Brain Switch for Reflex Micturition Control Detected by fMRI in Rats

Abstract

The functions of the lower urinary tract are controlled by complex pathways in the brain that act like switching circuits to voluntarily or reflexly shift the activity of various pelvic organs (bladder, urethra, urethral sphincter, and pelvic floor muscles) from urine storage to micturition. In this study, functional magnetic resonance imaging (fMRI) was used to visualize the brain switching circuits controlling reflex micturition in anesthetized rats. The fMRI images confirmed the hypothesis based on previous neuroanatomical and neurophysiological studies that the brain stem switch for reflex micturition control involves both the periaqueductal gray (PAG) and the pontine micturition center (PMC). During storage, the PAG was activated by afferent input from the urinary bladder while the PMC was inactive. When bladder volume increased to the micturition threshold, the switch from storage to micturition was associated with PMC activation and enhanced PAG activity. A complex brain network that may regulate the brain stem micturition switch and control storage and voiding was also identified. Storage was accompanied by activation of the motor cortex, somatosensory cortex, cingulate cortex, retrosplenial cortex, thalamus, putamen, insula, and septal nucleus. On the other hand, micturition was associated with: 1) increased activity of the motor cortex, thalamus, and putamen; 2) a shift in the locus of activity in the cingulate and insula; and 3) the emergence of activity in the hypothalamus, substantia nigra, globus pallidus, hippocampus, and inferior colliculus. Understanding brain control of reflex micturition is important for elucidating the mechanisms underlying neurogenic bladder dysfunctions including frequency, urgency, and incontinence.

INTRODUCTION

Storage and periodic elimination of urine can be controlled involuntarily in infants or voluntarily in adults by neural pathways in the brain and spinal cord (Barrington 1921; de Groat 1975; de Groat and Ryall 1969; de Groat et al. 1993; Kuru 1965). During storage when urine slowly accumulates in the bladder, bladder afferent activity in the pelvic nerve gradually increases and is transmitted via the sacral spinal cord to the brain to provide information about the extent of bladder filling. When the bladder is controlled involuntarily and volume is below the threshold level for triggering micturition, the brain switch for reflex micturition is turned off and the bladder is quiescent, thus promoting the storage of urine. However, when bladder volume reaches the micturition threshold, activation of a brain-switching circuit sends an excitatory signal through the spinal cord to the bladder to induce a sustained bladder contraction and an inhibitory signal to induce a reciprocal relaxation of the external urethral sphincter (EUS), leading to the release of urine.

Electrophysiological studies in animals indicate that the neural switching circuit controlling reflex micturition is located in the rostral brain stem (de Groat 1975; de Groat and Ryall 1969; Kuru 1965; Noto et al. 1991). After midcollicular decerebration reflex micturition is maintained, however, destruction of a region in the dorsolateral pons (termed the pontine micturition center [PMC]) or transection of the neuraxis at any level below the PMC blocks reflex micturition (Barrington 1921; Ruch and Tang 1956). Recordings of neural firing in the PMC in cats revealed all-or-none patterns of activity that correlate with the storage and voiding phases of bladder activity (de Groat et al. 1998; Sakakibara et al. 2002; Sasaki 2002, 2005; Sugaya et al. 2003; Tanaka et al. 2003; Willette et al. 1988). In addition microinjections of inhibitory agents into the PMC in decerebrate cats increase the micturition volume threshold or completely block micturition (Mallory et al. 1991).

Neuroanatomical studies in cats (Blok and Holstege 1994, 1996, 1998) indicate that afferent input from the bladder is received in the periaqueductal gray (PAG) and then transmitted to the PMC, which in turn sends motor signals back to the spinal cord to induce micturition. This spinobulbospinal switching circuit is modulated by inputs from the forebrain that control reflex micturition (de Groat et al. 1993; Kuru 1965; Ruch and Tang 1956; Yokoyama et al. 2002) and mediate voluntary voiding. Sensory input from the bladder to the forebrain very likely passes through relays in the PAG (Holstege 2005) as well as in the thalamus (Craig 1996, 2002; Mayer et al. 2006).

Functional brain imaging technologies including functional magnetic resonance imaging (fMRI) and positron emission tomography (PET) have been used in humans to identify brain regions activated during bladder filling and voluntary control of micturition (Athwal et al. 2001; Blok et al. 1997a,b, 1998, 2006; Dasgupta et al. 2005; Di Gangi Herms et al. 2006; Griffiths et al. 2005, 2007; Kitta et al. 2006; Kuhtz-Buschbeck et al. 2005, 2007; Matsuura et al. 2002; Mehnert et al. 2008; Nour et al. 2000; Sakakibara et al. 1999; Seseke et al. 2006, 2008; Zhang et al. 2005). Under these conditions the neural mechanisms controlling micturition are influenced not only by the peripheral sensory input from the bladder, but also by conscious brain processes including attention, expectation, decision, and emotion (Mayer et al. 2006). Since these processes can vary at different times in the same individual or between individuals this complicates the interpretation of brain imaging data (Mayer et al. 2006). To separate the basic micturition neural circuitry that mediates reflex control from the conscious control of bladder function, it was recently suggested that human brain fMRI studies might be performed under anesthesia (Naliboff and Mayer 2006). Information about the reflex control of voiding is important clinically because many dysfunctions of the lower urinary tract (e.g., urinary incontinence) are mediated by involuntary neural mechanisms.

