Introduction

Premature ejaculation (PE), is one of the most common male sexual dysfunction disorders and presents a diagnostic and treatment challenge for physicians with a prevalence of 75%1,2. Those with lower intravaginal ejaculation latency time (IELT) may consider themselves premature ejaculators, and suffer from quality of life issues normally associated with premature ejaculation, and even benefit from non-pharmaceutical treatment3. Based on standardized definitions of evidence, the International Society for Sexual Medicine (ISSM) classifies PE into LPE and acquired PE (APE). LPE was defined as ejaculation often or always occurring ≤ 1 min after vaginal insertion from the initiation of sexual intercourse. In contrast, a significantly shortened (IELT) (usually less than approximately 3 min) was defined as APE4. There is growing evidence that multiple physiological and biological factors are associated with PE, such as anxiety, depression, and abnormal testosterone secretion5,6. LPE may be underestimated due to embarrassment and lack of awareness.

Ejaculation is a complex physiological process regulated by a network involving the central and peripheral nervous systems, consisting of two main phases: emission and discharge, controlled by the autonomic nervous system. Three nervous system interactions regulate the spinal reflex: modulatory influences from the upper spinal column, interactions at the spinal cord level, and sensory inputs determining the sacral reflex. Key regions such as the mesodiencephalic transition zone, particularly the ventral tegmental area (VTA), are activated during ejaculation, highlighting its association with reward and pleasure. Additionally, the basal ganglia and specific cortical regions, including the parietal and temporal lobes, are involved in sensory and emotional processing. Dopamine plays a central role in mediating reward mechanisms, while serotonin modulates ejaculatory latency, forming the basis for therapeutic strategies like selective serotonin reuptake inhibitors (SSRIs)7. The human penis conveys behaviorally significant sensory information through the dorsal penile nerve, which plays a crucial role in initiating and maintaining an erection, this is thought to be one of the pathological mechanisms by which penile hypersensitivity may be responsible for PE8. Previous studies suggested that the organic basis of PE may be a somatic disease and/or neurobiological imbalance2,9. However, we know almost nothing about the mechanisms of processing sexually significant input to the penis in the brain, especially in PE patients. Moreover, no studies have fully elucidated the central neurological mechanisms underlying the pathophysiology of PE10,11. Therefore, it is very important to explore the changes in the central nervous system in PE patients under penile stimulation.

With the advantages of non-invasiveness and high spatial resolution, MRI technology can effectively identify functional and structural brain abnormalities, responses, and activity in patients with PE12,13,14,15. For instance, compared to healthy controls, the cortices of the PE patients in a wide range of areas (including the parietal and occipital lobes, frontal lobe, and limbic system) were thicker13. Structural alterations have been reported in patients with LPE, shown by an increase in the mean volume of the caudate nucleus16. The rs-fMRI study by Geng et al. found that in patients with LPE, FC between the amygdala and the precuneus, posterior cingulate cortex (PCC), and middle temporal cortex (MTC) was increased. In contrast, FC between the amygdala and the precuneus middle, insula, and inferior frontal gyrus was decreased17. Another fMRI study showed that FC between the motor cortex and bilateral ventral thalamus, between the prefrontal cortex and left dorsomedial thalamus, and between the temporal cortex and bilateral ventral thalamus was significantly reduced in patients with LPE15. Moreover, using visual task-based fMRI, several studies that examined brain responses to visual sexual stimulation for healthy men have indicated increased responses in the parietal, anterior cingulate, temporal lobe, amygdala, insula, and septal regions18,19,20. However, few studies have explored the brain response to penile stimulation.

