Detecting bladder volume and pressure from sacral nerve signals: the key to future artificial control

Lead Research Organisation: University of Bath
Department Name: Electronic and Electrical Engineering

Abstract

This project is about the treatment of urinary incontinence, consequent of damage to the spinal cord.

Managing the urinary bladder is of the first importance to clinicians and patients following trauma to the spinal cord. Historically, kidney damage due to high bladder pressures and/or infection was the usual cause of death resulting from such an injury. Infections still raise mortality and morbidity, exacerbated by the risk of antibiotic resistance. In order to achieve urination-i.e. complete voluntary micturition (CVM) and as an alternative to the expensive process of intermittent sterile catheterisation, a neuroprosthesis for controlling the bladder after spinal cord injury (SCI) was developed by GS Brindley at the MRC Neurological Prostheses Unit in London 30 years ago. The Brindley method employs sacral anterior root stimulation (SARS) but is not popular in Europe, in terms of the fraction of the SCI population treated, because implantation of the device is accompanied by cutting the sacral posterior (sensory) nerve roots (rhizotomy) to prevent reflex incontinence during bladder filling and improve stimulated voiding. Clearly there is a need for a new neuroprosthesis that is more widely acceptable (primarily because no rhizotomy is necessary) and which, in addition, reduces the lifetime cost of care. The aim of this project is to design and demonstrate such a device.

Since its introduction, the Brindley method has been improved in several ways in attempts to address the problems mentioned. However, in spite of these developments, at present (a) no satisfactory, practical method exists for detecting the onset of bladder contractions in a chronic implant and, (b) no method is available to inform the patient of the level of bladder fullness to indicate when the bladder should be emptied. These are critical obstacles to the design of a complete prosthesis and our proposed solution is to use the bladder neural signals themselves since surgically implanted electrodes are essential anyway (i.e. for stimulation). A suitable site for the electrodes is the extradural roots; this is surgically attractive and electrodes are routinely implanted here in the Brindley procedure. In order that the nerves are similar to those in man, it is essential to use a large experimental animal as a preclinical model and we propose to use sheep for these experiments.

Planned Impact

This project is about the treatment of urinary incontinence, consequent of damage to the spinal cord. Trauma to the spinal cord causes permanent paralysis and incontinence. In the UK, around 50,000 people live with SCI, with about 1,000 new cases every year amounting to a total annual cost of approximately £1 billion or approximately 1% of the total NHS budget [see www.spinal-research.org]. Worldwide, SCI affects about 2.5 million people with approximately 130,000 new cases each year. For patients, having a satisfactory method of bladder management is the highest priority for paraplegics and the second highest, after restoration of the hand, for tetraplegics.
At present, in Britain, the most popular method for bladder management following SCI is intermittent catheterisation supported by the use of anticholinergic drugs. This is expensive (£6k/annum per patient) and the drug side effects are unpleasant. As an alternative, the Brindley neuroprosthesis is manufactured by Finetech Medical Ltd of Welwyn Garden City and approximately 4,000 have now been implanted worldwide. However, due to the need for destructive surgery to ensure its successful function (cutting certain sensory neural pathways 'rhizotomy'), the method is not generally popular. The new prosthesis described in this proposal avoids the need for a rhizotomy and so removes the main obstacle to the device being widely adopted by patients and clinicians. In addition, the process of technology transfer will be greatly assisted by the innovative First-in-Man initiative at UCL, which is designed to help get new devices like this through the current regulatory process and into commercial production.

A successful outcome to the proposed project would allow the new neuroprosthesis to be designed, tested in SCI patients and become the treatment of choice for complete spinal lesions (~40% of SCI patients have no residual sensation from the sacral region). This will have a significant impact on society, as people living with SCI will benefit tremendously from the developed technology and experience a better quality of life. Economically, too, the impact of the research is likely to be high, as the amount of NHS budget going towards people living with SCI will be dramatically reduced.

