ON PHYSIOLOGICAL MARKERS

[0001] This application claims the benefit of U.S. Patent Application No. 16/744,784 filed on January 16, 2020, U.S. Provisional Patent Application No. 62/795,624 filed on Januar 23, 2019, and to U.S. Provisional Patent Application No. 62/926,012 filed on October 25, 2019, the entire content of all of which is incorporated herein by reference.

TECHNICAL FIELD

[0002] The disclosure relates to medical devices and, more particularly, medical devices that deliver therapy to a patient.

BACKGROUND

[0003] Disease, age, and injury may impair physiological functions of a patient. In some situations, the physiological functions are completely impaired. In other examples, the physiological function may operate sufficiently at some times or under some conditions and operate inadequately at other times or at other conditions. In one example, bladder dysfunction, such as overactive bladder, urgency, or urinary incontinence, is a problem that may afflict people of all ages, genders, and races. Various muscles, nerves, organs and conduits within the pelvic floor cooperate to collect, store and release urine. A variety of disorders may compromise urinary tract performance, and contribute to an overactive bladder, urgency, or urinary incontinence that interferes with normal physiological function. Many of the disorders may be associated with aging, injury or illness.

[0004] Urinary incontinence may include urge incontinence and stress incontinence. In some examples, urge incontinence may be caused by disorders of peripheral or central nervous systems that control bladder micturition reflexes. Some patients may also suffer from nerve disorders that prevent proper triggering and operation of the bladder, sphincter muscles or nerve disorders that lead to overactive bladder activities or urge incontinence. In some cases, urinary

incontinence may be attributed to improper sphincter function, either in the internal urinary sphincter or external urinary sphincter. [0005] Electrical nerve stimulation may be used for several therapeutic and diagnostic purposes, including the treatment of urinary incontinence. Electrical nerve stimulation may be delivered by devices with a limited power source (e.g., implantable devices that use a batter}'). Power consumption may be a limiting factor in the effectiveness and viability of such devices. Additionally, a body may adapt to continuous stimulation. Thus, it may desirable to deliver stimulation non-continuously.

SUMMARY

[0006] In general, the disclosure is directed to devices, systems, and techniques for managing therapy delivery based on physiological markers. A system may control delivery of

neurostimulation therapy to a patient based on one or more detectable physiological markers and time delivery relative to these physiological markers. Accordingly, the system may time deliver}' of a targeted neurostimulation therapy to a particular period after detecting the one or more physiological markers. In some examples, the system may withhold neurostimulation therapy delivery to a patient during certain times or phases of a physiological cycle until the system determines that neurostimulation should be delivered. Alternatively, the system may deliver different neurostimulation during the period of time between the detected physiological marker and when the targeted neurostimulation is to be delivered later in the physiological cycle, e.g., the system may modify or change one or more parameters defining the neurostimulation at a time after the physiological marker is detected. In this manner, the system may control delivery of neurostimulation in response to detecting a physiological marker or during a phase later in time from the physiological marker. In some examples, the system may dynamically adapt the timing of the delivery of the neurostimulation and/or one or more of the parameters defining the neurostimulation based upon data collected over time.

[0007] For example, the system may monitor a bladder fill cycle for a patient and control the time of deliver of neurostimulation to occur during a particular phase of the fill cycle. After a voiding event (e.g., a type of physiological marker) is detected, the system may withhold neurostimulation during a first phase of the fill cycle and begin delivery of neurostimulation for a later second phase of the fill cycle where neurostimulation may be more effective at reducing or eliminating a dysfunctional state of the bladder such as urinary incontinence. In this manner, the system may predict the phase of the fill cycle during which neurostimulation should be delivered and deliver neurostimulation during this phase.

[0008] In one example, the disclosure is directed to an adaptive system for providing electrical stimulation to a patient, the system comprising memory configured to store one or more programs, and processor circuitry coupled to the memory and configured to execute the one or more programs, the processor circuitry being configured to: monitor a sensor signal and classify a physiological marker of the patient based upon the sensor signal, the physiological marker indicative of a phase of a physiological cycle; generate a control signal based on the classified physiological marker of the patient, wherein the control signal controls an implantable stimulation device to provide the electrical stimulation, in accordance with one or more stimulation parameters of a stimulation program, at a target site within the patient to one more of: enable a sphincter to open to allow a voiding event or constrict the sphincter to inhibit the voiding event; and adapt one or more of a manner in which the processor circuitiy classifies the physiological markers, generates the control signal, or the one or more stimulation parameters of the stimulation program based on the classified physiological markers or patient input so as to automatically adjust one or more of timing of the delivery of the electrical stimulation or the stimulation parameters of the electrical stimulation.

[00Q9] In another example, the disclosure is directed to a method comprising monitoring a sensor signal and classifying a physiological marker of the patient based upon the sensor signal, the physiological marker indicative of a phase of a physiological cycle, generating a control signal based on the classified physiological marker of the patient, wherein the control signal controls an implantable stimulation device to provide the electrical stimulation, in accordance with one or more stimulation parameters of a stimulation program, at a target site within the patient to one more of: enable a sphincter to open to allow a voiding event or constrict the sphincter to inhibit the voiding event, and adapting one or more of a manner m which processor circuitiy classifies the physiological markers, generates the control signal, or the one or more stimulation parameters of the stimulation program based on the classified physiological markers or patient input so as to automatically adjust one or more of timing of the delivery of the electrical stimulation or the stimulation parameters of the electrical stimulation.

[0010] In a further aspect, the disclosure is directed to a non-transitory storage medium containing instructions which when executed by one or more processors, cause the one or more processors to monitor a sensor signal and classify a physiological marker of the patient based upon the sensor signal, the physiological marker indicative of a phase of a physiological cycle, generate a control signal based on the classified physiological marker of the patient, wherein the control signal controls an implantable stimulation device to provide the electrical stimulation, m accordance with one or more stimulation parameters of a stimulation program, at a target site within the patient to one more of: enable a sphincter to open to allow a voiding event or constrict the sphincter to inhibit the voiding event and adapt one or more of a manner in which the processor circuitry classifies the physiological markers, generates the control signal, or the one or more stimulation parameters of the stimulation program based on the classified physiological markers or patient input so as to automatically adjust one or more of timing of the delivery of the electrical stimulation or the stimulation parameters of the electrical stimulation.

[0011] The details of one or more examples are set forth in the accompanying drawings and the description below. Other features, objects, and advantages of the disclosure will be apparent from the description and drawings, and from the claims.

[0012] The above summary is not intended to describe each illustrated example or every implementation of the present disclosure.

BRIEF DESCRIPTION OF DRAWINGS

[0013] FIG. 1 is a conceptual diagram illustrating an example system that manages deliver} of neurostimulation to a patient to manage bladder dysfunction, such as overactive bladder, urgency, or urinary incontinence.

[0014] FIGS. 2A , 2B AND 2C are block diagrams illustrating example configurations of implantable medical devices (IMDs) which may be utilized in the system of FIG 1.

[0015] FIG. 3 is a block diagram illustrating an example configuration of an external programmer which may be utilized in the system of FIG 1.

[0016] FIG. 4 is a flow diagram that illustrates an example technique for determining a timing for therapy delivery based on a physiological marker.

[0017] FIG. 5 is a flow diagram that illustrates an example technique for determining phases to withhold and delivery neurostimulation to manage bladder dysfunction.