In this study, brain fMRI imaging was performed on anesthetized rats to identify neuronal circuitry involved in reflex micturition. In the rat the PAG and PMC appear to have roles in the reflex control of micturition similar to those identified in the cat. Electrophysiological studies showed that stimulation of bladder afferent nerves evoked action potentials in PAG at shorter latencies than the action potentials in the PMC, suggesting that bladder afferent input from the spinal cord is processed first in the PAG and then relayed to the PMC (Noto et al. 1991). However, ascending projections from the lumbosacral spinal cord have been detected with axonal tracing techniques both in the PMC and in the PAG (Ding et al. 1997), raising the possibility that bladder afferent input may also go directly to the PMC. Electrical stimulation in the PAG and in the PMC evoked firing in parasympathetic efferent pathways to the bladder and bladder contractions, demonstrating a connection of neurons in both of these areas with the efferent outflow to the urinary bladder (de Groat 1975; Kruse et al. 1990; Noto et al. 1989, 1991; Taniguchi et al. 2002). This was confirmed by retrograde transneuronal virus tracing in the rat (Grill et al. 1999; Marson 1997; Nadelhaft et al. 1992; Sugaya et al. 1997; Vizzard et al. 1995), which revealed that both the PMC and the PAG were labeled after injection of virus into the bladder or the urethra. Assuming that the micturition switch is located in the PMC in the rat then it is reasonable to hypothesize that fMRI imaging should detect a signal in the PAG area but not in the PMC in response to bladder distention during urine storage. During micturition, however, both regions should exhibit an increased signal.

How neurons in other brain centers participate in the switch from storage to micturition is largely unknown. A recent neuroanatomical study in the cat (Kuipers et al. 2006) showed that, with the exception of the hypothalamus and PAG, no other brain structures have direct monosynaptic access to the PMC. Thus control of the PMC switching circuit by the brain must be primarily indirect via relays through the PAG and hypothalamus. The goal of this study was to use fMRI to determine how activity in brain regions, which are potentially involved in micturition control, changes during filling of the bladder, and after the initiation of reflex micturition in anesthetized animals.

METHODS

Animal preparation

Nine male Sprague–Dawley rats (300–400 g) were used in this study. All protocols involving the use of animals were approved by the Animal Care and Use Committee at the University of Pittsburgh. The animals were initially anesthetized with 5% isoflurane. Following intubation for mechanical ventilation (RSP-1002; Kent Scientific, Torrington, CT) isoflurane was reduced to 2–3% during surgical preparation. The femoral artery and vein were catheterized for blood pressure monitoring, blood gas sampling, and fluid/drug administration. Via an abdominal incision a tube (PE 50) was inserted into the bladder through the bladder dome and secured by a ligature. Through a T-connector the tube was connected to a syringe pump and a pressure transducer to infuse the bladder with saline and record the bladder pressure. A ligature was tied around the base of the penis to prevent urine release into the MRI scanner. The animal was then placed in a custom-built plastic cradle, with the head secured by two ear bars and a bite bar to reduce motion. A neuromuscular blocking agent (pancuronium bromide, 0.6–0.8 mg·kg⁻¹·h⁻¹, intravenous [iv]) was administered during the experiment to further reduce head motion. After completing the surgery, urethane (1.2 g/kg initial dose followed by a supplemental dose of 0.4 g/kg every 4–5 h) was injected subcutaneously to replace isoflurane anesthesia. The arterial blood pressure and breathing pattern were continuously monitored with a multichannel recording unit (AcKnowledge; BioPak, Kerman, CA). End-tidal CO₂ was monitored (Stat profile pHOx; Nova Biomedical, Waltham, MA) and maintained at 3.5–4% by varying ventilation volume and frequency. The animal's body temperature was maintained at 37.5 ± 0.5°C with a feedback-controlled warm-water pad using a rectal thermal probe. The animal was given saline with 5% dextrose (1.5–2 ml·kg⁻¹·h⁻¹) iv.

MRI scanning protocol

All MRI experiments were performed on a 9.4 T/31 cm magnet (Magnex, Oxford, UK), interfaced to a Unity INOVA console (Varian, Palo Alto, CA). The actively shielded 12-cm-diameter gradient insert (Magnex) operates at a maximum gradient strength of 40 gauss/cm with a rise time of 130 μs. A rectangular surface coil (2.5 × 2 cm) was positioned on top of the animal's head for both excitation and reception. fMRI images were acquired using the spin-echo echo planar imaging (EPI) technique (Lee et al. 1999) with repetition time/echo time (TR/TE) = 600/32 ms, field of view (FOV) = 2.8 × 2.1 cm, and matrix size = 64 × 48. Scanning the entire brain required 5.4 s with a total of nine coronal slices (2-mm thickness of each slice). Anatomical images were acquired using TurboFLASH sequence with TR/TE = 20/5.5 ms and matrix size = 192 × 144.

For each animal 10–12 fMRI experimental trials of bladder filling (i.e., cystometrogram) were performed for blood oxygen level–dependent (BOLD) imaging. During each trial the rat brain was scanned continuously (see Fig. 1). The bladder was initially empty during the continuous fMRI scanning to collect the control data (control 1 in Fig. 1) that included ≥20 fMRI images (each image included nine coronal slices). Then, the bladder was slowly infused with saline (0.1–0.3 ml/min) until a large bladder contraction was induced (see Fig. 1). If the bladder contraction was not maintained for ≥2 min, additional bladder infusion was immediately started to maintain the micturition reflex so that ≥20 fMRI images could be acquired. At the end of the continuous fMRI scanning, the bladder was emptied by withdrawing the saline via the bladder catheter. Another 20 fMRI images were acquired after bladder emptying to collect control data (control 2 in Fig. 1). After a 10-min resting period for the reflex pathways to recover, the same fMRI experiment trial was repeated. In all, 105 experimental trials were performed on the nine animals.

