Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Jan 1.
Published in final edited form as: Int J Psychophysiol. 2021 Nov 29;171:48–54. doi: 10.1016/j.ijpsycho.2021.11.003

Hormonal contraceptive use moderates the association between worry and error-related brain activity

Courtney C Louis 1, Chelsea Kneip 1, Tim P Moran 2, Adriene M Beltz 3, Kelly L Klump 1, Jason S Moser 1
PMCID: PMC9128846  NIHMSID: NIHMS1763766  PMID: 34856221

Abstract

Uncovering mechanisms that can help explain the experience and impact of anxiety in women is important for improving etiological models and treatments to meet the needs of unique individuals. An enlarged error-related negativity (ERN) – an electrophysiological marker of cognitive control-related error monitoring– represents one indicator of neural processes more strongly related to anxiety in women than men. In this study, we further examined this association in women by testing the moderating effect of hormonal contraceptive (HC) use on the relationship between worry – i.e., a transdiagnostic cognitive dimension of anxiety – and the ERN. Results revealed that HCs moderated the worry-ERN association. Specifically, we found a significant and large relationship between worry and enlarged ERN in women using HCs, which was smaller and nonsignificant in naturally cycling women (i.e., those not using HCs). These findings suggest that the interplay among HC use, error-related cognitive control, and worry represents a novel mechanism for better characterizing the expression and impact of worry in women.

Keywords: Anxiety, worry, error-related negativity, hormonal contraceptives

Introduction

Anxiety 1is nearly twice as likely to affect women2 than men worldwide (Baxter et al., 2013; Kessler et al., 2005; Seedat et al., 2009). Women also suffer greater burden associated with anxiety, such as longer course and greater use of healthcare services (Baxter et al., 2014; McLean et al., 2011). Accordingly, anxiety is associated with impairments in cognitive function (Eysenck et al., 2007), and meta-analytic evidence from our group has found a stronger association between worry and error-related brain activity in women than men (Moser et al., 2016). Research investigating sex differences in anxiety and cognitive function has uncovered the crucial modulatory contributions of ovarian hormones (Barth et al., 2015; McEwen & Milner, 2017). Yet, there continues to be a poor understanding of the role of ovarian hormones in the association between anxiety and its related cognitive impacts in women. Uncovering factors that may contribute to the unique expression and impact of anxiety in women is critical for improving etiologic models and treatments anxiety and its related impacts. As such, it is not only important to understand the role of hormones in anxiety and cognition independently, but also how hormones may contribute to their association. That is how might ovarian hormones influence the association between and anxiety and cognitive function. Therefore, the current study aimed to investigate this by examining the role of hormonal contraceptives (HCs) in the association between worry and error-related cognitive control.

Across college, community and patient samples, anxiety is associated with an enlarged error-related negativity (ERN), an electrical brain signal that occurs less than 100 milliseconds after error commission. The ERN is generated by a neural circuit including the anterior cingulate cortex (ACC), prefrontal cortex (PFC), and motor areas (Gehring et al., 2012), and is understood to be involved in cognitive control-related performance monitoring (Yeung et al., 2004). This robust association indicates that the ERN has relevance across the anxiety spectrum (Cavanagh & Shackman, 2015; Moser et al., 2013). We have proposed that this enlarged signal reflects the added cognitive effort – or compensatory error monitoring – that anxious individuals require to correct and adapt performance because of the distracting effects of worry (Moser et al., 2013). Therefore, we predict that the prevalence of worry, a cognitive dimension of anxiety, has particular relevance for error monitoring that results in an enhanced ERN. Importantly, our group’s prior work identified that the worry-ERN association is larger in women than men (Moran et al., 2012; Moser et al., 2016). This suggests that the ERN may represent an important marker of the unique experience and impact of worry in women. Because ovarian hormones are known to impact anxiety-related neural processes and behavior in animal and human samples (Li & Graham, 2017), understanding their role in the worry-ERN association in women may be an important next step.

Ovarian hormones (e.g., estradiol and progesterone) provide a window into understanding the nature and impact of anxiety-related PFC function in women (McEwen & Milner, 2017). For instance, increasing work points to the critical role of ovarian hormones in brain structure and function, particularly in regions involved in cognitive control such as the prefrontal cortex (Beltz & Moser, 2020; Ter Horst et al., 2012; Van Wingen et al., 2011). In addition, ovarian hormones have modulatory effects on neural processes via their interaction with various neurotransmitter pathways involved in regulating anxiety and cognitive function (Barth et al., 2015). Most relevant to the current investigation, converging animal work has shown that hormones may moderate anxiety/stress-induced PFC dysfunction in female rats (Shansky & Lipps, 2013). Specifically, Shansky and Lipps (2013) propose that during times of high anxiety/stress, female rats with high estradiol levels may have exacerbated disruptions in PFC function than those with low estradiol levels. Although this modulatory role of hormones has been shown in rodent models, no studies have investigated how such findings may translate to humans.

One way to examine the influence of hormones on anxiety-related PFC function in women is by investigating the effects of HCs (Beltz & Moser, 2020; Hampson, 2020; Taylor et al., 2021). Combined HCs, which make up the majority of HCs used by women (Mosher & Jones, 2010), contain synthetic doses of estrogen and a progestin that inhibit endogenous ovarian production of estradiol and progesterone and suppress their variation across the course of the menstrual cycle (Fleischman et al., 2010; Hampson, 2020; Rivera et al., 1999). Approximately 80% of women use HCs at some point in their lives; this includes oral contraceptives, intrauterine devices, among other administration methods (Daniels & Jones, 2013; Mosher & Jones, 2010). This translates to more than 100 million active HC users worldwide (Petitti, 2003).