Somatosensory evoked potential (SEP) is a useful and non-invasive test for assessing somatosensory system function. We used it to investigate the brain’s response to penile inputs. Depending on the site of electrical stimulation, SEPs were divided into the dorsal nerve (DNSEP) and glans penis (GPSEP) tests. DNSEP is an objective neurophysiological test that assesses the integrity, peripherality, and centrality of the dorsal penile afferent pathway. GPSEP is a modification of DNSEP that sends stimulation to the glans penis21. The dorsal penile nerve consists of two distinct groups of axons that innervate the glans penis, shaft, and anterior urethra22. Previous studies found that compared to the control group, PE patients had a significantly greater mean amplitude for GPSEP but no significant amplitude for DNSEP, and the specificity of GPSEP was better than that of DNSEP (especially for patients with PE)21. Previous studies using SEPs have suggested that penile hypersensitivity from sensory-evoked potential testing may be one of the important pathological mechanisms of PE23. Georgiadis et al.'s study suggests that human sexual behavior follows a typical "pleasure cycle," encompassing multiple stages from sexual desire to orgasm and finally to satiation. This process is closely associated with various brain regions, such as the nucleus accumbens, orbitofrontal cortex, and anterior cingulate cortex24. Given that premature ejaculation (PE) shortens the duration of sexual experience, it may lead to abnormal activation patterns in these brain regions. Therefore, investigating the effects of penile electrical stimulation on brain function can contribute to a deeper understanding of the neural mechanisms underlying premature ejaculation. In our study, we chose fMRI to test the brain response and brain activity in patients with LPE using GPSEP electrical stimulation on the glans penis.

We hypothesized that PE patients have different patterns of brain responses to sensory-evoked potential tests. To test this hypothesis, we performed task-fMRI and rs-fMRI in LPE patients and healthy control subjects. To our knowledge, we are the first to introduce task-fMRI to estimate the changes using hemodynamic responses. DC was used to compare the difference in brain response between PE and healthy controls, especially for the dopamine pathway.

Participants and methods

Participants

This cross-sectional study’s design was based on the Declaration of Helsinki and was approved by the Research Ethics Committee of Nanjing Drum Tower Hospital. The IRB number of this study is 2013-083-01. After providing a detailed description of the study procedures to the patients, written informed consent was obtained from each participant. From 2013 to 2014, 28 PE patients and 18 age-matched healthy controls were recruited for this study. All subjects were right-handed. The lifelong premature ejaculation (PE) patients in our study were diagnosed according to the International Society for Sexual Medicine (ISSM) guidelines, which include the following criteria: (a) Ejaculation that always or nearly always occurs prior to or within about 1 min of vaginal penetration; (b) Inability to delay ejaculation on all or nearly all vaginal penetrations; (c) Negative personal consequences such as distress, bother, frustration, and/or the avoidance of sexual intimacy.For the control group, healthy subjects were enrolled with self-reported intravaginal ejaculatory latency time (IELT) > 3 min25,26. The study design was explained to all included subjects, who then signed an informed consent form. Four subjects were excluded since they moved their head extensively during the procedure (maximum translation distance > 3 mm, or maximum rotation angle > 3°) or showed poor spatial normalization to the EPI template. Finally, 26 patients with LPE (age: 27.42 ± 4.30 years) and 16 NCs (age: 27.63 ± 3.77 years) were included in our study.

Stimuli and the experimental design paradigm of task-fMRI

All subjects were stimulated by electrical shocks using an electromyograph and evoked potential equipment (Keypoint4, Alpine BioMed ApS, Copenhagen, Denmark) with a stimulation interval of 1 ms and a stimulation frequency of 3 Hz. A pair of circular surface electrodes were placed on the glans penis, and the stimulation intensity was set to 10.0 mA (at this level, the subject felt no pain)21in each subject. The electrical stimulation was performed in blocks of 30 s with a 30 s-rest interval in between. Each electrical stimulation block and rest block were alternated six times. The pattern lasted a total of 360 s (Fig. 1A).

Fig. 1
figure 1

Workflow of the experimental design and data analysis. (A) Paradigm of task-fMRI using electrical stimulation on the glans penis. The blue bars represent the rest state, and the green bars represent the electrical stimulation state. The participants received electrical stimulation on the glans penis. The stimuli intensity was 10.0 mA, the frequency was 3 Hz, and the duration of stimuli was 1 ms in each subject. The electric stimuli were given in 30-s blocks (green) separated by 30-s intervals of rest (blue). Each electric stimulation block and rest block were alternately presented six times. The paradigm lasted 360 s in total. (B) Brain regions with group differences in beta values under task-fMRI. (C) Calculation of the DC metric using the rs-fMRI dataset. (D) Brain regions with group differences in DC. (E) The abnormal functional connectivity within the brain regions founded based on the task and rest fMRI data. (F) Correlation between MRI parameters and IELT.