Fig. 1.

Functional magnetic resonance imaging (fMRI) experimental protocol for an individual animal. MRI images acquired during each continuous scanning were extracted and averaged according to the different box cars for detecting brain activation during bladder storage or contraction.

After acquiring BOLD images, monocrystalline iron oxide nanoparticles (MION, 10 mg/kg) were administered iv to the animal. Then the fMRI scanning protocol as shown in Fig. 1 was repeated for 10 experimental trials on each animal (total of 70 trials on seven animals) with a faster infusion rate of 1 ml/min. The cerebral blood volume (CBV)–weighted MION imaging has been shown to increase the functional sensitivity compared with the BOLD technique (Jones 2002; Mandeville et al. 1998; Zhao et al. 2006). The purpose of the MION fMRI experiment was to further confirm the brain activation detected by BOLD fMRI during a micturition reflex.

fMRI data processing and analysis

The MRI image series acquired during each experimental trial in an individual animal was extracted to form a new image series (see Fig. 1). Twenty images were extracted from each of the following periods: during the initial empty bladder period (control 1 in Fig. 1), during saline infusion before a bladder contraction (storage in Fig. 1), during micturition contraction (contraction in Fig. 1), and during the period after bladder emptying (control 2 in Fig. 1). The series of images extracted during the same time period in repeated trials were averaged to increase the signal-to-noise ratio. Then, the averaged image series was reorganized according to different box cars (see Fig. 1) for detecting brain activation during either the storage phase or contraction phase. The newly formed MRI image series from each individual animal was used to determine brain activation by SPM5 software (available at http://www.fil.ion.ucl.ac.uk/spm).

The MRI images were preprocessed in SPM5 software by: 1) realignment for subtle motion correction, 2) coregistration of functional and anatomical images, 3) spatial normalization to a standard rat brain template image, 4) high-pass-filtered (cutoff period of 128 s) to remove low-frequency drift, and 5) spatial smoothing. After the preprocessing of the MRI data, statistical analysis was performed in two steps. First, individual statistical analysis was performed on each animal using a box-car function (either for storage or for contraction) and a general linear model to calculate the statistical parametric map (SPM). A contrast threshold of P < 0.01 for each voxel and an extend threshold of at least five contiguous suprathreshold voxels per cluster were used. In the second step, group statistical analysis (i.e., one-sample t-test) was performed on the fMRI images from the nine animals. Different P values (0.002–0.0001) with a cluster size at least two contiguous voxels were used to threshold the fMRI images and display the significantly activated brain regions. Due to the weak MRI signal in the brain stem area, a region of interest (ROI) analysis was performed in this region to detect the activation of PAG and PMC neurons. A one-sample t-test was performed in the brain stem area using P < 0.03, with a cluster size at least two contiguous voxels.

The final fMRI images were superimposed on the anatomical MRI template images and on the standard rat brain template drawings (Paxinos and Watson 2005) that correspond to the Bregma coordinates in the anterior–posterior direction as 2.28, 0.24, −1.80, −3.84, −5.88, −7.80, and −9.84 mm (see Fig. 2). The standard template drawings were normalized to the corresponding anatomical MRI template images by aligning the brain midline and ventricles and by maximally fitting the brain outline curvatures. Bregma coordinates (Paxinos and Watson 2005) were used to indicate the center of the activated brain region. The size of the activated brain area was indicated by the maximal lengths in three directions—i.e., LR, left(−)/right(+); DV, dorsal/ventral; and AP, anterior/posterior. The activation intensity of each brain region was indicated by the peak t value.

Fig. 2.

Blood oxygen level–dependent (BOLD) images showing brain stem activation associated with switching from the bladder storage phase to the bladder contraction phase. The locations of coronal brain sections (A–G) are indicated in the sagittal brain image at the bottom, which correspond to the Bregma coordinates in the anterior–posterior direction as 2.28, 0.24, −1.80, −3.84, −5.88, −7.80, and −9.84 mm. Region of interest (ROI) analysis was performed on the brain stem at coronal sections F and G to detect the activation. The periaqueductal gray (PAG) and pontine micturition center (PMC) are indicated by the blue arrows. The color scale bars indicate the t value.

RESULTS

Brain stem switch for micturition reflex

Because of the small size of the neuronal groups and the weak fMRI signal, ROI analysis was performed on the brain stem area to detect changes in activity during the switch from storage to micturition. As shown in Fig. 2, the brain stem switch for the micturition reflex involved both PAG and PMC (P < 0.03), confirming previous results of neurophysiological and neuroanatomical studies (Blok and Holstege 1996, 1998; de Groat 1975; Noto et al. 1991; Sasaki 2005). During storage the PAG was activated, but the PMC was inactive (Fig. 2, left column). During the micturition reflex, the PMC was activated and PAG activation was enhanced further (Fig. 2, right column). In addition, the ROI analysis also identified activation of the parvicellular reticular nucleus (PCRt) and inferior colliculus during the micturition reflex (Table 1 and Fig. 2).

Table 1.