Despite their widespread use, the effects of HCs on the brain and behavior are only recently being recognized (Montoya & Bos, 2017; Taylor et al., 2021). Findings are emerging linking HC use with increased anxiety and mood symptoms (Deci et al., 1992; Porcu et al., 2019; Skovlund et al., 2016). For instance, studies have found that those using HCs evidence heightened neural activation and connectivity in areas such as the amygdala, thalamus, and prefrontal regions, including the ACC during fear extinction, and show reduced extinction learning (Graham & Milad, 2013; Merz et al., 2012). Cognitive neuroscience research further shows that the use of HCs is associated with altered structure and function of the ACC and PFC (Engman et al., 2018; Petersen et al., 2014; B. Pletzer et al., 2014), key regions implicated in anxiety and the generation of the ERN. Extant studies have also revealed differences in resting-state brain connectivity in HC users compared to non-users (Beltz & Moser, 2020). For instance, HC users show reduced connectivity in default mode (DMN) and executive control networks (ECN), which may have implications for anxiety and cognitive processing (Peterson et al., 2014). Also, during a cognitive task, HC users show increased activation in prefrontal regions compared to non-users, and this enhanced activation was related to performance (Pletzer et al., 2014). These findings indicate that HCs may systematically alter neurocognitive function.

Further, Studies generally suggest that the dampening of endogenous hormone fluctuations by HCs is associated with worse anxiety symptoms and disruptions of frontal control networks (Andreano et al., 2018; Montoya & Bos, 2017; Petersen et al., 2014). These findings imply that HCs affect anxiety- and cognitive- related processes. However, it is also important to identify whether HCs affect associations between anxiety and cognitive mechanisms. While a majority of women are likely to use HCs at some point in their lives, there continues to be a poor understanding of whether anxiety differentially impacts cognitive processing in the context of HC use.

In sum, the current study aimed to examine the role of HC use in the association between worry and the ERN in women. As stated above, there are known sex differences in anxiety and neural function that are at least partially related to the neuromodulatory role of ovarian hormones. Although research has identified that the worry-ERN association is larger in women, there continues to be a lack of clarity on what contributes to this enhanced association. Therefore, examining HC use might provide insight into a potential mechanism by which worry impacts the ERN in women.

Specifically, we examined whether the worry-ERN association differed between those using HCs and not (i.e., a moderation). Extant research indicates between-person differences in HC users and non-users, such that HC users evidence reduced connectivity at rest in DMN and ECN (Petersen et al., 2014), associations between neural activity and performance during cognitive tasks (Pletzer et al., 2014), and increased activity and connectivity in prefrontal areas during fear extinction (Graham & Milad, 2013). On the other hand, research has shown that HC use leads to increases in anxiety and reductions in PFC function (Montoya & Bos, 2017), and one study reported that hormones affect the ERN magnitude via increases in obsessive-compulsive symptoms (Mulligan et al., 2019). Importantly, rather than indicating that neurocognitive differences in those using HCs is attributed to increased worry levels, the extant literature indicates that HC use leads to between-person differences in neurocognitive function. Therefore, we predicted that the worry-ERN association would be moderated by HC use.

Method

Participants

Demographic information is presented in Table 1. Sixty-eight undergraduates recruited through the Michigan State University Psychology subject pool participated in the current study for partial course credit3. Thirty of these women were recruited because they were using HCs (24 using monophasic HCs; 2 using triphasic HCs; 1 using a progestin only HC; 3 using unspecified HCs) and the other 38 were recruited because they were not using any form of HC – hereafter referred to as “naturally cycling (NC) women.” For NC women, we used previously established procedures for coding menstrual cycle phase (see Klump et al., 2013; Lester et al., 2003)) such that we asked participants to provide menstruation start dates for their three most recent cycles to predict which cycle phase they were in at the time of the laboratory visit. This allowed us to estimate the cycle phase based on the average length of participants’ menstrual cycle. The first day of menstruation was considered day 1. The ovulation phase was considered days −15 to −12 days before the first day of menstruation (i.e., 12–15 days prior to Day 1). The luteal phase was coded as days −5 to −1 before the menstruation. Finally, the follicular phase was coded as days +3 up until +10 or day −16, depending on the length of an individual’s cycle. Sufficient data for calculating cycle phase were available for 29 (76%) of the 38 naturally cycling women. The number of women who were in each phase of the menstrual cycle during study visits was as follows: Nine (24%) were in their follicular phase, nine (24%) were in their ovulatory phase, 11 (28%) were in their luteal phase, and nine (24%) were unknown. All but three women (10%) in the HC group were in the active pill phase of their regimen. Women using HCs had been using their HCs for an average of 20.91 months (SD = 16.48; range .5 – 51.25 months). Women using HCs and NC women were similar in age (t (66) = 1.89, p = .07) and race (χ2= 4.25, p = .24; See Table 1).

Table1.

Descriptive statistics and tests of differences between HC groups on demographic, behavioral performance, and ERP measures.

Variable HC Group (n =30) NC Group (n = 38) HC vs. NC Comparison
Age 19.20 (1.03) 18.74 (0.98) .07
Race (n (%)) .24
 Caucasian 27 (90%) 30 (78.9%)
 Black/AA 0 (0%) 2 (5.3%)
 Asian 0 (0%) 3 (7.9%)
 Other 1 (6.7%) 2 (5.3%)
Missing 1 (3.3%) 1 (2.6%)
Accuracy .91 (.04) .90 (.06) .34
Error RT 362.54 (32.41) 364.42 (50.04) .85
Correct RT 431.42 (32.89) 433.11 (38.19) .85
ERN −2.09 (6.37) −2.88 (4.51) .57
CRN 1.22 (5.92) 0.38 (4.05) .51
ΔERN −3.32 (6.49) −3.26 (5.46) .97
Open in a new tab

Note. p values are presented in the “HC vs. NC Comparison” column. Racial composition was compared using a χ2 test. All other comparisons were t-tests. AA, African American.

