- Email: [email protected]
Motor Cortical Representation of the Pelvic Floor Muscles
Motor Cortical Representation of the Pelvic Floor Muscles A. Schrum, S. Wolff, C. van der Horst and J. P. Kuhtz-Buschbeck* From the Institute of Physiology, Christian-Albrechts-Universität (AS, JPKB), and Neuroradiology/Department of Neurosurgery (SW), and Departments of Urology and Pediatric Urology (CvH), Universitätsklinikum Schleswig-Holstein, Kiel, Germany
Purpose: Pelvic floor muscle training involves rhythmical voluntary contractions of the external urethral sphincter and ancillary pelvic floor muscles. The representation of these muscles in the motor cortex has not been located precisely and unambiguously. We used functional magnetic resonance imaging to determine brain activity during slow and fast pelvic floor contractions. Materials and Methods: Cerebral responses were recorded in 17 healthy male volunteers, 21 to 47 years old, with normal bladder control. Functional magnetic resonance imaging was performed during metronome paced slow (0.25 Hertz) and fast (0.7 Hertz) contractions of the pelvic floor that mimicked the interruption of voiding. To study the somatotopy of the cortical representations, flexion-extension movements of the right toes were performed as a control task. Results: Functional magnetic resonance imaging during pelvic floor contractions detected activity of the supplementary motor area in the medial wall and of the midcingulate cortex, insula, posterior parietal cortex, putamen, thalamus, cerebellar vermis and upper ventral pons. There were no significant differences in activation between slow and fast contractions. Toe movements involved significantly stronger activity of the paracentral lobule (ie the medial primary motor cortex) than did the pelvic floor contractions. Otherwise the areas active during pelvic floor and leg muscle contractions overlapped considerably. Conclusions: The motor cortical representation of pelvic floor muscles is located mostly in the supplementary motor area. It extends further ventrally and anteriorly than the representation of distal leg muscles. Key Words: motor cortex, magnetic resonance imaging, pelvic floor PELVIC floor muscle training consists of repetitive voluntary contractions that strengthen the striated urethral sphincter and other pelvic floor muscles, eg in the treatment of stress urinary incontinence. Training for several months results in neuroplastic changes in the accompanying brain activity.1 However, the physiological representation of pelvic floor muscles in the motor cortex has not yet been pinpointed unequivocally. The classic homunculus diagram of Penfield and Rasmussen does not explicitly indicate the region of the pelvic floor in
the primary motor cortex,2 whose location may be assumed to lie between the leg and trunk representations (“hip motor region”) in the superomedial precentral gyrus. Several neuroimaging studies of brain activity during voluntary contractions of the pelvic floor had divergent results.3–7 The seminal PET study of Blok et al reported activity in the superomedial and superolateral M1.3 Two fMRI studies revealed predominant activity in the supplementary motor area, whereas involvement of M1 was negligible.4,5 Seseke et al observed activity
0022-5347/11/1861-0185/0 THE JOURNAL OF UROLOGY® © 2011 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION
Vol. 186, 185-190, July 2011 Printed in U.S.A. DOI:10.1016/j.juro.2011.03.001
AND
RESEARCH, INC.
Abbreviations and Acronyms BA ⫽ Brodmann area EMG ⫽ electromyogram fMRI ⫽ functional magnetic resonance imaging FP ⫽ fast repetitive contractions of the pelvic floor FT ⫽ fast flexion-extension movements of the right toes FWE ⫽ family wise error M1 ⫽ primary motor cortex PET ⫽ positron emission tomography SMA ⫽ supplementary motor area SP ⫽ slow repetitive contractions of the pelvic floor ST ⫽ slow flexion-extension movements of the right toes Submitted for publication November 2, 2010. Study received local ethics committee approval. Supported by Deutsche Forschungsgemeinschaft (Grant KU 1937/2-1). Supplementary material for this article can be obtained at http://www.physiologie.unikiel.de/homepages_mitarbeiter/kuhtzHomepage MotorControl.pdf. * Correspondence: Institute of Physiology, Christian-Albrechts-Universität, Olshausenstr. 40, D-24098 Kiel, Germany (telephone: ⫹49 431 8803657; FAX: ⫹49 431 8804580; e-mail: [email protected]).
See Editorial on page 14. For another article on a related topic see page 334.
www.jurology.com
185
186
MOTOR CORTICAL REPRESENTATION OF PELVIC FLOOR MUSCLES
of the superolateral convexity of the sensorimotor cortex and additional involvement of the SMA.6,7 Some differences may be due to nonuniform experimental paradigms (ie different pelvic floor exercises). Possibly fast and short willful muscle contractions engage other brain regions than slow sustained contractions. Therefore, the present fMRI study examines brain activity during both types of voluntary contractions of pelvic floor muscles in healthy volunteers. As a control task, flexion-extension movements of the toes were performed to delineate the somatotopic layout of the neighboring cortical representation of the foot. Knowledge of the normal somatotopy is needed for the understanding of neuroplastic changes induced by pelvic floor muscle training in patients.1
Figure 1. EMG during fast and slow pelvic floor muscle contractions. Bipolar surface electrodes were used. Recordings were made of pelvic floor muscles (PF, electrodes on perineum) and of right leg muscles including flexor hallucis (FH), rectus femoris (RF) and gluteus maximus (GL). Black bars indicate sounds of metronome (Mt) that paced contractions.
