Changes in Brain Activity and Functional Connectivity During Spinal Nerve Stimulation in a Rat Model of Overactive Bladder
Article information
Abstract
Purpose
Sacral neuromodulation is widely used for refractory overactive bladder; however, its mechanism of action remains unclear. This study aims to investigate real-time changes in brain activity and functional connectivity (FC) during neuromodulation in an overactive bladder (OAB) rat model using functional magnetic resonance imaging.
Methods
Twelve female Sprague Dawley rats were implanted with fine bipolar electrodes adjacent to the L6 nerve root. Cystometry was performed on normal rats, acetic acid-induced OAB rats, and during spinal nerve stimulation (SNS) to confirm efficacy. Task-based functional magnetic resonance imaging (fMRI) was acquired using a 20-second rest-stimulus cycle, followed by T2-weighted anatomical imaging on a 9.4T MRI scanner. Comparative analyses examined changes in the amplitude of low-frequency fluctuations (ALFF) and FC between normal and OAB rats. Brain activity during SNS was further assessed using the general linear model (GLM) and FC analysis. Statistical significance was defined as P<0.05 (family-wise error–corrected).
Results
Compared with normal rats, OAB rats exhibited increased ALFF in the left prefrontal cortex, periaqueductal gray (PAG), and left primary somatosensory cortex. In addition, FC between the PAG and pons was enhanced (P=0.002). GLM analysis revealed that the left primary somatosensory cortex, left prefrontal cortex, corpus callosum, left secondary motor area, and right brainstem exhibited decreased activity during SNS (P<0.05). Significant FC changes were observed between several regions: the left cerebellum and left caudal zona incerta (P=0.024), right fasciculus retroflexus and left ventral orbital area (P=0.025), and between the pons and PAG (P=0.004). Seed-to-voxel analysis indicated altered FC between clusters identified in the GLM analysis and regions including the PAG, left cingulate area, left prefrontal cortex, left caudate putamen, and right granular insular cortex.
Conclusions
Our fMRI study identified several alterations in brain activity during SNS in rats. Specifically, activity in the left prefrontal cortex decreased during SNS, and FC between the PAG and pons was reduced. These changes may represent central mechanisms underlying sacral neuromodulation in OAB patients.
• HIGHLIGHTS
- Using task-based functional magnetic resonance imaging in an overactive bladder (OAB) rat model, this study demonstrates that spinal nerve stimulation (SNS) acutely inhibits neural activity in critical brain regions, including the left prefrontal cortex and primary somatosensory cortex.
- Functional connectivity analysis reveals heightened periaqueductal gray (PAG)-pons connectivity in OAB rats, which is significantly attenuated by SNS—suggesting a central regulatory mechanism of SNS on bladder function.
- These findings offer direct evidence for the central neural mechanisms of sacral neuromodulation in OAB, identifying the prefrontal cortex and PAG-pons network as potential targets to optimize neuromodulation therapies.
INTRODUCTION
Overactive bladder (OAB) is a common condition primarily characterized by urinary urgency, often accompanied by increased frequency and nocturia, with or without urgency urinary incontinence [1]. Sacral neuromodulation (SNM) is a therapeutic method for OAB that involves electrical stimulation of sacral nerves via an implanted electrode in the S3 foramen. By electrically stimulating the sacral nerves, SNM can inhibit overactive micturition reflexes, thereby reducing symptoms of urinary urgency and frequency [2].
The mechanism of SNM primarily involves regulation of the nervous system, including both afferent and efferent nerve fibers of the sacral nerve, which modulates bladder storage and voiding functions [3, 4]. Recent studies suggest that SNM not only exerts peripheral effects but also influences the central nervous system through complex neural networks. Blok employed positron emission tomography-computed tomography to detail the effects of SNM on the human brain, revealing decreased regional cerebral blood flow in the cingulate cortex, ventromedial orbitofrontal cortex, and adjacent midline thalamus, along with increased blood flow in the dorsolateral prefrontal cortex (DLPFC) [5]. The research of Weissbart et al. [6] using 1.5T resting-state functional magnetic resonance imaging (fMRI) in OAB patients observed decreased activation in the DLPFC, left anterior cingulate cortex (ACC-L), bilateral insula, and bilateral orbitofrontal cortex after 6 weeks of SNM therapy. Another study reported activation of the right inferior frontal cortex and deactivation of the pons and periaqueductal gray (PAG) during SNM [7].
