Kraepelin (1919) noted that there is a sex/gender difference in schizophrenia and wrote: “The male sex appears in general to suffer somewhat more frequently from the dementia praecox than the females.” For the past decade, “sex” has been used to refer to biological variables, and “gender” has been used for psychosocial variables (Mendrek and Mancini-Marie, 2016). In recent years, considerable attention has been given to the treatment of social cognition in schizophrenia, including social cognitive training programs and pharmacotherapies, which take into account biological aspects (Javed and Charles, 2018). We use the word “sex” instead of “gender” to indicate that we are discussing the biological aspect in schizophrenia through an assessment of social cognition.

Social cognition is defined as “the mental operations that underlie social interactions, including perceiving, interpreting, and generating responses to the intentions, dispositions, and behaviors of others” (Green et al., 2008). This multidimensional construct is composed of the subdomains of emotional processing, theory of mind (ToM), social perception, and attributional bias (Pinkham et al., 2014). In patients with schizophrenia, impairment of social cognition is considered one of the greatest obstacles to social participation, including that involving interpersonal relationships, education, and employment (Green et al., 2012; Kubota et al., 2021). Over the past decade, considerable attention has been given to the development of new therapies targeting social cognitive deficits as a means to improve functional outcomes in schizophrenia (Green et al., 2019). As some patients in the early course schizophrenia have been reported to respond to both pharmacological and psychosocial treatments (Robinson et al., 1999; Catalan et al., 2018), the optimal time to intervene to prevent future decline might be early in its course (Crumlish et al., 2009; Catalan et al., 2018).

Sex differences have been shown to affect not only functional outcomes (Seeman, 2019), but also social cognition in schizophrenia, and therefore any intervention in early course schizophrenia to improve social cognition should account for these sex differences (Chen et al., 2021). Research on sex differences in social cognition has been conducted to improve functional outcomes. According to previous studies (Buck et al., 2016; Roberts et al., 2016), impairment of social cognition in schizophrenia has a two-factor structure, social cognitive skills including ToM and attributional bias including hostility bias (Buck et al., 2016; Roberts et al., 2016). In terms of sex differences in ToM, there are mixed results indicating either no sex differences or better ToM in females (Abu-Akel and Bo, 2013; Ayesa-Arriola et al., 2014; Navarra-Ventura et al., 2018, 2021; Riecher-Rossler et al., 2018). These conflicting results might be due to differences in disease duration between the studies. A meta-analysis has not yet been carried out, and no conclusion about sex differences in ToM has been reached. To our knowledge, hostility bias has not yet been investigated separately for females and males, so there is a need to investigate potential sex differences.

Treatment of social cognitive functioning should take into account neurocognitive functioning (Deckler et al., 2018; Thibaudeau et al., 2020), and combining the treatment of neurocognitive impairment and social cognitive impairment may improve functional outcomes (Deckler et al., 2018). A number of studies have examined the association between ToM and neurocognition, with a meta-analysis showing that neurocognition such as attention, working memory, episodic memory, speed of processing, language, visuospatial/problem solving and executive functions influence ToM to the same extent, but with no specific neurocognitive area having a pronounced effect on ToM (Thibaudeau et al., 2020). It is possible that several areas influence ToM to the same extent, because that meta-analysis did not account for sex differences. Given that sex differences affect both social cognition and neurocognition (Longenecker et al., 2010; Chen et al., 2021), sex differences need to be taken into account when examining the association between ToM and neurocognition. However, to our knowledge, no studies have done this to date. As for hostility bias, although studies have examined its association with neurocognition (Donohoe et al., 2008; Buck et al., 2017, 2020; Lo and Siu, 2017), the analyses were not conducted separately for female and male subjects. Therefore, the association between hostility bias and neurocognition needs to be examined by sex.

