Pattern recognition reveals sex-dependent neural substrates of sexual perception. Vesa Putkinen, Sanaz Nazari-Farsani, Tomi Karjalainen, Severi Santavirta, Matthew Hudson, Kerttu Seppälä, Lihua Sun, Henry K. Karlsson, Jussi Hirvonen, Lauri Nummenmaa. Human Brain Mapping, February 11 2023. https://doi.org/10.1002/hbm.26229
Abstract: Sex differences in brain activity evoked by sexual stimuli remain elusive despite robust evidence for stronger enjoyment of and interest toward sexual stimuli in men than in women. To test whether visual sexual stimuli evoke different brain activity patterns in men and women, we measured hemodynamic brain activity induced by visual sexual stimuli in two experiments with 91 subjects (46 males). In one experiment, the subjects viewed sexual and nonsexual film clips, and dynamic annotations for nudity in the clips were used to predict hemodynamic activity. In the second experiment, the subjects viewed sexual and nonsexual pictures in an event-related design. Men showed stronger activation than women in the visual and prefrontal cortices and dorsal attention network in both experiments. Furthermore, using multivariate pattern classification we could accurately predict the sex of the subject on the basis of the brain activity elicited by the sexual stimuli. The classification generalized across the experiments indicating that the sex differences were task-independent. Eye tracking data obtained from an independent sample of subjects (N = 110) showed that men looked longer than women at the chest area of the nude female actors in the film clips. These results indicate that visual sexual stimuli evoke discernible brain activity patterns in men and women which may reflect stronger attentional engagement with sexual stimuli in men.
5 DISCUSSION
Our main finding was that sexual stimuli elicit discernible patterns of brain activation in men and women. The GLM analysis revealed that sexual movie clips and pictures elicited widespread activation across the brain in both sexes: Activations were observed in regions associated with reward and emotion (e.g., brainstem, basal ganglia, thalamus, ACC, amygdala, and medial prefrontal cortex) and in somatosensory and motor cortices (pre- and postcentral gyrus, SMA) implicated in sexual arousal (Georgiadis & Kringelbach, 2012). Activations were also observed in visual regions in the occipital and inferior temporal cortices and in the dorsal attention network (frontal eye fields, FEF and intraparietal sulcus, IPS). Men showed stronger responses than women particularly in visual regions in the occipital cortex and fusiform gyri, in the dorsal attention network as well as in various prefrontal regions. Notably, using multivariate pattern classification we were able to accurately predict the sex of the individual subjects. The classifier generalized across the movie and picture experiments, underlining the consistency of the sex-specific response patterns. These results indicate that, although visual sexual stimuli engage similar networks in men and women, brain activity patterns induced by such stimuli are different across sexes.
5.1 Brain activity patterns induced by visual sexual stimuli predict individuals' sex
Using multivariate pattern classification of the brain responses to the sexual movies and pictures, we were able to accurately classify the subjects as men or women. This indicates that sex differences in the brain responses to sexual signals are robust enough to differentiate men and women at the individual subject level. To our knowledge, there are no previous studies employing sex classification with visual sexual stimuli, but the classification accuracies achieved in the current study are comparable with those obtained in sex classification with resting-state fMRI (Satterthwaite et al., 2015; Weis et al., 2020; Zhang et al., 2018; Zhang et al., 2020) and functional connectivity during a semantic decision task (Xu et al., 2020). The accuracy was better in the movie (76%) than in the picture experiment (66%). This likely reflects the fact that audiovisual movies are more representative of the natural sociosexual environment, and consequently activate the brain more strongly and consistently than still photos (Hasson et al., 2010). Importantly, above chance level classification was achieved even with cross-classification where the classifier was trained on the data from one experiment and tested on data from the other indicating that the male/female-typical brain activation patterns evoked by sexual signals were consistent across the experiments. This indicates that the sex-specific brain responses reflected the processing of the sexual content shared across the dynamic videos and still pictures.
Virtually the same regions showing sex differences in the GLM analysis (see below) also contributed the most to the classification with SVM in both experiments as indexed by the high correlation between the SVM weights and the beta values for the sex difference 2nd level contrasts (r = .7 for the movie experiment and r = .6 for picture experiment). Namely, occipital cortex and fusiform gyrus and frontal regions showed strong weights indicative of male category while temporal regions showed the strong voxel weights indicative of female category (see Figure S4). Interestingly, above chance level sex classification was obtained even with the responses to the control dimension. However, the cross-classification between the responses to the control dimension and the sexual stimuli was at chance level indicating that different activity patterns contributed to the sex classification for sexual versus nonsexual stimuli. In line with this, the spatial distribution of the SVM weight across the brain for the sexual stimuli and the control dimension were dissimilar as illustrated by the low correlation between the voxel weight maps for the sexual stimuli and the control dimension (r = −.17 for the movie experiment and r = .07 for the pictures experiment).