METHODS A total of 17 right-handed healthy men with a mean age (⫾ SD) of 28.9 ⫾ 7.1 years old (range 21 to 47) volunteered to participate. The study received local ethics committee approval. Participants were staff members or medical students who were familiar with the anatomy and function of the pelvic floor. None of the subjects used medication that might have influenced bladder or brain activity. None had symptoms of incontinence or urinary tract infection according to their anamnesis and a questionnaire.8 Participants performed slow (SP) and fast (FP) repetitive contractions of the pelvic floor which mimicked interruption of voiding, and slow and fast flexion-extension movements of the right toes (ST, FT). Fast contractions were performed at 0.7 Hz (every 1.4 seconds), paced by brief sounds (0.2 seconds) of a metronome. Slow contractions occurred at 0.25 Hz (every 4 seconds). The slow contraction-relaxation cycles followed a sound that gradually increased and decreased in loudness. All contractions were performed with moderate effort (approximately 25% of maximum force) to avoid fatigue. The tasks were first practiced and EMG recordings of muscle activity were performed outside the magnetic resonance tomography scanner (fig. 1). Subjects emptied their bladder before neuroimaging started so that scanning was performed in a no desire to void condition. A 3 Tesla (Philips Achieva®) tomograph with a phased array head coil (Sense 8-channel) was used. The subjects lay still in a supine position with eyes closed. Head motion was restricted by foam pads. T2-weighted echo planar image volumes with blood oxygenation level dependent contrast (echo time 34.5 milliseconds, flip angle 90 degrees, field of view 224 mm, pixel size 3.5 mm) were collected continually every 2.5 seconds during 4 imaging runs. Each volume consisted of 35 axial slices (3 mm slice thickness, 0.3 mm interslice gaps) that included the brain from the vertex to the cerebellum. Fast (20 per period) and slow (7 per period) repetitive contractions of the pelvic floor (FP, SP) or of the right foot muscles (FT, ST), respectively, were performed in 4 separate runs in randomized order. Each run (144 image volumes) comprised 6 contraction periods of 28 seconds each, which alternated with
28-second periods of rest. Metronome sounds were audible during task and rest periods, and tasks were announced via headphones. Additional T1-weighted anatomical images with a high resolution were obtained from each subject. fMRI data were analyzed with SPM5 statistical parametric mapping software (www.fil.ion.ac.uk/spm). The image volumes were realigned and standardized to the stereotaxic space of the Montreal Neurological Institute brain template, sampled to a voxel size of 2 ⫻ 2 ⫻ 2 mm3 and smoothed with a 6 ⫻ 6 ⫻ 6 mm3 isotropic Gaussian filter. Differences in global brain activity among subjects were removed by proportional scaling. Blood oxygen level dependent responses during the 4 contraction tasks (FP, SP, FT, ST) were modeled with boxcar functions (block design) that had been convolved with the hemodynamic response function. Fast and slow contractions were compared (FP-SP, FT-ST). Activations during pelvic floor contractions and toe movements were contrasted ([FP⫹SP] minus [FT⫹ST] and vice versa). The contrast images of all 17 participants were entered into a random effects second level analysis. Statistical inferences were drawn at the voxel level with p ⬍0.05 corrected for multiple comparisons (controlling family wise error rate, FWE corrected). We described Montreal Neurological Institute spatial coordinates as given by SPM5. Active brain regions were superimposed on a mean normalized anatomical image of all 17 subjects. Published probabilistic cytoarchitectonic maps were used to identify the anatomical locations of active foci.9
RESULTS All participants reported that they could perform the slow and fast repetitive contractions of the pelvic floor without experiencing fatigue. The surface EMG showed rhythmical contractions of perineal muscles, whereas leg muscles and gluteal muscles were not active (fig. 1). Nine volunteers had more widespread brain activity during fast vs slow pelvic
MOTOR CORTICAL REPRESENTATION OF PELVIC FLOOR MUSCLES
floor contractions, whereas no differences or reverse patterns were found in the others. Thus, the group analysis of all subjects did not show any significant difference in brain activation between the slow and fast contractions. The data were combined to delineate brain activity during pelvic floor contractions per se (regardless of the pace). Bilateral activity of the medial wall was conspicuous and highly significant (fig. 2, table 1). More than 80% of the cluster of active voxels in the medial wall belonged to the SMA, ie to the medial aspect of Brodmann area 6 according to cytoarchitectonic maps.9 Further significant activations were found in left and right insula, frontal operculum, putamen, thalamus, posterior parietal cortex, upper ventral pons and in the cerebellar vermis. Blood oxygen level dependent signals of the precentral gyrus during pelvic floor contractions did not reach significance. Activity of the medial primary motor cortex was only detectable when a liberal statistical threshold was applied (p ⬍0.05 uncorrected). The brain regions that were active during the control task, namely toe movements, are listed in table 2. The activity was similar during slow and fast toe movements. Figure 3 illustrates the relevant brain regions (green), which are superimposed on
187
Table 1. Brain activity during slow and fast contractions of pelvic floor muscles Peak Activation Positions (mm)
Superior frontal gyrus (SMA) Cingulate gyrus Lt putamen Lt thalamus Lt insular cortex Lt superior temporal gyrus Lt posterior parietal cortex (BA 40) Rt putamen Rt anterior insular cortex Rt thalamus Rt parietal operculum Cerebellar vermis Lt upper ventral pons Rt upper ventral pons
x
y
z
Z Score
Cluster Size
2 ⫺6 ⫺26 ⫺10 ⫺44 ⫺56 ⫺52
2 6 ⫺14 ⫺18 0 ⫺32 ⫺40
62 42 4 4 ⫺4 22 48
6.49 5.78 5.79 5.72 6.45 5.77 5.34
492 Same cluster 523 Same cluster 616 15 10
34 34 8 50 ⫺4 ⫺8 4
2 18 ⫺22 ⫺30 ⫺58 ⫺20 ⫺20
4 2 ⫺2 16 ⫺32 ⫺18 ⫺16
6.34 6.16 6.09 6.61 5.03 5.79 5.32
556 Same cluster 240 162 2 8 4
Combined data of fast and slow contractions with significance at p ⬍0.05 (FWE corrected for multiple comparisons) at the voxel level. Minimum cluster size 10 voxels, 2 voxels for cerebellum and pons.
the activation map of the pelvic floor contractions (red). Overlapping active regions were found in the SMA, cingulate cortex, operculum, thalamus, basal ganglia and cerebellum. Despite this considerable overlap, the representation of the toes in the medial wall (fig. 3, A) extended further dorsally and posterior than the pelvic floor representation. The statistical comparison yielded differences in activation strength since the paracentral lobule, ie the mesial part of the precentral gyrus and a small region in bank of the upper left central sulcus, were significantly more active during toe movements than pelvic floor contractions (table 3). The reverse comparison yielded no significant task dependent differences at the chosen conservative statistical threshold (p ⬍0.05, FWE corrected). However, Table 2. Brain activity during fast and slow movements of the right toes Peak Activation Positions (mm)
Figure 2. Brain activity (yellow) during pelvic floor muscle contractions per se (regardless of pace) superimposed on mean images of all subjects, including midsagittal (A), frontal (B) and horizontal (C) brain slices, and parasagittal (D) and horizontal (E) brainstem slices. Stereotaxic x, y, z coordinates of sections are shown. 1, SMA. 2, midcingulate cortex. 3, insula and operculum. 4, basal ganglia. 5, thalamus. White broken line (A) marks central sulcus, green asterisk denotes paracentral lobule that was significantly more active during toe movements than pelvic floor muscle contractions. Random effects group analysis of all 17 subjects with significance at p ⬍0.05 (FWE corrected) at voxel level.
Superior frontal gyrus (SMA) Cingulate gyrus Lt paracentral lobule (BA 4) Lt frontal operculum (BA 44) Lt putamen Lt thalamus Rt frontal operculum (BA 44) Rt thalamus Cerebellar vermis
x
y
z
Z Score
Cluster Size
4 ⫺4 ⫺2 ⫺52 ⫺28 ⫺16 56 14 4
2 4 ⫺28 4 4 ⫺12 10 ⫺8 ⫺50
66 44 68 4 4 6 8 6 ⫺34
5.45 5.69 6.11 5.77 6.16 5.93 5.62 6.14 5.09
196 Same cluster 51 23 195 99 77 39 2
Combined data of fast and slow contractions with significance at p ⬍0.05 (FWE corrected for multiple comparisons) at the voxel level. Minimum cluster size 10 voxels, 2 voxels for cerebellum.
