Women's Sexual Function and Dysfunction. Irwin Goldstein MD

Blood flow: magnetic resonance imaging and brain imaging for evaluating sexual arousal in women

Kenneth R Maravilla


It is a fundamental concept that increased blood flow to the genital structures and pelvic tissues is closely associated with and is a major contributor to the sexual arousal response (see Chapters 6.1-5.6 of this book). While this is true for both men and women, it has been particularly difficult to observe the blood flow changes in women in a nonintrusive manner (see Chapters 10.1, 10.3, and 10.4). It is even more problematic to attempt to quantify these changes. Magnetic resonance imaging has been applied in recent years as a new method to evaluate changes in tissue signal intensity that are associated with sexual arousal in women.1, 2 These signal intensity changes are a direct result of increases in local blood flow and in regional blood volume.

In the past few years, magnetic resonance techniques have been applied to observation and analysis of two aspects of the female sexual arousal response that are located in two widely separate and functionally diverse anatomic locations: the pelvic genital area and the brain. This chapter describes the basic principles of using magnetic resonance for monitoring the female sexual arousal response, describes the basic methodologies used, and outlines some initial research results using these new techniques. The first section describes changes in genital engorgement observable with magnetic resonance imaging that document and quantify the physical arousal response in women.

Since the cerebral response to a sexual stimulus is tightly coupled with genital sexual arousal, the second section addresses functional magnetic resonance imaging of the brain as a tool to monitor regions of brain activation that occur during arousal, as well as highlighting the initial experience with this exciting area of research.

Magnetic resonance imaging

Magnetic resonance imaging is a revolutionary medical imaging technique that was introduced into clinical use approximately 20 years ago. It has rapidly gained widespread acceptance because it offers several advantages over traditional medical imaging techniques such as radiography, computed tomography, ultrasound, and radioisotope scanning. For example, magnetic resonance is exquisitely sensitive to subtle changes in tissue composition, is noninvasive, and uses no ionizing radiation. It can provide dynamic assessment of tissue changes over time, and also quantitative physiologic information on areas such as three-dimensional volumes of anatomic structures, regional blood flow changes, and regional metabolite composition (magnetic resonance spectroscopy). In recent years, a few research centers worldwide, including ours at the University of Washington, have begun to explore the application of magnetic resonance techniques to improve knowledge and understanding of female sexual arousal response. Different types of magnetic resonance techniques were devised to evaluate female sexual arousal response in two separate areas of the body: the pelvic genital region and the brain. Assessments in both of these anatomic areas have their fundamental basis in observational measurements in changes of regional blood flow and regional blood volume.

Historical perspective

Until now, it has been very difficult to observe and monitor female sexual arousal response in an objective, reliable, and nonintrusive manner. Methods used in the past, such as temperature probes,3,4 vaginal photoplethysmography,5-7 and ultrasound monitoring of intravascular flow in pelvic blood vessels,8-10 provide limited information, are difficult to reproduce, and, in some cases, such as vaginal photoplethysmography (which requires vaginal insertion of a measurement probe), may confound or interfere with the physiologic response one is trying to measure. Dynamic magnetic resonance evaluation, on the other hand, can provide quantifiable information that is reproducible and may be used to compare dynamic changes that occur within subjects at different time points.1,2,11-14

Initial attempts to employ magnetic resonance to observe arousal changes in pelvic genital structures in women utilized a gadolinium-based blood pool contrast agent.1,11 Since blood flow changes and engorgement are a major factor in physical sexual arousal, it was hypothesized that the best chance of observing such changes on the magnetic resonance scan would occur if a marker of blood pool volume in arteries, veins, capillaries, and cavernous structures was enhanced with a contrast agent. A new investigational contrast agent, MS-325 (EPIX Medical, Waltham, MA, USA), was used. This magnetic resonance contrast agent remains in the vascular system with a long half-life compared with conventional, extracellular, gadolinium contrast agents that rapidly “leak” from the vascular system into the interstitial tissues of the body.15 The contrast agent MS-325 also reversibly binds to serum albumin that serves both to retain the contrast agent in the vascular system and to increase significantly the enhancement effect provided by the contrast agent.15-17 These properties enabled excellent enhancement of the genital tissues that lasted for up to several hours of observation and imaging. In particular, the erectile tissue within the clitoris and vestibular bulb was especially well enhanced. Imaging of the target structures, both before and after intravenous contrast injection, also enabled quantitative calculation of relative regional blood volume changes that occurred during arousal.The major disadvantage of this technique is that it is not totally noninvasive, since it requires injection of a magnetic resonance contrast agent with the attendant risk, albeit small, of an adverse reaction.

Therefore, alternative methods for observing female sexual arousal response were explored with magnetic resonance techniques that did not require intravenous contrast media injection. It was found that T2-weighted images with fat suppression could also provide precise anatomic detail of the genital structures.2,12 In addition, changes related to engorgement with arousal that are similar in appearance to those shown to occur with contrast-enhanced images of the clitoris could also be observed by noncontrast, T2-weighted techniques.12 Magnetic resonance imaging is exquisitely sensitive to even small changes in water content, and this is probably the reason that engorgement can be visualized on noncontrast magnetic resonance. As blood and serum are mostly water, an increase in regional blood volume that accompanies arousal and engorgement results in increased regional water content and thus an increase in T2 signal intensity along with an observable change in anatomic volume. The change in clitoral volume can be measured and provides a quantifiable means by which it is possible to compare the arousal response at different time points or across different subjects, as with the types of clitoral volume measurements done with the investigational contrast agent MS-325.2,11,12 Thus, this method provides a totally noninvasive technique to observe physical genital arousal in women. One limitation of the noncontrast magnetic resonance technique is that it is no longer possible to obtain quantitative regional blood volume measurements. However, since this has already proved to be less robust than three-dimensional, anatomic clitoral volume measurements,1 it is not felt to be a significant limitation.

