting in a total of 24 trials. All conditioning trials were conducted within the conditioning context (CXT+).

During the extinction phase, the extinguished stimulus (CS+E) was repeatedly presented without the US to extinguish the conditioned response. The number of extinction trials varied based on participant age: younger participants (children, ages 6–12) completed eight trials each of CS+E and CS-, whereas older participants (adolescents, ages 10–17) completed 12 trials each. This adjustment was made to account for prior evidence suggesting that adolescents may be more resistant to extinction than children (Pattwell et al., 2012). All extinction trials were conducted in a second context (CXT-).

On the second day, participants completed a test of extinction retention during which they were presented with the CS+E, CS+U, and CS− stimuli in the CXT−, with younger participants completing eight trials per stimulus and older participants completing 12 trials.

2.4. Skin conductance responses (SCRs)

SCRs were recorded during CS presentation using two electrodes (EL509, BIOPAC Systems, Inc., Goleta, CA) filled with 0.5 % saline paste (GELE101, BIOPAC Systems) attached to the second and third fingers of the nondominant hand. A ground electrode was placed on the palm during the extinction retention session conducted in the fMRI scanner. SCR data were acquired at a sampling rate of 1000 Hz using AcqKnowledge software v.4.4 (BIOPAC Systems) and analyzed using event-related electrodermal response analysis. SCR values were calculated based on the maximum signal within a 0.5–4.5 s latency window following CS onset, normalized to a 2 s baseline average prior to CS onset. To improve comparability across age groups, an extinction retention index was used to account for individual differences in conditioned fear responses during the acquisition phase. This index was calculated as follows:

Extinction Retention Index= 100 −([Average SCR during the first two extinction retention trials/ Largest SCR during fear conditioning] x 100). A score of 0 indicates poor extinction retention, whereas a score of 100 indicates complete retention. While alternative methods of defining extinction retention exist (Lewis et al., 2023, Lonsdorf et al., 2019), our choice to calculate the Extinction Retention Index using only the first two retention trials is based on our a priori hypotheses and theoretical frameworks emphasizing that the early trials are the most sensitive indicators of extinction memory retention (Milad et al., 2007). This approach is also consistent with our prior work using this method (Marusak et al., 2018), enhancing comparability across findings. For completeness, we also conducted sensitivity analyses using an alternative extinction retention method (i.e., average SCRs during all retention trials). These analyses show that the main findings are largely consistent across metrics (see Supplemental Material). Further, although our primary focus was on extinction retention, effects of Sex and Puberty on SCRs during earlier phases (i.e., fear conditioning, extinction learning) are provided in the Supplemental Material.

2.5. FMRI analysis

2.5.1. Data acquisition

FMRI data were collected during extinction retention using a single research-dedicated Siemens 3 T MAGNETOM Verio System equipped with a 32-channel head coil. Whole-brain blood oxygen level-dependent (BOLD) signal was acquired using a multi-echo/multiband echo-planner imaging sequence (TR = 1500 ms, TEs = 15.0, 30.72, 46.44 ms, FA = 83 degrees, slices = 51, FOV = 209 ×209, voxel size = 2.9 mm isotropic, acceleration factor = 2). Before entering the scanner, all participants underwent mock scanner training to familiarize themselves with the task and reduce potential head motion. SCR data were recorded continuously using MRI-safe fiber optic cables (MECMRI-TRANS), an SCR amplifier (EDA100C-MRI), and disposable electrodes (LEAD119A-MRI; BIOPAC Systems). The total scan session lasted approximately 1 h, including a baseline anatomical scan followed by the functional scan during the extinction retention task. The extinction retention task was scheduled near the end of the session to minimize potential confounding effects of the novel scanning environment on task performance.

2.5.2. Data preprocessing and quality assurance

Preprocessing, denoising, and quality assurance were conducted using a custom preprocessing pipeline based on AFNI’s afni_proc.py. The preprocessing pipeline included: (1) skull-stripping and warping of the anatomical image to the Montreal Neurological Institute template (@auto_tlrc), (2) removal of the first 4.5 s of data to allow for signal equilibration (3dTcat), (3) time shifting of the echo timeseries to account for differences in slice acquisition timing (3dTshift), (4) co-registration of the first echo times, which has maximal signal, for motion correction and for anatomical-functional co-registration, with the saved transformation matrix applied to the second and third echoes (3dvolreg, 3dAllineate), (5) within-brain mask creation based on the first echo to restrict ICA to within-brain areas (3dAutomask), (6) ICA and optimal combination of the BOLD timeseries (31.1 ± 7.5 BOLD components retained) using tedana, including minimum image regression, and (7) 12-parameter affine anatomical-functional co-registration (align_epi_anat.py).