Full size image

Image acquisition

The fMRI image data were acquired on a 3 T MRI system (Achieva3.0 T TX, Philips Medical Systems, Eindhoven, Netherlands) using an 8-channel phased-array coil. Using an echo-planar imaging (EPI) method, task-fMRI image data were obtained from the whole brain. The detailed parameters of fMRI are as follows: section thickness = 4 mm with no section gap, field of view (FOV) = 192 × 192 mm2, matrix = 64 × 64, echo time (TE) = 30 ms, repetition time (TR) = 2000 ms, and flip angle (FA) = 90°. Each fMRI sequence contains 180 volumes. Each volume consists of 35 axial slices roughly aligned with the anterior commissure–posterior commissure line. Rs-fMRI was acquired by using the same parameters but using 230 volumes. High-resolution T1-weighted structural images were also acquired using the following parameters: FOV = 256 mm × 256 mm × 192 mm, slice thickness = 1 mm with no gap, TR = 7600 ms, TE = 3400 ms, and FA = 8°.

Task-fMRI

The task-fMRI data were processed using SPM12 software. The first six images of each fMRI dataset were discarded to eliminate the initial transit signal fluctuations. The subsequent images in the session were realigned to eliminate any slight head movements and were coregistered and spatially normalized T1WI high-resolution anatomical images to the Montreal Neurological Institute (MNI) brain template27. A spatial resolution of 3 mm × 3 mm × 3 mm was used for spatial normalization, and then a Gaussian smoothing kernel of 8 × 8 × 8 mm3 (full width at half maximum, FWHM) was used for smoothing in the EPI images. To generate statistical parametric maps for each subject, we matched stimulation paradigms with functional data from electrical stimulation conditions and fused them with HRF and time derivatives. Group analysis was performed to generate a mean activation map for each study group. In SPM12 software, HRF was applied to analyze bold task-fMRI data using a standard analysis pipeline. The hemodynamic response (HR) allowed for rapid blood delivery to active neuronal tissue. HRF was represented by a beta value.

Rs-fMRI

Image preprocessing

Preprocessing of the rs-fMRI data was performed by using the advanced version of the rs-fMRI Data Processing Assistant (DPARSFA, vision 2.3, http://www.restfmri.net) based on MATLAB platform (The Math Works, Inc., Natick, MA, USA). To equalize the signal, we allowed the participant to adjust to the noise from the scanner and discarded the first ten time points of each subject’s image. Slice-timing, head motion correction, and spatial normalization to the standard MNI EPI template were performed sequentially with a resampled voxel size of 3 × 3 × 3 mm3 conducted for the remaining 220 time-points. No global signal regression was performed to prevent the introduction of distortion in the time series data28,29. Afterwards, the data were detrended to remove linear trends in the time course and to regress out interfering covariates, including the white matter and CSF. Finally, bandpass filtering was performed (temporal bandpass filtering was 0.01–0.10 Hz voxel-by-voxel to reduce the effects of low-frequency drift and high-frequency respiratory and cardiac noise).

Degree centrality

The DC is the metric used to calculate the total number of connections. Therefore, it depends on the direct neighbors of the nodes, but it has nothing to do with the overall topology of the rest of the network. The REST software is used for DC processing30. Subsequently, an undirected adjacency matrix was achieved by thresholding each correlation at r > 0.2531,32. Then, DC is computed as the sum of the weights of the important connections (weighted) for each voxel31. Finally, individual-level voxel DCs were converted to a DC Z score (z-DC) map by subtracting the mean DC across the whole brain and dividing by the standard deviation of the whole-brain DC (Fig. 1C)33.