[0018] FIG. 6 is a flow diagram that illustrates an example technique for training a device to automatically associate a bladder void with physiological markers m sensor signals. [0019] FIG. 7 is a flow diagram that illustrates an example technique for automatically adapting parameters or timing of therapy.

[0020] FIG. 8 is an example timing diagram of a bladder fill cycle and delivery of neurostimulation timed to a phase of the bladder fill cycle.

[0021] FIGS. 9A and 9B are graphs showing example bladder volumes for neurostimulation delivered during different phases of a bladder fill cycle.

[0022] FIGS. 10A and 10B are graphs showing example bladder volumes prior to neurostimulation delivery shown in FIGS. 9 A and 9B.

[0023] FIGS. 11 A and 1 IB are graphs showing example bladder volumes for

neurostimulation delivered during different phases of a bladder fill cycle.

[0024] FIGS. 12A and 12B are graphs showing example bladder volumes prior to neurostimulation delivery shown in FIGS. 9 A and 9B.

[0025] FIG. 13 is a conceptual diagram of an experimental general-purpose neuromodulation system.

[0026] FIG. 14 is a graph showing example bladder volumes changes based on

neurostimulation delivered during different phases of a bladder fill cycle.

[0027] FIGS. 15A and 15B are graphs showing example physiological events during a physiological cycle and predictive delivery of neurostimulation to avoid a dysfunctional state of the physiological cycle.

DETAILED DESCRIPTION

[0028] The disclosure is directed to devices, systems, and techniques for managing the delivery of electrical stimulation to a patient such that selective delivery of neurostimulation based on one or more physiological markers may reduce or eliminate a dysfunctional state. The techniques may be used to provide therapy for a variety of dysfunctions, diseases or disorders. For purposes of illustration, but without limitation, use of the techniques will be described below with respect to bladder dysfunction. Bladder dysfunction generally refers to a condition of improper functioning of the bladder or urinary tract, and may include, for example, an overactive bladder, urgency, or urinary incontinence. Overactive bladder (QAB) is a patient condition that may include symptoms, such as urgency, with or without urinary incontinence. Urgency is a sudden, compelling urge to urinate, and may often, though not always, be associated with urinary incontinence. Urinary incontinence refers to a condition of involuntary loss of urine, and may include urge incontinence, stress incontinence, or both stress and urge incontinence, which may be referred to as mixed urinary incontinence. As used in this disclosure, the term“urinary incontinence” includes disorders in which urination occurs when not desired, such as stress or urge incontinence. Other bladder dysfunctions may include disorders such as non-obstructive urinary retention.

[0029] One type of therapy for treating bladder dysfunction includes delivery of continuous electrical stimulation to a target tissue site within a patient to cause a therapeutic effect during delivery^ of the electrical stimulation. For example, delivery of electrical stimulation from an implantable medical device (IMD) to a target therapy site, e.g., a tissue site that delivers stimulation to modulate activity of a spinal nerve (e.g., a sacral nerve), a pudendal nerve, dorsal genital nerve, a tibial nerve, an inferior rectal nerve, a perineal nerve, or branches of any of the aforementioned nerves, may provide an immediate therapeutic effect for bladder dysfunction, such as a desired reduction m frequency of bladder contractions. In some cases, electrical stimulation of the sacral nerve may modulate afferent nerve activities to restore urinary' function during the electrical stimulation. However, continuous electrical stimulation or other types of neurostimulation (e.g., drug delivery' therapy) may provide neurostimulation during unnecessary' phases of a physiological cycle that may cause undesirable side effects, accommodation, less focused therapy, and increased energy usage by the medical device delivering therapy.

[0030] In contrast to this type of continuous neurostimulation therapy, the example devices, systems, and techniques described in this disclosure are directed to managing delivery of neurostimulation therapy based on timing with one or more physiological marker such that neurostimulation is delivered and withheld based on the one or more physiological markers. For example, the physiological markers may be indicative of one or more points within a

physiological cycle. The system may machine learn to associate the physiological markers with the one or more points within the physiological cycle. The system may monitor the patient for occurrence of the physiological marker or markers as a trigger to automatically stop delivery of neurostimulation therapy, start neurostimulation therapy, or time the delivery of neurostimulation to begin at a later point within the physiological cycle. For example, the system may learn that a patient typically consumes coffee in the morning, leading to a shorter fill cycle and adjust the timing so as to start neurostimulation earlier than it otherwise would in the morning . [0031 ! Timing the delivery may involve predicting the appropriate phase within the physiological cycle for delivery of neurostimulation based on one or more prior physiological cycles. For example, the system may deliver neurostimulation during a phase that begins prior to, and may also terminate prior to, a dysfunctional phase during which a dysfunctional state typically occurs during the physiological cycle. In this manner, the system may effectively reduce or eliminate the dysfunctional phase by preemptively delivering neurostimulation and withholding neurostimulation during one or more phases of the physiological cycle at which neurostimulation is unnecessary or even detrimental to treating the dysfunctional state. By delivering and withholding neurostimulation based on one or more physiological markers indicative of the physiological cycle of the patient, the system may reduce undesirable side effects from neurostimulation delivered during phases of the cycle that do not benefit from neurostimulation and/or extended periods of neurostimulation delivery, increase efficacy of the neurostimulation therapy, increase durability' of the therapy, decrease tissue accommodation to the therapy, decrease energy usage (e.g , during electrical stimulation therapy), and/or decrease material usage (e.g., drug delivery' therapy).

[0032] In one example, a system may monitor, through a sensor such as a pressure sensor, a bladder fill cycle (e.g., a type of physiological cycle) for a patient and time delivery of neurostimulation to occur during a particular phase (e.g., one of a plurality' of phases) of the fill cycle. After a voiding event (e.g., a type of physiological marker) is detected, the system may withhold neurostimulation during a first phase of the fill cycle and begin delivery' of

neurostimulation for a later second phase of the fill cycle where neurostimulation may be more effective at reducing or eliminating a dysfunctional state of the bladder such as urinary incontinence. In this manner, the system may predict the phase of the fill cycle during which neurostimulation should be delivered and deliver neurostimulation during this phase.

[0033] The disclosure includes discussion of various examples, aspects, and features. Unless otherwise stated, the various examples, aspects, and features are contemplated as being used together in different combinations. For ease of discussion and as a practical matter, each possible combination of features is not expressly recited. For example, the disclosure refers to aspects relating to stimulation devices used in connection with sensors (e.g., pressure sensors). It is understood that the system may have different types of sensors (e.g., temperature sensors, or electrical sensors) and may have different combinations of these sensors.

[0034] Various examples of the present disclosure are directed toward a system that provides a closed-loop neuromodulation solution. For instance, the system may be configured to target the sacral nerve using electrical stimulation (sacral neuromodulation (SNM)). The stimulation of the sacral nerve may provide therapy for various pelvic dysfunctions, particularly disorders of pelvic floor function. Examples of pelvic dysfunctions include, but are not necessarily limited to, overactive bladder, non-obstructive urinary retention, fecal incontinence, constipation, pelvic pain, and sexual dysfunction.