In efforts to facilitate more individualized assessment and treatment of social cognition in patients with early course schizophrenia, this study sought to clarify the association between neurocognitive and social cognitive subdomains in early course schizophrenia by sex. Specifically, first we examined sex differences in ToM and hostility bias in early course schizophrenia and then investigated the contribution of neurocognition to ToM and hostility bias by sex.

Our examination of sex differences in the social cognitive subdomains of ToM and hostility bias revealed no such differences in patients with early course schizophrenia. Previous studies on sex differences in ToM in schizophrenia reported either no sex differences in ToM (Ayesa-Arriola et al., 2014; Navarra-Ventura et al., 2018, 2021; Riecher-Rossler et al., 2018) or better ToM in females (Abu-Akel and Bo, 2013). This inconsistency in findings might be related to differences in disease duration between the studies. To address this, in the present study, we examined patients with schizophrenia within 5 years of onset. Our study also had no significant sex differences in sociodemographic variables and thus might provide more accurate results for examining sex differences. Although no previous studies have compared sex differences in hostility bias, one study examining whether sex affects hostility bias (Pinkham et al., 2017) had a finding consistent with ours, namely, that sex does not affect hostility bias. We found that there were no sex differences in ToM and hostility bias. In contrast, healthy individuals in previous studies have shown sex differences. These differences can be explained by the effect of estrogen. In the healthy individuals, females have higher ToM (Baron-Cohen et al., 2005; Navarra-Ventura et al., 2021) and lower hostility bias (Baron-Cohen, 2004; Combs et al., 2007; Runions and Keating, 2007). Lower estrogen is associated with lower ToM (Oldershaw et al., 2010; Abu-Akel and Bo, 2013) and higher Hostility (Riecher-Rossler and Kulkarni, 2011). Gonadal hormone levels including estrogen is known to be lower in schizophrenia in 92% of females and 27.7% of males (Howes et al., 2007). Since the decrease in estrogen is greater in females, it is possible that the sex difference observed in the healthy individuals are not present in schizophrenia. Therefore, it was suggested that treatment with estrogen may be effective against ToM and hostility bias.

Interestingly, we found that verbal fluency contributed significantly to ToM in females (p < 0.01), whereas executive function contributed significantly to ToM in males (p < 0.05). This is the first study to examine the association between ToM and neurocognition in schizophrenia according to sex. To our knowledge, only one study has examined the association between ToM and brain function in schizophrenia by sex (Walsh-Messinger et al., 2019). ToM in females with schizophrenia was found to be associated with the limbic system, whereas ToM in males with schizophrenia was associated with the frontal lobe. In people with schizophrenia and healthy controls, previous studies have shown an association between verbal fluency and the limbic system and between executive function and the frontal lobe (Pihlajamaki et al., 2000; Zhu et al., 2010; Morimoto et al., 2018). Together, the results of our study and previous studies indicate that limbic-based verbal fluency is an important predictor of ToM in females and frontal-based executive function in males. These implications are useful in planning treatment for cognitive rehabilitation. Cognitive rehabilitation aimed at improving ToM should account for deficits in executive function in males and verbal fluency in females. In some cases, it might be helpful to provide treatment for executive function and verbal fluency skills before treatment for ToM.