Principles for evaluating the pelvic genitalia with dynamic magnetic resonance imaging

The magnetic resonance technique currently used for genital evaluation consists of obtaining high-resolution, threedimensional, T2-weighted images of the pelvic genital area. A technique is used that suppresses signal from subcutaneous and interstitial fat in this region to provide better visualization of the clitoris and adjacent genital structures. The suppression of magnetic resonance signal from fat also facilitates better visualization of signal changes within the genital structures themselves. Sequential, serial, three-dimensional images of the genital region are obtained at approximately 3-min intervals while the patient views an audiovisual presentation comprising first a 15-min neutral documentary video and then a 15-min segment of sexually stimulating content. Audiovisual stimulus is delivered to the patient in the magnet by a magnetic resonance- compatible, fiber-optic video display and headphones (Fig. 10.2.1). Changes in size and signal intensity of key genital structures caused by an increase in regional blood volume with tissue engorgement are readily observed. Images are analyzed, and quantitative, volumetric measurement of the clitoris in mm3 at each time-point in the magnetic resonance imaging series is performed.

Although several analysis techniques were explored, our early research studies showed measurement of total clitoral volume to be the most robust measure. Attempts at measuring changes in regional blood volume were also shown to be quite good but were not as robust as the clitoral volume measurements. Measurements of other structures, including analysis of vaginal mucosa, vaginal wall thickness, changes in the labia, or vestibular bulb changes, all proved either unsuccessful or far less reproducible than the clitoral volume measurements.1 High- detail (high-resolution and high signal-to-noise ratio) magnetic resonance imaging techniques are needed, since the clitoral volume is very small. In our studies, the clitoris ranged in volume from approximately 1.5 ml to 5.5 ml in the nonaroused state, increasing significantly in size with sexual arousal to a maximum of approximately 10 ml among our subject population. Thus, lower resolution imaging, employing a relatively large picture element volume (pixel) size, would result in unacceptably large standard deviations for measurements of this small structure.

Figure 10.2.1. Subject lying in the magnetic resonance magnet with head in head-imaging coil. Fiber-optically coupled goggles allow subject easily to view the video presentation while the head is being imaged.


Magnetic resonance imaging can define the normal anatomic relationships of the female genital structures in vivo. This provides a more accurate picture and a better understanding of the functional anatomy of women than cadaver dissection studies. Our studies focused mainly in the area of the introitus. Threedimensional images are acquired and displayed mainly in the axial (horizontal) plane.18 Beginning most superficially (inferior or caudally), one can define the major and minor labia along with the tip of the glans clitoris (Fig. 10.2.2). As we image slightly deeper (superiorly or rostrally), we begin to see the glans clitoris along with the frenulum, which forms the junction of the minor labia and can be seen just posterior to the base of the glans clitoris and just superficial to the body of the clitoris (Fig. 10.2.3).


Figure 10.2.2. All cross-sectional images in this chapter are oriented identically; the anterior portion of the subject's anatomy is at the top, and the right side of the subject's anatomy is at the reader's left. Image through the superficial portion of the introitus shows the major (L maj; labia majorus) and minor labia (L min; labia minora). Note the tip of the glans clitoris (GC), and the prepuce (Pr) or hood is also visible at this level.

Figure 10.2.3. Image slightly more cephelad in location now shows some of the structures slightly deeper in the introitus. The base of the glans clitoris (GC) is now visible along with the labia minora (L min), which come together at the frenulum (Fr).

Just deep to the glans clitoris is the body of the clitoris, which is only seen in the magnetic resonance imaging and is not visible upon direct visual examination of the introitus. The clitoral body is seen as a paired structure that gives rise to the two crura of the clitoris, which are also hidden from direct visual examination (Fig. 10.2.4). Beginning at about the level of the body of the clitoris are the vestibular bulbs, which are oblong-shaped structures on either side of the vaginal opening in the midline. They form a slight concavity in the center and track alongside the clitoral crura, forming an inverted V-shaped configuration (Fig. 10.2.5). O’Connell et al.,19 who has done extensive gross and microscopic studies of the female genitalia, believes that the vestibular bulbs should be considered part of the clitoris. The histologic appearance is similar and there are many vascular spaces, although fewer sensory nerve endings than in the clitoris itself. Our magnetic resonance studies support this concept, since the signal intensity changes in the bulbs roughly parallel those of the clitoris, and there is a similar, although lesser, increase in volume with arousal. The clitoris itself is a complex, wishbone-shaped structure (Fig. 10.2.6). The glans clitoris forms the head of the wishbone, followed by the body and then the two crura that diverge posteriorly on either side of the vaginal canal. The two elongated crura track along the inner and inferior surface of the ischial rami. Just posterior to the point where the two crura join to form the body of the clitoris and also where the two vestibular bulbs join are the urethra and periurethral tissues (Fig. 10.2.7).

Figure 10.2.4. Image at this level shows the paired structures that form the body of the clitoris (CB) along with the inferior-most aspects of the crura of the clitoris (CCr). Note also that the paired vestibular bulbs (Bu) that lie on either side of the vaginal opening are also visible at this level. All of these structures lie deep to the superficial tissues of the introitus and are not visible upon direct visual examination of the subject.

Figure 10.2.5. Anatomic section at this level shows the upper part of the clitoral body (CB) as it widens to form the base of the crura. The upper part of the vestibular bulbs (Bu) is well demonstrated at this level along with the vaginal opening (V).

