Decades-long gender gaps in women’s and men’s drinking patterns have narrowed in the United States such that what was once a five-fold, male-to-female preponderance in alcohol use disorder (AUD) prevalence has dwindled to a less than two-fold difference (Grant et al., 2017). Some of the gender convergence in problematic drinking is being driven by young women whose past month alcohol use and reports of drunkenness have eclipsed men’s (Agabio et al., 2016, Grant et al., 2017). The increased drinking among female emerging adults is alarming, in part, because of sex differences in alcohol’s distribution and metabolism (McCaul et al., 2019). The higher blood alcohol concentrations attained per an equivalent dose of alcohol in women compared with men contribute to more negative health outcomes (Jones, 2019, Karaye et al., 2023). Thus, heavy drinking women with AUD are more prone to develop several alcohol-related diseases than their male counterparts including cirrhosis of the liver and cardiomyopathy (McCaul et al., 2019, White, 2020).

Some of alcohol’s effects on the brain may also be sex-specific. For example, in alcohol-naïve adolescents with a family history of AUD, heavy drinking is associated with lower cortical volume and thickness in boys whereas the opposite pattern is observed in girls (Tapert and Eberson-Shumate, 2022). As aging occurs, heavy drinking young adults generally exhibit greater thinning of the dorsolateral prefrontal cortices compared with light drinking controls; however, the magnitude of this effect is greater in men than women (Mann et al., 2005, Medina et al., 2008). Adult women also appear more vulnerable than men to the damaging effects of alcohol on the brain, including volumetric loss in fronto-limbic regions despite, on average, having fewer years of drinking and consuming less alcohol in their lifetime (Nixon et al., 2014).

In some instances, alcohol’s sex-specific effects may interact with brain regions exhibiting a degree of sexual dimorphism. Of relevance to the present study is alcohol’s effects on prefrontal cortical - limbic system pathways that mediate emotion regulation. In particular, the amygdala in the limbic system is a sexually dimorphic brain region that plays a key role in emotional processing (Alarcón et al., 2015). The amygdala is typically engaged when viewing emotionally salient images and/or faces, reflecting its central connectivity to visual, subcortical, and cortical regions of the brain that are needed to decode and detect emotional cues required for social interactions (Pessoa and Adolphs, 2010, Fusar-Poli et al., 2009). Structural magnetic resonance imaging (MRI) studies have found sex differences in the volume of the amygdala with men having higher volumes than women (Ruigrok et al., 2014). Functional MRI (fMRI) studies on amygdala activation during the viewing of emotional human faces generally report involvement of several additional brain regions, in addition to the amygdala, including the prefrontal cortex (PFC), anterior cingulate cortex (ACC), insula, as well as areas in the temporal, parietal, and occipital lobes (Preckel et al., 2019, Fusar-Poli et al., 2009). Regarding sex differences, men typically have greater regional activation than women during passive viewing of emotional faces (Fusar-Poli et al., 2009, Sergerie et al., 2008). Further, women generally are more accurate at recognizing and expressing affective states than men, while men react faster during facial affective decision-making (Mann et al., 2005). These findings support a need for greater understanding of the neurobiology of sex differences regarding alcohol effects in order to better elucidate potential sex differences in biological targets for prevention and treatment of AUD and other drug use disorders (Sinha et al., 2022).

The PFC-limbic pathways are particularly sensitive to intoxicating doses of alcohol (Sullivan and Pfefferbaum, 2005), potentially due to overlap with stress pathophysiology (Logrip et al., 2018, Peltier et al., 2019). Studies comparing men with women have found that sex moderates the effects of heavy drinking on brain function in PFC-limbic pathways as measured with fMRI, with men having greater regional activation than women in the medial PFC, inferior and middle frontal gyri, and caudate during an emotion processing task (Padula et al., 2015). Additionally, in an MRI resting-state functional connectivity study, men showed a stronger relationship between decreased amygdala-dorsal ACC connectivity and problem drinking compared to women (Hu et al., 2018). However, neuroimaging studies have not routinely accounted for potential sex differences (Lind et al., 2017). As such, sex differences are generally understudied and findings of sex-dependent effects on neuronal responses are inconsistent across studies. Further, even fewer studies have examined sex differences in functional connectivity between brain regions linked with the sexually dimorphic amygdala -- an ideal region in which to examine the differential effects of alcohol between men and women.

