Women and men with distressing low sexual desire exhibit sexually dimorphic brain processing

Distressing low sexual desire, termed Hypoactive Sexual Desire Disorder (HSDD), affects approximately 10% of women and 8% of men. In women, the ‘top-down’ theory of HSDD describes hyperactivity in higher-level cognitive brain regions, suppressing lower-level emotional/sexual brain areas. However, it is unknown how this neurofunctional disturbance compares to HSDD in men. To investigate this, we employed task-based functional MRI in 32 women and 32 men with HSDD to measure sexual-brain processing during sexual versus non-sexual videos, as well as psychometric questionnaires to assess sexual desire/arousal. We demonstrate that women had greater activation in higher-level and lower-level brain regions, compared to men. Indeed, women who had greater hypothalamic activation in response to sexual videos, reported higher psychometric scores in the evaluative (r = 0.55, P = 0.001), motivational (r = 0.56, P = 0.003), and physiological (r = 0.57, P = 0.0006) domains of sexual desire and arousal after watching the sexual videos in the scanner. By contrast, no similar correlations were observed in men. Taken together, this is the first direct comparison of the neural correlates of distressing low sexual desire between women and men. The data supports the ‘top-down’ theory of HSDD in women, whereas in men HSDD appears to be associated with different neurofunctional processes.

www.nature.com/scientificreports/ to advance our understanding of the neural mechanisms that govern sexual responses in these patients, with the ultimate goal of informing sex-specific therapeutic approaches.To examine this, we employed fMRI during a visual sexual stimuli task designed to trigger underlying sexual-brain processing in 32 women with HSDD and 32 men with HSDD.The task consisted of sexual and control (exercise) videos.To provide functional and behavioural relevance for the brain activity changes, participants also completed well-established psychometric questionnaires.

Demographics
Baseline demographics for the participants can be found in Table 1.The women were younger than the men, however, the participants did not differ significantly in the number of times they engaged in intercourse per month, duration of relationship, or depression and anxiety measures.

Whole-brain analysis comparison
The overall effects of the sexual-brain processing task were similar in both the women and men with HSDD (Fig. 1A,B).When comparing the sexual videos compared to control videos, activation (shown in red/yellow) occurred in the striatum, visual cortex, cerebellum, and anterior cingulate (areas which are well-established to be involved in sexual-brain processing 10,11 ).In addition, when comparing sexual videos to control videos, there were areas where there was less activation when watching sexual compared to control videos, defined as deactivation (shown in blue/green).Deactivation was observed around the temporal cortex, posterior cingulate, and motor cortex (Fig. 1A,B).
In order to directly compare the similarities in the spatial pattern of neural responses, between the women and men to sexual compared to control videos, the data were binarised to produce masks of the task effects (task effects shown in Fig. 1A,B and for additional brain slices of task effects see Supplementary Fig. 1), meaning voxels where there is (de)activation were set to a value of 1 and all other voxels retained a value of 0. These binarised images from the women and men's data were then multiplied together.Since 1 × 1 = 1 and 1 × 0 = 0, voxels where there is activation in both women and men will retain a value of 1 (with all other voxels being 0) and so an overlap image of spatial (de)activation was produced (Supplementary Fig. 1).The overlap image shows the common areas in the spatial pattern of activation to sexual and control videos in both women and men.Dice coefficients were then calculated to quantify the overlap in each contrast: activation (sexual > control) = 0.601, deactivation (control > sexual) = 0.643.As stated by Wüthrich et al. 12 , there is a lack of consensus regarding the interpretation of Dice coefficients.However, as proposed by Wüthrich 12 , we followed the guidelines set out by Cicchetti 13 .Coefficients below 0.40 are considered poor; between 0.40 and 0.59, fair; between 0.60 and 0.74, good, and > 0.75, excellent.Therefore, our results demonstrate good replicability of task effect results and a similar general spatial pattern of responses in both sexes.
After comparing the spatial similarities in neural responses to sexual and control videos, differences between the women and men were then directly compared in their responses to the sexual compared to control videos, and important dimorphisms were observed.Women had significantly greater activation in limbic brain regions (such as the amygdala, striatum, and thalamus), compared to men (purple, Fig. 1C).Furthermore, women also had greater activation in higher-level cortical regions such as the inferior frontal gyrus and the superior frontal gyrus.In contrast, men had greater activation in areas within the visual cortex, such as the lingual gyrus and occipital pole (green, Fig. 1C).A summary table of cluster coordinates can be found in Supplementary Table 1.