Although the brain activity patterns evoked by visual sexual stimuli were predictive of subject sex, some subjects were misclassified demonstrating that these brain activity patterns were not fully sexually dimorphic (compare Joel & Fausto-Sterling, 2016). Classification accuracy is partly determined by methodological factors such as the training set sample size (Balki et al., 2019) but some previous studies suggest that sex misclassification may also reflect a characteristic cognitive or affective profile (Satterthwaite et al., 2015; Zhang et al., 2021). In the current study, the misclassified men reported lower negative emotions toward pornography than the correctly classified men, but no other differences were found between the correctly and incorrectly classified subjects (see Data S1). Interestingly, those subjects who identified themselves as bisexual were no more likely to be misclassified than those who identified themselves heterosexual suggesting that brain activity patterns evoked by sexual stimuli are not dependent on bisexual vs. heterosexual orientation. It is possible that factors such as sexual history or more nuanced sexual preferences may contribute to the misclassification.
5.2 Sex-dependent activation of visual and attentional circuits
Sexual stimuli activated occipitotemporal visual regions consistently in both experiments, suggesting attentional modulation of visual cortical activity for sexually salient stimuli. Event-related potential studies indicate that human bodies with visible (vs. hidden) sexual signals induce amplified temporocortical responses already <200 ms from stimulus onset demonstrating facilitated processing of visual sexual cues early in visual processing stream (Alho et al., 2015; Hietanen et al., 2014; Hietanen & Nummenmaa, 2011). Occipital activity was particularly strong in the putative body sensitive regions (the “extrastriate body area”) in the lateral occipital cortex (Downing et al., 2001) suggesting amplified processing of sexual information in the human body recognition systems (e.g., Ponseti et al., 2006). Men showed stronger activity than women in a cluster extended from V2 along the fusiform gyrus (compare Sabatinelli et al., 2004; Sylva et al., 2013; Wehrum et al., 2013) which suggests that sexual stimuli trigger stronger attentional amplification of visual cortical activity in men than in women. Heightened attention toward visual sexual cues facilitates sexual arousal (Dawson & Chivers, 2016) which may explain previous findings that the activation of occipitotemporal visual regions is positively associated with measures of penile erection and subjective sexual arousal (Arnow et al., 2002; Moulier et al., 2006).
Eye-tracking studies indicate that men show an attentional bias toward the explicitly sexual aspects of visual sexual stimuli (Nummenmaa et al., 2012; Rupp & Wallen, 2007). Our control experiment with eye tracking revealed that men looked longer at the chest area of the nude female actors in the movie clips than women did (approximately 7% vs. 5% of the video duration in men vs. women, respectively). Women, in turn, tended to look at the male actors faces slightly longer than men did. These subtle sex differences in the locus of attention may partly account for the sex differences in brain activation in the visual cortices (compare Dolcos et al., 2020; Ferri et al., 2013). Both the sexual videos and pictures activated intraparietal sulcus (IPS) and frontal eye fields (FEF) which are central nodes in the dorsal attention network supporting controlled, top-down attention (Corbetta & Shulman, 2002). In both experiments, men showed stronger activation in a parietal cluster that extended to the IPS as well as in middle frontal gyrus/precentral gyrus extending to the FEF suggesting that visual sexual stimuli engage dorsal attention network more strongly in men that in women. Interestingly, the eye tracking data also revealed that faces were the most attended regions in the sexual scenes. This accords with the well-known attentional bias towards faces and suggests that during sexual contact the partner's face conveys important information regarding enjoyment and sexual arousal thus warranting preferetntial attention over genitals and other erogenous zones.
Women showed stronger activation than men only in the movie experiment in auditory cortical regions. This result suggests that women responded more strongly to the audio track in the sexual video clips which consisted mostly of nonverbal female vocalizations communicating sexual pleasure. A number of studies have shown that affective vocalizations, including sexual ones (Fecteau et al., 2007), elicit stronger auditory cortical activity than neutral voices (Frühholz et al., 2016). Behavioral studies suggest a slight female advantage in emotion recognition from nonverbal emotional vocalizations (Thompson & Voyer, 2014) but sex differences in affective sound processing have not been studied extensively with neuroimaging (however, see Ethofer et al., 2007). Our results tentatively suggest that women respond stronger to nonverbal sexual vocalizations and thereby that the stronger male reactivity to sexual cues might be specific to visual domain. However, as attention toward visual stimuli attenuates auditory cortical activity (Johnson & Zatorre, 2006; Molloy et al., 2015), another explanation is that this group difference reflects stronger reduction in auditory cortical activity in men due to stronger attention toward the visual sexual cues in the videos in men compared to women. In line with this interpretation both men and women showed reduced auditory cortex activity for the sexual videos as indicated by the negative beta weights in auditory cortex (Figure 5).