188
MOTOR CORTICAL REPRESENTATION OF PELVIC FLOOR MUSCLES
Figure 3. Brain activity during pelvic floor contractions (red) and toe movements (green). Active regions projected onto surface of midsagittal slice (A) and on template brain image in view from above (B). a, SMA. b, paracentral lobule. c, insula and operculum. d, posterior parietal cortex. e, basal ganglia. Blue arrow denotes paracentral lobule, which was significantly more active during toe movements than during pelvic floor contractions. Random effects group analysis of all 17 subjects with significance at p ⬍0.05 (FWE corrected) at voxel level.
when a more liberal threshold was applied, it became evident that parts of the left prefrontal cortex, insula and brainstem tended to be more active during the volitional contractions of the pelvic floor than during the toe movements.
DISCUSSION Pelvic floor muscle training is the therapy of choice for stress urinary incontinence in women.10 Repeated pelvic floor exercises can reduce the degree and duration of urinary incontinence after radical prostatectomy,11 and may also improve venoocclusive erectile dysfunction.12 Protocols of pelvic floor physiotherapy include repetitive short and more sustained contractions performed several times daily.12,13 A recent fMRI investigation demonstrated neuroplastic changes in women with stress urinary incontinence who performed pelvic floor muscle training during a period of 3 months.1 Activation of the sensorimotor cortex became more focused, and the activity of the insula, frontal operculum and anterior cingulate cortex during sphincter contractions was diminished after the training. Repeated analyses of brain activity during the course of pelvic floor training have not yet been performed with male patients. As a foundation for such investigations the normal representation of the pelvic floor in the sensorimotor cortex should be known. Astonishingly its exact location has not yet been established. Pioneer studies explored the functional layout of the precentral and postcentral gyri with electrical stimulation during open brain surgery.14,15 The well-known classic diagrams of the sensory and motor homunculi were introduced by Penfield and Rasmussen.2 Yet the classic motor homunculus does not clearly indicate the site of the pelvic floor muscle representation
in the precentral gyrus. The somatotopic continuity of the figure implies that the pelvic floor is represented between the upper leg and trunk regions at the brim of the central fissure. However, Foerster reported that electrical stimulation of the inferomedial precentral gyrus adjacent to callosomarginal fissure evokes muscle contractions of the bladder and rectum, and suggested that pelvic floor muscles are represented in the mesial wall below the toes.14 More recently transcranial magnetic stimulation over the vertex has been used to test motor efferents to the pelvic floor muscles.16 –18 The responses typically occur together with leg muscle contractions, which reflects overlapping representations and the difficulty in stimulating pelvic floor motor neurons selectively.16,17 In a pioneer PET study Blok et al detected increased blood flow in the superomedial and right superolateral primary motor cortex during repetitive and sustained pelvic floor straining in healthy female volunteers.3 In contrast, Zhang et al found predominant SMA activity when male subjects contracted the pelvic floor repeatedly, particularly in a full bladder condition.4 They hypothesized that pelvic floor muscles may not be significantly represented in M1 since no conspicuous signals were found in this region. Accordingly another fMRI study demonstrated that short contractions of the pelvic floor, which mimicked the interruption of micturition, were associated mainly with activity of the posterior portion of the SMA.5 Seseke et al reported that short (2 seconds) contractions and relaxations were associated with activity of the superolateral primary sensorimotor cortex and of the SMA.7 Others examined sustained contractions (30 seconds) of the external anal sphincter, which resulted in M1 and SMA activity, and additional involvement of the Table 3. Task dependent differences in regional brain activity Peak Activation Positions (mm)
Toe movements greater than pelvic floor contractions [(ST ⫹ FT) minus (SP ⫹ FP)]: Paracentral lobule (BA 4) Lt central sulcus Pelvic floor contractions greater than toe movements [(SP ⫹ FP) minus (ST ⫹ FT)]:* Lt middle frontal gyrus Lt orbitofrontal cortex Lt insula Rt insula/frontal operculum Rt upper ventral pons
x
y
0 ⫺12
⫺30 ⫺42
⫺40 ⫺44 ⫺36 38 10
40 44 14 ⫺18 ⫺20
Z Score
Cluster Size
66 72
6.34 5.37
66 6
16 ⫺2 6 18 ⫺20
4.66 4.27 4.15 4.64 4.47
52 19 14 47 31
z
* Trend (p ⬍0.0001 uncorrected), not significant after correction for multiple comparisons. Otherwise significance at p ⬍0.05 at the voxel level (FWE corrected for multiple comparisons).
MOTOR CORTICAL REPRESENTATION OF PELVIC FLOOR MUSCLES
anterior and medial insula bilaterally.19 The techniques (PET, fMRI), statistical models and thresholds varied among studies. Moreover variations in effort and contraction force may have influenced the respective results. Kern et al demonstrated that the intensity and volume of the cerebral cortical activation during willful contractions of the external anal sphincter were commensurate with the level of contractile effort.20 Possible effects of the contraction speed and overlaps between the pelvic floor and foot representations in the sensorimotor cortex have not yet been studied. Therefore, we examined fast and slow pelvic floor contractions in the empty bladder state, and toe movements (control task). A random effects group analysis and a conservative statistical threshold were applied to detect the most reliable foci of activity. We had originally expected stronger signals during the fast task, in line with previous evidence that area and strength of sensorimotor cortical activation may depend on the rate of repeated muscle contractions.21,22 However, the group analysis did not reveal any significant differences. Some volunteers found the fast pace more challenging, while others reported that the slow gradual contractionrelaxation cycles were more difficult because they could be performed less automatically than the fast contractions. Despite these interindividual variations the medial wall was consistently active. As in 2 previous fMRI studies,4,5 more than 80% of the active voxels in the mesial wall lay in the SMA according to cytoarchitectonic probability maps.9 Intracortical microstimulation in nonhuman primates has revealed a rostrocaudal somatotopic arrangement of the forelimb, trunk, hindlimb and tail representations in the SMA.23 Fast and slow willful pelvic floor muscle contractions seem to activate primarily the posterior SMA, which borders the medial part of M1. Activity of the medial primary motor cortex itself was much weaker but could be detected when a liberal statistical threshold was applied. The activity of the insular region during pelvic floor straining may be related to processing of the accompanying visceral sensations.24 The posterior parietal cortex is involved when attention is focused on nonpainful visceral stimuli. A cerebellar representation of the pelvic floor may be located near the midline in the upper vermis.3,7 Toe movements (control task) involved brain regions that overlapped considerably with the representation of pelvic floor muscles (fig. 3). Since the paracentral lobule (medial part of M1) was more active during toe movements than pelvic floor contractions, the toe representation might extend fur-
189
ther posteriorly than the pelvic floor representation (fig. 3, A). A recent fMRI study of isolated lower limb joint movements revealed overlapping representations of the ankle, knee and hip joints in the primary sensory and motor cortices.25 It is known from microstimulation studies that M1 does not contain clearly segregated point-to-point representations of all body parts.26 Instead, beneath coarse gradual somatotopic gradients, the representations of muscles and smaller body parts are distributed over multiple discontinuous and overlapping zones. The pattern of overlapping activity found in the present study might concur with the segmental level of the respective motor neurons (fig. 3). Somatic motor nerves that supply the external urethral sphincter arise from motor neurons of the spinal segments S2 to S4,27,28 while the motor neurons that innervate foot muscles lie in the spinal segments L5 to S2. The classic sensory homunculus diagram indicates a representation of the genitalia in the mesial wall below the toes.2 However, Kell et al recently stated that this layout means a disruption of the somatotopic continuity.29 They established a mediolateral sequence of the sensory foot, penis and lower abdominal wall representations in the contralateral primary sensory cortex. Likewise, other researchers used electrical dorsal clitoral nerve stimulation in combination with fMRI to demonstrate that the somatosensory representation of the human clitoris is located on the lateral surface of the postcentral gyrus.30 Interestingly these recent neuroimaging studies did not find sensory representations of the genitalia in the medial part of the postcentral gyrus. Future work may locate the sensory and motor representations of the urogenital region in the same individuals to detect overlaps and dissociations.