[0035] As described above, m some techniques where constant therapy is delivered to address incontinence, there may be negative impact on power consumption of the medical device. For example, the amount of time that SNM is being actively delivered to a patient has a correlation to the power consumption of the delivering device. Various examples of the present disclosure are directed toward systems that are configured to provide SNM in a manner that is responsive to the needs of the patient at any given moment such that continuous power does not need to be delivered. The systems may be configured to provide adaptive stimulation in response to inputs from sensors that monitor appropriate biomarkers or physiological markers. For example, the system may monitor physiological markers of a patient associated with a physiological phase and automatically adjust the timing of the electrical stimulation or parameters of the electrical stimulation based on the physiological makers. In this manner, the systems may be considered a closed-loop SNM system. Without being limited by theory, such closed-loop systems may be useful relative to power usage, patient adaptation, and be more robust to moment- to-moment variations in patient conditions.

[0036] In some examples, a system may be configured to provide stimulation at a neural target located at a distant site from the end organ that is being affected. For example, sacral stimulation sites may be located a relatively large distance from the bladder or bowel. The system may include of multiple devices (e.g., implantable sensors and implantable stimulation devices) with wireless communication circuitry that allows for wireless communication of information between the devices, which may provide of sensing or therapeutic stimulation. For example, the wireless circuitry may be designed to communicate using Near Field

Communication, Bluetooth®, or other wireless protocols. [0037] For ease of discussion, various examples of are discussed in connection with bladder function. It is to be recognized that bladder function is but one possible application. Various aspects of the present disclosure may also be used in connection with urinary , bowel, and general pelvic floor dy sfunction. For the sake of brevity , each type of dysfunction is not repeated for each feature or example discussed herein.

[0038] As discussed above, constant stimulation may result in undesirable side effects, accommodation, less focused therapy, and increased energy usage by the medical device delivering therapy.

In certain instances, a reduction in the stimulating on-time results in less power consumption, which may translate into longer recharge intervals, longer replacement intervals, smaller devices, or combinations thereof.

[0039] While certain examples are discussed in connection with a distributed platform (see below) using wireless communications, other examples allow for a single (all-in-one) device that may sense a physiological marker, such as bladder pressure, classify voids, and adjust stimulation accordingly.

[0040] As will be discussed further below, neurostimulation delivered at certain times during the bladder full cycle are responsible for increased bladder capacity and urine retention.

Therefore, a system may withhold neurostimulation therapy, such as therapy configured to reduce bladder contractions, during a portion of the bladder fill cycle and target deliver _} ' during phases of bladder filling more receptive to neurostimulation therapy, such as the second half, or third or fourth quartile, of the bladder fill cycle. The system may use a detected void event to predict when these phases occur during the fill cycle and time neurostimulation delivery accordingly or directly detect the phases of the fill cycle using one or more sensors. Although neurostimulation therapy is generally discussed as including electrical stimulation therapy, neurostimulation therapy may alternatively, or also include, drug delivery therapy.

[0041] A medical device, such as an implantable medical device (IMD) may implement the techniques described in this disclosure to deliver stimulation therapy to at least one nerve (e.g., spinal nerve or a pelvic floor nerve) to modulate activity of the nerve via at least one electrode electrically connected to the IMD. The electrical stimulation may be configured to modulate contraction of a detrusor muscle of the patient to cause a decrease m frequency of bladder contractions (to reduce incontinence) or an increase m the frequency of bladder contractions (to

Q promote voiding). Reduction in frequency of bladder contractions may reduce urgency of voiding and may reduce urgency and/or urinary incontinence, and thereby at least partially alleviate bladder dysfunction.

[0042] The neurostimulation described herein may be targeted to manage bladder dysfunction, such as an overactive bladder, urgency, urinary incontinence, or even non obstructive urinary retention. For example, the stimulation may be delivered to target tissue sites normally used to alleviate these types of dysfunction. Although the techniques are primarily- described in this disclosure for managing bladder dysfunction, the techniques may also be applied to manage other pelvic floor disorders or disorders relating to other organs, tissues or nerves of the patient. For example, the devices, systems, and techniques described in this disclosure alternatively or additionally may be utilized to manage sexual dysfunction, pelvic pain, fecal urgency or fecal incontinence. Example nerves that may be targeted for therapy include sacral nerves, pudendal nerves, a dorsal nerve of the penis or clitoris, tibia! nerves, sural nerves, sciatic nerves, the inferior rectal nerve, and peroneal or perineal nerves. Example organ systems that may be treated for dysfunction may include the large and small bowel, stomach and/or intestines, liver, and spleen, which may be modulated by delivering neurostimulation directly to the organs, to one or nerves innervating the organ, and/or blood supplies reaching the organs.

[0043] In the example of fecal incontinence, the IMF) may deliver the neurostimulation therapy timed to the detection of a physiological marker indicative of an increased probability of an occurrence of a fecal incontinence (e.g , an increased patient activity level) or bowel fill level or activity level. The physiological marker may include, for example, a magnitude of contraction of the anal sphincter, a patient activity level or a patient posture state.

[0044] Various examples are discussed relative to one or more stimulation devices. It is recognized that the stimulation devices may include features and functionality in addition to electrical stimulation. Many of these additional features are expressly discussed herein. A few example features include, but are not limited to, different types of sensing capabilities and different types of wireless communication capabilities. For ease of discussion, the present disclosure does not expressly recite every conceivable combination the additional features, such as by repeating every feature each time different examples and uses of the stimulation devices are discussed. [0045] FIG. 1 is a conceptual diagram illustrating an example system 10 that manages deliver} of neurostimulation to patient 14 to manage bladder dysfunction, such as overactive bladder, urgency, or urinary incontinence. As described above, system 10 may be configured to deliver neurostimulation to the patient timed during a physiological cycle based on detection of one or more physiological markers. System 10 may terminate delivery of therapy, begin delivery of therapy, and/or withhold delivery of therapy based on one or more detected physiological markers. For example, system 10 may monitor one or more phy siological markers to track and/or predict different phases of a physiological cycle that recurs within the patient. System 10 may then control delivery ⁷ of neurostimulation to occur during the appropriate phase of the physiological cycle to reduce or eliminate one or more dysfunctional states related to the physiological cycle.

[0046] As shown m the example of FIG. 1, therapy system 10 includes an implantable medical device (IMD) 16 (e.g., an example medical device), which is coupled to leads 18, 20, and 28 and sensor 22. System 10 also includes an external programmer 24, which is configured to communicate with IMD 16 via wireless communication. IMD 16 generally operates as a therapy device that delivers neurostimulation (e.g., electrical stimulation in the example of FIG.

1) to, for example, a target tissue site proximate a spinal nerve, a sacral nerve, a pudendal nerve, dorsal genital nerve, a tibial nerve, an inferior rectal nerve, a perineal nerve, or other pelvic nerves, or branches of any of the aforementioned nerves. IMD 16 provides electrical stimulation to patient 14 by generating and delivering a programmable electrical stimulation signal (e.g., in the form of electrical pulses or an electrical waveform) to a target a therapy site near lead 28 and, more particularly, near electrodes 29A-29D (collectively referred to as“electrodes 29”) disposed proximate to a distal end of lead 28.