The fMRI-based emotional matching paradigms first introduced by Hariri et al. (Hariri et al., 2002, Hariri et al., 2000) are emotional reactivity paradigms that reliably activate the amygdala (Preckel et al., 2019, Foland-Ross et al., 2010). Participants are asked to match the affective state of faces standardized to portray either negatively valenced affective states, such as anger and fear, or positively valenced affective states, such as happiness. Decoding facial affect is vital when interacting with one’s social environment and successful navigation of social interactions that are important for mental and physical health. Paramount to such efforts is the ability to resolve the nature of socially ambiguous information, that is, social inputs that do not provide enough information about social actions to take such as behaviorally incongruent facial expressions.

Impairments in decoding basic and complex emotional facial expressions have been consistently reported in individuals with AUD, even during periods of lengthy abstinence (Bora and Zorlu, 2017, Le Berre, 2019, Castellano et al., 2015). As a potential forme fruste of this phenomenon, and using a modified Hariri paradigm of angry, fearful, and happy faces, we reported functional connectivity differences between low and high responders to alcohol in amygdala-based connections. A low level of response (LR) to alcohol was associated with lower functional connectivity in PFC-amygdala regions (McKenna et al., 2022). Although we did not find sex differences in the connectivity patterns of low- versus high-LR participants, we speculated that attenuated connectivity among low LR individuals when processing emotional faces may contribute to an impaired ability to recognize alcohol intoxication in social situations and appraise emotional states. While research has focused on the contributions of a cortico-limbic circuit supporting emotion processing in the emergence of AUD risk (e.g., Glahn et al., 2007), how social drinking women and men differ in their brain response while intoxicated is not known. At a time when alcohol misuse is on the rise, and binge drinking and AUD rates have substantially increased in women (Grant et al., 2017), there is a major gap in understanding the mechanisms and processes that specifically increase risks for the onset and development of AUD in women compared with men. Given that AUD is a heterogeneous disorder, elucidating sex-specific neuroimaging biomarkers underlying emotion processing could aid in developing sex-specific novel interventions.

The aim of the present study was to focus on an integrative perspective of the brain using amygdala-based functional connectivity analyses to investigate sex differences in response to alcohol with a modified Hariri paradigm in a sample of young social drinkers without AUD. This sample allowed us to examine potential sex-dependent functional connectivity changes in response to alcohol before more chronic alcohol use leads to significant neuroadaptive changes (Clapp et al., 2008) or changes neurobiology as observed during AUD (Gilpin and Koob, 2008). Once an individual has AUD it is more difficult to determine if any variable, including fMRI patterns, are caused by heavy drinking. Further, the vast majority of drinkers do not have AUD and we aimed to understand important information on sex differences among this population. Specifically, we employed generalized psychophysiological interactions (gPPI), a technique that examines the task-modulated connectivity between a seed region (e.g., the amygdala) and the whole brain (O’Reilly et al., 2012, McLaren et al., 2012). Hypothesis 1 predicts that the effect of modest alcohol intoxication would differentially impact PFC-amygdala functional connections in men and women such that women would have lower amygdala-frontal cortical connectivity values compared to men during the placebo condition. Hypothesis 2 predicts that women would have greater amygdala-frontal cortical connectivity values compared with men while under the influence of alcohol. That non-AUD women would have greater fronto-amygdalar connectivity compared to men under the influence of alcohol might indicate that women can better process affectively-laden stimuli when intoxicated.

The primary goal of this study was to expand upon our previous findings (McKenna et al., 2022) demonstrating that in young social drinkers, alcohol’s effects on amygdala-based functional circuits underlying facial affect matching varied independently with LR to alcohol. Focusing this time on potential sex differences in emotional processing, and as predicted in hypothesis 1, we found that women demonstrated lower amygdala-frontal cortical functional connectivity compared to men during the placebo condition across all three emotionally-valence conditions. We also observed a main effect of sex when processing angry and happy faces during both placebo and alcohol conditions. Thus, compared with men, women exhibited markedly decreased contralateral functional connectivity between the left amygdala seed and right middle insula for fearful faces, and right inferior frontal gyrus for happy faces. These findings are consistent with other reports that men exhibit increased bilateral amygdala activity than do women (Fusar-Poli et al., 2009, Sergerie et al., 2008), perhaps reflecting their need to work harder than women to decode emotions (Thompson and Voyer, 2014).