CONCLUSIONS This fMRI study of voluntary pelvic floor contractions suggests that the predominant motor representation of the pelvic floor is located in the SMA. There is considerable overlap with the areas that are active during toe movements. The physiological somatotopy should be considered when neuroplastic changes associated with pelvic floor muscle training are interpreted.
ACKNOWLEDGMENTS Dr. M. Hamann and Prof. K. P. Jünemann supported the research and experiments. Prof. O. Jansen provided access to the magnetic resonance scanner.
190
MOTOR CORTICAL REPRESENTATION OF PELVIC FLOOR MUSCLES
REFERENCES 1. Di Gangi Herms AM, Veit R, Reisenauer C et al: Functional imaging of stress urinary incontinence. Neuroimage 2006; 29: 267. 2. Penfield W and Rasmussen T: The Cerebral Cortex of Man. New York: Macmillan 1950. 3. Blok BF, Sturms LM and Holstege G: A PET study on cortical and subcortical control of pelvic floor musculature in women. J Comp Neurol 1997; 389: 535. 4. Zhang H, Reitz A, Kollias S et al: An fMRI study of the role of suprapontine brain structures in the voluntary voiding control induced by pelvic floor contraction. Neuroimage 2005; 24: 174. 5. Kuhtz-Buschbeck JP, van der Horst C, Wolff S et al: Activation of the supplementary motor area (SMA) during voluntary pelvic floor muscle contractions–an fMRI study. Neuroimage 2007; 35: 449. 6. Seseke S, Baudewig J, Kallenberg K et al: Voluntary pelvic floor muscle control–an fMRI study. Neuroimage 2006; 31: 1399. 7. Seseke S, Baudewig J, Kallenberg K et al: Gender differences in voluntary micturition control: an fMRI study. Neuroimage 2008; 43: 183. 8. Karantanis E, Fynes M, Moore KH et al: Comparison of the ICIQ-SF and 24-hour pad test with other measures for evaluating the severity of urodynamic stress incontinence. Int Urogynecol J 2004; 15: 111. 9. Eickhoff SB, Stephan KE, Mohlberg H et al: A new SPM toolbox for combining probabilistic cytoarchitectonic maps and functional imaging data. Neuroimage 2005; 25: 1325. 10. Freeman RM: The role of pelvic floor muscle training in urinary incontinence. BJOG 2004; 111: 37.
11. Cornel EB, de Wit R and Witjes JA: Evaluation of early pelvic floor physiotherapy on the duration and degree of urinary incontinence after radical retropubic prostatectomy in a non-teaching hospital. World J Urol 2005; 23: 353. 12. van Kampen M, de Weert W, Claes H et al: Treatment of erectile dysfunction by perineal exercise, electromyographic biofeedback, and electrical stimulation. Phys Ther 2003; 83: 536. 13. Dorey G, Glazener C, Buckley B et al: Developing a pelvic floor muscle training regimen for use in a trial intervention. Physiotherapy 2009; 95: 199. 14. Foerster O: Motorische Felder und Bahnen. In: Handbuch der Neurologie. Edited by O Bumke and O Foerster. Berlin: Springer 1936; vol 6, pp 48 – 56. 15. Penfield W and Boldrey E: Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation. Brain 1937; 60: 389. 16. Turnbull GK, Hamdy S, Aziz Q et al: The cortical topography of human anorectal musculature. Gastroenterology 1999; 117: 32. 17. Brostrom S, Jennum P and Lose G: Motor evoked potentials from the striated urethral sphincter and puborectal muscle: reproducibility of latencies. Clin Neurophysiol 2003; 114: 1891. 18. Lefaucheur JP: Excitability of the motor cortical representation of the external anal sphincter. Exp Brain Res 2005; 160: 268. 19. Bittorf B, Ringler R, Forster C et al: Cerebral representation of the anorectum using functional magnetic resonance imaging. Br J Surg 2006; 93: 1251.
20. Kern MK, Arndorfer RC, Hyde JS et al: Cerebral cortical representation of external anal sphincter contraction: effect of effort. Am J Physiol Gastrointest Liver Physiol 2004; 286: G304. 21. Rao SM, Bandettini PA, Binder JR et al: Relationship between finger movement rate and functional magnetic resonance signal change in human primary motor cortex. J Cereb Blood Flow Metab 1996; 16: 1250. 22. Schlaug G, Sanes JN, Thangaraj V et al: Cerebral activation covaries with movement rate. Neuroreport 1996; 22: 879. 23. Mitz AR and Wise SP: The somatotopic organization of the supplementary motor area: intracortical microstimulation mapping. J Neurosci 1987; 7: 1010. 24. Griffiths D, Derbyshire S, Stenger A et al: Brain control of normal and overactive bladder. J Urol 2005; 174: 1862. 25. Kapreli E, Athanasopoulos S, Papathanasiou M et al: Lower limb sensorimotor network: issues of somatotopy and overlap. Cortex 2007; 43: 219. 26. Schieber MH: Constraints on somatotopic organization in the primary motor cortex. J Neurophysiol 2001; 86: 2125. 27. Fowler CJ, Griffiths D and de Groat WC: The neural control of micturition. Nat Rev Neurosci 2008; 9: 453. 28. Vodusek DB: Anatomy and neurocontrol of the pelvic floor. Digestion 2004; 69: 87. 29. Kell CA, von Kriegstein K, Roesler A et al: The sensory cortical representation of the human penis: revisiting somatotopy in the male homunculus. J Neurosci 2005; 25: 5984. 30. Michels L, Mehnert U, Boy S et al: The somatosensory representation of the human clitoris: an fMRI study. Neuroimage 2010; 49: 177.