Fingerprint-based FC analyses

FC was calculated by using the conventional fingerprint-based correlation analysis method. The method involves the following steps: (1) selecting a specific voxel or voxel cluster in the brain, which is called the fingerprint, (2) extracting a time series of BOLD signals in the fingerprint region, and (3) using this time series to calculate the entire voxel FC map of the brain34. The fingerprint regions of interest were selected as spheres with a radius equal to 5.5 mm at the peak points of those clusters, which were significantly activated in the erotic picture stimulus task-fMRI in patients with PE compared to the NC group. FC calculation was implemented using REST software (http://restfmri.net/forum/index.php)(35). The altered FC value was used to distinguish PE and NC by receiver operating characteristic (ROC) curve analysis (Figs. 1E, 4C,D).

Evaluation tools

To evaluate sexual function and the severity of premature ejaculation, the following standardized tools were utilized: (1) International Index of Erectile Function (IIEF):

The IIEF is a validated 15-item questionnaire widely used to assess male sexual function across five domains: erectile function, orgasmic function, sexual desire, intercourse satisfaction, and overall satisfaction. Each item is scored on a scale of 0–5, with higher scores indicating better sexual function. This tool has been extensively applied in clinical and research settings to evaluate sexual health. (2) Chinese Index of Premature Ejaculation (CIPE): The CIPE is a validated scale developed for Chinese-speaking populations to specifically assess premature ejaculation severity. It evaluates dimensions such as ejaculatory control, satisfaction, and the impact of PE on sexual relationships, with lower scores indicating more severe PE symptoms. This tool is widely used in clinical studies and has demonstrated good reliability and validity in PE assessmet3,35. (3) Intravaginal Ejaculatory Latency Time (IELT): IELT is an objective measure of the time from vaginal penetration to ejaculation. Participants were instructed to use a stopwatch to record IELT during at least four separate sexual encounters over a four-week baseline period. The mean IELT was calculated and used as an indicator of ejaculatory control. This method is considered the gold standard for PE severity evaluation and has been widely employed in both clinical practice and research36.

Relationship between the beta and DC values and the IELT

We analyzed the correlations between fitted beta, DC values, and IELT to determine whether beta and DC values change with disease progression in NC subjects and patients with PE. As these analyses were exploratory, we used a statistical significance level of p < 0.05 (Fig. 1F).

Statistical analysis

Clinical data were statistically analyzed using SPSS software version 23.0 (Statistical Programme for the Social Sciences, SPSS Inc., Chicago, IL), and fMRI data were analyzed by using SPM12. Two-sample t-tests were used to compare brain activation between the patients with PE and NC groups (p < 0.001 at the lowest statistical threshold voxel level and p < 0.05 at the cluster level, AlphaSim correction)(Fig. 1B,D). A region-of-interest analysis of the BOLD response was performed for each condition, involving sex-related scores. Separate analyses were performed for the entire cohort and the patients in the PE and NC groups.

Results

Demographic and clinical data

The demographic, psychiatric, and behavioral characteristics of patients with LPE and NCs are presented in Table 1. As shown, there were no differences between the two groups in terms of age (NC: 27.63 ± 3.77 years, PE: 27.42 ± 4.30 years, P = 0.878), marital status (P = 0.750), or education level (P = 0.960). In addition, the International Index of Erectile Function-5 (IIEF-5) score did not differ between the two groups, suggesting that erectile function was normal in patients with LPE. However, the Chinese Index of Premature Ejaculation-5 (CIPE-5) scores and IELT showed significant differences between the two groups (NC: 21.88 ± 1.96, PE: 8.85 ± 2.17, P < 0.001; NC: 9.71 ± 4.37 min, PE: 0.80 ± 0.41 min, P < 0.001). Moreover, each question of CIPE also showed significant differences between LPE patients and NCs (Table 1).