The bulbourethral glands, also known as Bartholin’s glands, are also well visualized at this level; they are pea-shaped structures embedded in the posterolateral introitus near the posterior aspect of the vestibular bulbs. The periurethral tissues do not appear to change in signal intensity or size with arousal, so our studies do not confirm this area as the “G-spot”20-22 that some investigators believe to be associated with the sexual response of women.

Figure 10.2.6. Anatomic drawing shows the complex, three-dimensional structure of the clitoris forming a wishbone-shaped structure that lies close to and parallels the inferior surface of the pubic symphysis and the ischial rami. CB = clitoral body, CCr = clitoral crura, GC = glans clitoris.

Figure 10.2.7. At this level, one can see a good demonstration of the clitoral body (CB) and the clitoral crura (CCr) on either side. They surround vestibular bulbs (Bu) and the bulbourethral glands (BG), which are very well shown at this level. Note the position of the urethra (U) and periurethral tissues. The vaginal opening lies just posterior to this.

The vaginal mucosa appears as an H-shaped, irregular rugal pattern in most premenopausal women (Fig. 10.2.8). In the postmenopausal state, however, the vaginal mucosa shows a smoother appearance to the lining of the vaginal canal, presumably due to atrophy of the mucosa in the postmenopausal state (Fig. 10.2.9).

Magnetic resonance images show characteristic changes in key anatomic structures with engorgement during arousal. These changes are most evident in the clitoris, especially in the crura and body of the clitoris. During arousal, these structures demonstrate a very prominent increase in size and signal intensity that is easily discernible upon visual inspection of the images (Fig. 10.2.10). In addition, there are similar, although less robust, changes present within the vestibular bulbs. Obvious changes to visual observation are not present in the major and minor labia, the vaginal wall, or the vaginal mucosa, although careful anatomic measurements of the images do show a slight increase in size of the major and minor labia.23 The vaginal mucosa does not show any increase in size and signal intensity, most probably because it is only a few cell layers thick and is below the limit of resolution of the magnetic resonance imaging technique. Thus, the vaginal mucosa is not resolvable separate from the muscular wall of the vagina itself.

Initially, it was postulated that one might be able to see changes within the vaginal canal due to the watery secretions that form lubrication during arousal. However, such changes are not visible on the magnetic resonance images; this is probably due to the small amount of lubrication fluid that forms a thin film over the surface of the vaginal canal, and this thin film is also below the limit of resolution of magnetic resonance imaging.

Figure 10.2.8. Image of a premenopausal woman at the level of the vaginal canal (V). Note the irregular contours of the mucosal surface, which form an irregular rugal pattern that is characteristic of the premenopausal vaginal canal. U = Urethra and periurethral tissue.

Figure 10.2.9. Image of the vaginal canal (V) in a postmenopausal woman. Note the smooth, regular configuration of the vaginal canal in the postmenopausal state. The irregular rugal pattern has now atrophied due to the hormonal changes that occur with menopause. This smooth appearance of the vaginal opening is characteristic of the postmenopausal state in contrast to the irregular rugal pattern shown in Fig. 10.2.8. U = Urethra and periurethral tissue.

Quantitative analysis

As already stated, quantitative analysis of a number of genital structures was attempted, including measurements of the major and minor labia, vaginal wall thickness, and vestibular bulb changes. Of all the measurements attempted, the two that were found to be most robust were anatomic clitoral volume measurements and relative regional blood volume measurements. The latter set of measurements could be performed only on contrast-enhanced magnetic resonance scans, and these measurements were less reproducible than anatomic clitoral volume measurements. Thus, the primary measurement used in our studies has been comparison of changes in anatomic clitoral volume before, during, and after sexual arousal by an audiovisual sexual stimulus.

Figure 10.2.10. Images of the same subject taken at the same anatomic position illustrating the prearousal or neutral state (A) and the postarousal or stimulated state (B) after viewing of audiovisual sexual stimulus. Note the prominent change in both the size as well as the increased signal intensity of the clitoral body and crura. There is also a slight increase in signal intensity within the vestibular bulbs on either side, although the size of the vestibular bulbs is not significantly changed.

After acquisition of the magnetic resonance image series as previously described, quantitative measurements of the anatomic clitoral volume are obtained by the planometric technique to outline clitoral structures on each individual image at each time point. The various slices that contain portions of the clitoral anatomy are then summed, and a total volume of the clitoris at each imaging time point is determined. A curve of anatomic clitoral volume versus time is plotted, as illustrated in Fig. 10.2.11. An average value of all neutral time points is then compared with the average clitoral volume obtained during all stimulus time points, and a measure of percent change in clitoral volume is calculated. This provides a quantitative measure of arousal by which one can compare results within subjects as well as across different subjects and different subject populations. Results of these comparisons are presented in Table 10.2.1.

As can be seen, there is an excellent correlation within subject comparisons at two different imaging time points, with an r value equal to 0.95 (Fig. 10.2.12), and there is more variability across subjects than within subjects. The average increase in clitoral volume measured approximately 90% when compared with a baseline average for all subjects. Furthermore, in our various studies, there was also good correlation between clitoral volume changes observed at two different time-points within each subject in a premenopausal group compared with that seen in a postmenopausal group. The numbers of subjects in each group, however, are very small at this point in time, so these early results should be interpreted with caution.

Figure 10.2.11. Plot of anatomic clitoral volume versus time shows the time response curve of a typical, normal subject during the neutral video phase (time points labeled N) with the rapid increase in clitoral volume during the sexual stimulus portion of the video (time points marked E).

Figure 10.2.12. Correlation curve plotted from the data in Table 10.2.1 shows the high correlation coefficient of 0.95 obtained when comparing the anatomic clitoral volume change obtained with arousal at two different time points within each subject.