Purpose: Pelvic floor muscle training involves rhythmical voluntary contractions of the external urethral sphincter and ancillary pelvic floor muscles. The representation of these muscles in the motor cortex has not been located precisely and unambiguously. We used functional magnetic resonance imaging to determine brain activity during slow and fast pelvic floor contractions. Materials and Methods: Cerebral responses were recorded in 17 healthy male volunteers, 21 to 47 years old, with normal bladder control. Functional magnetic resonance imaging was performed during metronome paced slow (0.25 Hertz) and fast (0.7 Hertz) contractions of the pelvic floor that mimicked the interruption of voiding. To study the somatotopy of the cortical representations, flexion-extension movements of the right toes were performed as a control task. Results: Functional magnetic resonance imaging during pelvic floor contractions detected activity of the supplementary motor area in the medial wall and of the midcingulate cortex, insula, posterior parietal cortex, putamen, thalamus, cerebellar vermis and upper ventral pons. There were no significant differences in activation between slow and fast contractions. Toe movements involved significantly stronger activity of the paracentral lobule (ie the medial primary motor cortex) than did the pelvic floor contractions. Otherwise the areas active during pelvic floor and leg muscle contractions overlapped considerably. Conclusions: The motor cortical representation of pelvic floor muscles is located mostly in the supplementary motor area. It extends further ventrally and anteriorly than the representation of distal leg muscles. Key Words: motor cortex, magnetic resonance imaging, pelvic floor PELVIC floor muscle training consists of repetitive voluntary contractions that strengthen the striated urethral sphincter and other pelvic floor muscles, eg in the treatment of stress urinary incontinence. Training for several months results in neuroplastic changes in the accompanying brain activity.1 However, the physiological representation of pelvic floor muscles in the motor cortex has not yet been pinpointed unequivocally. The classic homunculus diagram of Penfield and Rasmussen does not explicitly indicate the region of the pelvic floor in
the primary motor cortex,2 whose location may be assumed to lie between the leg and trunk representations (“hip motor region”) in the superomedial precentral gyrus. Several neuroimaging studies of brain activity during voluntary contractions of the pelvic floor had divergent results.3–7 The seminal PET study of Blok et al reported activity in the superomedial and superolateral M1.3 Two fMRI studies revealed predominant activity in the supplementary motor area, whereas involvement of M1 was negligible.4,5 Seseke et al observed activity
0022-5347/11/1861-0185/0 THE JOURNAL OF UROLOGY® © 2011 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION
Vol. 186, 185-190, July 2011 Printed in U.S.A. DOI:10.1016/j.juro.2011.03.001
AND
RESEARCH, INC.
Abbreviations and Acronyms BA ⫽ Brodmann area EMG ⫽ electromyogram fMRI ⫽ functional magnetic resonance imaging FP ⫽ fast repetitive contractions of the pelvic floor FT ⫽ fast flexion-extension movements of the right toes FWE ⫽ family wise error M1 ⫽ primary motor cortex PET ⫽ positron emission tomography SMA ⫽ supplementary motor area SP ⫽ slow repetitive contractions of the pelvic floor ST ⫽ slow flexion-extension movements of the right toes Submitted for publication November 2, 2010. Study received local ethics committee approval. Supported by Deutsche Forschungsgemeinschaft (Grant KU 1937/2-1). Supplementary material for this article can be obtained at http://www.physiologie.unikiel.de/homepages_mitarbeiter/kuhtzHomepage MotorControl.pdf. * Correspondence: Institute of Physiology, Christian-Albrechts-Universität, Olshausenstr. 40, D-24098 Kiel, Germany (telephone: ⫹49 431 8803657; FAX: ⫹49 431 8804580; e-mail: [email protected]).
See Editorial on page 14. For another article on a related topic see page 334.
www.jurology.com
185
186
MOTOR CORTICAL REPRESENTATION OF PELVIC FLOOR MUSCLES
of the superolateral convexity of the sensorimotor cortex and additional involvement of the SMA.6,7 Some differences may be due to nonuniform experimental paradigms (ie different pelvic floor exercises). Possibly fast and short willful muscle contractions engage other brain regions than slow sustained contractions. Therefore, the present fMRI study examines brain activity during both types of voluntary contractions of pelvic floor muscles in healthy volunteers. As a control task, flexion-extension movements of the toes were performed to delineate the somatotopic layout of the neighboring cortical representation of the foot. Knowledge of the normal somatotopy is needed for the understanding of neuroplastic changes induced by pelvic floor muscle training in patients.1
Figure 1. EMG during fast and slow pelvic floor muscle contractions. Bipolar surface electrodes were used. Recordings were made of pelvic floor muscles (PF, electrodes on perineum) and of right leg muscles including flexor hallucis (FH), rectus femoris (RF) and gluteus maximus (GL). Black bars indicate sounds of metronome (Mt) that paced contractions.