For hostility bias, higher verbal fluency was significantly associated with lower hostility bias in females (p < 0.05), whereas none of the neurocognitive subdomains was significantly associated with hostility bias in males, although verbal fluency may have a marked contribution (p = 0.068). No sex differences were found in the association between neurocognition and hostility bias. As far as we know, no previous studies have examined the association between neurocognition and hostility bias in schizophrenia in terms of sex differences, although some studies have consistently reported that the neurocognitive subdomains that contribute to attributional bias are verbal-related ones, such as verbal fluency (Lo and Siu, 2017) and verbal IQ (Donohoe et al., 2008; Buck et al., 2017). Our results also showed that verbal-related neurocognitive subdomains tended to be associated with hostility bias. Previous studies examining the neural basis of hostility in schizophrenia have pointed out an association between hostility and the limbic system (Hoptman et al., 2010; Del Bene et al., 2016; Perlini et al., 2018). As mentioned above, an association also exists between verbal fluency and the limbic system. In fact, hostility and verbal fluency in schizophrenia have a common neural basis, which is in line with our results. Before conducting this study, we expected to find sex differences in the association between neurocognitive subdomains and hostility bias, but we found no such differences. Therefore, female and male patients alike might benefit from similar cognitive rehabilitation strategies to treat hostility bias. In addition to the lack of sex differences in ToM and hostility bias observed in schizophrenia as mentioned above, the associations between neurocognition and ToM and neurocognition and hostility bias also suggest that treatment with estrogen may benefit ToM and hostility bias in both females and males. Previous studies have found that estrogen modifies frontal lobe and limbic system activity during emotion perception tasks, a subdomain of social cognition (Ji et al., 2016; Javed and Charles, 2018). Estrogen is also associated with frontal-based executive function and limbic-based verbal fluency (Ko et al., 2006). Estrogen not only improves frontal-based executive function and limbic-based verbal fluency, but may also be effective in treating ToM and hostility bias, as well as emotion perception. Combining estrogen with neurocognitive and social cognitive rehabilitation may facilitate the treatment of ToM and hostility bias. Further studies focusing on sex differences in hormones might be worthwhile to improve the rehabilitation on social cognition (Sumiyoshi et al., 1997; Gonzalez-Rodriguez et al., 2020). Furthermore, female patients are said to have better prognosis. Our hope is that studying sex differences will uncover critical elements of good outcome thus leading to interventions that will benefit both females and males (Seeman, 2019). By considering associations with cognitive function, it may be possible to provide interventions that take account of sex differences in prognosis.

The limitations of this study are as follows. First, as this was not a longitudinal study, it cannot establish causality. Second, antipsychotic medications that our patients were taking might affect cognitive function; however, the recent literature generally does not report any significant effects of antipsychotics on social cognition (Hempel et al., 2010; Catalan et al., 2018). Third, we assessed ToM and hostility bias using the SCSQ, but it would have been beneficial to have examined other subdomains of social cognition as well. Fourth, in this study, the social cognitive assessment relies on SCSQ which is used as a screening tool. However, ToM and hostility bias domains scores of SCSQ were well-correlated with those of standard measures such as the Hinting task and AIHQ (r = 0.52, P = 0.0001 for ToM and r = 0.34, P = 0.05 for hostility bias) (Kanie et al., 2014). We therefore do not believe this undermines the validity of our results. To date, there is no consensus on what kind of scale is appropriate for measuring social cognitive function among the Japanese, and we are currently conducting research to clarify this (Kubota et al., 2021; Okano et al., 2021). We believe it is desirable to use tests that have reached consensus on this. Lastly, this study did not have a healthy control group to compare our results with. Therefore, it is difficult to state whether these results are specific to schizophrenia or reflect the sex differences found in the general population. In order to link our findings to rehabilitation, the differences need to be fully clarified.

In conclusion, we found sex differences in the associations between neurocognition and ToM and neurocognition and hostility bias in early course schizophrenia, despite the absence of sex differences in ToM and hostility bias. These results suggest that sex differences in neurocognition related to social cognition need to be taken into account to treat social cognition more effectively. These results can be adapted to patients with early course schizophrenia and contribute to the promotion of future treatments targeting social cognition.

The datasets presented in this article are not readily available because participants of this study did not agree for their data to be shared publicly. Requests to access the datasets should be directed to RK, ryo-okubo@ncnp.go.jp.

The studies involving human participants were reviewed and approved by the Institutional Review Board of the National Center of Neurology and Psychiatry. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

RK, RO, SI, and NY contributed to writing—original draft preparation. RO, SI, and NY contributed to supervision. All authors contributed to conceptualization and writing—review and editing, read and agreed to the published version of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.