Evolutionary accounts posit that men and women have evolved different mating strategies in domains where they have faced different adaptive challenges (Buss & Schmitt, 1993). Lower obligatory parental investment in men has presumably given rise to the stronger preference for short-term mating and sexual variation in men, as these have increased the probability of genetic success more for men than for women. Men may also have evolved a preference for physical features associated with youth since such cues signal fertility and many years of potential future reproduction (Buss & Schmitt, 2019). The type of pornography consumed by men often simulates short-term sexual encounters with novel young women (Malamuth, 1996; Salmon & Diamond, 2012). Thus, men's higher attentional engagement with sexual stimuli might reflect evolved preferences for sexual variety and physical cues of reproductive potential. Such biological biases probably interact with cultural norms in shaping sex-typical preferences as evidenced by cross-cultural variation and changes across time in the magnitude of these sex differences (Buss & Schmitt, 2019; Petersen & Hyde, 2011).
5.3 Emotion circuit activation in men and women
Both men and women experienced strong positive emotions and only weak negative emotions toward pornography although men reported slightly higher feelings of sexual arousal and joy and less shame than women. In accordance with the incentive value of sexual stimuli, both experiments activated limbic and mesolimbic regions associated with reward and emotion such as the ventral striatum and amygdala in both men and women. Unlike some previous studies, we did not observe sex differences in the activation of the amygdala (Hamann et al., 2004) or NAc (Wehrum-Osinsky et al., 2014) suggesting that activity evoked by visual sexual stimuli in these regions does not reliably differentiate men and women. The sexual stimuli also activated the primary and secondary somatosensory cortices and insula in both men and women in accordance with previous studies (Arnow et al., 2009; Ferretti et al., 2005). These regions contribute to emotion via the processing of bodily sensations and interoceptive feedback (Craig, 2002) and SII is more generally involved in the perception of touch (Keysers et al., 2010). We also observed activation in the ACC, which is a common finding in studies employing visual sexual stimuli (Stoléru et al., 2012) presumably because cingulate activity is coupled with autonomic arousal (Beissner et al., 2013). Overall, the sex differences observed in the GLM analysis were most consistent in cortical regions although activity in the brainstem and thalamus were also stronger in men than in women in the pictures experiment.
5.4 Sex differences in neural substrates of sexual perception: State of the evidence
The recent meta-analysis of Mitricheva et al. (2019) found no evidence of sex differences in brain responses to visual sexual stimuli. However, Poeppl et al. (2020) re-analyzed these data with a number of methodological improvements, such as the exclusion of ROI analyses, and found more consistent activation in ACC and hypothalamus in men and in lateral occipital cortex in women. We also found a sex difference in a frontal cluster partly overlapping with the ACC but did not replicate the hypothalamic and occipital effects.
One possible source of the discrepancies between our results and these meta-analyses is that the latter were not based on direct comparisons between sexes in the original studies. Instead, both meta-analyses compared separate Activation Likelihood Estimation (ALE) maps for men and women that were computed mostly from studies with only men or women as subjects. This approach can reveal sex differences in how consistently a given region is activated in different studies (i.e., sex difference in convergence), but cannot uncover the consistency of sex differences in brain activation across studies (i.e., convergence of sex differences) (Müller et al., 2018). Such meta-analyses may also be confounded by methodological differences between studies with only men or women as participants. A meta-analysis of direct comparisons between men and women is precluded by the scarcity of pertinent high-quality studies: Mitricheva et al included only 11 studies with both men and women, and most of these were underpowered for reliable group comparison (Hamann et al., 2004; Klucken et al., 2009) or did not perform a whole brain analysis of the sex differences (Strahler et al., 2018). Notably, the only study with a whole brain GLM analysis of sex differences and a sample size comparable to ours (Wehrum et al., 2013) revealed stronger responses to sexual pictures in men than women in the thalamus and occipital and parietal cortex in line with our results. Thus, we provide much-needed data on sex dependency of neural responses to sexual stimuli and pave way for robust meta-analyses of such sex differences.
5.5 Limitations
Subject-specific emotion ratings and physiological arousal responses were not acquired from the participants in the fMRI experiment; thus we could not directly link the hemodynamic data with direct indices of sexual arousal. The current study specifically focused sex differences in brain responses to stimuli representative of “mainstream” pornography typically consumed more by men than women and we did not attempt to balance how sexually arousing or interesting the stimuli were for the male and female participants (cf. Janssen et al., 2003; Laan et al., 1994). The majority of our subjects identified as exclusively heterosexual and thereby we were unable to test the effects of sexual preference irrespective of gender and our results may not generalize to individuals with non-heterosexual preference.
Posts
No comments:
Post a Comment