METHODS A total of 17 right-handed healthy men with a mean age (⫾ SD) of 28.9 ⫾ 7.1 years old (range 21 to 47) volunteered to participate. The study received local ethics committee approval. Participants were staff members or medical students who were familiar with the anatomy and function of the pelvic floor. None of the subjects used medication that might have influenced bladder or brain activity. None had symptoms of incontinence or urinary tract infection according to their anamnesis and a questionnaire.8 Participants performed slow (SP) and fast (FP) repetitive contractions of the pelvic floor which mimicked interruption of voiding, and slow and fast flexion-extension movements of the right toes (ST, FT). Fast contractions were performed at 0.7 Hz (every 1.4 seconds), paced by brief sounds (0.2 seconds) of a metronome. Slow contractions occurred at 0.25 Hz (every 4 seconds). The slow contraction-relaxation cycles followed a sound that gradually increased and decreased in loudness. All contractions were performed with moderate effort (approximately 25% of maximum force) to avoid fatigue. The tasks were first practiced and EMG recordings of muscle activity were performed outside the magnetic resonance tomography scanner (fig. 1). Subjects emptied their bladder before neuroimaging started so that scanning was performed in a no desire to void condition. A 3 Tesla (Philips Achieva®) tomograph with a phased array head coil (Sense 8-channel) was used. The subjects lay still in a supine position with eyes closed. Head motion was restricted by foam pads. T2-weighted echo planar image volumes with blood oxygenation level dependent contrast (echo time 34.5 milliseconds, flip angle 90 degrees, field of view 224 mm, pixel size 3.5 mm) were collected continually every 2.5 seconds during 4 imaging runs. Each volume consisted of 35 axial slices (3 mm slice thickness, 0.3 mm interslice gaps) that included the brain from the vertex to the cerebellum. Fast (20 per period) and slow (7 per period) repetitive contractions of the pelvic floor (FP, SP) or of the right foot muscles (FT, ST), respectively, were performed in 4 separate runs in randomized order. Each run (144 image volumes) comprised 6 contraction periods of 28 seconds each, which alternated with
28-second periods of rest. Metronome sounds were audible during task and rest periods, and tasks were announced via headphones. Additional T1-weighted anatomical images with a high resolution were obtained from each subject. fMRI data were analyzed with SPM5 statistical parametric mapping software (www.fil.ion.ac.uk/spm). The image volumes were realigned and standardized to the stereotaxic space of the Montreal Neurological Institute brain template, sampled to a voxel size of 2 ⫻ 2 ⫻ 2 mm3 and smoothed with a 6 ⫻ 6 ⫻ 6 mm3 isotropic Gaussian filter. Differences in global brain activity among subjects were removed by proportional scaling. Blood oxygen level dependent responses during the 4 contraction tasks (FP, SP, FT, ST) were modeled with boxcar functions (block design) that had been convolved with the hemodynamic response function. Fast and slow contractions were compared (FP-SP, FT-ST). Activations during pelvic floor contractions and toe movements were contrasted ([FP⫹SP] minus [FT⫹ST] and vice versa). The contrast images of all 17 participants were entered into a random effects second level analysis. Statistical inferences were drawn at the voxel level with p ⬍0.05 corrected for multiple comparisons (controlling family wise error rate, FWE corrected). We described Montreal Neurological Institute spatial coordinates as given by SPM5. Active brain regions were superimposed on a mean normalized anatomical image of all 17 subjects. Published probabilistic cytoarchitectonic maps were used to identify the anatomical locations of active foci.9
RESULTS All participants reported that they could perform the slow and fast repetitive contractions of the pelvic floor without experiencing fatigue. The surface EMG showed rhythmical contractions of perineal muscles, whereas leg muscles and gluteal muscles were not active (fig. 1). Nine volunteers had more widespread brain activity during fast vs slow pelvic
MOTOR CORTICAL REPRESENTATION OF PELVIC FLOOR MUSCLES
floor contractions, whereas no differences or reverse patterns were found in the others. Thus, the group analysis of all subjects did not show any significant difference in brain activation between the slow and fast contractions. The data were combined to delineate brain activity during pelvic floor contractions per se (regardless of the pace). Bilateral activity of the medial wall was conspicuous and highly significant (fig. 2, table 1). More than 80% of the cluster of active voxels in the medial wall belonged to the SMA, ie to the medial aspect of Brodmann area 6 according to cytoarchitectonic maps.9 Further significant activations were found in left and right insula, frontal operculum, putamen, thalamus, posterior parietal cortex, upper ventral pons and in the cerebellar vermis. Blood oxygen level dependent signals of the precentral gyrus during pelvic floor contractions did not reach significance. Activity of the medial primary motor cortex was only detectable when a liberal statistical threshold was applied (p ⬍0.05 uncorrected). The brain regions that were active during the control task, namely toe movements, are listed in table 2. The activity was similar during slow and fast toe movements. Figure 3 illustrates the relevant brain regions (green), which are superimposed on
187
Table 1. Brain activity during slow and fast contractions of pelvic floor muscles Peak Activation Positions (mm)
Superior frontal gyrus (SMA) Cingulate gyrus Lt putamen Lt thalamus Lt insular cortex Lt superior temporal gyrus Lt posterior parietal cortex (BA 40) Rt putamen Rt anterior insular cortex Rt thalamus Rt parietal operculum Cerebellar vermis Lt upper ventral pons Rt upper ventral pons
x
y
z
Z Score
Cluster Size
2 ⫺6 ⫺26 ⫺10 ⫺44 ⫺56 ⫺52
2 6 ⫺14 ⫺18 0 ⫺32 ⫺40
62 42 4 4 ⫺4 22 48
6.49 5.78 5.79 5.72 6.45 5.77 5.34
492 Same cluster 523 Same cluster 616 15 10
34 34 8 50 ⫺4 ⫺8 4
2 18 ⫺22 ⫺30 ⫺58 ⫺20 ⫺20
4 2 ⫺2 16 ⫺32 ⫺18 ⫺16
6.34 6.16 6.09 6.61 5.03 5.79 5.32
556 Same cluster 240 162 2 8 4
Combined data of fast and slow contractions with significance at p ⬍0.05 (FWE corrected for multiple comparisons) at the voxel level. Minimum cluster size 10 voxels, 2 voxels for cerebellum and pons.
the activation map of the pelvic floor contractions (red). Overlapping active regions were found in the SMA, cingulate cortex, operculum, thalamus, basal ganglia and cerebellum. Despite this considerable overlap, the representation of the toes in the medial wall (fig. 3, A) extended further dorsally and posterior than the pelvic floor representation. The statistical comparison yielded differences in activation strength since the paracentral lobule, ie the mesial part of the precentral gyrus and a small region in bank of the upper left central sulcus, were significantly more active during toe movements than pelvic floor contractions (table 3). The reverse comparison yielded no significant task dependent differences at the chosen conservative statistical threshold (p ⬍0.05, FWE corrected). However, Table 2. Brain activity during fast and slow movements of the right toes Peak Activation Positions (mm)
Figure 2. Brain activity (yellow) during pelvic floor muscle contractions per se (regardless of pace) superimposed on mean images of all subjects, including midsagittal (A), frontal (B) and horizontal (C) brain slices, and parasagittal (D) and horizontal (E) brainstem slices. Stereotaxic x, y, z coordinates of sections are shown. 1, SMA. 2, midcingulate cortex. 3, insula and operculum. 4, basal ganglia. 5, thalamus. White broken line (A) marks central sulcus, green asterisk denotes paracentral lobule that was significantly more active during toe movements than pelvic floor muscle contractions. Random effects group analysis of all 17 subjects with significance at p ⬍0.05 (FWE corrected) at voxel level.
Superior frontal gyrus (SMA) Cingulate gyrus Lt paracentral lobule (BA 4) Lt frontal operculum (BA 44) Lt putamen Lt thalamus Rt frontal operculum (BA 44) Rt thalamus Cerebellar vermis
x
y
z
Z Score
Cluster Size
4 ⫺4 ⫺2 ⫺52 ⫺28 ⫺16 56 14 4
2 4 ⫺28 4 4 ⫺12 10 ⫺8 ⫺50
66 44 68 4 4 6 8 6 ⫺34
5.45 5.69 6.11 5.77 6.16 5.93 5.62 6.14 5.09
196 Same cluster 51 23 195 99 77 39 2
Combined data of fast and slow contractions with significance at p ⬍0.05 (FWE corrected for multiple comparisons) at the voxel level. Minimum cluster size 10 voxels, 2 voxels for cerebellum.