The level of sexual arousal achieved in each subject was documented by a validated sexual arousal score question- naire.13,14 This questionnaire used a seven-point scale scoring system and presented queries that probed for overall feelings of sexual arousal, mental sexual arousal, and feelings of physical sexual arousal. Scores from the various questions were totaled, and an average arousal score between 1 (no feelings of sexual arousal) and 7 (most intense feelings of arousal) was calculated. Results from one group of subjects are shown in Table 10.2.1. While subjective, these scores provided a reasonable level of documentation regarding subject arousal status. These scores were also correlated with the measurement of percent change in clitoral volume, where it was found that there was a reasonably good correlation, with an r score greater than 0.7 (Fig. 10.2.13).

To date, all of our published data have involved evaluation of healthy, volunteer, pre- and post-menopausal women without reported sexual difficulties. We have initiated studies utilizing these magnetic resonance techniques to evaluate women with selected types of sexual arousal disorder, and these studies are underway. We are exploring the possibility that differences in the pattern and quantitative measurements of the arousal response and/or the degree of reproducibility within subjects may provide valuable information that would help to clarify the underlying dysfunctional physiology, and eventually may help to design and validate better treatment methods. These techniques also provide a possible basis for monitoring response to therapy. This should also prove useful in testing new drugs or other therapies to treat patients with sexual arousal disorders and to determine the effectiveness of proposed new therapies.

Table 10.2.1. Comparison of clitoral volume and arousal score data obtained at two time points for a normal volunteer group of women (n = 7)




Session 1


Session 2


CV (ml) Neutral/Erotic

% CV Increase



CV (ml) Neutral/Erotic

% CV Increase




















































CV = clitoral volume.

Figure 10.2.13. Correlation coefficient curve derived for a set of subjects listed in Table 10.2.1 shows a correlation between the average clitoral change in volume and the subjective arousal score. There was moderately good correlation with an r value of approximately 0.70.

Functional brain imaging during sexual arousal


In its most basic definition, functional brain imaging is the use of rapid, dynamic magnetic resonance imaging to observe anatomic sites of brain activation in response to a specific task performed by a subject. This technique is often referred to as “functional magnetic resonance imaging”, and in the few short years since its introduction, it has gained widespread acceptance in clinical and experimental brain imaging research. First applied as a method to define eloquent areas of the brain, such as activation of motor or visual cortex, the technique has been refined and become more sensitive and is now being applied to define brain activation that occurs in association with subtle, cognitive tasks. Thus, this technique is an ideal tool for exploring brain function associated with emotional attraction, feelings of pleasure, and even sites of activation associated with sexual arousal.

Figure 10.2.14. Functional magnetic resonance imaging (fMRI) (A) and time response signal intensity curve (B) illustrating the fMRI activation seen in association with tapping of the right fingers. (A) Axial image through the upper portion of the brain shows the area of activation in the left motor cortex highlighted in red. (B) Signal intensity time response curve shows the changes in signal intensity (irregular white line) and their correlation with the activation of rest versus finger tapping of the right hand (solid blue line). Note that the change in signal intensity correlates very nicely with the period of time that the subject is tapping the fingers of the right hand. Note also that the amount of signal change is very slight and measures only approximately 2% of the total signal intensity. This small amount of signal intensity change is not visible on visual inspection of the images but can only be discerned on accurate statistical analysis of the images.

Theory of method and techniques

Functional brain imaging is most easily understood by illustrating what happens when a subject performs a simple motor task. If a subject taps the fingers of one hand and functional magnetic resonance imaging images are obtained during that task, these images can be analyzed to demonstrate the area of brain activation in the contralateral motor cortex that is responsible for performance of this motor task (Fig. 10.2.14). However, if one looks at a single echo planar functional magnetic resonance image, the change in the magnetic resonance image is very slight, and there is no visual indication of this activation (Fig. 10.2.15). So how does one identify the area(s) of motor activation, and why does this work?

There are several key points to be aware of regarding functional magnetic resonance imaging. The first is that functional magnetic resonance imaging is based on changes in regional cerebral blood flow that occur when focal areas of the brain are activated. The second is that with changes in brain activity there are also local blood flow changes within the same region of the brain. These local blood flow changes result in a slight change in magnetic resonance signal intensity that is very subtle and is not discernible by eye on visual inspection of a single image. Third, the areas that are activated are identified by statistical analysis of a series of images that are rapidly acquired during rest and during performance of a task. By statistical comparison of the magnetic resonance signal pattern in a set of dynamic images acquired during a resting state with those acquired during performance of a task (finger tapping), the area of activation in the motor cortex can be identified.

Figure 10.2.15. Typical axial echo planar image of the brain done during performance of a functional activation task does not show any obvious areas of signal intensity change. As already stated, only detailed statistical analysis of an entire sequence of dynamic images can show the subtle changes that occur with brain activation.

Theory of activation

The functional magnetic resonance imaging technique is based upon changes that occur with blood flow due to changes that occur in the level of blood oxygenation. These changes in signal intensity that occur with changes in oxygen level are referred to as blood oxygen level-dependent tissue contrast. These changes were first defined by Ogawa,24 who demonstrated differences between images of oxyhemoglobin and deoxyhemoglobin. Blood containing oxyhemoglobin, that is, hemoglobin that is fully oxygenated, is diamagnetic and has a neutral effect on the magnetic resonance signal derived from oxygenated blood when compared with surrounding tissues. Deoxyhemoglobin, which represents the state of hemoglobin when oxygen has been extracted, is a paramagnetic substance and has a high magnetic susceptibility relative to surrounding tissues, leading to loss of signal intensity on T2-weighted (susceptibility-weighted) magnetic resonance imaging sequences. Thus, in the normal resting state, oxygenated blood entering the brain has a slightly higher baseline signal intensity, while blood in the venous system of the brain, because it has had much of the oxygen extracted as it passed through the capillary bed, has a subtly lower signal intensity.