However, resting-state fMRI primarily captures static changes in brain activity during rest, potentially overlooking dynamic and transient neural changes induced by SNM [8]. In contrast, task-based fMRI can provide more specific spatial and temporal information regarding the impact of SNM by applying various stimulation conditions within a single scanning session, thereby capturing dynamic responses in brain region activations [9].
Exploring how SNM alters lower urinary tract function by examining brain activity during stimulation can help elucidate its mechanisms of action and optimize treatment strategies. However, most fMRI studies have focused on resting-state activity because SNM devices can only be scanned in MRI systems with a magnetic field strength below 1.5T while powered off. The OAB rat model and spinal nerve stimulation (SNS) have been widely used to study the mechanisms of SNM in OAB [10, 11]. To investigate real-time brain activity during SNM, our study employed an acetic acid-induced OAB rat model and performed fMRI scans using a block-designed SNS protocol.
MATERIALS AND METHODS
Animals and Experimental Design
A total of 12 female specific-pathogen-free Sprague Dawley rats (8 weeks old, 240–280 g) were used in this study. All animals were housed in a controlled environment (25°C±1°C, 12-hour light/dark cycle, 50% relative humidity) with ad libitum access to food and water.
The experimental design is illustrated in Fig. 1. Each rat underwent the complete experimental procedure. For SNS, a fine bipolar electrode made of platinum-iridium alloy was placed adjacent to the L6 nerve root, which innervates the urinary bladder. Electrical stimulation of the L6 nerve can inhibit bladder activity in rats [12]. Cystometry was performed 3 days after electrode implantation. Subsequently, fMRI and anatomical MRI of the rat brain were conducted. The OAB model was established using a 0.25% acetic acid bladder infusion and verified by cystometry [11]. The effectiveness of SNS was confirmed by cystometry during L6 nerve electrical stimulation. fMRI was then performed first without stimulation and subsequently using a stimulation paradigm consisting of a 20-second rest period followed by a 20-second stimulation period. The stimulation parameters were set at a frequency of 10 Hz, a pulse width of 200 μsec, and an intensity corresponding to the motor threshold. Electrical pulses were generated using a PowerLab 4/26 (ADInstruments International Trading Co.).
Experimental workflow. SD, Sprague Dawley; CMG, cystometry; MRI, magnetic resonance imaging; fMRI, functional MRI; T2MI, T2-weighted image; AA, acetic acid; SNS, spinal nerve stimulation.
Animal Surgery
To administer SNS, a fine bipolar electrode made of platinum-iridium alloy was positioned adjacent to the right L6 spinal nerve root [12]. Initially, the rats were anesthetized with continuous inhalation of 3% isoflurane. The skin over the upper pelvic region was shaved, and the vertebral body was identified by palpation. An incision approximately 1.5 cm in length was made along the right edge of the vertebral body at the sacrumlumbar junction following the body’s longitudinal axis. The deep fascia was incised, and the muscles were bluntly dissected to expose the lamina. The electrode was placed in the anterior neural groove of the sacrum at the point where the L6 nerve exits the foramen and descends. If necessary, a portion of the lamina was removed to facilitate electrode placement. The other end of the wire was secured subcutaneously at the back of the rat’s neck. Upon initiating electrical stimulation, responses such as hind-toe twitches or pelvic floor muscle contractions were observed, confirming proper electrode placement. The electrode position is shown in Fig. 2A. The stimulation intensity was then adjusted to determine the threshold at which a motor response was elicited [10].
Electrodes implantation and cystometry. (A) The location of the implanted electrode. (B) Cystometry with 0.9% infusion, 0.25% AA infusion, 0.25% AA infusion + SNS. AA, acetic acid; SNS, spinal nerve stimulation.