188
MOTOR CORTICAL REPRESENTATION OF PELVIC FLOOR MUSCLES
Figure 3. Brain activity during pelvic floor contractions (red) and toe movements (green). Active regions projected onto surface of midsagittal slice (A) and on template brain image in view from above (B). a, SMA. b, paracentral lobule. c, insula and operculum. d, posterior parietal cortex. e, basal ganglia. Blue arrow denotes paracentral lobule, which was significantly more active during toe movements than during pelvic floor contractions. Random effects group analysis of all 17 subjects with significance at p ⬍0.05 (FWE corrected) at voxel level.
when a more liberal threshold was applied, it became evident that parts of the left prefrontal cortex, insula and brainstem tended to be more active during the volitional contractions of the pelvic floor than during the toe movements.
DISCUSSION Pelvic floor muscle training is the therapy of choice for stress urinary incontinence in women.10 Repeated pelvic floor exercises can reduce the degree and duration of urinary incontinence after radical prostatectomy,11 and may also improve venoocclusive erectile dysfunction.12 Protocols of pelvic floor physiotherapy include repetitive short and more sustained contractions performed several times daily.12,13 A recent fMRI investigation demonstrated neuroplastic changes in women with stress urinary incontinence who performed pelvic floor muscle training during a period of 3 months.1 Activation of the sensorimotor cortex became more focused, and the activity of the insula, frontal operculum and anterior cingulate cortex during sphincter contractions was diminished after the training. Repeated analyses of brain activity during the course of pelvic floor training have not yet been performed with male patients. As a foundation for such investigations the normal representation of the pelvic floor in the sensorimotor cortex should be known. Astonishingly its exact location has not yet been established. Pioneer studies explored the functional layout of the precentral and postcentral gyri with electrical stimulation during open brain surgery.14,15 The well-known classic diagrams of the sensory and motor homunculi were introduced by Penfield and Rasmussen.2 Yet the classic motor homunculus does not clearly indicate the site of the pelvic floor muscle representation
in the precentral gyrus. The somatotopic continuity of the figure implies that the pelvic floor is represented between the upper leg and trunk regions at the brim of the central fissure. However, Foerster reported that electrical stimulation of the inferomedial precentral gyrus adjacent to callosomarginal fissure evokes muscle contractions of the bladder and rectum, and suggested that pelvic floor muscles are represented in the mesial wall below the toes.14 More recently transcranial magnetic stimulation over the vertex has been used to test motor efferents to the pelvic floor muscles.16 –18 The responses typically occur together with leg muscle contractions, which reflects overlapping representations and the difficulty in stimulating pelvic floor motor neurons selectively.16,17 In a pioneer PET study Blok et al detected increased blood flow in the superomedial and right superolateral primary motor cortex during repetitive and sustained pelvic floor straining in healthy female volunteers.3 In contrast, Zhang et al found predominant SMA activity when male subjects contracted the pelvic floor repeatedly, particularly in a full bladder condition.4 They hypothesized that pelvic floor muscles may not be significantly represented in M1 since no conspicuous signals were found in this region. Accordingly another fMRI study demonstrated that short contractions of the pelvic floor, which mimicked the interruption of micturition, were associated mainly with activity of the posterior portion of the SMA.5 Seseke et al reported that short (2 seconds) contractions and relaxations were associated with activity of the superolateral primary sensorimotor cortex and of the SMA.7 Others examined sustained contractions (30 seconds) of the external anal sphincter, which resulted in M1 and SMA activity, and additional involvement of the Table 3. Task dependent differences in regional brain activity Peak Activation Positions (mm)
Toe movements greater than pelvic floor contractions [(ST ⫹ FT) minus (SP ⫹ FP)]: Paracentral lobule (BA 4) Lt central sulcus Pelvic floor contractions greater than toe movements [(SP ⫹ FP) minus (ST ⫹ FT)]:* Lt middle frontal gyrus Lt orbitofrontal cortex Lt insula Rt insula/frontal operculum Rt upper ventral pons
x
y
0 ⫺12
⫺30 ⫺42
⫺40 ⫺44 ⫺36 38 10
40 44 14 ⫺18 ⫺20
Z Score
Cluster Size
66 72
6.34 5.37
66 6
16 ⫺2 6 18 ⫺20
4.66 4.27 4.15 4.64 4.47
52 19 14 47 31
z
* Trend (p ⬍0.0001 uncorrected), not significant after correction for multiple comparisons. Otherwise significance at p ⬍0.05 at the voxel level (FWE corrected for multiple comparisons).
MOTOR CORTICAL REPRESENTATION OF PELVIC FLOOR MUSCLES
anterior and medial insula bilaterally.19 The techniques (PET, fMRI), statistical models and thresholds varied among studies. Moreover variations in effort and contraction force may have influenced the respective results. Kern et al demonstrated that the intensity and volume of the cerebral cortical activation during willful contractions of the external anal sphincter were commensurate with the level of contractile effort.20 Possible effects of the contraction speed and overlaps between the pelvic floor and foot representations in the sensorimotor cortex have not yet been studied. Therefore, we examined fast and slow pelvic floor contractions in the empty bladder state, and toe movements (control task). A random effects group analysis and a conservative statistical threshold were applied to detect the most reliable foci of activity. We had originally expected stronger signals during the fast task, in line with previous evidence that area and strength of sensorimotor cortical activation may depend on the rate of repeated muscle contractions.21,22 However, the group analysis did not reveal any significant differences. Some volunteers found the fast pace more challenging, while others reported that the slow gradual contractionrelaxation cycles were more difficult because they could be performed less automatically than the fast contractions. Despite these interindividual variations the medial wall was consistently active. As in 2 previous fMRI studies,4,5 more than 80% of the active voxels in the mesial wall lay in the SMA according to cytoarchitectonic probability maps.9 Intracortical microstimulation in nonhuman primates has revealed a rostrocaudal somatotopic arrangement of the forelimb, trunk, hindlimb and tail representations in the SMA.23 Fast and slow willful pelvic floor muscle contractions seem to activate primarily the posterior SMA, which borders the medial part of M1. Activity of the medial primary motor cortex itself was much weaker but could be detected when a liberal statistical threshold was applied. The activity of the insular region during pelvic floor straining may be related to processing of the accompanying visceral sensations.24 The posterior parietal cortex is involved when attention is focused on nonpainful visceral stimuli. A cerebellar representation of the pelvic floor may be located near the midline in the upper vermis.3,7 Toe movements (control task) involved brain regions that overlapped considerably with the representation of pelvic floor muscles (fig. 3). Since the paracentral lobule (medial part of M1) was more active during toe movements than pelvic floor contractions, the toe representation might extend fur-
189
ther posteriorly than the pelvic floor representation (fig. 3, A). A recent fMRI study of isolated lower limb joint movements revealed overlapping representations of the ankle, knee and hip joints in the primary sensory and motor cortices.25 It is known from microstimulation studies that M1 does not contain clearly segregated point-to-point representations of all body parts.26 Instead, beneath coarse gradual somatotopic gradients, the representations of muscles and smaller body parts are distributed over multiple discontinuous and overlapping zones. The pattern of overlapping activity found in the present study might concur with the segmental level of the respective motor neurons (fig. 3). Somatic motor nerves that supply the external urethral sphincter arise from motor neurons of the spinal segments S2 to S4,27,28 while the motor neurons that innervate foot muscles lie in the spinal segments L5 to S2. The classic sensory homunculus diagram indicates a representation of the genitalia in the mesial wall below the toes.2 However, Kell et al recently stated that this layout means a disruption of the somatotopic continuity.29 They established a mediolateral sequence of the sensory foot, penis and lower abdominal wall representations in the contralateral primary sensory cortex. Likewise, other researchers used electrical dorsal clitoral nerve stimulation in combination with fMRI to demonstrate that the somatosensory representation of the human clitoris is located on the lateral surface of the postcentral gyrus.30 Interestingly these recent neuroimaging studies did not find sensory representations of the genitalia in the medial part of the postcentral gyrus. Future work may locate the sensory and motor representations of the urogenital region in the same individuals to detect overlaps and dissociations.