Table 1 Demographic and clinical characteristics of LPE patients and NC subjects.
Full size table

Differences in activation between groups based on task-MRI results

Brain activation in response to electrical stimulation was assessed by using the resting state as a baseline. Comparison between LPE and NC groups showed that brain activation in the bilateral thalamus (THA. R, x, y, z = 12, − 24, 6, T = 2.89, p < 0.05; THA. L, x, y, z = − 24, − 12, 3, T = 3.43, p < 0.05, AlphaSim corrected) and the bilateral inferior frontal gyrus, triangulation area (IFGtriang. R, x, y, z = 42, 21, 6, T = 2.99, p < 0.05; IFGtriang. L, x, y, z = − 42, 30, 24, T = 3.51, p < 0.05, AlphaSim corrected) of the LPE patients significantly decreased when the patients were stimulated with electricity (Fig. 2A, axial view; 2B, coronal view and 2C). The beta values were extracted from imaging of the bilateral thalamus and inferior frontal gyrus and triangulation of the voxels (shown in Fig. 2D and Table 2).

Fig. 2
figure 2

Brain responses to electrical stimulation on the glans penis of subjects. The results of the independent samples t-test showed a significant increase in activation in the LPE patients compared to the NC group (A, axial view; B, coronal view. T maps threshold at p < 0.05, AlphaSim corrected); The 3-D render of the bilateral thalami and IFGtriang (C); The beta values for the bilateral thalami and IFGtriang (D, each dot represents the value of each subject, the shaded area represents the standard deviation, and the black horizontal line represents the mean value). **p < 0.01; IFGtriang: inferior frontal gyrus, triangulation area.

Full size image
Table 2 Significantly different brain areas in task-fMRI and rs-fMRI between the patients with LPE and NC.
Full size table

Degree centrality

The independent t-test showed increased DC values in the bilateral SMA (SMA, x, y, z = 9, − 9, 45, T = 4.83, p < 0.05, AlphaSim corrected) and reduced DC values in the bilateral precuneus (PCUN, x, y, z = -3, -78, 48, T = − 4.88, p < 0.05, AlphaSim corrected) in LPE patients compared with NCs. This suggests that patients with LPE show brain dysfunction during ejaculation and exhibit pre-existing variations in spontaneous brain activity patterns in the resting state compared to NCs (Fig. 3A–D). Table 2 shows the DC values in the SMA and PCUN.

Fig. 3
figure 3

The difference in DC from rs-fMRI between PE and NC. Independent sample t-test results showed that the patients with LPE exhibited significantly altered voxels of DC compared to the NC group (A, axial view; B, coronal view. T maps threshold at p < 0.05, AlphaSim corrected, red areas mean increased FC and blue areas mean decreased FC.); The 3-D render of the bilateral SMA and PCUN (C); The DC values for the bilateral SMA and PCUN (D, each dot represents the value of each subject, the shaded area represents the standard deviation, and the black horizontal line represents the mean value). **p < 0.01; SMA: SMA; PCUN: precuneus.

Full size image

Functional connectivity altered in the rs-fMRI

We used the spherical ROI of each peak point using the significant difference between LPE patients and controls in DC as a fingerprint. Voxelwise FC maps were calculated for the whole brain using GLM analysis. The results showed that the SMA was enhanced using the FC of the right inferior frontal gyrus, triangulation area (IFGtriang. R)(Fig. 4A, p < 0.05), and the PCUN had a weakened FC with the bilateral thalamus (Fig. 4B, p < 0.05).

Fig. 4
figure 4

Using the bilateral SMA and PCUN from the DC map as fingerprints, fMRI showed that FC was altered between groups. An independent sample t-test was performed between the LPE patients and NC groups (T maps threshold at p < 0.05, AlphaSim corrected). (A) The brain areas where FC was enhanced between the bilateral SMA and the right inferior frontal gyrus; (B) The brain areas where FC was dropped off between the bilateral PCUN and bilateral thalamus. (C) The ROC curves for distinguishing the two groups using the FC between the inferior frontal gyrus and SMA. The AUC was 0.82 (P = 0.002). (D) The ROC curves for distinguishing the two groups using the FC between the precuneus and thalamus. The AUC was 0.83 (p < 0.001). ROI: region-of-interest SMA: SMA; PCUN: precuneus.