2.5.3. First-level fMRI analysis

Preprocessed and fully denoised BOLD timeseries were analyzed using general linear modeling (GLM) in SPM8. Trial onset times for experimental conditions were modeled and convolved with the hemodynamic response function. First-level models were created with the conditions (CS+E, CS+U, CS- and context), and BOLD activity was isolated to the contrast CS+E > CS+U during extinction retention, following prior work (e.g., Fullana et al., 2018).

2.5.4. Regions of interest (ROI) analysis

A priori ROIs were selected based on their roles in extinction learning and/or retention in adults (Fullana et al., 2018, Lang et al., 2009): left and right amygdala and hippocampus, and the ACC (Brodmann Area [BA] 32). ROIs were anatomically defined using the Automated Anatomical Labeling (AAL) Atlas (see Fig. 2F). Primary eigenvariates (i.e., first principal components) were extracted from each ROI for the contrast CS+E > CS+U during extinction retention and submitted to SPSS for analysis.

Fig. 2.

Associations between Sex, Puberty, and neural activation during extinction retention. (a, b) Hippocampal activation (left and right, respectively) during Pre/Early and Mid/Late puberty between males and females, showing no significant main effects or interactions. (c, d) Amygdala activation (left and right, respectively) showing significant interaction between puberty and sex, driven by significantly lower activation in Mid/Late puberty females compared to Pre/Early puberty (p = 0.024). (e) ACC activation demonstrating significantly lower activity during Mid/Late puberty compared to Pre/Early puberty (p = 0.018). (f) Neuroanatomical locations of the hippocampus, amygdala, and ACC highlighted on a template brain, as defined using the Automated Anatomical Labeling (AAL) atlas. Brain activation isolated during extinction retention to the contrast: CS+E > CS+U and shown in arbitrary units (a.u.).

2.6. Statistical analysis

Statistical analyses were performed using SPSS Statistics (Version 29.0.2.0 (20)). Descriptive statistics and correlations were computed for sex, age, pubertal (Tanner) stage (continuous), pubertal (Tanner) group, and anxiety symptoms (SCARED scores). A chi-square test was performed to examine relationships between sex, gender, puberty, household income, and race/ethnicity. We conducted three a priori, theory-driven analyses of variance (ANOVAs) focused on our primary outcomes of interest: anxiety symptoms, extinction retention, and neural activity during extinction retention. ANOVA was conducted with Sex (Male, Female) and Puberty (Pre/Early, Mid/Late), and their interaction (Sex × Puberty) as fixed factors and the following dependent variables: (1) anxiety symptoms (SCARED scores), (2) extinction retention (SCR retention index), (3) brain activity during extinction retention (ROI response to CS+E > CS+U). In addition to the ANOVAs, we conducted exploratory t-tests comparing male and female participants in the Pre/Early and Mid/Late pubertal stages, separately. Follow-up analyses examined specificity of effects to Puberty rather than Age Group (median split: Younger, ages 6–11; Older, ages 12–17). P-values were calculated and reported with a significance value of 0.05.

3. Results

3.1. Associations between sex, puberty, and demographic variables

As expected, the Pre/Early and Mid/Late pubertal groups differed in age, with more advanced pubertal stages associated with older age. However, the pubertal groups did not significantly differ in biological sex, gender, race/ethnicity, or household income. There were no sex differences in age, race/ethnicity, or household income. Notably, the difference in pubertal development between males and females did not reach significance (p = 0.095; see Table 1). Additionally, results did not differ when using self-reported gender instead of parent-reported sex assigned at birth (see Supplemental Material).

3.2. Sex × puberty associations with anxiety symptoms

A 2 × 2 (Sex: Male, Female × Pubertal Group: Pre/Early, Mid/Late) ANOVA revealed: (1) a significant main effect of sex, F(1,97) = 17.216, p < 0.001, with females reporting higher anxiety symptoms than males (see Figs. 1a), (2) a significant Sex × Puberty interaction, F(1,97) = 11.204, p = 0.001 (see Fig. 1a), and (3) no significant main effect of Puberty, F(1,97) = 2.322, p = 0.131. Follow-up comparisons showed that Mid/Late pubertal females reported significantly higher anxiety symptoms than Pre/Early pubertal females, t(46) = 2.756, p = 0.008. Additionally, Mid/Late pubertal females had higher anxiety symptoms than Mid/Late pubertal males, t(61) = 6.366, p < 0.001 (see Fig. 1a). Anxiety symptoms did not differ between Pre/Early pubertal males and Mid/Late pubertal males (p = 0.094) or between Pre/Early pubertal females and Pre/Early males (p = 0.630). Follow-up analyses tested for specificity of effects to Puberty compared to Age Group (younger, older). There was no significant Sex × Age interaction, F(1,23) = 1.656, p = 0.100, and no main effect of Age on anxiety symptoms. However, there was a significant main effect of Sex, F(1,23) = 20.149, p < 0.001, again with females reporting higher anxiety symptoms than males.