Note. The age range for both the HC and NC group was 18–21 years old.

Task

Participants completed a letter version of the Eriksen Flankers task (Eriksen & Eriksen, 1974). They were instructed to respond to the center letter of a five-letter array in which the target was either congruent (e.g., MMMMM) or incongruent (e.g., MMNMM) with the surrounding letters. During each trial, flanking letters were presented 35 ms prior to target letter onset, after which all five letters remained on the screen for an additional 100 ms. Participants were given 1000 ms to respond before the trial was terminated and excluded from analysis. A fixation cross (+) was presented during the intertrial interval, which varied from 1200 ms to 1700 ms. The experimental session included 480 trials grouped into 12 blocks of 40 trials. Participants were instructed to respond as quickly and as accurately as possible using either their left or right index finger to select the respective left (“A”) and right (“L”) keyboard buttons that corresponded to the flanker target. Response mappings were switched between blocks (e.g., left button for a target “N” in Block 1, right button for a target “N” in Block 2) to increase difficulty. Letters making up the arrays differed across block pairs.

Measures

Upon completion of the flanker task, participants filled out a series of self-report questionnaires including the Penn State Worry Questionnaire (PSWQ; Meyer et al., 1990) and the Anxious Arousal (AA) subscale of the Mood and Anxiety Symptom Questionnaire (MASQ; Watson & Clark, 1991)). The PSWQ was used as a measure of anxious apprehension/worry and the MASQ-AA was used as a measure of anxious arousal (Nitschke et al., 2001). The MASQ-AA scale was primarily utilized as a control variable in an attempt to replicate our previous findings that the ERN shows a specific relationship with worry (i.e., PSWQ) and not anxious arousal (i.e., MASQ-AA; Moser et al., 2012). Reliability estimates were computed for PSWQ using Cronbach’s alpha(α =.95)4.

Psychophysiological Recording and Data Reduction

Continuous electroencephalographic activity was recorded using the ActiveTwo BioSemi system (BioSemi, Amsterdam, The Netherlands) via 64 Ag-AgCl active electrodes placed in accordance with the 10/20 system. Two additional electrodes were placed on the left and right mastoids. Electrooculogram (EOG) activity generated by eye movements and blinks was recorded at FP1 and at three electrodes placed inferior to the left pupil and on the left and right outer canthi approximately 1 cm from the pupil. During data acquisition, the Common Mode Sense active electrode and Driven Right Leg passive electrode formed the ground, as per BioSemi’s design specifications. All signals were digitized at 512 Hz using ActiView software (BioSemi). In addition, an online low pass filter was applied (fifth order since filter) with a −3dB cutoff point at 102.4.

Offline analyses were conducted using BrainVision Analyzer 2 (BrainProducts, Gilching, Germany). Scalp electrode recordings were referenced to the numeric mean of the mastoids and a Butterworth filter was used for offline filtering with cutoffs of 0.1 and 30 Hz (12 dB/oct rolloff). Ocular artifacts were corrected using the regression method developed by Gratton et al. (1983). Physiological artifacts were detected using a computer-based algorithm such that trials in which the following criteria were met were rejected: a voltage step exceeding 50 μV between contiguous sampling points, a voltage difference of more than 200 μV within a trial, or a maximum voltage difference less than 0.5 μV within a trial. Trials were also removed from ERP and behavioral analyses if the RT fell outside of a 200–800 ms time window or if accuracy was less than 50% in a block, suggesting the participant confused the stimulus-response mappings. The response-locked data were segmented into individual epochs beginning 200 ms before response onset and continuing for 800 ms following the response, for errors and corrects and right and left-handed responses. To quantify response-locked ERPs, the average activity in the 200-ms window preceding response onset was subtracted from each data point subsequent to the response.

The ERN and its correct counterpart, the correct-related negativity (CRN), were quantified as the average activity at fronto-central recording sites – FC1, FC2, FCz, Fz, and Cz for error and correct trials respectively. Interpolation was completed by following the spherical splines method as reported in (Perrin et al., 1989). Trials only focused on responses made with the right hand (described below). Approximately 87% of trials for right-handed responses were retained after artifact rejection. The number of trials included in the final analysis ranged from 2–61 for error trials (M = 19, SD = 13), and ranged from 29–227- 475 for correct trials (M = 182; SD = 44). A split-half reliability analysis (using odd and even trials) was performed with a Spearman correction. The reliabilities was as follows for the sample – ERN: r = .25, p = .04; CRN: r = .96, p <.001, similar to those reported in other samples of women (Swainston et al., 2021).

Overview of Analyses

Multi-level models were completing using the “lme4” package (Bates et al., 2007) in R Version 3.5.1. Analyses focused on behavioral and ERP measures elicited by the right hand given recent evidence from our lab indicating that the ERN-PSWQ relationship is only apparent when individuals respond with their right hand in bi-manual response paradigms (see (Lin et al., 2015). Specifically, worry is related to left-lateralized frontal activity, and this may be due of its verbal nature (Heller et al., 1997; Sharp et al., 2015). Because right-handed responses are also controlled by the left hemisphere, the worry-ERN association may be larger for right-handed errors due the enhanced conflict of verbal processing and error monitoring (Lin et al., 2015). Indeed, correlations between ERNs elicited by left hand errors and PSWQ were small and non-significant across the whole sample and within each group in the current study (rs < −.16, ps > .35).