Publications

10 25 50
publication icon
Sadrafshari S (2024) CMOS Analogue Velocity-Selective Neural Processing System in Electronics

publication icon
Sadrafshari S (2022) The Design of a Low Noise, Multi-Channel Recording System for Use in Implanted Peripheral Nerve Interfaces. in Sensors (Basel, Switzerland)

publication icon
Metcalfe BW (2021) Array processing of neural signals recorded from the peripheral nervous system for the classification of action potentials. in Journal of neuroscience methods

publication icon
Metcalfe B (2021) Urinary Bladder Innervation within the Sacral Roots of a Sheep

publication icon
Metcalfe B (2020) Selective Recording of Urinary Bladder Fullness from the Extradural Sacral Roots. in Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference

publication icon
Metcalfe B (2020) A dataset of action potentials recorded from the L5 dorsal rootlet of rat using a multiple electrode array. in Data in brief

publication icon
Taylor J (2020) The Effects of the Presence of Multiple Conduction Velocities in the Analysis of Electrically-Evoked Compound Action Potentials (eCAPs). in Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference

 
Description The essence of this project is to investigate if bladder pressure and volume can be estimated using electrodes implanted at an extradural site - sacral nerves - acutely and chronically in a large animal experimental model. The conduction velocities of the sensory (myelinated) afferent fibres from mechanoreceptors of the genital region in sheep are similar to those of fibres conveying bladder pressure and volume information in man (approximately 38 and 41 m/s respectively). In order to discriminate between these signals, we propose to use a method that we have been developing called velocity selective recording (VSR) that may make this possible without using nerve interfaces that are difficult to implant or that endanger these important nerves. The principle of VSR is based on the fact that action potentials (APs) propagate at specific velocities and that these velocities are closely related to the nerve fibre diameter (for myelinated nerves). If, therefore, a recording of neural traffic can be characterised in terms of its velocity spectrum, it will provide a signature relating the recorded signals to their function (VSR can also discriminate between afferent and efferent traffic on a mixed nerve, such as extradural bladder nerves). The method requires multi-electrode cuffs (MECs) and we have used such devices with as many as 11 electrodes along the length of a nerve cuff.

Since the signals appearing at the electrodes of an MEC in response to naturally-occurring (physiological) ENG are very small (order of 1 µV) a very low noise environment is required. This requires not only specially-designed recording instrumentation but also an electrically quiet environment with well-characterised and understood background interference sources. This latter requirement is particularly hard to achieve in a modern operating room (OR), which tends to contain many electrically-based devices. Most of our effort to date has therefore been directed towards preparing and characterising the experimental set-up in a designated OR at the Royal Veterinary College (RVC) so that successful neural recordings can be made that make maximum use of our various resources and that do not require expensive additional interference reduction equipment.

Initially, we have designed a new nerve interface, multi-electrode cuff (from silicone with steel electrodes), connected to a Cooper cable for signal extraction (photograph 1). The cuff design is based on the anatomy of sheep sacral nerves. Using this implant, we have conducted impedance measurements following stimulation of explanted pig vagal nerves and then live recordings in sheep. This showed impedance values of ~5kO for frequencies ranging 100-1,000Hz, which fulfils theoretical requirements for afferent bladder sacral nerve recording.

Then, four acute implantations have been carried out on sheep (15 June 2018, 30 October 2018, 13 December 2018 and 1 March 2019 - photograph 2). The first three experiments were significantly compromised by the presence of electrical interference and only compound action potentials (CAPs) trafficking through sacral nerves, evoked by sacral nerve electrical stimulation were recordable. In response to this, two extra days were spent in the OR (3 January and 1 February 2019) with the instrumentation alone, in an attempt to understand these interfering sources. These efforts were successful and in the most recent critical experiment (1 March 2019) we were able to record physiological ENG in response to touch stimulation of the skin in the region of the animal's tail, innervated by sacral nerves. These recordings were clear above the noise and repeatable. Recording CAPs from both sacral nerve electrical stimulation and skin stimulation validates our new cuff electrode design and brings confidence that the interface nerve / cuff electrode is exploitable. Sacral nerves were sampled in two animals for morphometry to serve as controls against future chronic experiments.