Fig. 1.

Significant Sex × Puberty interactions for anxiety symptoms (a) and extinction retention (b). (a) Anxiety symptoms by Pre/Early and Mid/Late puberty group in males and females. There was a significant difference between Pre/Early pubertal females and Mid/Late pubertal females, and between Mid/Late pubertal males and Mid/Late pubertal females. (b) Extinction retention index, measured by skin conductance response (SCR), by Pre/Early and Mid/Late puberty group in males and females. There was a significant difference between Mid/Late pubertal males and Mid/Late pubertal females and between Pre/Early pubertal males and Mid/Late pubertal males. 0 % indicates poor extinction retention (i.e., return of fear); 100 % indicates complete extinction retention (i.e., low fear).

3.3. Sex × puberty associations with extinction retention

Extinction retention ranged from 0 % to 100 % (M = 52 %, SD = 42.97 %). There were no main effects of Sex or Puberty on extinction retention. Similar to anxiety symptoms, there was a Sex × Puberty interaction; however, this did not reach significance (F(1,97) = 3.26, p = 0.074). Exploratory t-tests within Pre/Early and Mid/Late groups separately revealed significantly poorer extinction retention in Mid/Late females compared to Mid/Late males (t(57) = 2.746, p = 0.008; see Fig. 1B), and no significant difference between Pre/Early females compared to Mid/Late females (t(45) = 0.531, p = 0.598). Additionally, among males, extinction retention was significantly better among Mid/Late males compared to Pre/Early males, t(49) = 2.13, p = 0.038. Similar to anxiety symptoms, there was no significant Sex × Age interaction on extinction retention (F(1,97) = 0.95, p = 0.76), and also no main effect of Sex (F(1,97) = 3.154, p = 0.079). There was, however, a significant main effect of Age Group (F(1,97) = 6.906, p = 0.01), with older youth (ages 12–17) showing better extinction retention compared to younger youth.

3.4. Sex × puberty associations with neural activation during extinction retention

There were no main effects or interactions of Sex or Puberty in the left hippocampus, right hippocampus, or left amygdala (p’s > 0.05; see Fig. 2A-C). However, a significant Sex × Puberty interaction was observed for right amygdala response to the CS+E > CS+U (F(1,97) = 7.406, p = 0.027). This was driven by Mid/Late pubertal females showing lower amygdala response compared to Pre/Early pubertal females (t(46) = 2.335, p = 0.024; see Fig. 2D). No main effects of Sex or Puberty were observed on right amygdala activity. In the ACC, a main effect of Puberty was found, with lower activity observed in Mid/Late puberty compared to Pre/Early puberty (F(1,97) = 5.789, p = 0.018; see Fig. 2E). There was no significant main effect of Sex or significant Sex × Puberty interaction on ACC activity. An alternate model using Age Group instead of Puberty showed no significant main effects or interactions between Sex and Puberty on neural activation in any ROI (p’s> 0.05).

4. Discussion

This study examined the effects of sex assigned at birth, puberty, and their interaction on anxiety symptoms, extinction retention, and the underlying neural circuitry in youth spanning puberty. Our findings align with prior research, showing that Mid/Late pubertal females exhibit higher anxiety symptoms than both Mid/Late pubertal males and Pre/Early pubertal females. This supports evidence that advanced pubertal status in female (but not male) adolescents is associated with increased anxiety symptoms (Dalsgaard et al., 2020, Khanal et al., 2022, Wesselhoeft et al., 2015).

Beyond heightened anxiety symptoms, although the Sex × Puberty interaction did not reach significance for extinction retention, exploratory within-group t-tests showed poorer extinction retention in Mid/Late pubertal females compared to their male counterparts. Given the non-significant interaction, these behavioral findings should be interpreted with caution and considered preliminary. In addition, we observed differences in key neural regions implicated in fear and extinction processing. Specifically, Mid/Late pubertal females exhibited reduced right amygdala activation during extinction retention compared to Pre/Early pubertal females, suggesting that puberty-related amygdala changes could contribute to sex-specific extinction retention differences. Additionally, puberty-related differences in ACC activation were observed, with both Mid/Late pubertal males and females showing lower ACC activation during extinction retention compared to their Pre/Early pubertal counterparts. Together, these findings provide initial evidence that puberty-related alterations in extinction retention and underlying fear circuitry may contribute to the well-documented sex differences in fear-based disorder risk, although further research is needed to confirm these behavioral patterns.

