Trust is one of the most prominent facets of human relationships on a social and institutional scale for economic and social advancement (Bjørnskov, 2017; Fehr, 2009; Nannestad, 2008; Zak & Knack, 2001). Although sex differences in trust behavior exist (J. Byrnes, D. Miller, & W. Schafer, 1999; Van Den Akker, Van Assen, Van Vugt, & Wicherts, 2020) and the neuropsychological mechanisms of trust are known (Baumgartner, Heinrichs, Vonlanthen, Fischbacher, & Fehr, 2008; Bellucci, Chernyak, Goodyear, Eickhoff, & Krueger, 2017; Krueger & Meyer-Lindenberg, 2019), the underlying anatomical structure of sex differences is still obscure. Trust is a mental state that represents a social dilemma consisting of the intention of one person (i.e., trustor) to accept vulnerability and having positive expectations about the preferences of another person (i.e., trustee) to behave reciprocally (Krueger & Meyer-Lindenberg, 2019; Rousseau, Sitkin, Burt, & Camerer, 1998). Sex differences in trust have been an area of broad interest, and various theories have been proposed to understand those differences.

According to parental investment theory (Trivers, 1972), derived from Darwin's sexual selection theory (Darwin, 1871), sex differences in social behavior may be driven by an evolutionary asymmetry in the costs of parental investment of men and women in producing and raising offspring (Trivers, 1972). On the one hand, women must spend a large amount of energy and time raising a limited number of children during their reproductive lifecycle and, therefore, must be selective in choosing a mate since the fitness of the child is influenced by the quality of the father (Dreber & Hoffman, 2010). On the other hand, although men's investment requires a contribution of their sperm minimally, men must engage in intense competition to attract potential mating partners due to the higher selectivity pressure of women (Buss & Schmitt, 2017).

Higher risk-taking in men than in women may result from parental investment theory (Apicella, Demiral, & Mollerstrom, 2017; J. P. Byrnes, D. C. Miller, & W. D. Schafer, 1999; Fischer & Hills, 2012). Whereas women avoid excessive physical and social risks to maximize their reproductive potential, men, in contrast, engage in those risky activities to achieve more resources such as goods and money or a higher social status—often seen by women as an indicator of a man's high-quality genes and potential aid in raising a child (Wilson & Daly, 1985). A plethora of evidence exists that men take more social risks compared to women across different ages, cultures, and societies (Apicella, Crittenden, & Tobolsky, 2017; J. P. Byrnes, Miller, & Schafer, 1999; Cross, Cyrenne, & Brown, 2013; Eckel & Grossman, 2008). Those examples of differences in risk-taking might also explain sex differences in trust behavior since trust, defined as social risk, involves a willingness to be vulnerable to a stranger's potential adverse behavior (Croson & Gneezy, 2009; Fetchenhauer & Dunning, 2012; Molm, Takahashi, & Peterson, 2000; Thielmann, Spadaro, & Balliet, 2020). As a result, men exhibit higher risk-taking that leads to more trusting behavior (Apicella, Demiral, & Mollerstrom, 2017; Buser, van den Assem, & van Dolder, 2023; Chaudhuri & Gangadharan, 2007; Fischer & Hills, 2012).

In addition to evolutionary explanations, sociocultural theories assume that men and women internalize cultural expectations about how they should behave due to their traditional sex roles determined by parental investment theory (L. M. Dinella, M. Fulcher, & E. S. Weisgram, 2014; Dulin, 2007; Eagly, 1997; Wood & Eagly, 2012). Social role theory, for example, stresses the expectation of the fulfillment of social norms, demanding that women behave in a more nurturing, communal, and caring way and men in a more risky, competitive, and self-confident way (Diekman & Eagly, 2008; Dulin, 2007; Eagly, 1997; Wood & Eagly, 2012). The proposed sex differences in trusting behavior based on parental investment theory (Trivers, 1972; Darwin, 1871) and social role theory (Wood & Eagly, 2012) can be tested using economic exchange games.

The two-stage sequential trust game (TG) (also widely known as the investment game) is a well-established experimental paradigm for examining interpersonal trust behavior (Berg, Dickhaut, & McCabe, 1995). The TG involves two anonymous players—player 1 (trustor) and player 2 (trustee)—who receive an endowment (e.g., $10) at the start of the game. First, the trustor decides whether to keep the endowment or give any amount to the trustee, which the experimenter triples and sends to the trustee. Then, the trustee returns any amount of the received money to the trustor. A trustor who expects the trustee to be selfish should send nothing to the trustee because of the realization (via backward induction) that the trustee should have no incentives to send anything back (Berg et al., 1995; Johnson & Mislin, 2011). However, based on a previous meta-analysis performed on 162 replications of the standard TG involving more than 23,000 participants, trustors sent in an average of about 50 percent of the initial endowment to trustees (Johnson & Mislin, 2011)—indicating that trustors are willing to start a social interaction with strangers by trusting them. The TG can be played either in a one- or multi-shot format. The one-shot TG measures trust propensity (TP), defined as the general willingness to trust strangers. In contrast, the multi-shot TG measures the dynamic of a developing trust relationship between the trustor and trustee (Mayer, Davis, & Schoorman, 1995). Looking at sex differences, a recent meta-analysis on the standard one-shot TG (including 77 behavioral studies, 174 effect sizes, and 17,082 participants from 23 countries) revealed an overall sex effect (g = .22) in which men are more trusting in strangers than women (Van Den Akker et al., 2020). These behavioral variations might be driven by inherent anatomical structural differences between men and women related to their TP.