Activated brain stem regions during storage or bladder contraction detected by BOLD images and ROI analysis (P < 0.03)

	Bregma Coordinates, mm			Activation Size, mm
Activated Brain Region	LR	DV	AP	ΔLR	ΔDV	ΔAP	Peak t Value
A. During storage
PAG	−0.5	5.9	−7.80	1.0	0.8	2	3.50
PCRt	2.6	8.5	−9.84	0.4	0.7	2	3.10
B. During micturition contraction
PAG	0.8	6.5	−7.80	1.8	2.0	2	5.20
PMC	1.3	7.8	−9.84	1.3	0.8	2	2.76
	−0.9	7.4	−9.84	0.5	0.3	2	2.54
Inferior colliculus	−2.0	5.9	−7.80	3.0	0.8	2	2.57

Brain network switch for micturition reflex

During bladder filling (i.e., the storage phase) multiple sites in the brain (Fig. 3) were activated including: motor cortex, primary and secondary somatosensory cortices, cingulate cortex, restrosplenial cortex, thalamus, putamen, insula, and septal nucleus (Table 2 and Fig. 3, left column, P < 0.002). Activation of motor cortex, as well as primary and secondary somatosensory cortices occurred only on the right side of the brain (Fig. 3, A–C) during storage.

Fig. 3.

BOLD images showing brain activation associated with switching from the bladder storage phase to the bladder contraction phase. The locations of coronal brain sections (A–G) are indicated in the sagittal brain image at the bottom, which have the same Bregma coordinates as those in Fig. 2. The color scale bars indicate the t value.

Table 2.

Activated brain regions during storage detected by BOLD images (P < 0.002)

	Bregma Coordinates, mm			Activation Size, mm
Activated Brain Region	LR	DV	AP	ΔLR	ΔDV	ΔAP	Peak t Value
Motor cortex	1.6	2.2	2.28	0.5	0.3	2	4.65
	1.7	2.3	0.24	0.4	0.4	2	4.71
Primary	3.4	2.2	0.24	1.4	0.6	2	5.35
somatosensory
cortex
Secondary	6.7	5.5	0.24	0.9	0.7	2	5.01
somatosensory	5.8	6.2	−1.80	1.6	2.0	2	8.06
cortex
Cingulate cortex	0.4	2.3	0.24	1.5	0.6	2	4.60
Retrosplenial cortex	−0.2	2.2	−1.80	1.4	1.4	2	6.11
	−0.2	2.1	−3.84	0.5	0.5	2	4.85
Thalamus	1.8	5.0	−1.80	0.9	0.5	2	5.21
	−0.7	5.6	−1.80	0.5	0.4	2	4.93
Putamen	2.0	5.2	0.24	0.7	0.7	2	5.62
	−1.8	5.2	0.24	0.8	0.3	2	5.20
Insula	5.8	6.5	−1.80	1.6	0.9	2	7.98
Septal nucleus	−0.6	6.0	0.24	0.8	1.0	2	6.41
	1.3	5.2	0.24	0.5	0.6	2	5.62
	−1.4	5.4	0.24	0.4	0.3	2	5.33

When the bladder volume was increased to the micturition threshold (0.65 ± 0.23 ml) and induced a large-amplitude bladder contraction (48 ± 10 cmH₂O), many brain regions that were activated during storage exhibited enhanced activation. The most strongly activated brain regions during micturition are shown in Fig. 3, right column (P < 0.0001). A comparison of the left and right columns in Fig. 3 reveals that activation of the motor cortex, thalamus, and putamen was enhanced. In addition the hippocampus, which was not activated during storage, was activated during micturition (also see Tables 2 and 3). Other brain regions that were activated during storage (Fig. 3, left column) but showed more modest activation during micturition (Fig. 3, right column) were identified by ROI analysis (Fig. 4, P < 0.002). These regions include the primary somatosensory cortex, cingulate cortex, retrosplenial cortex, and septal nucleus. Activation in the posterior area of the right secondary somatosensory cortex was decreased, but activation in the right anterior area was increased (Fig. 4, B and C).

Table 3.

Activated brain regions during bladder contraction detected by BOLD images (P < 0.0001)

	Bregma Coordinates, mm			Activation Size, mm
Activated Brain Region	LR	DV	AP	ΔLR	ΔDV	ΔAP	Peak t Value
Motor cortex	1.8	2.0	2.28	0.9	1.3	2	11.33
	1.9	1.7	0.24	0.9	1.0	2	7.65
	−1.6	1.7	2.28	0.9	0.7	2	9.18
	−1.8	2.1	0.24	1.3	1.3	2	7.89
Secondary	6.0	5.5	0.24	0.4	1.1	2	7.25
somatosensory
cortex
Cingulate cortex	0.9	2.4	2.28	0.9	0.7	2	7.52
	−0.9	2.1	2.28	0.5	0.6	2	6.95
	−0.3	3.2	0.24	0.9	0.7	2	8.67
Retrosplenial cortex	1.3	2.7	−5.88	1.0	0.7	2	8.21
Thalamus	2.3	4.9	−1.80	1.8	1.0	2	9.82
	−3.2	5.5	−1.80	1.4	0.7	2	9.25
	1.5	6.0	−3.84	0.8	1.1	2	7.87
	−3.1	5.3	−3.84	1.4	1.1	2	7.31
Putamen	2.8	4.8	0.24	1.8	1.1	2	10.35
	−2.0	4.4	0.24	0.9	0.6	2	7.31
	4.6	5.5	−1.80	1.4	0.7	2	9.51
Insula	5.6	6.0	2.28	0.9	0.6	2	8.17
Hippocampus	−0.8	4.0	−1.80	1.0	0.6	2	7.85
	2.2	3.4	−5.88	0.9	0.8	2	9.74

Fig. 4.