Full size image

Distinguishing PE from NC by FC value altered in rs-fMRI

ROC curve analysis indicated that the FC between task-fMRI and rs-MRI can significantly separate PE and NC. The sensitivity was 73.08%, the specificity was 75.00%, and the AUC was 0.82 (P = 0.002) to differentiate the two groups using the FC between the precuneus and thalamus (Fig. 4C). The sensitivity was 80.77%, the specificity was 81.25%, and the AUC was 0.83 (p < 0.001) to differentiate the two groups using the FC between the inferior frontal gyrus and SMA (Fig. 4D).

Correlation between IELT scores and brain activity

The activation beta and DC values of the bilateral thalamus, bilateral inferior frontal gyrus, triangulation area, SMA, and PCUN in the LPE patients and NCs were extracted. We found that the IELT had a significantly positive correlation with the DC value at PCUN and negatively correlated with the DC value at SMA (Fig. 5A and B). Significant negative correlations existed between IELT scores and the bilateral thalami and bilateral inferior frontal gyrus, triangulation area (Fig. 5C–F).

Fig. 5
figure 5

Correlation between clinical scale score and brain activity (beta values and DC). The DC value in SMA (A) and PCUN (B) correlated with the IELT. The beta value of the bilateral thalamus and IFGtriang correlated with the IELT (C–F). IELT: intravaginal ejaculatory latency time; SMA: SMA; PCUN: precuneus; IFGtriang: inferior frontal gyrus, triangulation area; DC: DC.

Full size image

Discussion

Our study combined task-fMRI using electric stimuli on the glans penis and DC from rs-fMRI to explore aberrant intrinsic activity in LPE patients. Our results showed that LPE patients have significant impairments within several brain regions that were largely consistent with previous studies of the dopamine pathway26,37,38. In the task conditions, LPE patients showed significantly higher activation in the dopamine pathway compared to NC groups, such as the thalami and frontal gyrus. In the resting state, FC analysis showed that FC in these brain regions of the dopamine pathway and non-dopamine pathway were altered in LPE patients. ROC curve analysis indicated that using the changed FC value between task-fMRI and re-fMRI can differentiate LPE and NC. Correlation analysis revealed a significant correlation between brain functional changes and clinical behavior tests.

In this study, LPE patients exhibited hyperactivation in the nucleus accumbens, inferior frontal gyrus, and supplementary motor area, while showing reduced function in the precuneus. These abnormalities can be explained by Georgiadis et al.'s “pleasure cycle” model24, which suggests that the pathway for sexual pleasure transmission in LPE patients is shortened, leading to a faster transition to orgasm. The prefrontal cortex (which is a reward and emotional center considered important for decision-making) receives projections from the medial thalamic macronucleus. IFGtriang is located in the orbitofrontal cortex (OFC), which is a component of the prefrontal cortex39. The researchers proposed the reinforcement sensitivity theory (RST), which suggests the existence of neurobehavioural systems involved in processing appetitive stimuli40. The primary function of this system is to integrate individuals with biological rewards (e.g., sex and food). Moreover, some studies found that the brain has different responses to erotic picture stimulation in PE patients18,19,20,41 in the OFC using fMRI. The OFC, located in the dopamine reward system, is considered to have a crucial impact on assessing the sexual relevance of stimuli37,42. The thalamus, also located in the dopaminergic reward system, has been shown to process sensory information while also relaying it; each major sensory relay area receives strong feedback connections from the cerebral cortex. The higher activation in THA and the OFC indicated hypersensitivity in the dopaminergic reward system. This is consistent with previous reports indicating that dopamine has a crucial impact on mating43. Studies have demonstrated that activation of the excitatory areas of the brain by dopamine or dopamine agonists can promote ejaculation. Moreover, dopamine has been shown to affect the regulation of male sexual behavior44. This evidence may be the neural circuit foundation of the hyperexcitable ejaculatory reflex neurological factor in PE patients.