Cystometry and OAB Model Construction
Three days after electrode implantation, cystometry was performed under 3 conditions: saline infusion, 0.25% acetic acid infusion, and 0.25% acetic acid infusion with SNS. The rats were anesthetized with an intraperitoneal injection of 20% urethane (120–150 mg/100-g body weight) and placed in a supine position, secured appropriately. A PE-50 catheter was inserted into the urethra, and a T-shaped catheter connected the PE-50 catheter to both the bladder perfusion device and a pressure sensor for recording bladder pressure and infusion. Bladder pressure was continuously monitored throughout the infusion process. Initially, cystometry was performed with saline infusion. After the first fMRI scan, 0.25% acetic acid was continuously infused into the bladder and cystometry was recorded. Finally, cystometry was conducted concurrently with SNS and acetic acid infusion [11]. The stimulation parameters were set at 10 Hz, 200 μsec, and at the motor threshold intensity.
MRI Data Acquisition
All MRI data were acquired using a 9.4T horizontal animal MRI system (Biospec 94/30 USR, Bruker). During scanning, a thermoregulation system maintained the rats’ body temperature, and heart rate and respiration were monitored. Blood oxygenation level-dependent functional images were obtained using a gradient-echo echo-planar imaging sequence with the following parameters: field of view (FOV)=30×24 mm, echo time (TE)=17 msec, repetition time (TR)=2,000 msec, flip angle= 90°, voxel size=0.3×0.3×0.5 mm, slice thickness=0.5 mm, and 47 slices covering the entire brain. Shimming was performed prior to scanning to ensure image quality, followed by 3 dummy scans before formal data acquisition. A total of 300 volumes were collected over 600 seconds, with each rat undergoing 15 cycles of stimulation and rest.
High-resolution T2-weighted structural scans were obtained using a fast spin echo sequence with the following parameters: FOV=30×24 mm, TR=2,500 msec, TE=33 msec, flip angle= 8°, voxel size=0.1×0.1×0.5 mm, and 47 slices covering the entire brain.
fMRI Data Preprocessing
Preprocessing was performed in MATLAB R2014a (The Math-Works, Inc.) using the SPM8 toolbox (Statistical Parametric Mapping 8, Wellcome Trust Institute of Neurology, University College London, UK, www.fil.ion.ucl.ac.uk/spm) and the SPMMouse toolbox (http://spmmouse.org) [13]. Both functional and anatomical images were converted from DICOM to NIfTI format using the dcm2nii toolbox (https://www.nitrc.org/plugins/mwiki/index.php/dcm2nii). The first 3 dummy scans were excluded, and slice timing correction was applied. Motion correction was performed based on the first volume, and subjects with head displacement greater than 1 mm or rotation exceeding 1° were excluded. Segmentation and normalization were conducted using the Sigma rat atlas, which is based on Waxholm space [14]. Rat brain anatomical images were segmented into gray matter, white matter, and cerebrospinal fluid using tissue probability maps, and a mapping matrix was computed to align the fMRI data with the standard template. To enhance data quality and statistical analyses, smoothing was applied using a Gaussian kernel with a full width at half maximum of 0.9×0.9×0.9 mm.
Amplitude of Low-Frequency Fluctuation
Amplitude of low-frequency fluctuation (ALFF) analysis was performed to compare changes in brain region activity between normal and OAB rats. A fast Fourier transform was applied to the preprocessed time series data to extract the power spectrum. The square root of the power within the low-frequency band (0.01–0.1 Hz) was computed for each voxel to obtain the ALFF values. Statistical significance was determined using a voxel-wise threshold of P<0.001 and a cluster-level correction for multiple comparisons (P<0.05, family-wise error [FWE]-corrected).
General Linear Model
General linear model (GLM) implemented in SPMMouse was used to assess brain activation during SNS [15]. Head motion parameters obtained during preprocessing were included in the design matrix as covariates to mitigate motion artifacts. Regressors for the experimental conditions were modeled as boxcar functions convolved with the canonical hemodynamic response function. Contrast images were generated to compare the stimulation condition with baseline. Group-level analysis was performed using a 1-way t-test with a contrast matrix comparing stimulation and baseline conditions. Statistical significance was defined as P<0.001 at the voxel level and P<0.05 at the cluster level (FWE-corrected).
Functional Connectivity
Functional connectivity (FC) analysis was performed for both normal and OAB rats. The brain was parcellated using the Sigma rat atlas [14], and Pearson correlation coefficients were calculated between the time series of each pair of brain regions.