Let us now consider what happens in the motor cortex when a subject begins to tap fingers. In this case, the neurons in the cortex become activated and begin to fire rapidly in order to produce motion in the fingers. With activation of the neurons, there is an increased oxygen demand in order to supply energy to the neurons. This increased oxygen demand is met by a signal from the brain to increase blood flow to that local area of the cortex with the neurons that have become active. This increased flow brings increased oxyhemoglobin. However, the mechanism described many years ago by Fox et al.,25 termed “decoupling”, results in an excess of blood flow to the area of neuronal or brain activation. That is, the amount of increased blood going to the tissues is greater than what is needed to supply the increased oxygen requirements of the active neurons and, thus, an excess of oxyhemoglobin is delivered. The venous blood draining from the area of activation then carries both the deoxygenated hemoglobin that has supplied oxygen to the neurons and the excess oxyhemoglobin that was not needed by the tissues. This results in a local decrease in the concentration of deoxyhemoglobin during the activation state when compared with the resting state. Since deoxyhemoglobin normally causes a mild decrease in signal intensity on susceptibility-weighted (T2) images, the decreased concentration in deoxyhemoglobin results in a slight increase in signal intensity on these activated images compared with the resting (baseline) state. The magnitude of this signal increase is very small, but it is reproducible and generally measures 1-5% of baseline signal intensity. Thus, such a small change in intensity is not detectable by merely looking at the magnetic resonance images.

However, if one were to subtract an image acquired during the resting state with the subject lying quietly with both hands by their sides from an image acquired during the activation state when the subject was tapping their fingers, we would be able to see a very slight signal increase in the motor cortex of the opposite hemisphere responsible for the finger tapping. In practice, however, rather than doing a simple image subtraction, a rapid series of images is acquired over a few minutes and then statistically analyzed to show this subtle change in signal that occurs over time.

Each magnetic resonance image is a digital image that is composed of a number of picture elements, or pixels. A pixel is a single, digitized “dot” that has a measurable gray level, that is, a certain level of whiteness or blackness that contributes to the picture. Pictures are made up of a matrix of pixels that is defined by the number of pixels on the longitudinal axis versus the number of pixels on the vertical axis. Thus, a 256 x 256 image is composed of 256 pixel rows in the vertical direction and 256 pixel columns in the longitudinal or horizontal direction. The larger the number of pixels in the matrix, the higher the resolution of the image and the more detail it can provide (Fig. 10.2.16). Unlike a digital picture taken with a camera, which is a two-dimensional representation of the surfaces that reflect light, in magnetic resonance each image of the brain is a “slice” though the brain of a certain thickness - generally 5-10 mm. Because of this slice thickness or depth of the tissue, each picture element is more often referred to as a “voxel” (the threedimensional equivalent of a pixel).

Rapid dynamic images called echoplanar images are acquired of the entire brain and then repeated every 2 or 3 s for the time required to perform a repetitive task (such as finger tapping) that is alternated with a control time interval (rest) - usually about 2-5 min. During these few minutes of echoplanar image acquisition, the subject alternately lies still and performs the finger-tapping task at specified intervals in response to a verbal or visual signal. Each voxel in each image of this series of images is then statistically analyzed to determine voxels that change in signal intensity over time. In comparing voxels that change signal intensity with the periods of finger tapping and rest, only those voxels that show a pattern of signal change that matches or correlates with the times of task performance are selected or highlighted (Fig. 10.2.16B). This statistical image- analysis technique is referred to as a correlation coefficient, and is far more powerful than a simple subtraction of two images, since it averages signal intensity from a large number of images to reduce interference from the random signal fluctuation due to background noise that normally occurs with all magnetic resonance images.

Once the voxels responsible for finger tapping are statistically identified, they can be highlighted in color and their location superimposed (or mapped) over a detailed anatomic magnetic resonance image of the brain acquired at the same magnetic resonance session with the subject’s head in the same position (Fig. 10.2.16A). This allows the specific anatomic areas associated with the activation task (finger tapping) to be visually displayed, and produces the resulting functional magnetic resonance imaging image.

These same principles can be applied to cognitive tasks.

However, in general, the amount of associated signal change is greater for motor tasks than for cognitive tasks. In addition, while eloquent areas of brain associated with motor function, vision, or speech production are well defined and well localized, most cognitive functions are widely dispersed across multiple areas of the brain that act in association with each other in complex, and often incompletely understood, patterns to produce the cognitive task. Thus, the combination of very low signal change and widely distributed areas of activation makes it difficult to analyze and interpret functional magnetic resonance imaging scans done for complex cognitive tasks. Nevertheless, through a painstaking and methodical series of research experiments, scientists are slowly unraveling some of the mysteries of the brain that are associated with learning, problem solving, and emotional responses.