To test the hypotheses of the current study, we conducted multi-level model to account for the nesting of Accuracy and electrodes within participants. To test whether HCs moderated the association between worry and the ERN, we included Accuracy (Error vs Correct) as an effects-coded, within-participants factor, Hormonal Contraceptive Status (HC vs. NC) as a between-participants, effects-coded factor and PSWQ as a grand mean-centered between-participants continuous predictor. The random effects followed a cross-classified structure with two intercepts for participants and channel. Accuracy was specified as a random slope nested within participant allowing for the association between accuracy and ERP amplitude to vary by person (Volpert-Esmond et al., 2021). The variance-covariance matrix was unstructured and the intraclass coefficient indicated there was substantial variability between participants (ICC = .54).

F values were estimated from these linear models using Type III ANOVA tests, and partial eta squared (η2p) is reported as an estimate of effect size in ANOVA models where .05 represents a small effect, .1 a medium effect, and .2 a large effect (Cohen & Taylor, 1973; Volpert-Esmond et al., 2021). When significant effects of worry scores were detected, we examined the effect of PSWQ by dummy-coding the categorical variables of interest. In addition, a post-hoc sensitivity analysis was conducted using G Power to estimate the size of the effect we could detect given our sample size. The results revealed that given our sample size, we could detect a small interaction effect (f =.04). Further, we conducted a series of robustness tests to evaluate the stability of the effect of interest under different conditions.

Results

Anxiety Measures

Women using HCs were not significantly different from naturally cycling women on PSWQ scores (M = 52.37, SD = 14.30 vs. M = 52.84, SD = 13.94, respectively; t (66) =.14; p = .89). However, women using HCs reported more anxious arousal symptoms on the MASQ-AA scale (HC M = 38.30, SD = 9.44 vs. NC M = 33.05, SD = 9.90, respectively; t (66) = 2.23, p = .03, d = .54).

Behavioral Performance Measures

See Table 1 for a summary of descriptive statistics and tests of differences between HC groups on primary behavioral performance measures. Women using HCs did not differ from NC women with respect to overall accuracy or RT in the flanker task. Neither PSWQ nor MASQ-AA scores were related to error rate in either group (rs < |.24|, ps > .14). A 2 (Group: HC vs. NC) X 2 (Accuracy: Error vs Correct) X PSWQ multi-level model conducted on reaction times replicated the well-documented finding that errors are faster than corrects in the flanker task (M = 363.59, SD = 42.87 vs. M = 432.36, SD = 35.69, respectively; F(1, 64) = 4.63, p = .04, η2p = .07). None of the other effects were significant, thus there were no HC group differences in task performance (Fs < 2.67, ps > .10).

HC use as a moderator of the association between worry and the ERN

Table 1 also displays a summary of descriptive statistics and tests of differences between HC groups on ERN, CRN and ERN – CRN difference (ΔERN). Figure 1 displays grand averaged waveforms for the ERN across the two HC groups. The 2 (Group: HC vs. NC) X 2 (Accuracy: Error vs. Correct) X PSWQ multi-level model5 revealed a significant effect of PSWQ scores (F(1, 64) = 5.89, p = .02, η2p = .08) and Accuracy (F(1, 64) = 24.99, p < .001, η2p = .28). The results revelated that PSWQ predicted more negative activity overall (b = −.09, p = .02), and error trials predicted more negative activity than correct trials (b = −3.32, p < .001). Both of these effects were qualified by a significant PSWQ X Accuracy interaction (F(1, 64) = 6.84, p = .01, η2p = .10). This interaction was examined by dummy-coding Accuracy, revealing that there was no significant effect of PSWQ on correct trials (b = −.03, p = .56), but there was a significant negative effect on error trials (b = −.15, p = .001). Important to the main aims of the current study, the Group X Accuracy X PSWQ interaction was significant (F(1, 64) = 4.48, p = .03, η2p = .07). To decompose this 3-way interaction, the Accuracy X PSWQ interaction was examined in each group separately. The Accuracy X PSWQ interaction did not approach significance in NC women (b = −.02; p = .71). Consistent with our prediction, however, the Accuracy X PSWQ interaction was significant for women using HCs (b = −.23; p = .002).

Figure 1.

Figure 1.

Open in a new tab

Response-locked grand averaged ERP waveforms for error and correct responses across HC groups. Time point 0 indicates response onset. ERN emerges as sharp negative deflection within 100ms of response onset.

Therefore, we further broke down the PSWQ x Accuracy interaction for those using HCs. We found that there was a significant effect of PSWQ on error trials (b = −.23, p < .001), but not on correct trials (b = −.003; p = .95), as shown in Figure 1. That is, worry predicted a larger (i.e., more negative) ERN, but had no effect on the CRN6. These results indicate that they are likely attributable to active ingredients of the HCs, as all but three women in the HC group were in the active pill phase of their regimen7.

A series of robustness checks were conducted to examine whether the three-way interaction held when including anxious arousal, age and race. Importantly, this interaction remained significant with a similar effect size when controlling for MASQ-AA scores (F(1, 64) = 4.47, p = .03, η2p = .07), age (F(1, 64) = 4.47, p = .03, η2p = .07), and race, (F(1, 62) = 3.74, p = .05, η2p = .06) between groups. This interaction also remained significant after removing those who were not on the active phase of their pill use (F(1, 61) = 4.07, p = .04, η2p = .06). In addition, we found that in HC users, the two-way interaction between PSWQ and accuracy remained significant and of similar effect size when controlling for HC type (F(1, 25) = 7.47, p = .01, η2p = .23) and length of use (F(1, 28) = 7.47, p = .003, η2p = .27).

Exploratory analyses examining length of HC use

Given that the worry-ERN relationship was moderated by HCs, we wanted to further explore how HC use may impact this relationship. We therefore conducted an exploratory, follow-up median split analysis by months of HC use such that HC users were split into women who used their HCs fewer than 16 months (n = 15) versus women who used their HCs 16 months and longer (n = 15). The relationships between worry and ERN and ΔERN tended to be larger in women who had been using HCs longer (longer: worry-ERN r = −.83, p < .001 vs. shorter: worry-ERN r = −.37, p = .18; longer: worry-ΔERN r = −.58, p = .02 vs. shorter: worry-ΔERN r = −.46, p = .09).