The BASCO (BetA Series COrrelation) toolbox was used for task-based FC analysis [16]. Mean beta-series for each region were extracted from the GLM under both stimulus and baseline conditions. FC was computed by calculating the Pearson correlation between the beta-series of each region pair. A seed-to-voxel FC analysis was conducted to identify brain regions exhibiting FC changes with clusters of altered brain activity identified by the GLM analysis. BASCO was used to extract mean beta-series from predefined seed clusters and correlate them with the beta-series of individual voxels across the brain.
For group-level comparisons, correlation coefficients were Fisher z-transformed, and the paired t-test was performed on these z-scores. The significance level was set at P<0.05 (FWE-corrected).
RESULTS
All 12 rats successfully underwent electrode implantation and completed the fMRI scans. No displacement or rotation exceeding 1 mm or 1° was observed in the fMRI images. Cystometry confirmed the successful establishment of the OAB model. Compared with saline infusion, bladder contraction intervals were reduced by more than 50% following acetic acid infusion. SNS inhibited abnormal contractions and significantly increased bladder contraction intervals (Fig. 2B).
ALFF and FC Changes Between Normal and OAB Rats
Compared with normal rats, the left prefrontal cortex (PFC), PAG, and left primary somatosensory cortex (PSC-L) exhibited increased ALFF in OAB rats (P<0.05, FWE-corrected) (Table 1, Fig. 3). In addition, FC between the PAG and pons was enhanced in OAB rats (P=0.002, FWE-corrected) (Fig. 4). These findings indicate significant alterations in brain functional activity in OAB rats.
Brain regions with altered ALFF between normal and overactive bladder rats
Regions with amplitude of low-frequency fluctuation change between normal and overactive bladder rats. PAG, periaqueductal gray; PSC-L, left primary somatosensory cortex; PFC-L, left prefrontal cortex.
Altered functional connectivity between normal and overactive bladder (OAB) rats. (A) Brain surface axial view of PAG and pons. (B) The mean functional connectivity of normal and OAB rats. PAG, periaqueductal gray. **P<0.01, family-wise error–corrected.
Brain Regions With Altered Activation Levels During SNS
GLM analysis revealed a significant decrease in activation levels in several brain regions during the stimulus condition compared with baseline. Specifically, the PSC-L, left PFC (PFC-L), right corpus callosum (CC-R), left secondary motor area (SMAL), and right brainstem (Brainstem-R) exhibited reduced activity (Table 2, Fig. 5) (P<0.05, FWE-corrected). No regions showed increased activity after correction for multiple comparisons. These results suggest that SNS primarily exerts an inhibitory effect on various brain regions.
Brain regions with altered activation level during SNS in GLM analysis
Brain regions with altered activation level during spinal nerve stimulation. PSC-L, left primary somatosensory cortex; PFC-L, left prefrontal cortex; CC-R, right corpus callosum; SMA-L, left secondary motor area.
Brain Regions With Altered FC During SNS
Compared with baseline, FC between several regions during the stimulus condition exhibited a decreasing trend (Fig. 6). Specifically, FC between the left cerebellum and the left caudal zona incerta decreased (P=0.024, FWE-corrected); between the right fasciculus retroflexus and the left ventral orbital area (VOA-L) decreased (P=0.025, FWE-corrected); and between the pons and PAG decreased (P=0.004, FWE-corrected). These changes, consistent with the GLM analysis, further indicate an inhibitory effect of SNS in OAB rats.
Brain regions with altered functional connectivity during SNS. (A) Brain surface axiel view. (B) Brain surface sagittal view. (C) Cerebellum-L and ZI-Ca-L. (D) FR-R and VOA-L. (E) Pons and PAG. VOA-L, left ventral orbital area; ZI-Ca-L, left zona incerta caudal part; PAG, periaqueductal gray; FR-R, right fasciculus retroflexus. Correlation coefficients, Fisher z-transformed. *P<0.05, **P<0.01, family-wise error–corrected.