Correlational analysis
To provide behavioural relevance for the brain activity data, post-hoc correlations were performed separately in women and men.The post-scan Sexual Arousal and Desire Inventory (SADI) domain psychometric ratings were correlated with the degree of activation in sex-network regions (in response to sexual > control videos).Significant positive correlations were observed between hypothalamic activation and the evaluative (P FDR < 0.05 *, r = 0.55, P = 0.001, adjusted threshold α < 0.0042), motivational (P FDR < 0.05 *, r = 0.56, P = 0.003, adjusted threshold α < 0.0063), and physiological (P FDR < 0.05 *, r = 0.57, P = 0.0006, adjusted threshold α < 0.0021) SADI domains in the women, (Fig. 3).No significant correlations were found between the men's SADI domain scores and the degree of activation in any of the sexual-brain ROIs.

Discussion
This study presents the first direct comparison of the neural correlates of distressing low sexual desire between women and men, identifying sexual dimorphisms in the neurobiological network underlying the abnormal sexual responses in HSDD.The findings are somewhat consistent with previous research in individuals with normal sexual desire, suggesting that women and men exhibit similar general overall patterns of activation to visual sexual stimuli 11,14 .However, notable differences were observed in the activation of limbic brain regions in women and men with HSDD, particularly the hypothalamus, amygdala, and thalamus, which are key structures associated with emotional processing and sexual motivation 15,16 .Indeed, previous fMRI studies have identified the hypothalamus 14,17,18 , amygdala 14 , and thalamus 17 as regions more strongly activated in response to sexual stimuli in healthy men compared to healthy women.By contrast, here we found that women with HSDD demonstrate a greater magnitude of activation in these three regions than the men with HSDD.Furthermore, we confirmed that these findings were specifically in response to visual sexual stimuli as no differences were identified in the control comparison (exercise > baseline).Therefore, given that these brain regions are known to play a pivotal role in the motivational and emotional components of sexual arousal 19 , this provides possible evidence for their involvement in the neural bases underlying HSDD in men.Furthermore, previous work has identified www.nature.com/scientificreports/women with HSDD as having lower activation in limbic regions in response to sexual stimuli compared to healthy women 7 .Our data therefore presents a possible hierarchy of activation in limbic regions to sexual stimuli: healthy men, healthy women, women with HSDD, and finally men with HSDD having the lowest activation in emotional brain regions to sexual stimuli.This sexual dimorphism highlights the need for more research to be conducted into men with HSDD.The theoretical model of HSDD presented by Cacioppo suggests that women with HSDD have hyperactivation in higher-level cortical regions, which then leads to hypoactivation in limbic regions of the brain in a 'top-down' inhibition of the sexual response 7 .The present correlations from our data in women support this hypothesis, since women with lower activation in the hypothalamus reported lower scores on sex-positive scales in the SADI questionnaires.This suggests that women with more hypoactivation in this limbic sex-brain-region feel less positive about sex, and perhaps have more severe HSDD.We also identified significantly greater activation in the left inferior-frontal gyrus in the women with HSDD relative to the men with HSDD.Consistent with this, the left inferior frontal gyrus is an area associated with internal monologuing 20 , feelings of guilt 21 , and has been previously identified as a region which is hyperactive in women with HSDD 7 .