Our observation of poorer extinction retention in Mid/Late pubertal females aligns with previous research indicating that more advanced pubertal status in females (but not males) is associated with higher anxiety symptoms (Reardon et al., 2009). Prior studies also suggest that individuals with anxiety and PTSD symptoms exhibit heightened physiological fear responses during extinction learning and retention (e.g., Abend et al., 2020; Graham and Milad, 2011b). Therefore, lower extinction retention in Mid/Late pubertal females may reflect a unique vulnerability to the development of fear-based disorders.

Animal studies provide further insights into these findings. Research on adolescent rats has shown that female rats exhibit impaired extinction learning and retention when estradiol levels are high, specifically during proestrus or met/diestrus phases (Perry et al., 2020). Importantly, gonadectomy prior to puberty improved extinction retention in female adolescent rats, suggesting that puberty-related increases in estradiol may be detrimental to extinction retention during adolescence. These findings contrast with adult rat studies in which high estrogen has been associated with enhanced extinction retention (Milad et al., 2009). This discrepancy underscores the significant neural reorganization that occurs during adolescence, particularly in females, which may contribute to their heightened sensitivity to estrous/menstrual cycling and sex hormones in fear extinction (Juraska and Drzewiecki, 2020). Taken together, our findings suggest that the onset of puberty may negatively impact extinction retention in some individuals assigned female at birth, potentially contributing to their increased risk for fear-based disorders. Further, while menstrual cycle variability in adolescence complicates direct comparisons, future research should investigate whether fluctuations in estradiol at the time of extinction learning relate to retention.

Interestingly, only among males (not females), extinction retention ability improved from Pre/Early to Mid/Late puberty. This finding contradicts prior studies in male rodents, which have reported extinction retention impairments in adolescent compared to preadolescent and adult rats (see Malik et al., 2023 for review). These findings build on our prior work in predominantly prepubertal youth, which demonstrated overall poor extinction retention and laid the groundwork for exploring developmental trajectories in humans (Marusak et al., 2018). Indeed, we previously observed overall low extinction retention (retention index: 13.56 %; Marusak et al., 2018) in a predominantly prepubertal sample compared to what is typically reported in adults (74–85 %; Holt et al., 2009; Milad et al., 2007; Milad, Pitman, et al., 2009; Rabinak et al., 2014). The current results extend these findings by showing that, although prepubertal youth of both sexes exhibit poor extinction retention, males—but not females—demonstrate improved recall with advancing pubertal development, suggesting a sex-specific developmental trajectory. Consistent with this, recent findings suggesting that extinction retention impairments that are commonly found in rats during adolescents may not occur in human extinction retention models (Widegren et al., 2025). Additionally, evidence from animal studies indicate that testosterone enhances extinction retention in adolescent rats (Malik et al., 2023, Perry et al., 2020); therefore, it is possible that pubertal testosterone may facilitate extinction learning in human adolescents males, contributing to this observed developmental trajectory. Longitudinal studies are needed to evaluate when individuals exhibit more adult-like patterns of extinction retention and brain activation, as adolescence is a significant period of neurobiological reorganization—possibly extending into emerging adulthood (e.g., ages 18–24).

Further, we identified a sex-specific pattern in which extinction retention improved with pubertal development in males but remained low in females, which may correspond with the greater vulnerability of females to fear-based disorders. Future longitudinal studies will be needed to determine the precise timing of these developmental changes within individuals and to assess whether females eventually "catch up" in extinction retention ability during adulthood. Notably, the observed effects of sex and puberty on extinction retention were specific to pubertal status rather than chronological age, and were not evident during earlier learning phases (i.e., fear conditioning, extinction learning). This distinction highlights that, although puberty was associated with SCRs during earlier phases, sex differences in extinction retention appear to emerge specifically in relation to pubertal maturation, rather than younger age alone.