During in vivo experiments in sheep, we have acquired new data on sheep's physiological bladder function (during anaesthesia) and found: (i) peak bladder pressures ranging from 23 to 31cmH2O during S2 and S3 electrical stimulation - image 1; (ii) maximal bladder pressure during filling cystometry ranging from 6 to 59cmH2O for infused volumes of 187 to 727mL - image 2; and (iii) a drop in bladder compliance after 4 hours of anaesthesia.

So far, we have used two methods for detecting bladder afferent signals: (i) monitoring the running RMS amplitude of the wide-band amplifier output, and (ii) listening to this raw recording during the experiment. Despite filling the bladder to very large volume (1 litre) and the consequent high pressure (up to 59cmH2O), there was no apparent neural signal. This might be due to the long period of anaesthesia and the observed reduction in bladder compliance, but it shows that signal processing is essential. The challenge now is, therefore, to see whether the bladder afferent signals can be extracted using VSR.

In summary, our findings to date have been:

1. A custom 5-channel integrated amplifier system (CMOS asic) has been constructed (configurable for 10 channels as needed) and its performance verified;
2. A new multi-electrode cuff (MEC) has been designed and built (4 electrodes initially but with a capability to increase to 10 or more);
3. A special tool has been designed to aid the implantation of the MECs on extradural roots of sheep;
4. The OR has been characterised for electrical interference and counter measures taken so that the recording instrumentation is not compromised by the environment;
5. The cuff impedances (2-and 4-wire configurations) have been extensively characterised in vitro and in vivo;
6. Repeatable 4-electrode (configured as 3 dipole channels) in vivo measurements have been made of naturally-occurring (physiological) ENG evoked by (a) stroking and (b)
pricking with needles the tissue close to the animal's tail;
7. New data on sheep's physiological bladder function (during anaesthesia) have been acquired for future reference; in parallel nerve morphometry has been done on two controls.


Revised version 5 March 2020:

1. An improved 10-channel integrated amplifier system (CMOS asic) has been constructed and its performance verified. The total input referred voltage noise of this amplifier,
including the RF filters connected at the inputs is approximately 800 nV, corresponding to a noise density of 3.5 nV/vHz at 1 kHz;
2. A new 10 electrode multi-electrode cuff (MEC) has been designed and built;
3. A special tool has been designed to aid the implantation of the MECs on extradural roots of sheep;
4. The OR has been characterised for electrical interference and counter measures taken so that the recording instrumentation is not compromised by the environment, at
least for acute experiments (animal anaesthetised);
5. The cuff impedances (2-and 4-wire configurations) have been extensively characterised in vitro and in vivo;
6. Repeatable 10-electrode (configured as 9 dipole channels) in vivo measurements have been made of naturally-occurring (physiological) ENG evoked by (a) stroking and (b)
pricking with needles the tissue close to the animal's tail;
7. A preliminary demonstration of the correlation of ENG activity to bladder pressure has been made;
8. An initial demonstration of VSR has been made but this is currently very speculative and requires a lot more thought;
9. The new head connector has been implanted and the sheep successfully recovered. Significant problems from interference make awake experiments difficult at present and
some improved method of screening needs to be devised;
10. Three papers have been submitted to an international conference and we await the reviewers' comments.
Exploitation Route We intend to publish several papers in the next period to assist future researchers and clinicians with this difficult technique
Sectors Electronics