Discussion

The current study’s primary aim was to examine the role of HC use in the association between worry and error-related brain activity. We did this by examining whether HC use moderated the worry-ERN association. Consistent with our prediction, HC use moderated the worry-ERN association, such that women taking HCs demonstrated a large association between worry and the ERN. In contrast, NC women showed a much smaller and non-significant relationship. This effect was observed in the absence of any behavioral differences between groups and even when controlling for co-occurring anxious arousal symptoms, ruling out possible confounding effects of behavioral performance and related symptoms. In addition, our results did not reveal evidence that worry mediated the association between HC use and the ERN (see Figure S1). Finally, length of HC use seemed to be a pertinent factor to consider, as exploratory analyses revealed that women who were using HCs for a longer amount of time (i.e., 16 months or more) showed a greater worry-ERN association compared to those who were using HCs for shorter amounts of time.

Why would the worry-ERN association be enhanced in HC users? Some suggest that the psychobiological effects of HC use are attributable to suppression of endogenous ovarian hormone production, which may mitigate the regulatory effects of estradiol in particular (Fleischman et al., 2010; Graham & Milad, 2013; Montoya & Bos, 2017). Our finding that women taking HCs show a larger worry-ERN relationship than NC women is supported by previous research indicating that HC use exacerbates anxiety symptoms and frontal control network disruptions, including the ACC (Montoya & Bos, 2017). In addition, research has also indicated that HC use leads to differences in functional connectivity of frontal networks that include the ACC, such as the salience network (Sharma et al., 2020). Another interpretation, however, is that although HCs reduce endogenous hormones, they deliver a fixed level of exogenous hormones (Marečková et al., 2014). Thus, although levels of endogenous ovarian hormones are reduced, levels of active synthetic hormones are stable and of moderate magnitude across time. The larger worry-ERN relationship in women taking HCs observed in the current study may therefore be due to fixed delivery of moderate levels of exogenous hormones or to a combination of the two processes (i.e., suppressing endogenous while fixing exogenous). Nonetheless, these results signify that ovarian hormones may play a critical role in the association between worry and error-related brain activity.

We have proposed that the ERN is enhanced in worriers to signal increased effort due to the taxing effects of worry on cognitive control in women, specifically (Moran et al., 2012; Moser et al., 2013; Moser et al., 2016). Interestingly, while HC use moderated the association between worry and the ERN, it did not impact behavior on the task. This dovetails with research suggesting that worriers may recruit enhanced effort to perform comparably to non-worriers, and suggests that HCs may aid in that process. Furthermore, our exploratory analyses revealed that longer HC use is associated with a stronger worry-ERN association. This finding is similar to another study that found that those who started using HCs during puberty had stronger brain functional connectivity in frontal networks than those who started using HCs during adulthood (Sharma et al., 2020). Research investigating the effects of HCs on emotion, cognition, and brain structure and function is still in its infancy (Beltz & Moser, 2020; Pletzer & Kerschbaum, 2014; Taylor et al., 2021), and therefore more research in this area is needed to understand the longitudinal effects of HC use.

Finally, we also found that those using HCs reported higher levels of anxious arousal. This indicates that those using HCs may have increased levels of somatic anxiety, rather than worry. This is consistent with work indicating that the dampening of endogenous hormone levels may result in increased fear and stress (Li & Graham, 2017; Montoya & Bos, 2017). Therefore, HCs’ increase of anxious arousal may result in other related neurocognitive differences that remains a valuable area rea of future research.

The current study is not without its limitations. First, we did not directly assess levels of ovarian hormones. Although the relationship between worry and the ERN was smaller in the NC women, it was not zero. Women in the NC group were in various phases of their menstrual cycle, which would obscure any effects of hormones. We are currently conducting an intensive longitudinal study in which ovarian hormones, anxiety and error-monitoring brain activity are assessed several times throughout the menstrual cycle in NC women. Future research should also examine the worry-ERN relationship and assay endogenous hormones in women taking HCs to tease apart influences of exogenous versus endogenous hormones. While one other study has investigated the interactive effects of ovarian hormones and the ERN in OCD symptoms (Mulligan et al., 2019), no studies have investigated this in the context of worry. We were also unable to more precisely evaluate the influences of different types of HCs given that most of our participants were taking combined monophasic HCs. Future research should include a larger sample of heterogeneous HC users to better account for variation in pill constituents and therefore relative contributions of different hormone levels (see Beltz et al., 2015). Related to this, this study did not randomly assign participants to HC and non-HC use., Importantly however, work that has probed whether those using HCs vary from NC women across personality dimensions or gender self-concept have found no differences (Beltz et al., 2019; Nielson & Beltz, 2021), suggesting it may be unlikely that such factors influence HC use and the reported differences. Nonetheless, a repeated measures design that allows for assessment of change of the worry-ERN association pre and post HC use would benefit our understanding of the HC use on the worry-ERN association. Finally, larger samples will be needed to replicate the current findings.

The current findings add to a growing body of research indicating that HC use and ovarian hormone levels are important to consider in studies of anxiety and related cognitive and affective functioning (Li & Graham, 2017; Montoya & Bos, 2017). They suggest that HC use may enhance worry’s effects on cognitive control-related error monitoring. Practically, the present results indicate that HC use should be assessed in studies of the association between worry and the ERN in women. Clinically, these findings highlight the importance of assessing HC use, as clients suffering from worry-related concerns may be at demonstrate differences in cognitive function if taking HCs. Additional research will surely be needed to evaluate further how HCs and ovarian hormones impact cognitive and affective functioning in anxiety as accumulating evidence continues to point to this critical but understudied area of women’s health.