There are several limitations to functional brain imaging that the reader should be aware of. The most problematic of these are artifacts caused by head motion. Since each voxel in a series of images is compared over time, even small amounts of head motion in the range of 1-2 mm during the functional magnetic resonance imaging acquisition session can result in severe artifacts or completely uninterpretable results. In addition, air has a very high magnetic susceptibility difference from brain tissue. Thus, artifacts from local areas of magnet field disturbance at the skull base caused by air in the adjacent paranasal sinuses may also produce uninterpretable results due to local areas of signal void on the image (Fig. 10.2.17). Air, like deoxyhemoglobin, has high magnetic susceptibility differences from adjacent tissues, leading to this loss in signal. Fortunately, most of the brain is well away from the areas affected by this magnetic field disturbance. However, tasks related to the inferior frontal lobes of the brain adjacent to large frontal and ethmoid sinuses or the inferior portions of the posterior temporal lobes in subjects with large mastoid sinuses can be affected. Disturbances in magnetic field leading to image distortion and/or signal loss can also be caused by small amounts of metal in the skin, hair, or teeth. Past traumatic injuries occasionally result in small fragments of embedded metal, or a subject may forget to remove a hair ornament or earring. While most dental fillings do not interfere with magnetic resonance, there are a few - especially certain types of root canal device or bridge work - that can interfere with the images.

Finally, paradigm design is a major consideration that must be carefully formulated in order to produce the intended result. The brain is a very complex organ and responds in multiple different ways. Trying to isolate a specific function without interference from other tasks that may be performed simultaneously by the brain is a difficult undertaking.

Results from functional magnetic resonance imaging studies

There have been a number of studies applying functional magnetic resonance imaging techniques to the evaluation of sexual arousal or emotional response in both men and women. Initial imaging studies analyzing the functional neuroanatomic correlates of visually evoked sexual arousal in human males were performed by positron emission tomography scanning, and results from these studies were first reported by Stoleru et al.26 In these studies, the brain activation response in eight male subjects to sexually explicit video clips was compared with emotionally neutral control film clips and humorous control film clips. The researchers found visually evoked areas of cerebral activation that correlated with the sexual arousal response in the posterior temporal-occipital cortices (a visual association area); the right insula and right inferior frontal cortex (paralimbic areas related to processing of sensory information with motivational states); and the left anterior cingulate cortex, another paralimbic area known to control autonomic and neuroendocrine functions. Thus, this study identified for the first time the ability to define brain regions whose activation was associated with visually evoked sexual arousal in men.

Figure 10.2.16. Illustration showing changes in resolution that occur with increasing matrix size (decreasing size of imaging pixels): (A) 16 x 16, (B) 32 x 32, (C) 64 x 64, (D) 128 x 128 (continued overleaf)

Figure 10.2.16. (continued) (E) 256 x 256 matrix sizes, respectively. Note that there is no usable anatomic information at matrix sizes of 16 or 32. Limited resolution is obtained at matrix sizes 32 and 64 and improves at levels of 128, but is best shown on a 256 x 256 matrix.

Park et al.,27 using blood oxygen level-dependent functional magnetic resonance imaging techniques, such as those described above, published the first study evaluating regions of cerebral activation associated with the female sexual arousal response. This group studied six healthy, premenopausal women volunteers with a mean age of 33. Using a real-time video presentation with alternatively presented erotic and nonerotic film clips, they were able to identify several brain regions activated in association with the sexual arousal response. These areas included the inferior frontal lobes, anterior cingulate gyrus, insula, corpus callosum, thalamus, caudate nucleus, globus pallidus, and inferior temporal lobes. Thus, they were able to demonstrate that functional magnetic resonance imaging can also show sexual arousal response within the brain in women undergoing visual sexual stimulus.

Figure 10.2.17. Coronal image of the brain at the level of the sphenoid sinus. (A) Anatomic image (B) echoplanar image obtained for fMRI study. Note the prominent loss of signal at the base of the brain surrounding the area of the sphenoid sinus (arrows) in image (B). This is due to the high susceptibility difference of air in the sphenoid sinus compared with the surrounding brain tissue as described in the text.

A group from Stanford University led by Arnow28 used a novel paradigm and analysis method to define areas of brain activation associated with sexual arousal in healthy, heterosexual males. Rather than correlating the brain response associated with sexual arousal simply with the time of presentation of neutral versus erotic film clips, these researchers used a novel technique to correlate also the cerebral arousal response with erection. To do this, they constructed a device that measured the presence and degree of penile turgidity while the subject was in the magnetic resonance machine during presentation of the video material (alternating arousing and nonarousing video clips) while collecting functional magnetic resonance imaging data. They subsequently analyzed the data by two statistical methods: the traditional cross-correlation analysis already described that uses the time of presentation of the neutral versus arousing video clips, and a second cross-correlation using penile turgidity measures as the covariant. The team found a stronger activation response with the penile turgidity method than with the timing of the film clips, probably because penile turgidity correlated more precisely with the actual onset of arousal. Nevertheless, they defined areas of activation that were similar to several previous studies that included the right insular region, claustrum, left caudate, right middle occipital and middle temporal gyri, and bilateral cingulate gyri.

Another interesting study was reported by Karama et al.29 These authors compared the responses of men and women during viewing of erotic video film clips. This study is unique in that it used identical techniques and the same team of researchers to study groups of 20 male and 20 female subjects. This provided direct comparison of the cerebral responses in men and women. Their results demonstrated that the observed levels of cerebral activation that correlated with sexual arousal, as measured by the blood oxygen level-dependent functional magnetic resonance imaging technique, was significantly higher in the male group of subjects than the female group of subjects. They also found that there were many similar areas of brain activation between both genders, including the anterior cingulate, prefrontal, orbital frontal, insular, and occipitotemporal cortices, as well as the amygdala and the ventral striatum in both groups. They did find, however, that activation of the thalamus and hypothalamus was significantly greater in male subjects. Furthermore, the greater levels of arousal observed in male subjects were also correlated with greater levels of sexual arousal among the male subjects as reported by subjective questionnaire.