[0047] IMD 16 may be surgically implanted in patient 14 at any suitable location within patient 14, such as near the pelvis. In some examples, IMD 16 may be implanted in a subcutaneous location in the side of the lower abdomen or the side of the lower back or upper buttocks. IMD 16 has a biocompatible housing, which may be formed from titanium, stainless steel, a liquid crystal polymer, or the like. The proximal ends of leads 18, 20, and 28 are both electrically and mechanically coupled to IMD 16 either directly or indirectly, e.g., via respective lead extensions. Electrical conductors disposed within the lead bodies of leads 18, 20, and 28 electrically connect sense electrodes (e.g., electrodes 19A, 19B, 21 A, and 21B) and stimulation electrodes, such as electrodes 29, to sensing circuitry and a stimulation delivery circuitry (e.g., a stimulation generator) within IMD 16. in the example of FIG. 1, leads 18 and 20 carry electrodes 19 A, 19B (collective referred to as“electrodes 19”) and electrodes 21 A, 21 B

(collectively referred to as“electrodes 21”), respectively. As described in further detail below, electrodes 19 and 21 may he positioned for sensing an impedance of bladder 12, which may increase as the volume of urine within bladder 12 increases. In some examples, system 10 may include electrodes (such as electrodes 19 and 21), a strain gauge, one or more accelerometers, ultrasound sensors, optical sensors, or any other sensor capable of detecting contractions of bladder 12, pressure or volume of bladder 12, or any other indication of the fill cycle of bladder 12 and/or possible bladder dysfunctional states.

[0048] In other examples, system 10 may use sensors other than electrodes 19 and 21 for sensing bladder volume, or not use any sensors at all. For example, external programmer 24 may receive user input identifying a voiding event, perceived level of fullness, or any other indication of a physiological marker associated with a phase of the physiological cycle. The user input may be in the form of a voiding journal analyzed by external programmer 24 or IMD 16 or individual user inputs associated with respective voiding events, leakage, or any other event related to a phase of the physiological cycle. External programmer 24 and/or IMD 16 may use this user input to generate estimated fill cycles and determine phases of the fill cycle to deliver neurostimulation and withhold stimulation. In other words, one or more physiological markers may be identified from user input. The user input may be in addition to or instead of sensors such as electrodes 19.4 and 21 A for detecting a physiological marker.

[0049] One or more medical leads, e.g., leads 18, 20, and 28, may be connected to IMD 16 and surgically or pereutaneously tunneled to place one or more electrodes carried by a distal end of the respective lead at a desired nerve or muscle site, e.g., one of the previously listed target therapy sites such as a tissue site proximate a spinal (e.g., sacral) or pudendal nerve. For example, lead 28 may be positioned such that electrodes 29 deliver electrical stimulation to a spinal, sacral or pudendal nerve to reduce a frequency and/or magnitude of contractions of bladder 12. Additional electrodes of lead 28 and/or electrodes of another lead may provide additional stimulation therapy to other nerves or tissues as well. In FIG. 1, leads 18 and 20 are placed proximate to an exterior surface of the wall of bladder 12 at first and second locations, respectively. In other examples of therapy system 10, IMD 16 may be coupled to more than one lead that includes electrodes for delivery of electrical stimulation to different stimulation sites within patient 14, e.g., to target different nerves.

[0050] In the example shown in FIG 1, leads 18, 20, 28 are cylindrical. Electrodes 19, 20,

29 of leads 18, 20, 28, respectively, may be ring electrodes, segmented electrodes, partial ring electrodes or any suitable electrode configuration. Segmented and partial ring electrodes each extend along an arc less than 360 degrees (e.g., 90-120 degrees) around the outer perimeter of the respective lead 18, 20, 28. In some examples, segmented electrodes 29 of lead 28 may be useful for targeting different fibers of the same or different nerves to generate different physiological effects (e.g., therapeutic effects). In examples, one or more of leads 18, 20, 28 may be, at least in part, paddle-shaped (e.g., a“paddle” lead), and may include an array of electrodes on a common surface, which may or may not be substantially flat.

[0051] In some examples, one or more of electrodes 19, 20, 29 may be cuff electrodes that are configured to extend at least partially around a nerve (e.g., extend axially around an outer surface of a nerve). Delivering electrical stimulation via one or more cuff electrodes and/or segmented electrodes may help achieve a more uniform electrical field or activation field distribution relative to the nerve, which may help minimize discomfort to patient 14 that results from the delivery' of electrical stimulation. An electrical field may define the volume of ti ssue that is affected when the electrodes 19, 20, 29 are activated. An activation field represents the neurons that will be activated by the electrical field in the neural tissue proximate to the activated electrodes.

[0052] The illustrated numbers and configurations of leads 18, 20, and 28 and electrodes earned by leads 18, 20, and 28 are merely exemplary. Other configurations, e.g., numbers and positions of leads and electrodes are also contemplated. For example, in other implementations, IMD 16 may be coupled to additional leads or lead segments having one or more electrodes positioned at different locations proximate the spinal cord or in the pelvic region of patient 14. The additional leads may be used for delivering different stimulation therapies or other electrical stimulations to respective stimulation sites within patient 14 or for monitoring at least one physiological marker of patient 14.

[0053] In accordance with some examples of the disclosure, IMD 16 delivers electrical stimulation to at least one of a spinal nerve (e.g., a sacral nerve), a pudendal nerve, dorsal genital nerve, a tibial nerve, an inferior rectal nerve, or a perineal nerve to provide a therapeutic effect that reduces or eliminates a dysfunctional state such as overactive bladder. The desired therapeutic effect may be an inhibitory physiological response related to voiding of patient 14, such as a reduction in bladder contraction frequency by a desired level or degree (e.g., percentage). In particular, IMD 16 may deliver stimulation via at least one of electrodes 29 during a second phase of the bladder fill cycle during which the nerves or targeted tissue respond efficaciously to stimulation therapy. IMD 16 may determine this second phase based on one or more physiological markers such as a bladder fill level (e.g., volume or pressure) or tune since the previous voiding event. IMD 16 may then control the therapy delivery circuitry to withhold stimulation delivery during other phases, such as a first phase subsequent to the voiding event and prior to the second phase during which stimulation is delivered. IMD 16 may monitor the one or more physiological markers and automatically adjust timing of the delivery of stimulation or parameters of the stimulation.

[0054] The stimulation program may define various parameters of the stimulation waveform and electrode configuration which result in a predetermined stimulation intensity being delivered to the targeted nerve or tissue. In some examples, the stimulation program defines parameters for at least one of a current or voltage amplitude of the stimulation signal, a frequency or pulse rate of the stimulation, the shape of the stimulation waveform, a duty cycle of the stimulation, a pulse width of the stimulation, and/or the combination of electrodes 29 and respective polarities of the subset of electrodes 29 used to deliver the stimulation. Together, these stimulation parameter values may be used to define the stimulation intensity (also referred to herein as a stimulation intensity level). In some examples, if stimulation pulses are delivered in bursts, a burst duty cycle also may contribute to stimulation intensity. Also, independent of intensity, a particular pulse width and/or pulse rate may be selected from a range suitable for causing the desired therapeutic effect after stimulation is terminated and, optionally, during stimulation. In addition, as described herein, a period during which stimulation is delivered may include on and off periods (e.g., a duty cycle or bursts of pulses) where even the short inter-pulse durations of time when pulses are not delivered are still considered part of the deliver} of stimulation. A period during which system 10 withholds stimulation delivery is a period in which no stimulation program is active for IMD 16 (e.g., IMD 16 is not tracking pulse durations or inter-pulse durations that occur as part of the electrical stimulation deliver} scheme). In other words, the withholding period is based on one or more physiological markers instead of a pre-defined pulse frequency, burst frequency, or duty cycle for an electrical stimulation signal or set of pulses. Typically, a period during which system 10 withholds neurostimulation is on the order of minutes or hours, not tenths of a second or several seconds.