Healthcare

Pharmaceuticals and Medical Biotechnology
 
Description We are doing this research to find out whether it is possible to detect the volume and pressure in the urinary bladder from the natural nerve activity in the corresponding nerve roots. Finetech Medical Ltd is the UK company that makes bladder-control neuroprostheses for spinal cord injured patients. So far, our findings suggest that this idea is unlikely to be feasible because although neural signals are detectable, them are very small indeed, close to the noise floor, and prone to interference. We keep the company up to date with our results and we continue to explore alternative approaches, especially more biological approaches encouraging neuroplasticity.
First Year Of Impact 2020
Sector Healthcare
Impact Types Societal

Economic
 
Title Detecting bladder volume and pressure from sacral nerve signals 
Description The essence of this project is to investigate if bladder pressure and volume can be estimated using electrodes implanted at an extradural site - sacral nerves - acutely and chronically in a large animal experimental model. The conduction velocities of the sensory (myelinated) afferent fibres from mechanoreceptors of the genital region in sheep are similar to those of fibres conveying bladder pressure and volume information in man (approximately 38 and 41 m/s respectively). In order to discriminate between these signals, we propose to use a method that we have been developing called velocity selective recording (VSR) that may make this possible without using nerve interfaces that are difficult to implant or that endanger these important nerves. The principle of VSR is based on the fact that action potentials (APs) propagate at specific velocities and that these velocities are closely related to the nerve fibre diameter (for myelinated nerves). If, therefore, a recording of neural traffic can be characterised in terms of its velocity spectrum, it will provide a signature relating the recorded signals to their function (VSR can also discriminate between afferent and efferent traffic on a mixed nerve, such as extradural bladder nerves). The method requires multi-electrode cuffs (MECs) and we have used such devices with as many as 11 electrodes along the length of a nerve cuff. Since the signals appearing at the electrodes of an MEC in response to naturally-occurring (physiological) ENG are very small (order of 1 µV) a very low noise environment is required. This requires not only specially-designed recording instrumentation but also an electrically quiet environment with well-characterised and understood background interference sources. This latter requirement is particularly hard to achieve in a modern operating room (OR), which tends to contain many electrically-based devices. Much effort to date has therefore been directed towards preparing and characterising the experimental set-up in a designated OR at the Royal Veterinary College (RVC) so that successful neural recordings can be made that make maximum use of our various resources and that do not require expensive additional interference reduction equipment. Initially, a new nerve interface was designed, consisting of a multi-electrode cuff (MEC-using silicone with steel electrodes), connected to a Cooper cable for signal extraction. This cuff design (initially containing 4 electrodes) was based on the anatomy of sheep sacral nerves and was used in conjunction with a specially-designed neural amplifier. Using this combination, impedance measurements were made following stimulation of explanted pig vagal nerves and then live recordings in sheep. This showed impedance values of ~5kO for frequencies ranging 100-1,000Hz, which fulfils theoretical requirements for afferent bladder sacral nerve recording. Four acute implantations were carried out on sheep in the previous reporting period (15 June 2018, 30 October 2018, 13 December 2018 and 1 March 2019). The cuff design has been subsequently improved and enlarged to 10 electrodes, including a specialised tool to aid surgical implantation. In addition, an improved 10-channel neural amplifier with significantly lower noise than the original 5 channel model has been designed and constructed. This amplifier compares favourably with the best low noise instrumentation amplifiers currently available and because of its relatively low power consumption (about 10 mW) is potentially implantable. Ten further experiments including and four recovery experiments have been carried out in the current period, the last being on 27 February 2020. The recovery experiments involved the fitting of a head connector to allow the cables from the implanted electrodes to be terminated for connection to an external amplifier array. The first recovery experiment took place on 25 September 2019 and the animal recovered well with no major adverse effects. The early experiments were significantly compromised by the presence of high-amplitude electrical interference and, in particular, by RF interference in the MW band (approximately 1 MHz) from the nearby Brookman's Park broadcast transmitter. In response to this, intensive efforts were made to improve the earthing arrangements in the OR (see previous report) and the inclusion of RF filters between the amplifiers and the electrodes. These efforts were successful and repeatable acute experiments (animal anaesthetised) were possible with little or no interference perceptible above the noise floor of the amplifiers. These experiments also revealed a correlation between bladder pressure and rms ENG output, which was one of the objectives of the project. However, the results are not yet reliable or repeatable and there is uncertainty about which is the optimum root to use for this activity. Another major objective of the project was to demonstrate the application of velocity selective recording (VSR) in this application. VSR has been successful in the past in classifying electrically-evoked compound action potentials (eCAPs), where individual spikes are observable. In the case of this project, with naturally evoked ENG, spikes are not visible, even when the signal-to-noise ratio is increased using VSR. As a result an rms value was calculated at the output of the delay-and-add process and subsequently converted into velocity spectra, although the exact effect of these calculations on the velocity spectrum is currently unclear. Additional problems were caused by the relatively small spacing of the electrodes (1 mm), compounded by a restricted sampling rate (100 kS/s/channel). Whilst the electrode spacing cannot be changed within the scope of the current project, the recent purchase of a new recording pc will allow operation at the full sampling rate of the ADCs (500 kS/s/channel). In the recovery experiments with the head connector in place (involving a much longer wire connection to the amplifiers and working in a pen distant from the OR), interference is once again a critical problem that remains to be solved at the time of writing. Finally, it should be mentioned that three papers have been submitted to the 2020 annual International Conference of the IEEE Engineering in Medicine and Biology Society (IEMBC) to be held in Montreal, Canada in July. The titles are: 1. The Effects of the Presence of Multiple Conduction Velocities in the Analysis of Electrically-Evoked Compound Action Potentials (eCAPs); 2. The Design of a Low Noise, Multi-Channel Interface System for use in Noisy Recording Environments; 3. Selective Recording of Urinary Bladder Fullness from the Extradural Sacral Roots. 
Type Of Material Physiological assessment or outcome measure 
Year Produced 2020 
Provided To Others? Yes  
Impact None yet 
 