Amyg
By contrast, men exhibited significantly greater visual activation in response to sexual videos (compared to control exercise videos) than women.This heightened activation in the visual cortex aligns with previous findings in healthy men 22 , indicating their heightened sensitivity to visual sexual cues and a propensity to allocate more attention to such cues.This observation offers a fresh perspective on the potential neurobiological model for HSDD in men.In this proposed model, men demonstrate adequate visual attention to sexual cues; however, these cues might not be effectively relayed to the limbic system, which is responsible for emotional processing.This attentional response to sexual cues therefore seems not to develop into sexual arousal, perhaps suggesting a functional disconnection between sensory/attentional and emotional systems worthy of future study.This is the first study to directly compare sexual-brain processing in women and men with HSDD and had a number of strengths.All participants were biochemically eugonadal, thereby removing the confounding effect of hypogonadism on our neural and behavioural findings.Regarding the women, all participants were premenopausal with study visits undertaken on days 1-7 of the menstrual cycle (follicular phase) to ensure consistent reproductive hormone levels.This is important as it is well-established that brain activity is altered by fluctuations in reproductive hormones across the menstrual cycle 26 .Moreover, a sample of 64 participants is large for fMRI studies 23 .The participants all interacted with a same sex physician throughout the study, minimising bias associated with investigator sex 24 .The neural response to low-level visual features was adequately controlled for through the use of couple exercise videos.Importantly, this ensured all videos contained one woman and one man with similar levels of movement in both types of videos.The study employed multiple analysis methods, including whole-brain, ROI, and correlations to explore and understand the distinct differences between women and men with HSDD.
A potential limitation of the study is the age difference between the groups, with women on average 9 years younger than the male participants.Although this may have some effect on neural activation, the control > baseline comparison confirmed there was no difference in activation in the ROIs to the control stimuli, suggesting baseline brain activation was similar between the two groups.Moreover, this sample more accurately represents the HSDD population, since younger women tend to be more distressed by their HSDD than older women and so seek help/trials 25 , while HSDD becomes more prevalent as men get older 26 .
The lack of a non-HSDD comparator group limits our ability to extrapolate distinct differences between HSDD and individuals with normal sexual desire (especially in men where there is very limited previous data).Therefore, an important area for future investigation would be an fMRI study comparing sexual-brain processing in women and men with HSDD with individuals with normal sexual desire under the same protocol.Moreover, the present data originated from two previous studies primarily designed to discern variations in brain activity between placebo and hormone administration study visits.Consequently, our power calculation was not initially formulated to detect differences between women and men.Future studies should prioritise incorporating a healthy control group (with normal sexual desire).In addition, all the women were premenopausal which limits the generalisability of our results to post/perimenopausal women.However, this study was the first to directly compare sexual-brain processing in women and men with HSDD under the same protocol, and even with these confounds taken into consideration, the data has shown compelling direct differences between women and men with HSDD.The data provides further evidence and support for the top-down theory of HSDD in women, and the correlational data suggests therapies which may target the hypoactive-limbic system may be effective at increasing positive feelings towards sex in women with HSDD.In addition, the comparatively lower limbic activation in the men is of interest and warrants further study.
Taken together, the observed sexual dimorphism in behavioural and neuroimaging data have significant implications for our understanding of disrupted sexual desire and arousal in women and men.In women, HSDD is likely driven by top-down inhibition of the sexual response.This was shown in the hyperactive inferior frontal gyrus and the correlations implying lower activation in sexual brain regions is associated with lower sexual function on psychometric assessment.Together this suggests therapies targeted at reducing hyperactivity in higher-cortical regions or boosting activation in lower-limbic regions could improve sexual function in women with HSDD.In men, there may be a different model of HSDD, whereby the visual attention to the sexual cues is not effectively relayed to emotional centres involved in the sexual response.Unlike in women, more activation in limbic brain regions in men did not translate to greater sexual functioning on the SADI questionnaire.This suggests that therapies which may target functional connectivity could benefit men with HSDD.Psychedelic therapy has recently been shown to reduce brain modularity 27 , and is thought to enable the formation or restoration of adaptive functional connectivity patterns 28 .This could perhaps be a useful potential intervention for men with HSDD to strengthen the connectivity between the visual and emotional centres of the brain.Identifying fMRI data analysis fMRI data processing was performed using FEAT (fMRI Expert Analysis Tool), part of the Oxford Centre for Functional MRI of the Brain (FMRIB) Software Library (FSL) version 6.0 (www.fmrib.ox.ac.uk/ fsl).Registration on to high resolution structural images was carried out using FMRIB Linear Image Registration Tool (FLIRT) 36 .Registration from the high resolution T1 structural image of each participant to standard Montreal Neurological Institute (MNI) 152 space was then further refined using FMRIB's Nonlinear Image Registration Tool (FNIRT) 37,38 .The following pre-statistic processing was applied: motion-corrected FLIRT 36 , non-brain removal using the Brain Extraction Tool (BET) 39 , spatial smoothing (6.0 mm) and 90-s high-pass temporal filtering.All first level models included the extended of head motion parameters regressors (original parameters, plus derived temporal derivatives and quadratic functions).White matter and cerebral spinal fluid masks were created from each participant's anatomical scans using the FMRIB's Automated Segmentation Tool (FAST), and the time series from each functional scan was extracted from these masks for use as a regressor of no interest for each participant in each task to further denoise the data.Time-series statistical analysis was carried out using FMRIB's Improved Linear Model (FILM) with local autocorrelation correction 40 .
The regressors of interest were derived from the onset times of the sexual and control video conditions and were convolved with a gamma function to simulate the hemodynamic response function (HRF).These were used as the main regressors of interest in the general linear model (GLM) with the denoising methods mentioned above as regressors of no interest.The contrasts were defined by each stimulus condition compared to baseline and then also comparing two stimulus conditions of interest, contrasts comparing the sexual and control conditions were our main outcome.A between-subjects model was used to investigate differences in whole brain activation between women and men with HSDD in the sexual > control contrast.Statistical images were set at threshold using clusters determined by Z > 2.3 and a corrected cluster significance threshold at P = 0.05.