Region of interest (ROI) analysis of brain regions that did not show enhanced activation in Fig. 3. The 5 brain regions analyzed are marked by arrows. The locations of coronal brain sections (A–G) are the same as those in Fig. 2. Referencing the color scale bar in Fig. 3 left column for the t value.

In some brain regions that exhibited activity during storage and micturition the area of activation shifted during micturition (summarized in Table 4). For example, activation of cingulate cortex extended from a posterior area to more anterior area when switching from storage to micturition, whereas activation in the retrosplenial cortex expanded from an anterior area into a more posterior area (Table 4 and Fig. 4). The activation of insula shifted from a posterior area to an anterior area (Table 4 and Fig. 3). Although the predominant activation of both primary and secondary somatosensory cortices remained on the right side during the micturition reflex (Table 4 and Fig. 4), the motor cortex was activated bilaterally (Table 4 and Fig. 3).

Table 4.

Changes of the brain network activations when switching from storage to bladder contraction

Brain Region	Storage	Bladder Contraction
Motor cortex	× (right)	↑ (left + right)
Primary somatosensory cortex	× (right)	↑ (right)
Secondary somatosensory cortex	× (right anterior)	↑ (right anterior)
	× (right posterior)	↓ (right posterior)
Cingulate cortex	× (posterior)	↑ (anterior + posterior)
Retrosplenial cortex	× (anterior)	↑ (anterior + posterior)
Thalamus	×	↑
Putamen	×	↑
Insula	× (right posterior)	↑ (right anterior)
Septal nucleus	×	↑
Hippocampus	—	×

Activated (×); not activated (−); increased (↑); decreased (↓).

Brain network activation during micturition reflex: MION images

The activated brain regions during a micturition reflex were further confirmed by MION fMRI images (Fig. 5). Similar to BOLD images, brain activations were observed in motor cortex, primary somatosensory cortex, cingulate cortex, retrosplenial cortex, thalamus, putamen, insula, and hippocampus (Tables 3 and 5). In addition, MION fMRI also detected activation of hypothalamus, substantia nigra, and globus pallidus during a micturition contraction (Table 5).

Fig. 5.

Monocrystalline iron oxide nanoparticle (MION) images showing brain activation during the bladder contraction phase. The locations of coronal brain sections (A–E) are the same as those in Fig. 2. The color scale bar indicate the t value.

Table 5.

Activated brain regions during bladder contraction detected by MION images (P < 0.005)

	Bregma Coordinates, mm			Activation Size, mm
Activated Brain Region	LR	DV	AP	ΔLR	ΔDV	ΔAP	Peak t Value
Motor cortex	1.4	2.6	0.24	1.3	1.6	2	6.87
Primary	−6.6	4.0	2.28	1.0	0.4	2	5.56
somatosensory	5.3	4.5	0.24	1.4	1.0	2	5.93
cortex	−5.8	4.3	0.24	0.9	0.4	2	5.43
	−6.0	4.3	−1.80	0.3	0.3	2	4.95
Cingulate cortex	−1.1	3.9	2.28	1.7	1.0	2	16.39
	−0.4	3.6	0.24	0.4	0.4	2	6.15
	1.0	2.8	0.24	0.6	0.4	2	4.62
Retrosplenial cortex	1.0	0.9	−5.88	1.5	0.4	2	4.16
	−0.9	0.9	−5.88	1.3	0.4	2	4.53
	0.1	2.4	−1.80	1.3	1.1	2	4.18
Thalamus	2.6	5.7	−1.80	1.4	0.9	2	6.12
	−2.0	5.0	−1.80	2.0	1.3	2	6.33
	1.0	7.1	−1.80	0.5	0.6	2	5.14
	2.4	6.0	−3.84	2.5	2.7	2	9.33
	−2.9	5.4	−3.84	1.2	1.1	2	6.99
	−1.5	7.0	−3.84	0.9	0.7	2	5.56
Hypothalamus	0.3	7.1	−3.84	1.7	1.3	2	5.49
	1.6	8.3	−1.80	0.7	0.3	2	5.14
Putamen	5.0	4.9	−1.80	0.8	0.7	2	5.23
	5.2	5.4	−3.84	0.8	0.3	2	4.40
Hippocampus	0.4	3.4	−3.84	2.0	1.0	2	6.77
	2.9	2.0	−3.84	0.9	0.7	2	5.10
Insula	5.4	6.0	0.24	0.7	0.4	2	4.30
Substantia nigra	−3.0	7.7	−5.88	1.5	0.6	2	6.91
Globus pallidus	3.7	6.9	−1.80	1.8	0.7	2	5.51
	−3.8	6.0	−1.80	0.5	0.4	2	4.12

Average MRI signal intensity

During a micturition contraction the average MRI signal from the activated brain regions changed about 2% (Fig. 6). The BOLD signal change that was positive was opposite of the MION signal change (negative) due to the underlying physics. Compared with the average BOLD signal during a micturition contraction, the BOLD signal during storage was weaker and had a larger variation (compare black and red lines in Fig. 6), indicating that the brain activation during storage was less intense than the activation during micturition (Figs. 3 and 4).