[0055] in addition to the above stimulation parameters, the stimulation may be defined by other characteristics, such as a time for which stimulation is delivered, a time for which stimulation is terminated, and times during which stimulation is withheld. These times may be absolute or linked to the phy siological cycle and/or one or more physiological markers associated w th the physiological cycle. In some examples, IMD 16 may be configured to deliver different types of stimulation therapy at different times during the physiological cycle.

For example, IMF) 16 may deliver stimulation configured to reduce or eliminate bladder contractions to promote urine retention and/or increased bladder capacity and then deliver stimulation configured to promote urination (e.g., increased frequency or magnitude of bladder contractions) for a user requested voiding event or once a voiding event is detected to have begun.

[0056] System 10 may also include an external programmer 24, as shown in FIG 1. External programmer 24 may be a clinician programmer or patient programmer. In some examples, external programmer 24 may be a wearable communication device, with a therapy request input integrated into a key fob or a wristwatch, handheld computing device, smart phone, computer workstation, or networked computing device. External programmer 24 may include a user interface that is configured to receive input from a user (e.g., patient 14, a patient caretaker or a clinician). In some examples, the user interface includes, for example, a keypad and a display, which may for example, be a liquid crystal display (LCD) or light emitting diode (LED) display. The keypad may take the form of an alphanumeric keypad or a reduced set of keys associated with particular functions. External programmer 24 may additionally or alternatively include a peripheral pointing device, such as a mouse, via which a user may interact with the user interface. In some examples, a display of external programmer 24 may include a touch screen display, and a user may interact with external programmer 24 via the display. It should be noted that the user may also interact wath external programmer 24 and/or 1CD 16 remotely via a networked computing device.

[0057] A user, such as a physician, technician, surgeon, electrophysiologist, or other clinician, may also interact with external programmer 24 or another separate programmer (not shown), such as a clinician programmer, to communicate with 1MD 16. Such a user may interact with a programmer to retrieve physiological or diagnostic information from IMD 16. The user may also interact with a programmer to program IMD 16, e.g., select values for the stimulation parameter values with which IMD 16 generates and delivers stimulation and/or the other operational parameters of IMD 16, such as magnitudes of stimulation energy, user requested periods for stimulation or periods to prevent stimulation, or any other such user customization of therapy. As discussed herein, the user may also provide input to external programmer 24 indicative of physiological events (for physiological markers) such as bladder fill level perception and void events.

[0058] For example, the user may use a programmer to retrieve information from IMD 16 regarding the contraction frequency of bladder 12 and/or voiding events. As another example, the user may use a programmer to retrieve information from IMD 16 regarding the performance or integrity of IMD 16 or other components of system 10, such as leads 18, 20, and 28, or a power source of IMD 16. In some examples, this information may be presented to the user as an alert if a system condition that may affect the efficacy of therapy is detected.

[0059] Patient 14 may, for example, use a keypad or touch screen of external programmer 24 to request IMD 16 to deliver or terminate the electrical stimulation, such as when patient 14 senses that a leaking episode may be imminent or when an upcoming void may benefit from terminating therapy that promotes urine retention. In this way, patient 14 may use external programmer 24 to provide a therapy request to control the delivery of the electrical stimulation “on demand,” e.g., when patient 14 deems the second stimulation therapy desirable. This request may be a therapy trigger event used to terminate electrical stimulation. Patient 14 may also use external programmer 24 to provide other information to IMD 16, such as information indicative of a phase of a physiological cycle, such as the occurrence of a voiding event.

[0060] External programmer 24 may provide a notification to patient 14 when the electrical stimulation is being delivered or notify patient 14 of the prospective termination of the electrical stimulation. In addition, notification of termination may be helpful so that patient 14 knows that a voiding event may be more probable and/or the end of the fill cycle is nearing such that the bladder should be emptied (e.g., the patient should visit a restroom). In such examples, external programmer 24 may display a visible message, emit an audible alert signal or provide a somatosensory alert (e.g., by causing a housing of external programmer 24 to vibrate). In other examples, the notification may indicate when therapy is available (e.g., a countdown in minutes, or indication that therapy is ready) during the physiological cycle. In this manner, external programmer 24 may wait for input from patient 14 prior to terminating the electrical stimulation that reduces bladder contraction or otherwise promotes urine retention. Patient 14 may enter input that either confirms termination of the electrical stimulation so that the therapy stops for voiding purposes, confirms that the system should maintain therapy delivery until patient 14 may void, and/or confirms that patient 14 is ready for another different stimulation therapy that promotes voiding during the voiding event.

[0061] In the event that no input is received within a particular range of time when a voiding event is predicted, external programmer 24 may wirelessly transmit a signal that indicates the absence of patient input to IMD 16. IMD 16 may then elect to continue stimulation until the patient input is received, or terminate stimulation to avoid tissue damage, based on the programming of IMD 16. As described herein, the termination or continuation of electrical stimulation may be responsive to other physiological markers.

[0062] IMD 16 and external programmer 24 may communicate via wireless communication using any techniques known in the art. Examples of communication techniques may include, for example, low frequency or radiofrequency (RF) telemetry, but other techniques are also contemplated. In some examples, external programmer 24 may include a programming lead that may be placed proximate to the patient’s body near the IMD 16 implant site in order to improve the quality or security of communication between IMD 16 and external programmer 24.

[0063] In one example described herein, system 10 includes one or more processors (e.g., processor or control circuitry) contained in external programmer 24 and/or IMD 16) configured to monitor a fill cycle of bladder 12 of patient 14, where the fill cycle starts after completion of a first voiding event and ends at the completion of a second voiding event. In other examples, the fill cycle may be described as starting upon the beginning of a voiding event and ending upon detection of the end of the next fill cycle. A processor may control IMD 16 to withhold delivery of neurostimulation therapy to patient 14 for a first phase including the start of the fill cycle, wherein the neurostimulation therapy is configured to inhibit contraction of bladder 12. Since, as is discussed hereinafter with respect to FIGS. 9-14, bladder 12 (or the associated detrusor muscle or related nerves) may not be reactive to neurostimulation during the first phase of the fill cycle, IMD 16 may withhold the stimulation during this first phase. The processor may then determine, based on one or more physiological markers associated with the fill cycle (e.g., the previous voiding event or an indication of the current bladder fill level), a second phase of the fill cycle during which the neurostimulation therapy will be delivered to patient 14. This determination may include the duration of the second phase and when the second phase is to begin within the fill cycle (e.g., the timing of the second phase after the voiding event or prior to the next predicted voiding event). As described herein, the second phase for neurostimulation delivery may immediately follow the first phase where IMD 16 withholds stimulation. The processor of system 10 may then control IMD 16 to deliver neurostimulation therapy to patient 14 during the second phase of the fill cycle.