When interpreting these findings, it is important to consider that differences in regional activation and functional connectivity patterns depend upon the nature of the task and the novelty of the stimulus. For example, in men more than women, amygdala activation is attenuated with repeated presentations of the same stimulus, perhaps due to greater connectivity between the amygdala and medial frontal regions that inhibit amygdala activation (Andreano et al., 2014, Lungu et al., 2015). Alcohol may interfere with this process, leading to greater frontal lobe involvement to inhibit amygdala activation. Interestingly, the main effect of alcohol observed in the present study was limited to connections between the right amygdala and lateral postcenteral gyrus, as well as the contralateral insular cortex and superior frontal gyrus. These focal effects suggest that alcohol itself may not impact all amygdala-based functional networks and other functional networks should be examined in future studies. For example, among heavy drinkers, functional neuroimaging studies have found that brain areas in the salience network, including the insula and anterior cingulate cortex (ACC), or regions associated with reward processing, such as the ventral striatum, are activated by alcohol-related cues (Seo et al., 2011, Heinz et al., 2009, Wrase et al., 2007).

Consistent with hypothesis 2, we observed several sex-by-alcohol interactions in fronto-amygdalar functional connectivity. When moderately intoxicated, women exhibited increased fronto-amygdalar functional connectivity compared to men while processing negatively-valenced faces. Specifically, when processing angry faces, alcohol increased functional connectivity between the contralateral right ventral anterior cingulate and the left amygdala seed region in women, whereas in men under the influence, functional connectivity between these brain regions was decreased. A similar sex-by-alcohol interaction was observed for the right amygdala seed region and its functional connectivity with the unilateral right dorsal anterior cingulate cortex region; men’s functional connectivity decreased, while women’s functional connectivity increased, when under the influence of alcohol. Similar to the processing of angry faces, women and men exhibited differential patterns of functional connectivity when viewing fearful faces under the influence of alcohol versus placebo. Specifically, functional connectivity between the contralateral right angular gyrus and left amygdala seed region increased in women after imbibing alcohol while the opposite pattern was observed in men. This same pattern was observed in functional connectivity between the right amygdala seed region and a plethora of regions including bilateral medial and middle frontal gyri; lateral anterior and posterior cingulate cortex, caudate, superior and inferior parietal lobe; contralateral posterior cingulate, postcentral gyrus, putamen, and insular cortex.

The brain regions demonstrating a sex-by-alcohol interaction in the present study overlap with brain areas activated and reported by other investigators who used similar emotional processing paradigms. For example, others have found involvement of the ventrolateral and medial PFC, inferior frontal cortex, ACC, insula, superior temporal sulcus, temporal gyrus, superior parietal lobe, and caudate when processing facial emotions (Preckel et al., 2019, Sergerie et al., 2008, Fusar-Poli et al., 2009). Additionally, these regions, especially the PFC, ACC, and amygdala, are altered by alcohol use. For example, one study found that men with AUD had lower activation compared with men without AUD in frontal regions, whereas women with AUD had greater activation than women without AUD in superior frontal and supramarginal gyri in response to happy and aversive stimuli (Sawyer et al., 2019). Interestingly, we did not observe any sex-by-alcohol interactions while processing happy faces. While meta-analyses of fMRI studies have shown that the amygdala is involved in processing both negative and positive emotions, negative-valence emotional material appears to have particularly strong effects on amygdala reactivity (Morris et al., 1996). Engagement of amygdala-based networks in response to negative emotions may have greater salience for action, such as potential threat, to prepare an individual to react quickly (Fernandes Jr et al., 2013, Sinke et al., 2012). In support of this idea, we found the largest number of interaction effects when participants were presented with fearful faces, a negatively charged emotion that indicates potential danger. Positive emotions represent the converse, allowing psychological growth and well-being (Fredrickson, 2001). Our findings suggest that alcohol moderates underlying sex differences in functional connectivity through differential responses to negatively-valence emotions, which may be more amygdala-dependent than processing positive emotions.