Fig. 6.

MRI signal change during the bladder storage phase or during the bladder contraction phase.

DISCUSSION

This study provides support for the hypothesis based on previous neuroanatomical and neurophysiological studies (Barrington 1921; Blok and Holstege 1996, 1998; de Groat 1975; Noto et al. 1991) that in the anesthetized state a micturition switching circuit involving both the PAG and the PMC exists in rostral brain stem (Fig. 2). During the storage phase when the bladder was filling, the PMC was inactive but the PAG was activated, indicating that this region receives afferent input from the bladder prior to micturition. However, when bladder volume reached the micturition threshold, the PMC was activated and PAG activation was increased further (Fig. 2, right column), consistent with the idea that the PMC is the site for initiation of the micturition reflex (Barrington 1921; de Groat 1975; de Groat et al. 1993; Kuru 1965). Our studies also revealed that various sites in the forebrain were activated in parallel with activation of the PAG before micturition, indicating that a complex brain network processes afferent signals from the bladder, some of which may be relayed first through the PAG (Blok and Holstege 1994; Blok et al. 1995; Holstege 2005; Noto et al. 1991). Activation of other forebrain areas occurred in parallel with the activation of the PAG/PMC during micturition. These areas in the forebrain may subserve multiple functions related to the reflex control of the lower urinary tract including 1) processing sensory input from the bladder, 2) modulating efferent storage mechanisms, 3) regulating the initiation of micturition, and 4) regulating the coordination between bladder and urethra activity during micturition. The data are consistent with the conclusions from human fMRI and PET brain imaging studies (Athwal et al. 2001; Blok et al. 1997a,b, 1998, 2006; Dasgupta et al. 2005; Di Gangi Herms et al. 2006; Griffiths et al. 2005, 2007; Kitta et al. 2006; Kuhtz-Buschbeck et al. 2005, 2007; Matsuura et al. 2002; Mehnert et al. 2008; Nour et al. 2000; Sakakibara et al. 1999; Seseke et al. 2006, 2008; Zhang et al. 2005) that various forebrain regions participate in the control of urine storage and voiding.

However, in contrast to previous human fMRI brain imaging experiments, the present study imaged brain activation during a continuous slow infusion of the bladder in an attempt to mimic the physiological accumulation of urine. In addition differences in brain activity were evaluated in three conditions: 1) empty bladder, 2) partially filled bladder, and 3) reflexly active bladder to determine in the same animal which brain regions participated in storage and voiding functions. In some human fMRI studies (Griffiths et al. 2005, 2007; Kuhtz-Buschbeck et al. 2005; Mehnert et al. 2008) the bladder was filled to a volume causing a strong desire to void and micturition was voluntarily inhibited because the subjects were not allowed to urinate in the MRI scanner. In other experiments bladder afferents were activated by rapidly and repeatedly infusing and withdrawing a small amount of saline from the bladder to obtain multiple measurements during urine storage that could in turn be averaged to detect small changes in brain activity (Griffiths et al. 2005, 2007; Mehnert et al. 2008). However, these techniques did not mimic physiological distention of the bladder and measured brain activation only during the storage phase.

Early PET studies (Blok et al. 1997b, 1998; Nour et al. 2000) investigated brain activation during both storage and micturition in the same subject, but very few brain regions were activated during the storage phase and PAG activation was not significant. Later PET studies imaged the brain only either during the bladder storage phase (Athwal et al. 2001; Blok et al. 2006; Dasgupta et al. 2005; Matsuura et al. 2002) or during overactive bladder contractions (Kitta et al. 2006). Most fMRI studies investigated brain activation during voluntary pelvic floor muscle control (Di Gangi Herms et al. 2006; Kuhtz-Buschbeck et al. 2005, 2007; Seseke et al. 2006, 2008; Zhang et al. 2005). A few fMRI studies investigated bladder distention, but the bladder was infused only to the volume causing a strong desire to void and micturition was not allowed (Griffiths et al. 2005, 2007; Kuhtz-Buschbeck et al. 2005; Mehnert et al. 2008). An fMRI brain imaging study to reveal how brain activation switches from storage to micturiton in humans is currently not available.

Human studies have been performed under awake conditions, whereas our studies in rats were conducted under urethane anesthesia to examine reflex mechanisms. Under urethane anesthesia, rats exhibit normal urine storage at low intravesical pressures and coordinated activity of the bladder and urethral sphincter during voiding (Maggi et al. 1986; Yoshiyama et al. 1994). It has been proposed that urethane-anesthetized rats might be a useful model for neurogenic detrusor overactivity (NDO) because C-fiber bladder afferents that have been implicated in NDO in humans act in combination with A-fiber afferents to modulate reflex voiding in these animals. On the other hand, only A-fiber bladder afferents initiate voluntary voiding in awake rats (Chuang et al. 2001). Reflex control of micturition is clinically relevant because neurogenic bladder dysfunctions including bladder overactivity, urgency, and incontinence caused by brain disorders (de Groat et al. 1993; Fowler 1999) are often generated by reflex mechanisms that are resistant to voluntary control. Studies in awake humans are complicated by the contribution and interaction of a variety of brain processes, including attention, expectation, decision, and emotion (Mayer et al. 2006), that very likely influence voiding function. These processes may also influence voiding in awake rats but should be minimized under anesthesia.