Neuroimaging analyses revealed significantly lower right amygdala activation during extinction retention in Mid/Late pubertal females compared to Pre/Early pubertal females. While the amygdala is well established as a key node in fear acquisition and conditioned responding (Quirk and Mueller, 2008), its role in extinction retention is more complex, particularly across development. In adult studies, heightened amygdala activation during extinction retention is typically associated with impaired extinction memory and persistent threat responsivity (Suarez-Jimenez et al., 2020). However, our findings diverge from this pattern, with Mid/Late pubertal females exhibiting lower amygdala activation yet showing lower extinction retention behaviorally (i.e., elevated SCRs). This dissociation suggests that reduced amygdala activation in adolescent females may not reflect successful extinction retrieval, but rather a developmental reorganization of fear circuitry. One possibility is that this pattern reflects a shift toward increased reliance on other regulatory regions, such as the prefrontal cortex or hippocampus, as the amygdala undergoes structural and functional maturation. Indeed, adolescence is characterized by synaptic pruning and reductions in amygdala gray matter volume (Zimmermann et al., 2019), with evidence of sex-specific trajectories—such as earlier and more pronounced pruning in females relative to male (Giedd et al., 1996, Koss et al., 2014). These developmental changes may produce transient patterns of amygdala responsivity that reflect ongoing neural reorganization. Our findings highlight the need to consider sex and developmental stage as moderators of extinction neurocircuitry, and suggest that altered amygdala function in adolescent females may contribute to heightened anxiety risk during puberty.

We also observed puberty-related reductions in ACC activation. Reduced frontal activation during fear extinction has been linked to anxiety pathology (Marin et al., 2017), but our findings suggest that despite comparable decreases in ACC activation across sexes, females exhibit greater anxiety vulnerability. This discrepancy underscores the potential influence of puberty or sex-specific stress responses in shaping anxiety risk rather than brain influences alone.

4.1. Strengths and limitations

A key strength of this study is its novel focus on extinction retention and its potential role in risk for fear-based disorders in adolescents. Additionally, stratifying findings by pubertal status rather than chronological age may provide a more precise understanding of developmental changes. The study also benefits from a diverse, urban sample drawn from the Detroit metropolitan area, enhancing generalizability. However, several limitations should be noted. The sample size was relatively small, which may limit the generalizability of findings. Furthermore, this is a cross-sectional study. A longitudinal study would allow us to more precisely track changes throughout puberty within a single individual. Additionally, due to the variability of both estradiol and testosterone in both sexes during puberty (Collaer and Hines, 1995), future studies should incorporate hormone assays to directly assess estradiol and testosterone levels and their potential impact on extinction learning and retention. Additionally, differences in extinction retention across pubertal groups may have been influenced by the greater number of extinction trials administered to older participants. Further, although our prior work found no differences in fear responding between training completed inside versus outside the scanner (Marusak et al., 2018), we cannot rule out the possibility that physical context contributed to variability in extinction retention. Finally, these findings are based on cisgender youth; future research should explore these associations in transgender and gender-diverse populations to ensure broader applicability.

4.2. Potential clinical implications

These findings have important implications for understanding sex differences in fear behavior development and their relationship to pubertal progression. The results highlight puberty—particularly in females—as a critical period for the emergence of anxiety symptoms, underscoring the need for increased anxiety screening in pubertal females to facilitate early intervention. Additionally, the distinct developmental trajectories of males and females, and the contrast between findings in adults and adolescents, suggest that treatment approaches for fear-based disorders in adolescents should not be extrapolated from adults and should closely consider sex differences, pubertal stage, hormone levels, and/or menstrual stage at the time of extinction-based treatments.

5. Conclusions

This study highlights the critical role of puberty in the emergence of sex differences in anxiety symptoms, extinction retention, and neural activation. Given the global trend toward earlier puberty onset (Eckert-Lind et al., 2020), these findings emphasize the need for longitudinal studies to investigate the influence of pubertal and sex hormones on fear extinction and the development of interventions specific to developmental stage, hormone levels, and the unique neurobiology of youth.

CRediT authorship contribution statement

Marusak Hilary: Writing – review & editing, Writing – original draft, Visualization, Supervision, Resources, Project administration, Investigation, Funding acquisition, Formal analysis, Data curation. Clara G. Zundel: Writing – review & editing, Supervision, Project administration, Methodology, Investigation, Data curation. MacKenna Shampine: Writing – review & editing, Data curation. Samantha Ely: Writing – review & editing, Data curation. Carmen Carpenter: Writing – review & editing, Data curation. Jennifer Losiowski: Writing – review & editing, Data curation, Conceptualization. Shravya Chanamolu: Writing – review & editing, Data curation. Jovan Jande: Writing – review & editing, Data curation. Reem Tamimi: Writing – review & editing, Validation, Methodology, Data curation. Kamakashi Sharma: Writing – review & editing. Emilie O’Mara: Writing – review & editing. Alaina M. Jaster: Writing – review & editing. Sneha Bhargava: Writing – review & editing, Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Data curation, Conceptualization.

Declaration of Competing Interest

The authors have nothing to declare.

Appendix A. Supplementary material

Supplementary material

Data availability

Data will be made available on request.