Voxels With Altered FC to Deactivation Regions in GLM Analysis During SNS
Based on the GLM analysis, 5 clusters with decreased activity were identified and used as seeds for a whole-brain seed-to-voxel analysis (Table 3). The analysis revealed that cluster 1 exhibited decreased FC with the left cingulate area (CA-L) (P=0.007, FEW-corrected). Cluster 2 showed reduced FC with the ipsilateral PSC-L (P=0.003, FEW-corrected). FC between cluster 3 and the PFC-L also decreased (P=0.004, FWE-corrected). Cluster 4 displayed decreased FC with the PAG (P=0.002, FWE-corrected). Cluster 5 demonstrated increased FC with the pons (P<0.001, FWE-corrected), left caudate putamen (CP-L) (P<0.001, FWE-corrected), and PAG (P<0.001, FWE-corrected), as well as with the right granular insular cortex (GIC-R) (P<0.001, FWE-corrected), while it exhibited decreased FC with another region within the ipsilateral brainstem (P=0.006, FWE-corrected). The visualization of these clusters’ FC is presented in Fig. 7, representing more specific alterations in brain functional activity induced by SNS.
Voxels with altered FC to deactivation regions in GLM during SNS
Voxels with altered functional connectivity to deactivation regions in general linear model during spinal nerve stimulation. CA-L, left cingulate area; PSC-L, left primary somatosensory cortex; PFC-L, left prefrontal cortex; PL-L, prelimbic area; CC-R, right corpus callosum; SMA-L, left secondary motor area; PAG, periaqueductal gray; CP-L, left caudate putamen; GIC-R, right granular insular cortex; Brainstem-R, right brainstem.
DISCUSSION
SNM is widely used for treating OAB patients who are refractory to conventional therapies. Although SNM has demonstrated significant efficacy in improving urinary symptoms, the exact mechanisms by which it modulates bladder function remain unclear. Understanding the mechanisms of SNM in treating OAB will aid in selecting appropriate patients and refining clinical treatment methods. An increasing number of studies on lower urinary tract dysfunction are now focusing on the central nervous system [17, 18]. Consequently, task-based fMRI was employed in our study to explore changes in brain activity during SNS in an OAB rat model.
Our results indicated that OAB rats exhibited increased ALFF in the left PFC, left PSC-L, and PAG. Additionally, FC between the PAG and pons was enhanced in OAB rats. During SNS, the PSC-L, PFC-L, CC-R, SMA-L, and right brainstem exhibited decreased activity compared with baseline. Weissbart et al. [6] reported a significant reduction in brain activity after 6 weeks of SNM treatment in OAB patients—particularly in the PFC-L, ACC-L, bilateral insula, and bilateral orbitofrontal cortex—and the fMRI study of Gill et al. [7] observed decreased activity in the pons and PAG during SNM. Although the specific brain regions identified in our study do not entirely overlap with these previous findings, the overall trend of decreased functional activity is consistent, suggesting that SNM may modulate bladder function by inhibiting activity in specific brain regions.
Several regions associated with SNS were identified in our study, including the PSC-L, PFC-L, SMA-L, ACC-L, and PAG. The PSC is implicated in bladder fullness and other visceral sensations [19]; a reduction in PSC activity during SNS may indicate modulation of sensory input and altered perception of bladder sensation [20, 21]. The PFC is involved in advanced cognitive processes that govern voluntary urination control, and studies have shown that interaction between the anterior cingulate cortex (ACC) and the PFC is essential for regulating bladder activity [22, 23]. Disorders in the ACC, as seen in conditions such as multiple sclerosis, can lead to urinary dysfunction [24]. Moreover, SMA activity correlates with motor control during micturition, and altered activity in this region has been reported in individuals with urinary dysfunction [20, 25]. The deactivation observed in the right brainstem in our study is not confined solely to the pons or PAG. Previous transneuronal retrograde labeling studies have identified multiple brain regions involved in bladder innervation—including the midline raphe (magnus and obscurus), locus coeruleus, subcoeruleus, reticularis gigantocellularis, and nucleus paragigantocellularis [26, 27]—suggesting that our observed brainstem changes may overlap with these areas.