In our study we used BOLD and MION techniques to identify activated brain areas during micturition, whereas the BOLD technique is routinely used in human studies. Although these two methods identified similar brain centers the exact locations of activation were not always same. This is probably due to differences in sensitivity and the underlying physics of the BOLD and MION techniques. MION, which has a long half-life in the blood, can be detected by fMRI due to increased local cerebral blood volume that causes an increase in the local concentration of iron oxide (i.e., MION) (Jones 2002; Mandeville et al. 1998; Zhao et al. 2006). In contrast, a BOLD signal results from a decrease in the concentration of deoxyhemoglobin as venous blood becomes more oxygenated in brain areas of increased blood flow and volume. MION can enhance fMRI sensitivity approximately fivefold compared with BOLD. However, the fMRI images obtained with either technique show a change in hemodynamics that reflects only a net change of total neuronal activity. The change in activity could be due to alterations in afferent information arising in the bladder or due to changes in motor commands resulting from the initiation of a micturition reflex. Thus fMRI imaging does not provide insights into the functions of activated regions with regard to regulation of bladder activity.

Possible functions of the brain regions identified in our study can be inferred from previous electrophysiological and neuroanatomical experiments of other investigators. For example, it is well known that bladder afferent signals can be processed both in the PAG and in the thalamus and then pass to more distal sites in the anterior cingulate cortex and the insula before reaching the orbitofrontal cortex (Craig 1996, 2002; Mayer et al. 2006). These areas were activated during bladder filling in our experiments. The primary and secondary somatosensory cortical areas were also activated during bladder filling. Activation of these areas was unexpected because they are not assumed to play a role in visceral sensation.

The afferent signals generated in the bladder during filling occur in low-threshold (nonnociceptive) mechanoreceptive afferents; however, during micturition they could also arise in high-threshold (nociceptive) as well as low-threshold afferents because the bladder outlet was occluded in our experiments, thereby generating high pressures during the reflex bladder contractions. The high bladder pressure (40–60 cmH₂O) might have activated nociceptive bladder afferents during the micturition reflex, although the bladder was maintained at a volume just slightly above the micturition threshold volume (see Fig. 1). Measurements of immediate early gene (IEG) expression (c-fos) in the brain after noxious visceral stimulation (Clement et al. 1996) or bladder irritation with cyclophosphamide, which activates nociceptive bladder afferents (Bon et al. 1996), identified IEG expression in several brain regions that were also identified in the present experiments during micturition. Thus these regions, including the thalamus, periaqueductal gray, inferior colliculus, and hypothalamus (identified using the MION method only) could have been activated by nociceptive bladder afferent nerves.

Many brain regions identified in this study have been identified in transneuronal pseudorabies virus (PRV) tracing studies in which PRV was injected into the rat urinary bladder or urethra and then transported in a retrograde manner along peripheral and central efferent pathways to the spinal cord and then to the brain (Grill et al. 1999; Marson 1997; Nadelhaft et al. 1992; Sugaya et al. 1997; Vizzard et al. 1995). It is believed that the PRV initially labels neurons in the motor limb of the micturition reflex and then later labels neurons in sensory and modulatory circuits that send information to the motor pathways. Areas identified by both PRV labeling and using fMRI in the present studies include the PMC, PAG, hypothalamus, substantia nigra, and cerebral cortex. Thus these areas may be part of the efferent limb of the micturition reflex pathway.

Sites in the motor cortex, basal ganglia (putamen, globus pallidus), and substantia nigra, which may play a role in bladder-striated sphincter coordination, were also activated during the micturition reflex. Motor cortex, somatosensory cortex, cingulate cortex, insula, putamen, thalamus, and hypothalamus (Figs. 3–5 and Tables 2–5) are the brain regions most frequently reported to be involved in human voluntary micturition control (Fowler and Griffiths 2009; Griffiths and Tadic 2008; Kavia et al. 2005). It is worth noting that the activated region of cingulate cortex extended from a posterior region to a more anterior region when storage switched to micturition (Table 4 and Figs. 3 and 4). A similar posterior–anterior shift in the activation of the cingulate cortex was also observed in previous human studies during micturition (Athwal et al. 2001; Blok et al. 1997b, 1998; Matsuura et al. 2002). It was proposed that in humans the posterior region of cingulate cortex might be involved in processing afferent signals during storage and in turn contribute to the generation of bladder filling sensations, whereas the anterior region might be involved in the perception of changes in bladder volume rather than the process of micturition (Blok et al. 1997b, 1998; Kavia et al. 2005). In this study the micturition contraction occurred under isovolumetric conditions; therefore it is reasonable to conclude that activation of the anterior cingulate cortex was not related to the change of bladder volume, but rather directly related to the micturition reflex. This speculation is also in agreement with the proposal suggesting an “evaluation” function for the posterior cingulate cortex but an “executive” function for the anterior cingulate cortex (Vogt et al. 1992).

A similar posterior–anterior shift in activation was also observed in right insula (Tables 2–4 and Fig. 3), indicating that different regions of the insula might play different roles in processing sensory input during storage or micturition. Activation of the right insula was identified by human brain imaging studies during both storage and micturition (Blok et al. 1998; Griffiths et al. 2005, 2007; Kitta et al. 2006; Matsuura et al. 2002; Nour et al. 2000), but the posterior–anterior shift was not reported.