CONCLUSIONS This fMRI study of voluntary pelvic floor contractions suggests that the predominant motor representation of the pelvic floor is located in the SMA. There is considerable overlap with the areas that are active during toe movements. The physiological somatotopy should be considered when neuroplastic changes associated with pelvic floor muscle training are interpreted.
ACKNOWLEDGMENTS Dr. M. Hamann and Prof. K. P. Jünemann supported the research and experiments. Prof. O. Jansen provided access to the magnetic resonance scanner.
190
MOTOR CORTICAL REPRESENTATION OF PELVIC FLOOR MUSCLES
REFERENCES 1. Di Gangi Herms AM, Veit R, Reisenauer C et al: Functional imaging of stress urinary incontinence. Neuroimage 2006; 29: 267. 2. Penfield W and Rasmussen T: The Cerebral Cortex of Man. New York: Macmillan 1950. 3. Blok BF, Sturms LM and Holstege G: A PET study on cortical and subcortical control of pelvic floor musculature in women. J Comp Neurol 1997; 389: 535. 4. Zhang H, Reitz A, Kollias S et al: An fMRI study of the role of suprapontine brain structures in the voluntary voiding control induced by pelvic floor contraction. Neuroimage 2005; 24: 174. 5. Kuhtz-Buschbeck JP, van der Horst C, Wolff S et al: Activation of the supplementary motor area (SMA) during voluntary pelvic floor muscle contractions–an fMRI study. Neuroimage 2007; 35: 449. 6. Seseke S, Baudewig J, Kallenberg K et al: Voluntary pelvic floor muscle control–an fMRI study. Neuroimage 2006; 31: 1399. 7. Seseke S, Baudewig J, Kallenberg K et al: Gender differences in voluntary micturition control: an fMRI study. Neuroimage 2008; 43: 183. 8. Karantanis E, Fynes M, Moore KH et al: Comparison of the ICIQ-SF and 24-hour pad test with other measures for evaluating the severity of urodynamic stress incontinence. Int Urogynecol J 2004; 15: 111. 9. Eickhoff SB, Stephan KE, Mohlberg H et al: A new SPM toolbox for combining probabilistic cytoarchitectonic maps and functional imaging data. Neuroimage 2005; 25: 1325. 10. Freeman RM: The role of pelvic floor muscle training in urinary incontinence. BJOG 2004; 111: 37.
11. Cornel EB, de Wit R and Witjes JA: Evaluation of early pelvic floor physiotherapy on the duration and degree of urinary incontinence after radical retropubic prostatectomy in a non-teaching hospital. World J Urol 2005; 23: 353. 12. van Kampen M, de Weert W, Claes H et al: Treatment of erectile dysfunction by perineal exercise, electromyographic biofeedback, and electrical stimulation. Phys Ther 2003; 83: 536. 13. Dorey G, Glazener C, Buckley B et al: Developing a pelvic floor muscle training regimen for use in a trial intervention. Physiotherapy 2009; 95: 199. 14. Foerster O: Motorische Felder und Bahnen. In: Handbuch der Neurologie. Edited by O Bumke and O Foerster. Berlin: Springer 1936; vol 6, pp 48 – 56. 15. Penfield W and Boldrey E: Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation. Brain 1937; 60: 389. 16. Turnbull GK, Hamdy S, Aziz Q et al: The cortical topography of human anorectal musculature. Gastroenterology 1999; 117: 32. 17. Brostrom S, Jennum P and Lose G: Motor evoked potentials from the striated urethral sphincter and puborectal muscle: reproducibility of latencies. Clin Neurophysiol 2003; 114: 1891. 18. Lefaucheur JP: Excitability of the motor cortical representation of the external anal sphincter. Exp Brain Res 2005; 160: 268. 19. Bittorf B, Ringler R, Forster C et al: Cerebral representation of the anorectum using functional magnetic resonance imaging. Br J Surg 2006; 93: 1251.
20. Kern MK, Arndorfer RC, Hyde JS et al: Cerebral cortical representation of external anal sphincter contraction: effect of effort. Am J Physiol Gastrointest Liver Physiol 2004; 286: G304. 21. Rao SM, Bandettini PA, Binder JR et al: Relationship between finger movement rate and functional magnetic resonance signal change in human primary motor cortex. J Cereb Blood Flow Metab 1996; 16: 1250. 22. Schlaug G, Sanes JN, Thangaraj V et al: Cerebral activation covaries with movement rate. Neuroreport 1996; 22: 879. 23. Mitz AR and Wise SP: The somatotopic organization of the supplementary motor area: intracortical microstimulation mapping. J Neurosci 1987; 7: 1010. 24. Griffiths D, Derbyshire S, Stenger A et al: Brain control of normal and overactive bladder. J Urol 2005; 174: 1862. 25. Kapreli E, Athanasopoulos S, Papathanasiou M et al: Lower limb sensorimotor network: issues of somatotopy and overlap. Cortex 2007; 43: 219. 26. Schieber MH: Constraints on somatotopic organization in the primary motor cortex. J Neurophysiol 2001; 86: 2125. 27. Fowler CJ, Griffiths D and de Groat WC: The neural control of micturition. Nat Rev Neurosci 2008; 9: 453. 28. Vodusek DB: Anatomy and neurocontrol of the pelvic floor. Digestion 2004; 69: 87. 29. Kell CA, von Kriegstein K, Roesler A et al: The sensory cortical representation of the human penis: revisiting somatotopy in the male homunculus. J Neurosci 2005; 25: 5984. 30. Michels L, Mehnert U, Boy S et al: The somatosensory representation of the human clitoris: an fMRI study. Neuroimage 2010; 49: 177.