Title Nerve cuffs for sacral nerve roots 
Description Multiple electrode cuff designed for use in particular surgical space yet convenient to implant. The cuff has been shown to fit into the space provided by the laminectomy. A prototype implantaion tool has be made and tested. The cuff can be closed round the nerve root and sealed. Impedances have been measured after implantation. The recording environment has been characterised and recording set-up improved so that now microvolt level differential recording possible on multiple channels without significant interference. Cutaneous afferent nerve signals have been recorded when rubbing the sheep's rump (S3 dermatome). March 2020. We have improved the recording set-up in the operating theatre so that usually there is no apparent interference and the RMS input voltage (random) is about 1uV. We started with a 4-electrode cuff and progressed to 10 electrodes on two 5-wire implant-grade cables. The method of fabricating the cuffs has been improved. We changed the method of closing the cuff during implantation and that has meant that we now get higher impedances within the cuff which will have increase the signal amplitudes and reduced the interference. Special tools are needed to implant the cuffs and three designs have been made and tested during the operations. Having done 9 non-recovery sheep, in which the animal is euthanized and the end of the experiment, we did 2 in which the animal was allowed to recover and further measurements were made a week later under anaesthesia before death. We designed and made a special 'head connector' around a military grade hermetic socket with a hydroxy-apatite-coated titanium flange which should bod to the skin. These sockets have 13 pins, 10 for the cuff electrodes and one for the internal reference electrode. Two sheep have now been implanted with cuffs joined o these head connectors. Special trocars were made to enable the cables to be tunnelled from the sacrum and head to implantable Craggs Connectors in the middle of the sheep's back. The first of these connectors looks good after a month. We have not yet managed to detect bladder signals in the awake animals: a major difficulty is the proximity of the powerful Medium Wave transmitter at Brookman's Park (only 3km away) so we hear too much of Radio 5 Live. Many of these methods will be relevant to the development of Bioelectronic Medicines, even if not for artificial control of the bladder. 
Type Of Material Physiological assessment or outcome measure 
Year Produced 2019 
Provided To Others? Yes  
Impact none yet