Supplementary Material

Supp.Materials

Figure 2.

Figure 2.

Open in a new tab

Line plots illustrating differential correlation between worry (PSWQ) and the ERN- between HC users and NC participants. The shaded area represents the 95% confidence interval.

Note. HC = hormonal contraceptive (HC) users; Non-HC = Non HC users

Table 2.

Worry, Hormonal Contraceptive (HC) Status and Accuracy predicting ERPs. Results for cross-classified multi-level model investigating the three-way interaction between PSWQ × HC Status × Accuracy.

Fixed Effect Estimate Standard Error t P
Intercept −.82 .53 −1.56 .12
PSWQ −.09 .04 −2.43 .01*
HC Status −.40 .50 −.79 .43
Accuracy 1.66 .33 4.99 .00*
PSWQ × HC Status .02 .03 .79 .43
PSWQ × Accuracy .06 .02 2.62 .01*
HC Status × Accuracy −.02 .33 −.07 .94
PSWQ × HC Status × Accuracy −.05 .02 −2.12 .03*
Variance Components Variance Standard Deviation Correlation --
Intercept (Participant) 16.86 4.11 --
Slope (Accuracy) 7.19 2.68 .03
Intercept (Channel) .12 .34 --
Residual 2.09 1.44 --
Open in a new tab

Note. For the presented results, accuracy and HC status are included as effects-coded predictors.

Acknowledgments

Funding Information

National Institutes of Health K12 (grant HD065879) (to J. S. M.)

Footnotes

1

In the text, we use the term “worry” to refer to a specific cognitive dimension of anxiety and “anxious arousal” to refer to somatic anxiety (Heller et al., 1997; Nitschke et al., 2001). We use the term “anxiety” as a broad term to capture both worry and anxious arousal symptomology.

2

It is important to make the distinction between sex and gender, neither of which are binary nor should be used interchangeably. In the text, we use the term “women” when referring to the literature and participants in the current study who have self-identified that way. However, it is important to note that this only represents a subset of women and is not representative of all those who identify that way.

3

A subset (29%) of women reported on in this study were also included in Lin et al. (2015), however, the Lin et al. study did not evaluate the influence of HC use

4

Due to server error, we were not able to calculate item-level data for the MASQ-AA. We predict that the reliability resembles those reported in other college samples from our lab (Schroder et al., 2013).

5

We evaluated model fit of different random effect structures using Akaike Information Criteria (AIC). Specifically, we examined (a) a cross-classified model with channel and participant as random intercepts, accuracy nested within participants as a random slope, and their intercept-slope covariance; (b) a model with participant as the random intercept, accuracy as a random slope, and their covariance, and (c) an intercept-only model with participants specified as the random intercept. The results revealed the first model was a better fit (AIC = 2999.97) than the intercept only model (AIC = 3685.78), as well as the model that only contained participant as a random intercept and accuracy as a random slope (AIC =3018.92).

6

In addition, mediation analyses were conducted to examine whether (1) worry mediated the association between the HC and the ERN or (2) the ERN mediated the association between HC use and worry. Our results revealed that the indirect paths of both models did not reach significance.

7

Moreover, we found that correlations between PSWQ scores and ERN and ΔERN remained of comparable size and significance (rs > -.44, ps < .05) in the 24 women taking monophasic HCs who comprised the majority of the HC group. Likewise, correlations between PSWQ scores and ERN and ΔERN were unaffected when removing one HC user who had been on her HC less than 1 month (rs > -.52, ps < .01).