Previous research has established a connection between the frontal cortex and the mechanisms underlying OAB [28]. PFC lesions may reduce suprapontine inhibition of the pontine micturition center [29]. Although several studies implicate the frontal lobe in SNM, there are inconsistencies regarding the specific regions and directional changes. For example, the study of Gill et al. [7] reported increased activation in the right inferior frontal cortex during SNM, while studies by Pang et al. [30] and Blok et al. [5] identified increased activity in the left DLPFC during SNM, and Weissbart et al. [6] observed decreased activation in the left DLPFC. These discrepancies underscore the need for further validation of the PFC response to SNM. In our study, the left PFC exhibited increased ALFF during acetic acid infusion, which then decreased during SNS, and FC between the PFC and PSC was reduced during SNS. These findings support the hypothesis that reducing PFC activation and its connectivity with other brain regions may be one mechanism by which SNM alleviates OAB.
Currently, no published studies have examined FC during SNM. Our findings suggest that SNM influences FC among several brain regions involved in bladder control. Notably, the PAG and pons are key regions in micturition [31-33]. As the bladder fills, the PAG activates, triggering the micturition reflex that induces detrusor contraction and suppresses urethral sphincter activity, thereby allowing urine elimination. Conversely, inhibiting the PAG suppresses the reflex, preventing detrusor contraction and maintaining sphincter contraction to ensure continence [34]. Our study found increased FC between the PAG and pons in OAB rats and a decrease in FC during SNS, implying that modulation of this connectivity may represent a key mechanism by which SNM reduces urinary frequency and urgency. Additionally, studies on healthy individuals during bladder filling and emptying have reported increased FC between the default mode network (DMN) and the basal ganglia network (BGN), as well as between the DMN and the salience network (SN) [30, 35]. The brain regions exhibiting FC changes in our study, such as the pons, CA-L, and GIC-R, are parts of the DMN, while the CP-L belongs to the BGN and the PAG, VOA-L, and PSC-L are parts of the SN.
In animal fMRI studies, anesthesia is essential to minimize stress. In our study, urethane was used for both cystometry and fMRI scanning because it preserves a relatively normal micturition reflex [36]. Urethane is a well-established anesthetic for fMRI studies focusing on hemodynamic changes [37, 38]; however, it is important to note that anesthesia can affect the relationship between brain activity and hemodynamic responses.
The strength of our study lies in the comprehensive recording of brain functional activity changes in normal rats, OAB rats, and during SNS within the same group of animals. This approach enabled dynamic observation of brain activity fluctuations during the development of OAB and during SNS, providing a more accurate and reliable elucidation of the central mechanisms underlying both OAB and its treatment. Nonetheless, our study has limitations. Animal models may not fully represent human brain activity, necessitating further validation in human studies. Moreover, although the rise time of the brain hemodynamic response typically ranges from 6 to 8 seconds [19], the aftereffects of SNM on brain activity remain unclear, and our 20-second interval block design may not be optimal.
In conclusion, our fMRI study demonstrated several changes in brain region activity during SNS in OAB rats, including alterations in the PSC, PFC, pons, and PAG. Specifically, activity in the left PFC increased during acetic acid infusion and decreased during SNS, while FC between the PAG and pons increased in OAB rats and decreased during SNS. These changes may represent central mechanisms by which SNM alleviates OAB symptoms. As our findings are based on animal studies, further validation in larger human populations is required before translating these results into clinical practice—a prospect that will be enhanced once SNM devices become compatible with high-quality fMRI scanning.
Notes
Grant/Fund Support
This study was supported by National Key Research and Development Program of China (2023YFC3605301), National Natural Science Foundation of China (82300874), Fundamental Research Funds for Central Public Welfare Research Institutes (2023CZ-1), Beijing Natural Science Foundation (7222234) and Beijing Natural Science Foundation (7254461).
Research Ethics
The experiment was approved by the Institutional Animal Care and Use Committee of Capital Medical University (AEEI-2024-334).
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
ACKNOWLEDGMENTS
We express our gratitude to the Non-human Primate Research Center at Tsinghua University for providing the 9.4T MRI scanning services that were instrumental in completing this research.
AUTHOR CONTRIBUTION STATEMENT
· Conceptualization: HS, XL, LL
· Data curation: HS, YZ, LL
· Formal analysis: HS, LL
· Funding acquisition: LL
· Methodology: HS, LL
· Visualization: HS, YZ, QL, XL
· Writing - original draft: HS, YZ, QL
· Writing - review & editing: YZ, QL, XL, LL