Our group also studied a group of 10 healthy, premenopausal women without reported sexual difficulties. Analysis was done of the blood oxygen level-dependent functional magnetic resonance imaging studies, and two types of analyses were performed. The first analysis was the conventional blood oxygen level-dependent activation analysis to define areas of positive (increased) activation associated with the sexual arousal response. A second analysis was done to assess areas of the brain that showed decreased activation during sexual arousal. It was found that positive activation occurred in association with sexual arousal in many brain areas previously reported by several other groups. These areas include the orbital frontal region, the amygdala, the right caudate, the anterior cingulate, and the posterior temporal-occipital cortex on both sides (Fig. 10.2.18). Analysis of areas showing decreased activation revealed that there were localized areas of decreased regional cerebral blood flow during sexual arousal in both temporal lobes, predominately in the superior and middle temporal gyri. Additionally, the areas of decreased activation on the right were significantly greater than those observed on the left (Fig. 10.2.19). Interpretation of these data supports the hypothesis that the temporal lobe areas represent areas of normally active inhibition that underwent a decrease in their level of activation during the sexual arousal response. In support of this theory, similar anatomic sites in the superior and middle temporal gyri on the right side have been associated with moral judgment in previous cognitive functional magnetic resonance imaging studies.30-32

While the preceding studies demonstrate that the functional magnetic resonance imaging, blood oxygen level- dependent technique can define areas of activation associated with the sexual arousal response, this suggests that application of this technique could answer several questions about the cerebral component of the sexual arousal response. By defining normal areas of activation, one can begin to define the multiple associated cerebral components of the sexual arousal response. Through comparison of these arousal-related regions with other functional magnetic resonance imaging studies of the brain, we can begin to construct a model of the normal cerebral sexual arousal response and the various components that are associated with it. One can also begin to see the similarities and differences between male and female brain responses. It is interesting, although not unexpected, that early studies suggest that the major regions of activation are similar for both men and women, although there is a greater magnitude of response in males. Thus, these techniques show promise of eventually providing a better understanding of previously elusive information regarding the brain components related to sexuality.

Figure 10.2.18. Axial functional magnetic resonance imaging of brain showing results of a group analysis of 10 normal subjects that demonstrate areas of activation. Specific areas of activation are highlighted in red and their anatomic location is discussed in the text.

Figure 10.2.19. Functional magnetic resonance imaging study showing results of a group analysis of areas of decreased levels of activation during sexual arousal response. Note the dominant areas of decreased activation present in both temporal lobes. The decreased level of activation on the right is significantly greater than that seen on the left.

Of even greater importance, however, is the fact that these techniques may prove useful to analyze possible new methods of treatment, whether they are pharmacologic treatments or new approaches to psychotherapy. For example, a study done by Montorsi et al.33 analyzed brain activation patterns in men during video sexual stimulation both before and after administration of apomorphine, a known stimulator of the erectile process. The role of apomorphine in initiating the erectile process has been defined in animal models, though there has not been any direct data confirming its effectiveness in humans. The study by Montorsi utilized sublingual apomorphine versus a placebo control agent to study 10 male patients with psychogenic arousal (erectile) dysfunction. Their response was assessed by functional magnetic resonance imaging of the brain during visual stimulation by a video of alternating neutral and erotic content. Six volunteers with no history of erectile dysfunction were used as controls. In the patients with psychogenic erectile dysfunction, the sublingual apomorphine produced an increase in extent of the activated networks of the brain plus additional activation in deep structures, including the nucleus accumbens, hypothalamus, and mesencephalon, that was greater than in the placebo control group of patients. In addition, the pattern of increased responses among the apomor- phine group of patients more closely resembled the functional magnetic resonance imaging response pattern seen in control subjects without erectile dysfunction. This study demonstrates the potential utility of functional magnetic resonance imaging techniques to help confirm the positive response to a pharmacologic therapy as well as to define possible anatomic sites within the brain at which the pharmacologic therapy acts to produce its effect.

Finally, another area of functional magnetic resonance imaging activity that is helping to define normal and abnormal brain responses to sexual arousal has been explored by Komisaruk et al.34 This group studied the brain response to vaginal-cervical self-stimulation in women with complete spinal cord injury. These researchers were able to define the ability of women with a complete spinal cord injury to achieve orgasm by self-stimulation. They found that cervical self-stimulation increased activity in the region of the nucleus of the solitary tract, which is the brainstem nucleus to which the vagus nerves project. Their findings suggest that the vagus nerve can convey genital sensory input directly to the brain in women with spinal cord injury, thus completely bypassing the spinal cord injury.


Magnetic resonance imaging both in the genitalia and in the brain is a powerful tool that can provide unique information unobtainable by any other method. Its exquisite sensitivity and total noninvasiveness allow us to apply this technique in large numbers of subjects and even to perform repetitive serial studies without difficulty. Our understanding of the pelvic genital responses and the ability to quantify certain measures related to engorgement in women may eventually lead to a better understanding of the genital physiology during sexual arousal. This, in turn, may enable discoveries for new avenues of research into treatment of sexual dysfunction and to evaluate the effect of new therapies to improve genital sexual response in women.

Understanding of brain function associated with sexual arousal and response represents a truly unique and exciting area of application of functional magnetic resonance imaging . This previously mysterious and inaccessible region is now open to research and analysis by large numbers of cognitive brain scientists and human sexuality researchers. No doubt these and similar techniques will be applied in as yet unimagined ways to answer questions about normal human sexuality and dysfunction. The hope is that such increased understanding will inevitably lead to better methods of treatment. Thus, these exciting technologies promise to improve the quality of life for many in the coming years.


1. Deliganis AV, Maravilla KR, Heiman JR et al. Female genitalia: dynamic MR imaging with use of MS-325 initial experiences evaluating female sexual response. Radiology 2002; 225: 791-9.