[0064] System 10 may terminate the second phase of the fill cycle prior to the next voiding event. This termination may help patient 14 complete the voiding event and void urine from bladder 12. However, the second phase and stimulation delivery may continue until the end of a voiding even in some examples. The second phase may be determined to at least start prior to a predicted start of dysfunctional events related to the fill cycle, such as frequent, coordinated, and/or strong detrusor muscle contractions. In this manner, the stimulation delivered during the second phase may preemptively reduce or even eliminate these dysfunctional events that may result in incontinence. System 10 may determine the second phase to overlap during a dysfunctional phase when dysfunction is predicted to occur. In other example, system 10 may deliver stimulation during the second phase and terminate the second phase prior to predicted dysfunction when the stimulation may effectively quiet bladder activity

[0065] As discussed further below with respect to FIGS. 9A-14, the placement of the second phase within the fill cycle may be made for effective delivery of stimulation. In one example, the third quartile of the fill cycle may include or comprise the second phase. In another example, the fourth quartile of the fill cycle may include or comprise the second phase. Therefore, the second phase may be located at least partially within or fully within these quartiles of the fill cycle. In other examples, the first half of the fill cycle may include the first phase during which system 10 withholds stimulation, and the second half of the fill cycle may include the second phase during which system 10 delivers stimulation therapy to patient 14. System 10 may, in some examples, deliver additional types of neurostimulation after the second phase, such as stimulation configured to promote bladder contraction during a voiding event. [0066] In some examples, system 10 is configured to detect at the voiding events and, responsive to the detection, control IMD 16 to terminate delivery of the neurostimulation therapy. Therefore, the first phase of the fill cycle would not include neurostimulation delivered to patient 14. However, as discussed above, external programmer 24 and/or IMD 16 may control delivery of different types of neurostimulation after the second phase stimulation has been terminated. For example, IMD 16 may deliver therapy during the voiding event that promotes voiding, such as promotion of detrusor muscle contraction.

[0067] System 10 may determine the different phases of a physiological cycle (e.g., the bladder fill cycle) based on different factors and using different inputs in different examples. In this manner, system 10 may predict durations and times of the physiological cycle and durations and times of the different phases for future cycles in order to time delivery and withholding of stimulation therapy to physiological markers. In one example, system 10 determines the various phases of the physiological cycle based on one or more previous physiological cycles. This historical data from the previous physiological cycles may allow the system to predict when to withhold delivery' of stimulation and when to deliver stimulation in order to treat one or more dysfunctional events associated with the cycle. For example, a processor of system 10 may be configured to determine when the second phase (e.g., when stimulation is delivered) of the bladder fill cycle begins by tracking a time period from the last voiding event that was detected. System 10 compares the time period to a fill time threshold which is the threshold in the fill cycle at which the second phase is to begin. Responsive to the time period exceeding the fill time threshold, system 10 may initiate the second phase of the fill cycle. The fill time threshold may be a percentage or fraction of the total fill time of the fill cycle or an absolute amount of time from the beginning of the cycle or to the predicted end of the fill cycle.

[0068] In some examples, system 10 is configured to estimate the fill time threshold based on respective durations of one or more previous fill cycles of patient 14. The previous fill cycles may be used to establish typical cycle durations, from which the fill time threshold may be calculated. For example, system 10 may calculate the fill tune threshold by calculating an average of the respective durations of the pluralit of previous fill cycles, determining an estimated second phase duration based on the average, and determining the fill time threshold as an initiation point for the second phase based on the average of the respective durations of the plurality of previous fill cycles. In other examples, a median, rolling average, weighted average, or some other estimation of the typical duration of a fill cycle may be used by system 10. The second phase duration may be calculated as a percentage of the estimated duration of the fill cycle and placed within the future cycle based on percentages of time or absolute times from the beginning of the cycle, end of the cycle, or predicted dysfunctional phase during which a dysfunctional event is predicted to occur. System 10 may also estimate durations and timing of one or more dysfunctional phases within the physiological cycle based on previously detected dysfunctional events.

[0069] In some examples, system 10 may lengthen the estimated fill time as there may be a carry-over effect that patient’s under stimulation therapy fill cycle times increase over time. For example, when stimulating for relatively less time, for example 50% of a fill cycle rather than 90% of a fill cycle, a patient may experience increasingly longer times between voids. In this manner, the therapeutic benefits of the stimulation during one fill cycle may carry-over into the next fill cycle. System 10 may lengthen the estimated fill time based on lengths of previous fill cycles, for example.

[0070] As an alternative to predicting future phases of a physiological cycle based on previous cycles of patient 14, system 10 may utilize direct detection of a fill level of bladder 12 as physiological markers (e.g., instead of or in addition to voiding events). System 10 may be configured to determine the second phase of the bladder fill cycle for neurostimulation delivery by detecting a magnitude of the fill level, comparing the magnitude of the fill level to a threshold, and, responsive to the magnitude of the fill level exceeding the threshold, initiating the second phase of the fill cycle. For example, if the second phase is to begin when at the midpoint of the bladder fill cycle, system 10 may determine the threshold to be half of the magnitude change for the fill cycle during the entire fill cycle. The threshold may be a percentage of a total fill level such as a percentage of a bladder volume, pressure, singular dimension such as diameter. The threshold may also be linked to a physiological parameter indicative of the fill level such as muscle activity via an electromyogram (EMG), nerve activity, or other indication.

[0071] The magnitude of the fill level may be a physiological marker for the bladder fill cycle. In one example, system 10 may detect the magnitude of the fill level by detecting a pressure level of bladder 14 (e.g., via sensor 22). For example, one or more pressure or stretch sensors may be attached to the exterior of bladder 14 or implanted within the bladder. As another example, system 10 may detect the magnitude of the fill level by detecting an impedance level of bladder 14, such as by monitoring the impedance between electrodes 19 and 21 of FIG.

1

[0072] IMD 16 may detect a contraction of bladder 12 using any suitable technique, such as based on a sensed physiological parameter that may be a physiological marker for the physiological cycle. In one example, a physiological marker is an impedance of bladder 12. In the example shown in FIG 1 , IMD 16 may determine impedance of bladder 12 using a four-ware (or Kelvin) measurement technique. In other examples, IMD 16 may measure bladder impedance using a two-wire sensing arrangement. In either case, IMD 16 may transmit an electrical measurement signal, such as a current, through bladder 12 via leads 18 and 20, and determine impedance of bladder 12 based on the transmitted electrical signal. Such an impedance measurement may be utilized to determine response of contractions of bladder 12 during the electrical stimulation or after termination of the electrical stimulation, to determine a fullness of bladder 12, or the like. Although fullness may be a physiological marker indicative of the need for the desired therapeutic effect, fullness may also indicate that the frequency of bladder contractions will increase to void bladder 12.

[0073] In the example four- wire arrangement shown in FIG. 1, electrodes 19 A and 21 A and electrodes I9B and 2IB, may be located substantially opposite each other relative to the center of bladder 12 For example, electrodes 19A and 21 A may be placed on opposing sides of bladder 12, either anterior and posterior or left and right. In FIG. 1, electrodes 19 and 21 are shown placed proximate to an exterior surface of the wall of bladder 12. In some examples, electrodes 19 and 21 may be sutured or otherwise affixed to the bladder wall. In other examples, electrodes 19 and 21 may be implanted within the bladder wall. To measure the impedance of bladder 12, IMD 16 may source an electrical signal, such as current, to electrode 19A via lead 18, while electrode 21 A via lead 20 sinks the electrical signal. IMD 16 may then determine the voltage between electrode 19B and electrode 21B via leads 18 and 20, respectively IMD 16 determines the impedance of bladder 12 using a known value of the electrical signal sourced the determined voltage.

[0074] In other examples, electrodes 19 and 21 may be used to detect an EMG of the detrusor muscle. This EMG may be used to determine the frequency of bladder contractions and the physiological marker of patient 14. The EMG may also be used to detect the strength of the bladder contractions in some examples. As an alternative, or in addition, to an EMG, a strain gauge or other device may be used to detect the status of bladder 12, e.g., by sensing forces indicative of bladder contractions.