Statistical analysis
Overlap in the similarities of activation patterns were calculated using Dice Coefficients 41 .The thresholded results of the group analyses were binarised and multiplied together to create image masks containing only the voxels that overlap in the female and male groups.Dice coefficients were then calculated using the following equation:  where C is the number of voxels in the overlap image, A is the number of voxels in the female group average image and B is the number of voxels in the male group average image.Dice coefficients give an estimate of the spatial replicability of the task across the two different sexes.Dice coefficients are between 0 and 1, where 0 is no overlap and 1 would mean the images are identical.
ROIs employed in the current study were defined using the term 'sexual' in the meta-analytic site Neurosynth (available at: https:// neuro synth.org/ analy ses/ terms/ sexual/) for a separate and independent study 42 .MNI152 space masks for the ROIs are available at: https:// osf.io/ t7g5x/.ROIs are shown in Fig. 5. ROIs were treated as single regions and were not lateralised to left/right.Data was extracted from the sexual > control contrast within the ROIs and an ANOVA was used to confirm a main effect of sex followed by post-hoc unpaired t-tests to compare activation in the ROIs between women and men.To control for general brain activation differences between the sexes, data was extracted from the control > baseline and sexual > baseline contrasts as well.The data from the ROIs (sexual > control) was then correlated using Spearman's correlation for each of the four SADI domains (using the post-scan ratings).All results were FDR corrected using Benjamini-Hochberg correction 35,[43][44][45][46] , treating men and women as separate families.
Power calculations were performed on the original projects based on previous data of BOLD responses in key sexual brain regions.For full power calculations see 29,30 .The sample sizes used are consistent with previous fMRI research conducting between-subjects analyses 23 .

Figure 1 .
Figure 1.Women with HSDD have greater limbic activation to sexual videos than men.(A) The female and (B) male HSDD group average results showing the brains activation (red/yellow) and deactivation (blue green) to sexual compared to control (exercise) videos.(C) Brain regions more activated in women (relative to men) to sexual compared to control videos are shown in purple.Brain regions more activated in men (relative to women) to sexual compared to control videos are shown in green.Results are cluster corrected and thresholded to Z = 2.3, P < 0.05, N = 64 (32 women, 32 men).