The mechanisms underlying sex differences in neuronal functional networks are complex and not fully understood. Sex differences in human brain structure and function are likely due to interactions between social, genetic, and hormonal factors (Cosgrove et al., 2007, McCarthy and Arnold, 2011). Most research into sex differences has focused on sex steroid hormones on brain development. For example, convergent findings suggest that sex hormones influence the size and function of the amygdala, with prenatal androgens appearing particularly important (Beltz et al., 2023). There is also evidence that sex hormones affect gray matter volume, especially during adolescence (Bramen et al., 2012). Finally, there is evidence for both endogenous and exogenous effects of sex hormones on the adult brain, particularly with how estrogen effects the frontal lobes (Maki and Resnick, 2001).

In viewing the current results, it is important to keep some limitations in our research protocol in mind. First, participants were well educated European-American university students without an AUD, and it is not clear whether the current results generalize to other groups. Second, reflecting the fact that laterality findings are unclear in many fMRI studies of emotion processing (Davidson and Irwin, 1999), we did not have formal hypotheses on laterality and analyzed the combination of both the left and right amygdala seeds and their respective connectivity patterns. Thus, future work will need to further evaluate the importance of laterality influencing connectivity patterns associated with sex. Third, to better simulate the way that people consume alcohol in real-life situations, including oral stimulation and the relatively slow rise in BrAC values, our paradigm used oral alcohol as opposed to intravenous (IV) alcohol administration. Oral dosing is a common, affordable, and portable method and is useful when subjective experiences of alcohol are part of the research question, but as compared to IV administration, the oral approach might produce greater inter-individual variability in BrAC values (Cyders et al., 2020). However, as demonstrated in data presented by Paulus et al., (2012) and others (e.g., Schuckit, 2018) when the alcohol administration paradigm controls for the rate of drinking, percent body water, as well as recent drinking and drug use, blood alcohol levels within males and within females are remarkably similar across participants. In addition, when the current analyses examined the effect of BrAC at 60 minutes (the timepoint closest to when the Hariri task was given) by covarying for BrAC at 60 minutes in our analyses no effect on functional connectivity patterns were observed. Nonetheless, MRI studies using IV alcohol dosing methods would be powerful to control the alcohol concentrations in the brain at a steady state measure changes in fMRI patterns. With regards to our analytic plan, Woo et al. (2014) and others have suggested that a more stringent voxel-level primary threshold of p = 0.001 in cluster threshold approaches. This would reduce the possibility of large activation clusters and improve confidence in inferences about specific locations. However, importantly, the primary threshold level should be chosen a priori to reduce potential bias towards findings in brain regions investigators want to find. We sought to control the family-wise error rate at 0.01 consistent with a large body of neuroimaging studies on alcohol effects and power analyses based on that literature in our original analytic plan from which these data are derived in an a priori manner. As demonstrated in our Results and Figures, we observed medium to large effects sizes using this approach suggesting meaningful findings in anatomically constrained regions of interest. Nonetheless, future replication studies that use more stringent thresholds to better elucidate functional networks that differ between sexes in response to alcohol are needed. Lastly, there are physiological effects when consuming alcohol that directly impact the BOLD signal, such as cerebral blood flow changes (Courtney et al., 2019). With advances in neuroimaging methodologies, studies are being done isolating mechanisms that drive BOLD signal changes observed in the literature including blood flow, volume, and oxygenation. Our analyses did not covary these physiological effects and included multiple comparisons across hemispheres. Replication studies integrating multi-modal imaging with mode advance echo-planer collection techniques are needed to confirm our findings and expand upon them to better understand whether brain regions demonstrating sex differences are predictive of future alcohol problems and onset of AUD. Finally, we used shapes as our control condition, as opposed to neutral faces which may offer additional important information on emotion processing.

In conclusion, findings suggest that sex-specific brain pathways differentially modulate emotional processing responses in women and men. Women exhibited higher functional connectivity than men in cortico-amygdalar circuits subserving emotion processing following a moderate dose of alcohol. That women had higher functional connectivity than men following exposure to a moderate dose of alcohol might indicate women are more capable than men to synchronize signaling across these specific brain regions to process affective stimuli when intoxicated. Changes in the evaluation of emotional, especially aversive, stimuli, play a crucial role in the transition to AUD or alcohol relapse (Maleki and Oscar-Berman, 2019; Witkiewitz et al., 2015). Future prospective research is needed to understand if sex differences in emotional processing among social drinkers predicts subsequent problematic drinking behavior and onset of AUD in the future, as opposed to sex differences that develop as a result of chronic heavy drinking.