[0075] In the example of FIG. 1, IMD 16 also includes a sensor 22 for detecting changes in the contraction of bladder 12. Sensor 22 may include, for example, a pressure sensor for detecting changes in bladder pressure, electrodes for sensing pudendal or sacral afferent nerve signals, electrodes for sensing urinary sphincter EMG signals (or anal sphincter EMG signals in examples in which system 10 provides therapy to manage fecal urgency or fecal incontinence), or any combination thereof. In examples in which sensor 22 is a pressure sensor, the pressure sensor may be a remote sensor that wirelessly transmits signals to IMD 16 or may be carried on one of leads 18, 20, or 28 or an additional lead coupled to IMD 16. In some examples, IMD 16 may determine whether a contraction frequency of bladder 12 has occurred based on a pressure signal generated by sensor 22.

[0076] In examples m which sensor 22 includes one or more electrodes for sensing afferent nerve signals, the sense electrodes may be carried on one of leads 18, 20, or 28 or an additional lead coupled to IMD 16. In examples in which sensor 22 includes one or more sense electrodes for generating a urinary sphincter EMG the sense electrodes may be carried on one of leads 18, 20, or 28 or additional leads coupled to IMD 16. In any case, m some examples, IMD 16 may control the timing of the delivery of the electrical stimulation based on input received from sensor 22.

[0077] Sensor 22 may comprise a patient motion sensor that generates a signal indicative of patient activity level or posture state. In some examples, IMD 16 may terminate the delivery of the electrical stimulation to patient 14 upon detecting a patient activity level exceeding a particular threshold based on the signal from the motion sensor. The patient activity level that is greater than or equal to a threshold (which may be stored in a memory of IMD 16) may indicate that there is an increase in the probability that an involuntary voiding event will occur, and, therefore, system 10 should deliver electrical stimulation during the second phase or even begin the second phase early in some examples. In other examples, IMD 16 may use sensor 22 to identify posture states known to require the desired therapeutic effect. For example, patient 14 may be more prone to an involuntary voiding event when patient 14 is in an upright posture state compared to a lying down posture state. In any event, electrodes 19 and 21 and sensor 22 may ^¬ be configured to detect voiding events and/or the magnitude of a fill level of bladder 12 during the fill cycle. Any of these detected features from patient 14 may be a physiological marker used by system 10 to determine when to deliver and withhold stimulation therapy.

[0078] As discussed above, system may monitor the fill cycle of bladder 12 by detecting subsequent voiding events over time. In some examples, system 10 may detect voiding events by receiving an indication of a user input (e.g., via external programmer 24) representative of an occurrence of a voiding event. In other words, external programmer 24 may receive input from the user identifying that a voiding event occurred, the beginning of a voiding event, and/or the end of the voiding event. In other examples, system 10 may automatically detect voiding events without receiving user input via external programmer 24. System 10 may instead detect voiding events by detecting at least one of a pressure of the bladder, a flow of urine from the bladder, a wetness of an article external of the patient, a volume of the bladder, an electromyogram (EMG) signal, a nerve recording, a posture change, a physical location of the patient within a structure such as a house or care facility, or a toilet use event. Some sensors external to patient 14 may communicate with external programmer 24 and/or IMD 16 to provide this information indicative of likely voiding events. For example, wetness may be detected by a moisture sensor (e.g., electrical impedance or chemical sensor) embedded in an undergarment worn by the patient and transmitted to IMD 16 or external programmer 24. Similarly, a toilet may include a presence sensor that detects when a patient is using the toilet (e.g., an infrared sensor, thermal sensor, or pressure senor) and tran smits a signal indicating the presence of the patient to IMD 16 or external programmer 24. In this manner, non-invasively obtained data may provide information indicative of voiding events without implanted sensors.

[0079] These examples of bladder therapy described in FIG. 1 are examples of therapy delivered based on a physiological marker associated with a phase of a physiological cycle to treat a dysfunctional state of the physiological cycle. However, such processes may also be used by system 10 to treat other dysfunctions and conditions of patient 14. In one example, one or more processors of system 10 may be configured to detect a physiological marker that occurs prior in time to a dysfunctional phase of a physiological cycle, wherein a dysfunctional state of the physiological cycle occurs during the dysfunctional phase without treatment (e.g., absent stimulation therapy the dysfunctional state may occur during the dysfunctional phase).

Responsive to detecting the physiological marker, system 10 may determine a first phase of the physiological cycle having a duration of time and withhold delivery of neurostimulation therapy for the duration of time of the first phase. Responsive to the first phase elapsing, a processor of system 10 may control a therapy delivery circuitry (e.g., a therapy delivery circuitry of IMD 16) to deliver neurostimulation therapy during a second phase that begins prior to the dysfunctional phase. In this manner, the neurostimulation therapy is configured to treat the dysfunctional state.

[0080] In some examples, the second phase at least partially overlaps with the dysfunctional phase. In other examples, the second phase ends prior to the dysfunctional phase, and responsive to the second phase ending, system 10 is configured to terminate delivery of the neurostimulation therapy. Generally, delivery of the neurostimulation therapy during the second phase of the physiological cycle one or reduces or eliminates the dysfunctional state of the physiological cycle. In some examples, delivery of the neurostimulation therapy during a first physiological cycle may even reduce or eliminate the dysfunctional state of a second physiological cycle subsequent to the first physiological cycle without delivery of the neurostimulation therapy during the second physiological cycle. In other words, delivery of the neurostimulation therapy during an appropriate time in the physiological cycle, such as during the second phase, may retrain organs and/or nerves to function properly again.

[0081] As discussed above, the dysfunctional state may include a bladder dysfunction. For bladder dysfunctions, the physiological marker may include a fill level of the bladder, a detrusor contraction during the first phase of the physiological cycle, and/or a voiding event. In other examples, the dysfunctional state may include a colon dysfunction. As described herein, neurostimulation therapy includes at least one of electrical stimulation or drug therapy

[0082] FIGS. 2A , 2B and 2C are block diagram illustrating example configurations of different types of implantable medical devices (IMDs) which may be utilized in the system of FIG 1. As shown in FIG 2 A, IMD 16 includes sensor 22, processor circuitry 53, therapy delivery circuitry 52, impedance circuitry 54, memory 56, telemetry circuitry 58, and power source 60. In other examples, IMD 16 may include a greater or fewer number of components.

For example, in some examples, such as examples in which IMD 16 deliver the electrical stimulation m an open-loop manner, IMD 16 may not include sensor 22 (e.g., a pressure sensor or electrical signal sensors) and/or impedance circuitr 54. In some examples, physiological markers may be provided via patient input on an external programmer if no sensors (e.g., sensor 22 and/or impedance circuitry 54) are included with IMD 16. [0083] Certain examples are directed toward a system with processor circuitry 53 that utilizes a closed-loop algorithm, such as classifier algorithm 34. Classifier algorithm 34 may be stored in memory 56. Classifier algorithm 34 may be configured to be responsive to sensor input (sensor 22 and/or sensor(s) external to IMD 16) when executed by processor circuitry 53. The sensor input may provide information about a physiological marker that is linked to a phase of a physiological cycle of the patient.