Declaration if interest. None.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Andreano JM, Touroutoglou A, Dickerson B, & Barrett LF (2018). Hormonal cycles, brain network connectivity, and windows of vulnerability to affective disorder. Trends in Neurosciences, 41(10), 660–676. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Barth C, Villringer A, & Sacher J. (2015). Sex hormones affect neurotransmitters and shape the adult female brain during hormonal transition periods. Frontiers in Neuroscience, 9, 37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bates D, Sarkar D, Bates MD, & Matrix L. (2007). The lme4 package. R Package Version, 2(1), 74. [Google Scholar]
  4. Baxter AJ, Scott KM, Vos T, & Whiteford HA (2013). Global prevalence of anxiety disorders: A systematic review and meta-regression. Psychological Medicine, 43(5), 897. [DOI] [PubMed] [Google Scholar]
  5. Baxter AJ, Vos T, Scott KM, Ferrari AJ, & Whiteford HA (2014). The global burden of anxiety disorders in 2010. Psychological Medicine, 44(11), 2363–2374. [DOI] [PubMed] [Google Scholar]
  6. Beltz AM, Loviska AM, & Kelly D. (2019). No personality differences between oral contraceptive users and naturally cycling women: Implications for research on sex hormones. Psychoneuroendocrinology, 100, 127–130. 10.1016/j.psyneuen.2018.09.034 [DOI] [PubMed] [Google Scholar]
  7. Beltz AM, & Moser JS (2020). Ovarian hormones: A long overlooked but critical contributor to cognitive brain structures and function. Annals of the New York Academy of Sciences, 1464(1), 156–180. [DOI] [PubMed] [Google Scholar]
  8. Cavanagh JF, & Shackman AJ (2015). Frontal midline theta reflects anxiety and cognitive control: Meta-analytic evidence. Journal of Physiology-Paris, 109(1), 3–15. 10.1016/j.jphysparis.2014.04.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cohen ER, & Taylor BN (1973). The 1973 least-squares adjustment of the fundamental constants. Journal of Physical and Chemical Reference Data, 2(4), 663–734. [Google Scholar]
  10. Daniels K, & Jones J. (2013). Contraceptive methods women have ever used: United States, 1982–2010. US Department of Health and Human Services, Centers for Disease Control and …. [Google Scholar]
  11. Deci PA, Lydiard RB, Santos AB, & Arana GW (1992). Oral contraceptives and panic disorder. Journal of Clinical Psychiatry, 53(5), 163–165. [PubMed] [Google Scholar]
  12. Engman J, Poromaa IS, Moby L, Wikström J, Fredrikson M, & Gingnell M. (2018). Hormonal cycle and contraceptive effects on amygdala and salience resting-state networks in women with previous affective side effects on the pill. Neuropsychopharmacology, 43(3), 555–563. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Eriksen BA, & Eriksen CW (1974). Effects of noise letters upon the identification of a target letter in a nonsearch task. Perception & Psychophysics, 16(1), 143–149. [Google Scholar]
  14. Eysenck MW, Derakshan N, Santos R, & Calvo MG (2007). Anxiety and cognitive performance: Attentional control theory. Emotion, 7(2), 336. [DOI] [PubMed] [Google Scholar]
  15. Fleischman DS, Navarrete CD, & Fessler DM (2010). Oral contraceptives suppress ovarian hormone production. Psychological Science, 21(5), 750–752. [DOI] [PubMed] [Google Scholar]
  16. Gehring WJ, Liu Y, Orr JM, & Carp J. (2012). The error-related negativity (ERN/Ne). [Google Scholar]
  17. Graham BM, & Milad MR (2013). Blockade of Estrogen by Hormonal Contraceptives Impairs Fear Extinction in Female Rats and Women. Biological Psychiatry, 73(4), 371–378. 10.1016/j.biopsych.2012.09.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gratton G, Coles MG, & Donchin E. (1983). A new method for off-line removal of ocular artifact. Electroencephalography and Clinical Neurophysiology, 55(4), 468–484. [DOI] [PubMed] [Google Scholar]
  19. Hampson E. (2020). A brief guide to the menstrual cycle and oral contraceptive use for researchers in behavioral endocrinology. Hormones and Behavior, 119, 104655. 10.1016/j.yhbeh.2019.104655 [DOI] [PubMed] [Google Scholar]
  20. Heller W, Nitschke JB, Etienne MA, & Miller GA (1997). Patterns of regional brain activity differentiate types of anxiety. Journal of Abnormal Psychology, 106(3), 376. [DOI] [PubMed] [Google Scholar]
  21. Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR, & Walters EE (2005). Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry, 62(6), 593–602. [DOI] [PubMed] [Google Scholar]
  22. Klump KL, Keel PK, Racine SE, Burt SA, Neale M, Sisk CL, Boker S, & Hu JY (2013). The interactive effects of estrogen and progesterone on changes in emotional eating across the menstrual cycle. Journal of Abnormal Psychology, 122(1), 131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lester NA, Keel PK, & Lipson SF (2003). Symptom fluctuation in bulimia nervosa: Relation to menstrual-cycle phase and cortisol levels. Psychological Medicine, 33(1), 51. [DOI] [PubMed] [Google Scholar]
  24. Li SH, & Graham BM (2017). Why are women so vulnerable to anxiety, trauma-related and stress-related disorders? The potential role of sex hormones. The Lancet Psychiatry, 4(1), 73–82. 10.1016/S2215-0366(16)30358-3 [DOI] [PubMed] [Google Scholar]
  25. Lin Y, Moran TP, Schroder HS, & Moser JS (2015). The role of hand of error and stimulus orientation in the relationship between worry and error-related brain activity: Implications for theory and practice. Psychophysiology, 52(10), 1281–1292. [DOI] [PubMed] [Google Scholar]
  26. Marečková K, Perrin JS, Nawaz Khan I, Lawrence C, Dickie E, McQuiggan DA, Paus T, & Consortium I. (2014). Hormonal contraceptives, menstrual cycle and brain response to faces. Social Cognitive and Affective Neuroscience, 9(2), 191–200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. McEwen BS, & Milner TA (2017). Understanding the broad influence of sex hormones and sex differences in the brain. Journal of Neuroscience Research, 95(1–2), 24–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. McLean CP, Asnaani A, Litz BT, & Hofmann SG (2011). Gender differences in anxiety disorders: Prevalence, course of illness, comorbidity and burden of illness. Journal of Psychiatric Research, 45(8), 1027–1035. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Merz CJ, Tabbert K, Schweckendiek J, Klucken T, Vaitl D, Stark R, & Wolf OT (2012). Neuronal correlates of extinction learning are modulated by sex hormones. Social Cognitive and Affective Neuroscience, 7(7), 819–830. 10.1093/scan/nsr063 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Montoya ER, & Bos PA (2017). How Oral Contraceptives Impact Social-Emotional Behavior and Brain Function. Trends in Cognitive Sciences, 21(2), 125–136. 10.1016/j.tics.2016.11.005 [DOI] [PubMed] [Google Scholar]