To test the relationship between task and resting state, we found that the DC value was higher in SMA and lower in PCUN. DC, as one of the graph-based measures of network organization, is used to measure the number of direct connections between specific nodes in the entire connectivity matrix of the brain and the rest of the brain45. This metric has been widely used to detect variations in the functional network at the rest state. The SMA, as part of the somatosensory association cortex, interprets touch and is involved in the perception of space and limb position46. Therefore, the SMA (part of the mirror neuron system) may contribute to directing sexual desire to behavior25. As part of the superior parietal lobule in front of the occipital lobe (cuneus), the precuneus is involved in visuospatial processing47, self-reflection48, episodic memory49, and aspects of consciousness50. In addition, the precuneus is thought to be the hub of a well-connected central "small world network" between the parietal and prefrontal regions51. Further, FC found that the SMA had a strengthened FC using the right IFGtriang and that the PCUN had a weakened FC using the bilateral thalamus. These results could indicate that the FC between the dopaminergic reward system and other brain areas was destroyed in PE patients.

The ROC curve is one of the most critical evaluation metrics to check the performance of any classification model. ROC is a probability curve that tells how well the model can distinguish between categories52. FC was found to have strengthened FC using the right IFGtriang, and the PCUN had weakened FC using the bilateral thalamus, which ROIs were found using task-fMRI and rs-fMRI. We used the altered FC value between task-fMRI and re-fMRI to classify PE and NC using significantly high accuracy. These FCs can link task-fMRI and re-fMRI together and reflect the dopaminergic reward system integrative competence to handle external information. This further indicated that the interpreted FC between the dopaminergic reward system and other brain areas could be a potential biomarker for recognizing LPE patients.

IELT (a screening tool that is one of the criteria for the diagnosis of LPE) is the time a man takes before ejaculation during vaginal penetration36,53. IELT may be related to perceptions of sexual performance and actual satisfaction. In our study, the beta values of the bilateral thalamus, bilateral inferior frontal gyrus, and triangulation area, which means the lower IELT scores the patients gain, the more hypersensitivity and hyperexcitability in the dopaminergic reward system. This is in line with the results we have found. Furthermore, in the PCUN, the DC value was positively correlated with the IELT, indicating that PE patients have a lower IELT, a lower DC value, and more activity in THA. This means that the PCUN may be an important brain region that has an inhibitory effect on THA, whereas the inhibitory effect is broken in LPE patients. Our results may indicate a relationship between disease severity and imaging biomarkers, which may help us better understand the mechanisms of PE occurrence using some simple, accessible tools.

This study had some limitations. First, the main limitation of this study is the self-reported IELT. The IELT in this study was measured at a 4-week baseline period, during which LPE patients and NCs were asked to have intercourse at least 4 times to reduce error. The stopwatch technique provides a more accurate IELT than the self-reporting system. Second, the small sample size of this study is also a limitation. Finally, it is recommended to obtain more detailed clinical and demographic data of the participants, including blood parameters and biomarkers associated with the dopamine reward system, such as serotonin (5-HT) and dopamine (DA) levels. Additionally, this study lacks detailed information on pharmacological and psychological factors, which are critical in the mechanisms of premature ejaculation (PE). Data on the use of selective serotonin reuptake inhibitors (SSRIs) could provide valuable insights into how pharmacological interventions influence neural mechanisms. Similarly, psychological factors such as anxiety and depression, which are strongly associated with PE, were not assessed. Future research should incorporate comprehensive and validated questionnaires instead of solely relying on self-reported techniques. Furthermore, integrating psychological assessments and pharmacological data would enhance the understanding of the multifaceted nature of PE and improve the reliability and interpretability of the findings.

Conclusion

Our study demonstrated hypersensitivity and hyperexcitability in the dopaminergic reward system in LPE patients by electrical stimulation using task-fMRI and alternant FC between dopaminergic reward systems and other brain areas in rs-fMRI. We hope these findings will enhance the understanding of the pathogenesis of LPE and provide new imaging biomarker approaches for future research and assist in the development of new therapies.