Decoding micturition from the spinal cord using functional ultrasound imaging (fUSI)

Introduction:: Micturition characterizes the process in which neural pathways coordinate the activity of smooth muscles in the bladder to expel urine. Disease or injury of the nervous system can cause the re-emergence of involuntary urination, leading to urinary incontinence (UI). Today UI constitutes a major health problem that affects more than 25 million Americans. Elucidating the functional circuities in normal bladder control is a key to identify abnormalities in UI patients. Our understanding of the brain circuits involved in micturition has grown immensely over the past decades. Yet, assessing the functional attributes of spinal cord in relation to bladder function has proven to be a challenge. This may partly stem from the technical challenges that arise when imaging the spinal cord – i.e., susceptibility artifacts due to local magnetic field inhomogeneities created by interfaces between the surrounding bone, ligaments, soft tissues, and cerebrospinal fluid (CSF). To overcome these limitations, we utilized a novel neuroimaging modality, functional ultrasound imaging (fUSI), to characterize the spinal cord hemodynamics in response to micturition in rats. We acquired power Doppler images of the spinal cord while filling the bladder with saline. We showed that fUSI can detect region-specific changes in spinal cord hemodynamics associated with micturition, and can predict with about 90% accuracy an impending voiding. Our study provides the first proof-of-concept that fUSI is a viable modality for developing spinal cord machine interface (SCMI) technologies to restore bladder function in patients with UI.


Materials and Methods:: We used the Iconeus One scanner (Iconeus, Paris, France) to acquire fUSI images from the rat spinal cord. The system is equipped with a 128 elements linear array probe with 15 MHz center frequency and 0.1 mm pitch that generates power Doppler (pD) images with spatial resolution of 100 μm x 100 μm in-plane, temporal resolution 400ms, and penetration depth up to 20 mm. After inducing anesthesia, 12-14 week-old female Sprague-Dawley rats (N=8) underwent multi-level laminectomy at T12-L2 to expose the lumbosacral spinal cord. This window allowed the position of the 15 MHz ultrasound probe in a sagittal plane (Fig. 1A). A transducer was inserted and secured to the bladder dome to record the pressure dynamics. A catheter connected to an infusion pump was inserted in the same location to fill the bladder with saline at a constant rate of 0.1 ml/min. We acquired pD images for 50 minutes under the following protocol: [0-10] min baseline, [10-20] min continuous filling the bladder, [20- 30] min rest (infusion pump was stopped), [30-40] min continuous filling and [40-50] min rest. The animals were spontaneously voiding during the 50 minutes imaging. Voiding is associated with typical abrupt changes in bladder pressure, which were directly verified by observing (through camera) outflow of urine from urethra.


Results, Conclusions, and Discussions:: We measured % power Doppler (pD) changes of the spinal cord with respect to baseline – 10 min before starting to fill up the bladder. The results show that the pD signal in the spinal cord increases after starting to fill up the bladder (Fig. 1B). We further explored how the spinal cord encodes changes on the bladder by generating statistical parametric maps (SPMs). SPM is generated by performing a one-sided t test for each voxel individually with false discovery rate (FDR) correction based on the number of voxels tested. A typical SPM for the regional pD changes between baseline (first 10 minutes of fUSI recordings) and filling the bladder is illustrated in Fig. 1C. The results revealed region-specific hemodynamical changes that encode bladder pressure changes during filling. We further extended these results to test whether fUSI could predict an impending voiding within a signal trial – an essential component for developing spinal cord machine interface (SCMI) technologies for patients with bladder dysfunctions. We used principal component analysis (PCA) as a feature extraction to optimally discard shared information between the two classes (class 0: pre-voiding, class 1: voiding) (Fig. 1D). We then used linear discriminant analysis (LDA) and found that the PCA-transformed images predict an impending voiding with cross-validated accuracy 89.4% ± 3.3% (MEAN ± SE) about 4 seconds before the actual voiding (Fig. 1E).

Overall, we present the first functional neuroimaging study that aims to characterize the spinal cord hemodynamics in response to micturition in rats using the novel fUSI technology. We showed that fUSI is capable of detecting region-specific changes in spinal cord hemodynamics associated with micturition. We further extended the results showing that power Doppler signal encodes information about an impending voiding. These results provide the first proof-of-concept that fUSI is a viable modality for developing ultrasonic spinal cord machine interface (SCMI) technologies to restore bladder function in patients with urinary incontinence (UI) – a technology that currently does not exist.


Acknowledgements (Optional): : University of Southern California Neurorestoration Center, Hellman Foundation

References (Optional): : 1. Macé E, Montaldo G, Cohen I, Baulac M, Fink M, Tanter M. Functional ultrasound imaging of the brain. Nat Methods. 2011 Jul 3;8(8):662-4. doi: 10.1038/nmeth.1641. PMID: 21725300.
2. Agyeman KA, Lee DJ, Russin J, Kreydin EI, Choi W, Abedi A, Edgerton VR, Liu C, Christopoulos VN, Functional Ultrasound Imaging of the Human Spinal Cord, bioRxiv 2022.08.06.503044