Although the behavioral aspects of TP have been widely studied using the one-shot TG in behavioral economics and social psychology (Van den Bos, Van Dijk, Westenberg, Rombouts, & Crone, 2009; Yamagishi, Li, Takagishi, Matsumoto, & Kiyonari, 2014), social neuroscience has only recently started to investigate its underlying anatomical structure by implementing voxel-based morphometry (VBM). The first VBM study investigated the association between individual differences in the tendency to trust in a sample of healthy participants as measured by a self-report questionnaire and regional gray matter volume (GMV) (Haas, Ishak, Anderson, & Filkowski, 2015). The tendency to trust is associated with an increased GMV in the dorsomedial prefrontal cortex (dmPFC) extending into the ventromedial PFC and lateral orbitofrontal gyrus, inferior frontal gyrus (IFG), inferior temporal gyrus (ITG), precentral gyrus (PrCG), superior frontal gyrus (SFG), anterior insula, and thalamus.

The second VBM study applied a prediction framework implementing machine learning in two independent groups of healthy participants to investigate the association between individual TP differences (as measured with two versions of the TG) and multimodal measures of the brain as collected from structural magnetic resonance imaging (sMRI, measuring GMV) and resting-state functional MRI (rsfMRI, measuring resting-state functional connectivity, rsFC) (Feng et al., 2021). The multivariate prediction analyses demonstrated that individual differences of TP for the first sample playing the standard trust game (internal validation) are predicted by GMV across multiple regions, including superior temporal gyrus (STG), supramarginal gyrus (SMG), superior parietal lobule (SPL), PrCG, postcentral gyrus (PoCG), SFG, IFG, middle frontal gyrus (MFG), precuneus (PreC), and middle occipital gyrus (MOG). Further, the GMV of those regions allowed the classification of individuals from an independent sample with the propensity to trust or distrust as measured with the binary trust game (i.e., external validation). Finally, modular and functional decoding analyses uncovered that the predicted regions are part of three identified brain modules, for which the psychological functions have been linked with domain-general large-scale brain networks: default-mode network (DMN: SFG, STG, PreC), central-executive network (CEN: SMG, IFG, MFG, PrCG) and action-perception network (APN: SPL, MOG, PoCG, PrcG) (Feng et al., 2021).

These findings go together with a neuropsychoeconomic model of trust, assuming that trust builds on the interchange of psychological components that employ domain-general large-scale brain networks (Krueger & Meyer-Lindenberg, 2019). The one-shot trust game measuring an individual's TP represents a social dilemma in which the risk of betrayal contrasted with the anticipation of reward creates uncertainty. To convert the risk of betrayal into positive expectations of reciprocity, the CEN can be utilized to implement a calculus-based trust strategy, employing the APN to perform a cost–benefit calculations. At the same time, the DMN simulates the anonymous partner's trustworthiness.

However, a gap in knowledge exists regarding the underlying anatomical structure of sex differences in TP. Although the first VBM study explored how the tendency to trust is reflected in the human brain structure; however, it only used a self-report measure (i.e., determining a trustor's expectation about the partner's trustworthiness) but not a behavioral measure (i.e., determining a trustor's vulnerability due to betrayal and expectation about the partner's reciprocity) of trust and did not investigate sex differences in TP (Haas et al., 2015). Whereas the second study explored the association between TP and GMV measuring trust behavior with different versions of the TG, it also did not study the structural differences in sex differences of TP (Feng et al., 2021).

The primary objective of this study was to explore the underlying anatomical basis of sex differences in trust behavior towards strangers (referred to as TP) by integrating the one-shot TG (for measuring TP) and VBM (for measuring GMV). At the behavioral level, we employed inferential statistics on the TG data to identify sex differences in TP. Simultaneously, at the neural level, we employed a Region of Interest (ROI) analysis as our primary approach and a Whole-Bran (WB) analysis as our exploratory approach to identify GMV patterns associated with sex differences in TP. We tested two hypotheses about sex differences in TP based on the reviewed mechanisms proposed by parental investment theory (Darwin, 1871; Robert L Trivers, 2017; R. L. Trivers & Willard, 1973) and social role theory (Wood & Eagly, 2012) and the anatomical structure of TP as identified by a recent VBM study utilizing a behavioral measure of trust (Feng et al., 2021) and interpreted based on a recently proposed neuropsychoeconomic model of trust (Krueger & Meyer-Lindenberg, 2019). At the behavioral level, we hypothesized that men demonstrate a higher TP than women in the risky first move of the TG because men are willing to take more social risks on average than women to acquire resources while interacting with others (Darwin, 1871; Robert L Trivers, 2017; R. L. Trivers & Willard, 1973). At the neural level, we predicted that men have greater GMV compared to women in regions of the DMN (simulating the trustworthiness of the anonymous partner), CEN (implementing a calculus-based trust strategy), and APN (performing cost–benefit calculations), because those networks have been recently implicated in the neuropsychological mechanisms of trust (Krueger & Meyer-Lindenberg, 2019) men show more riskier, competitively, and self-confident behavior compared to women in acquiring resources during social interactions (Darwin, 1871; Eagly, 1997; Robert L Trivers, 2017; R. L. Trivers & Willard, 1973; Van Den Akker et al., 2020).

We report how we determined our sample size, all data exclusions, all inclusion/exclusion criteria, whether inclusion/exclusion criteria were established prior to data analysis, all manipulations, and all measures in the study.