2. Maravilla KR, Heiman JR, Garland PA et al. Dynamic MR imaging of the sexual arousal response in women. J Sex Marital Ther 2003; 29(Suppl 1): 71.

3. Levin RJ, Wagner G. Orgasm in women in the laboratory - quantitative studies on duration, intensity, latency, and vaginal blood flow. 

4. Wagner G, Levin RJ. Effect of atropine and methylatropine on human vaginal blood flow, sexual arousal and climax. Acta Pharmacol Toxicol (Copenh) 1980; 46: 321-5.

5. Beck JG, Barlow DH, Sakheim DK. Operating characteristics of the vaginal photoplethysmograph. Arch Sex Behav 1984; 13: 595-6.

6. Henson DE, Rubin HB, Henson C. Labial and vaginal blood volume responses to visual and tactile stimuli. Arch Sex Behav 1982; 11: 23-31.

7. Laan E, Everaerd W. Physiological measures of vaginal vasocon- gestion. Int J Impot Res 1998; 10(Suppl 2): S107.

8. Farquhar CM, Rae T, Thomas DC et al. Doppler ultrasound in the nonpregnant pelvis. J Ultrasound Med 1989; 8: 451-7.

9. Lavoisier P, Aloui R, Schmidt MH et al. Clitoral blood flow increases following vaginal pressure stimulation. Arch Sex Behav 1995; 24: 37-45.

10. Taylor KJ, Burns PN, Woodcock JP et al. Blood flow in deep abdominal and pelvic vessels: ultrasonic pulsed-Doppler analysis. Radiology 1985; 154: 487-93.

11. Maravilla KR, Cao Y, Heiman JR et al. Serial MR imaging with MS-325 for evaluating female sexual arousal response: determination of intrasubject reproducibility. 12. Maravilla K, Cao Y, Heiman J et al. Noncontrast dynamic magnetic resonance imaging for quantitative assessment of female sexual arousal. J Urol 2005; 173: 162-6.

13. Meston CM, Heiman JR. Ephedrine-activated physiological sexual arousal in women. Arch Gen Psychiatry 1998; 55: 652-6.

14. Heiman JR. A psychophysiological exploration of sexual arousal patterns in females and males.15. Grist TM, Korosec FR, Peters DC et al. Steady-state and dynamic MR angiography with MS-325: initial experience in humans. Radiology 1998; 207: 539-44.

16. Lauffer RB, Parmelee DJ, Ouellet HS et al. MS-325: a small- molecule vascular imaging agent for magnetic resonance imaging. Acad Radiol 1996; 3(Suppl 2): S356.

17. Prasad PV, Cannillo J, Chavez DR et al. Contrast-enhanced MR angiography and first-pass renal perfusion imaging using MS-325, an intravascular contrast agent. Acad Radiol 1998; 5(Suppl 1): S219.

18. Suh DD, Yang CC, Cao Y et al. Magnetic resonance imaging anatomy of the female genitalia in premenopausal and postmenopausal women.  j ural           

19. O’Connell HE, Hutson JM, Anderson CR et al. Anatomical relationship between urethra and clitoris.j ural

20. Grafenberg E. The role of the urethra in female orgasm. Int J Sexol 1950; 3: 145-8.

21. Goldberg DC, Whipple B, Fishkin RE et al. The Grafenberg spot and female ejaculation: a review of initial hypotheses. J Sex Marital Ther 1983; 9: 27-37.

22. Whipple B, Perry JD. The G-spot: a modern gynecologic myth. Am J Obstet Gynecol 2002; 187: 519; reply, 520.

23. Suh DD, Yang CC, Cao Y et al. MRI of female genital and pelvic organs during sexual arousal. J Psychosom Obstet Gynecol 2004; 25: 153-62.

24. Ogawa N. [Central acetylcholinergic systems in the normal aged and in the patient with Alzheimer-type dementia (ATD)]. Rinsho Mintun MA, Raichle ME et al. A noninvasive approach to quantitative functional brain mapping with H2 (15)O and positron emission tomography. J Cereb Blood Flow Metab 1984; 4: 329-33.

26. Stoleru S, Gregoire MC, Gerard D et al. Neuroanatomical correlates of visually evoked sexual arousal in human males. Arch Sex Behav 1999; 28: 1-21.

27. Park K, Kang HK, Seo JJ et al. Blood-oxygenation-level-dependent functional magnetic resonance imaging for evaluating cerebral regions of female sexual arousal response. Urology 2001; 57:1189-94.

28. Arnow BA, Desmond JE, Banner LL et al. Brain activation and sexual arousal in healthy, heterosexual males. Brain 2002; 125: 1014-23.

29. Karama S, Lecours AR, Leroux JM et al. Areas of brain activation in males and females during viewing of erotic film excerpts. Hum Brain Mapp 2002; 16: 1-13.

30. Greene JD, Nystrom LE, Engell AD et al. The neural bases of cognitive conflict and control in moral judgment. Neuron 2004; 44: 389-400.

31. Heekeren HR, Wartenburger I, Schmidt H et al. An fMRI study of simple ethical decision-making.  2003; 14: 1215-19.

32. Moll J, de Oliveira-Souza R, Bramati IE et al. Functional networks in emotional moral and nonmoral social judgments. Neuroimage 2002; 16: 696-703.

33. Montorsi F, Perani D, Anchisi D et al. Brain activation patterns during video sexual stimulation following the administration of apomorphine: results of a placebo-controlled study. Eur Urol 2003; 43: 405-11.

34. Komisaruk BR, Whipple B, Crawford A et al. Brain activation during vaginocervical self-stimulation and orgasm in women with complete spinal cord injury: fMRI evidence of mediation by the vagus nerves.