  31. Moran TP, Taylor D, & Moser JS (2012). Sex moderates the relationship between worry and performance monitoring brain activity in undergraduates. International Journal of Psychophysiology, 85(2), 188–194. [DOI] [PubMed] [Google Scholar]
  32. Moser J, Moran T, Schroder H, Donnellan B, & Yeung N. (2013). On the relationship between anxiety and error monitoring: A meta-analysis and conceptual framework. Frontiers in Human Neuroscience, 7, 466. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Moser JS, Moran TP, & Jendrusina AA (2012). Parsing relationships between dimensions of anxiety and action monitoring brain potentials in female undergraduates. Psychophysiology, 49(1), 3–10. [DOI] [PubMed] [Google Scholar]
  34. Moser JS, Moran TP, Kneip C, Schroder HS, & Larson MJ (2016). Sex moderates the association between symptoms of anxiety, but not obsessive compulsive disorder, and error-monitoring brain activity: A meta-analytic review. Psychophysiology, 53(1), 21–29. [DOI] [PubMed] [Google Scholar]
  35. Mosher WD, & Jones J. (2010). Use of contraception in the United States: 1982–2008. Vital and Health Statistics. Series 23, Data from the National Survey of Family Growth, 29, 1–44. [PubMed] [Google Scholar]
  36. Mulligan EM, Hajcak G, Klawohn J, Nelson B, & Meyer A. (2019). Effects of menstrual cycle phase on associations between the error-related negativity and checking symptoms in women. Psychoneuroendocrinology, 103, 233–240. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Nielson MG, & Beltz AM (2021). Oral contraceptive use is not related to gender self-concept. Psychoneuroendocrinology, 129, 105271. 10.1016/j.psyneuen.2021.105271 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Nitschke JB, Heller W, Imig JC, McDonald RP, & Miller GA (2001). Distinguishing dimensions of anxiety and depression. Cognitive Therapy and Research, 25(1), 1–22. [Google Scholar]
  39. Perrin F, Pernier J, Bertrand O, & Echallier JF (1989). Spherical splines for scalp potential and current density mapping. Electroencephalography and Clinical Neurophysiology, 72(2), 184–187. [DOI] [PubMed] [Google Scholar]
  40. Petersen N, Kilpatrick LA, Goharzad A, & Cahill L. (2014). Oral contraceptive pill use and menstrual cycle phase are associated with altered resting state functional connectivity. NeuroImage, 90, 24–32. 10.1016/j.neuroimage.2013.12.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Petitti DB (2003). Combination estrogen–progestin oral contraceptives. New England Journal of Medicine, 349(15), 1443–1450. [DOI] [PubMed] [Google Scholar]
  42. Pletzer BA, & Kerschbaum HH (2014). 50 years of hormonal contraception—Time to find out, what it does to our brain. Frontiers in Neuroscience, 8. 10.3389/fnins.2014.00256 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Pletzer B, Kronbichler M, Nuerk H-C, & Kerschbaum H. (2014). Hormonal contraceptives masculinize brain activation patterns in the absence of behavioral changes in two numerical tasks. Brain Research, 1543, 128–142. 10.1016/j.brainres.2013.11.007 [DOI] [PubMed] [Google Scholar]
  44. Porcu P, Serra M, & Concas A. (2019). The brain as a target of hormonal contraceptives: Evidence from animal studies. Frontiers in Neuroendocrinology, 55, 100799. 10.1016/j.yfrne.2019.100799 [DOI] [PubMed] [Google Scholar]
  45. Rivera R, Yacobson I, & Grimes D. (1999). The mechanism of action of hormonal contraceptives and intrauterine contraceptive devices. American Journal of Obstetrics and Gynecology, 181(5), 1263–1269. [DOI] [PubMed] [Google Scholar]
  46. Schroder HS, Moran TP, Infantolino ZP, & Moser JS (2013). The relationship between depressive symptoms and error monitoring during response switching. Cognitive, Affective, & Behavioral Neuroscience, 13(4), 790–802. [DOI] [PubMed] [Google Scholar]
  47. Seedat S, Scott KM, Angermeyer MC, Berglund P, Bromet EJ, Brugha TS, Demyttenaere K, De Girolamo G, Haro JM, & Jin R. (2009). Cross-national associations between gender and mental disorders in the World Health Organization World Mental Health Surveys. Archives of General Psychiatry, 66(7), 785–795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Shansky RM, & Lipps J. (2013). Stress-induced cognitive dysfunction: Hormone-neurotransmitter interactions in the prefrontal cortex. Frontiers in Human Neuroscience, 7, 123. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Sharma R, Fang Z, Smith A, & Ismail N. (2020). Oral contraceptive use, especially during puberty, alters resting state functional connectivity. Hormones and Behavior, 126, 104849. 10.1016/j.yhbeh.2020.104849 [DOI] [PubMed] [Google Scholar]
  50. Sharp PB, Miller GA, & Heller W. (2015). Transdiagnostic dimensions of anxiety: Neural mechanisms, executive functions, and new directions. International Journal of Psychophysiology, 98(2), 365–377. [DOI] [PubMed] [Google Scholar]
  51. Skovlund CW, Mørch LS, Kessing LV, & Lidegaard Ø (2016). Association of hormonal contraception with depression. JAMA Psychiatry, 73(11), 1154–1162. [DOI] [PubMed] [Google Scholar]
  52. Taylor CM, Pritschet L, & Jacobs EG (2021). The scientific body of knowledge – Whose body does it serve? A spotlight on oral contraceptives and women’s health factors in neuroimaging. Frontiers in Neuroendocrinology, 60, 100874. 10.1016/j.yfrne.2020.100874 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Ter Horst JP, de Kloet ER, Schächinger H, & Oitzl MS (2012). Relevance of stress and female sex hormones for emotion and cognition. Cellular and Molecular Neurobiology, 32(5), 725–735. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Van Wingen GA, Ossewaarde L, Bäckström T, Hermans EJ, & Fernández G. (2011). Gonadal hormone regulation of the emotion circuitry in humans. Neuroscience, 191, 38–45. [DOI] [PubMed] [Google Scholar]
  55. Volpert-Esmond HI, Page-Gould E, & Bartholow BD (2021). Using multilevel models for the analysis of event-related potentials. International Journal of Psychophysiology, 162, 145–156. 10.1016/j.ijpsycho.2021.02.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Watson D, & Clark LA (1991). The mood and anxiety symptom questionnaire. Unpublished Manuscript, University of Iowa, Department of Psychology, Iowa City. [Google Scholar]
  57. Yeung N, Botvinick MM, & Cohen JD (2004). The neural basis of error detection: Conflict monitoring and the error-related negativity. Psychological Review, 111(4), 931. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp.Materials
NIHMS1763766-supplement-Supp_Materials.docx (36.5KB, docx)

RESOURCES