Activation of the retrosplenial cortex, observed in our studies during storage and micturition, has not been reported in human imaging experiments nor identified as an area of interest for bladder function in physiological or anatomical experiments in animals. In humans the retrosplenial cortex, which is considered to be part of the posterior cingulate cortex, sends projections to the hippocampus and is thought to be involved in emotional processes and memory (Maddock 1999; Wyss and Van Groen 1992). During micturition the area of activation in the retrosplenial region shifted from anterior to posterior (Table 4 and Fig. 4), which coincided with the activation of the hippocampus (using both BOLD and MION imaging), an area that was inactive during filling (Tables 3–5). Hippocampal activation during micturition was identified in human imaging studies (Blok et al. 2006; Griffiths and Tadic 2008; Matsuura et al. 2002; Zhang et al. 2005). The hippocampus was also labeled in a tracing study after injection of the rat bladder with pseudorabies virus (Grill et al. 1999). The relatively small size of hippocampal activation in this study (Figs. 3 and 5) might reflect the effect of anesthesia. It is probable that only the reflex component of the emotional brain system was activated under anesthesia.

Although rats do not have a prefrontal cortex and their frontal cortex is not as fully developed as that in monkeys and humans (Preuss 1995), behavior observations suggest that rats do exhibit voluntary control of micturition that can be influenced by their environment. For example, normal Sprague–Dawley rats urinate ubiquitously throughout their cage, but rats exposed to chronic social stress urinate less frequently and always at the corner of the cage (Wood et al. 2009), indicating that they do exhibit to some extent “social continence.” Social continence mechanisms have been linked with activation of the frontal cortex, which has frequently been reported in human brain imaging studies of micturition control (Fowler and Griffiths 2009; Griffiths and Tadic 2008; Kavia et al. 2005). Activation of the frontal cortex was not detected in our study of reflex micturition, although this might reflect the fact that voluntary control of micturition was lost due to anesthesia. On the other hand, primary and secondary somatosensory cortices were activated during both storage and micturition (Table 4 and Fig. 4). Only a few human brain imaging studies (Athwal et al. 2001; Di Gangi Herms et al. 2006; Nour et al. 2000) detected the activation of somatosensory cortex during micturition control. These studies attributed the activation to either the presence of a urethral catheter or the contraction of pelvic floor muscle. In this study, a ligature was applied to the base of the penis to prevent fluid release from the bladder during micturition contraction. The increased urethral pressure during micturition contraction might contribute to the enhanced activation of the somatosensory cortex in this study. Neuronal responses to visceral stimulation have also been recorded in the somatosensory cortex in experimental animals (Bruggemann et al. 1994, 1997). Another possibility is that the concomitant contraction of the external urethral sphincter (EUS) during both bladder filling and contraction in the rat might cause the activation. The EUS of rats contracts intermittently during micturition to facilitate voiding. This is different from the inhibition of EUS activity in humans and cats and may explain the enhanced activation of motor cortex during micturition in the rat (Table 4 and Fig. 3).

The septal nucleus, which was activated during storage, was less activated during micturition. Electrical stimulation in this area in animals (Hess 1947) evokes bladder contractions, whereas injury to this area in humans due to aneurysms (Andrews and Nathan 1964; Nathan 1976) produces urgency, frequency, and incontinence. Thus neurons in the septal area may play a role in the control of motor pathways to the bladder.

MION imaging identified more brain regions than BOLD imaging, including substantia nigra and globus pallidus (Table 5). Parkinson's disease associated with selective degeneration of dopaminergic neurons in substantia nigra causes urinary disorders (Araki et al. 2000). Lesions in rat substantia nigra cause bladder overactivity (Yoshimura et al. 2003). In the cat electrical stimulation of substantia nigra (Lewin et al. 1967; Yoshimura et al. 1992) or the globus pallidus (Lewin et al. 1965, 1967) inhibited bladder activity, whereas activation of the globus pallidus has been observed in human brain imaging during micturition (Nour et al. 2000).

Predominant activation on the right side of the brain was noticeable (Fig. 3 and Table 4). Previous human brain imaging studies (Blok et al. 1997b, 1998; Mehnert et al. 2008) also reported right-side predominance. Although only male rats were used in this study, PET imaging studies have shown that the right-side predominance exists in both men and women (Blok et al. 1997b, 1998). Gender difference of voluntary micturition control was not observed with human brain fMRI (Kuhtz-Buschbeck et al. 2007; Seseke et al. 2008).

fMRI has been used previously in several animal studies (Angenstein et al. 2009; Colonnese et al. 2008; Westlund et al. 2009). However, our study is the first to apply fMRI to imaging brain control of micturition in an animal model. Although fMRI is noninvasive and easily applicable to investigate the human brain, the use of fMRI in an animal model of reflex micturition provides a wide range of opportunities for invasive physiological and pharmacological studies. The model could be used to investigate questions such as how brain switching for reflex micturition is influenced by damaging a specific brain region and how it is changed by centrally acting drugs. Animal brain fMRI is a very useful tool to monitor neuronal activation in a large brain area compared with traditional single-unit electrical recordings.

This study confirmed the hypothesis based on previous neuroanatomical and neurophysiological studies that the brain stem switch for reflex micturition control involves both PAG and PMC. It also revealed a complex brain network that probably modulates the activation of the brain stem switch to initiate micturition. Understanding brain control of reflex micturition is very important for elucidating the underlying mechanisms of neurogenic bladder dysfunctions, including bladder overactivity, frequency, urgency, and incontinence caused by brain disorders.

GRANTS

This work was supported by National Institutes of Health Grants DK-068566, DK-077783, NS-045078, and EB-003375.