[0084] According to some examples, processor circuitry 53 (e.g., utilizing classifier algorithm 34) identifies changes to the patient’s physiological state that are relevant to desired changes in neurostimulation (e.g., that are indicative of a different phase of a physiological cycle). For example, voiding of the bladder may indicate that stimulation is not required for a certain time or until sensor input indicates otherwise. The system may include one or more sensors that sense biomarkers indicative of a change in the relevant physiological state(s), e.g., sensor 22 and/or sensors external to IMD 16. For example, a pressure sensor may detect the amount of bladder pressure while classifier algorithm 34 is configured to identify bladder voiding (a patient state change) from the sensed pressure. Thus, processor circuitry' 53 may be configured to classify certain changes in bladder pressure as corresponding to a void (e.g., when the sensed signals match one or more sets of parameters). Processor circuitry 53 may also classify the relative fullness of the bladder from subsequently detected pressure levels. Processor circuitry 53 may utilize a control policy 36 that dictates what action is taken based on classified physiological markers as determined by processor circuitry 53. Control policy 36 may be stored in memory 56. The actions may include when to enable the stimulation and when to disable the stimulation, or how to change the stimulation parameters, such as intensity, frequency or pulse width.

[0085] Specified parameters of the bladder pressure signal may be used to inform processor circuitry 53 to identify when voiding events have occurred. For instance, and without limitation, processor circuitry 53 may monitor one or more of bladder pressure, the amount of change in bladder pressure, the duration of change m bladder pressure, and the rate of change in bladder pressure. This data may serve as the physiological biomarker(s) for triggering changes in therapy relative to a voiding event. It should be noted that other nerve targets may alter urinary function in a manner that is similar to sacral nerves, such as the tibial nerve, pudendal nerve, dorsal nerve of the penis, and the dorsal nerve of the clitoris. Systems may be configured to provide stimulation at such the stimulation targets as part of a closed-loop stimulation solution.

[0086] According to various examples, the system is configured to apply a control policy that is based upon the classification of the sensed data. Control policy 36 may specify what action(s) the system takes. With respect to bladder dysfunction, physiological data from anesthetized rodents (rats) demonstrates that SNM applied during late, but not early, phases of the bladder filling cycle lead to an increase in bladder capacity that is equivalent to continuous therapy delivery' as is discussed further later in this disclosure with respect to FIGS. 9A-12B. The physiological data suggests that SNM applied during the period immediately after a void has little or no effect on bladder capacity. In some examples, processor circuitry 53 turns off (or reduces) SNM after a void. Processor circuitiy 53 then turns the therapy on after a certain period of time, or m response to a sensed physiological marker, or a combination thereof. The period of time could be set, for example, as a percentage or fraction of the inter-void interval that is determined to be effective, e.g., relative to a continuous therapy delivery' solution. The period of time could be determined, for example, through machine learning algorithm 68.

In some examples, processor circuitiy 53 may be configured to determine the relative fill state of the bladder based upon input from a sensor and an amount of elapsed time since a void.

[0087] As the stimulation may be therapeutic for patient 14, the time between voids may generally increase over time. So further power savings may be gained by adjusting the timing of the delivery of stimulation and the withholding of the delivery of stimulation. For example, the stimulation may reduce the feeling of urgency to void that patient 14 may experience. The reduction in the feeling of urgency may lead to long times between voids. Therefore, the increase in the duration between voids may further reduce the power consumption of IMD 16.

As such, IMD 16 may contain machine learning algorithm 68. Machine learning algorithm 68 may learn to associate information in sensor signals with physiological states of patient 14 and may adapt one or more of classifier algorithm 34, control policy 36, or therapy programs 66 so as to change the timing of the of the delivery of stimulation and the withholding of the deli very of stimulation.

[0088] Some examples of this disclosure are directed to a system that is configured to use an adaptive algorithm, such as machine learning algorithm 68. The adaptive algorithm may allow for adjustment of the therapy-off times (when to withhold deliver of stimulation) based on data that is collected over time. For example, IMD 16 may have an adaptive algorithm (e.g., machine learning algorithm 68) and may adapt one or more of classifier algorithm 34, control policy 36 or therapy programs 66, such that the therapy is tailored to the specific patient and/or that patient’s current situation. The tailoring might, for example, allow for therapy to be adjusted based upon fluctuations in the sensed physiological state due to circadian rhythms, changing environment, activity level, or other factors. For example, the adaptive algorithm may learn that a patient typically consumes coffee in the morning, leading to a shorter fill cycle and adjust one or more of the classifier algorithm 34, the control policy 36 or therapy programs 66 to provide stimulation earlier than it otherwise would in the morning.

[0089] According to some examples, the system may be configured to use a baseline control policy that is used when the system is first used by a patient. This baseline control policy may be control policy 36, for example. Machine learning algorithm 68 may adapt the baseline control policy, for example, based upon input relevant to the specific patient. In some examples, rather than adjusting the baseline control policy, machine learning algorithm 68 may adapt one or more of stimulation programs 66 or classifier algorithm 34. The baseline control policy may, for example, represent an optimal set of control policy settings for a typical or average patient (whether mean, median or mode). The baseline control policy might also be set to account for a worst-case situation (e.g., a situation where continual stimulation is desired). Machine learning algorithm 68 may then adapt the baseline control policy for the specific patient (e.g., by gradually increasing the threshold for the detection of a void state in response to the adjusted baseline control policy becoming more accurate). Machine learning algorithm 68 may also make the adjustments based on user input, e.g., from external programmer 24, and/or information in sensor signals.

[0090] In general, IMD 16 may comprise any suitable arrangement of hardware, alone or in combination with software and/or firmware, to perform the techniques attributed to IMD 16 and processor circuitry 53, therapy delivery circuitry 52, impedance circuitry 54, and telemetry circuitry 58 of IMD 16. In various examples, IMD 16 may include one or more processors, such as one or more microprocessors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), or any other equivalent integrated or discrete logic circuitry, as well as any combinations of such components. IMD 16 also, in various examples, may include a memory 56, such as random access memory (RAM), read only memory (ROM), programmable read only memory (PROM), erasable programmable read only memory (EPROM), electronically erasable programmable read only memory (EEPROM), flash memory, comprising executable instructions for causing the one or more processors to perform the actions attributed to them. Moreover, although processor circuitry 53, therapy deliver _} circuitry 52, impedance circuitry 54, and telemetry circuitry 58 are described as separate circuitry, in some examples, processor circuitry 53, therapy delivery circuitry 52, impedance circuitry 54, and telemetry circuitry 58 are functionally integrated. In some examples, processor circuitry 53, therapy delivery circuitry 52, impedance circuitry 54, and telemetry ⁷ circuitry ⁷ 58 correspond to individual hardware units, such as microprocessors, ASICs, DSPs, FPGAs, or other hardware units. In further examples, any of processor circuitry 53, therapy delivery circuitry ⁷ 52, impedance circuitry ⁷ 54, and telemetry ⁷ circuitry ⁷ 58 may correspond to multiple individual hardware units such as microprocessors, ASICs, DSPs, FPGAs, or other hardware units.

[0091] Memory 56 stores therapy programs 66 that specify stimulation parameter values for the electrical stimulation provided by IMD 16. Therapy programs 66 may also store information regarding determining and using physiological markers, information regarding physiological cycles and/or dysfunctional states, or any other information required by IMD 16 to deliver stimulation therapy based on one or more physiological markers. In some examples, memory 56 also stores bladder data 69, which processor circuitry 53 may use for controlling the timing of the delivery of the electrical stimulation (e.g , phases of physiological cycles that define when to deliver and withhold stimulation). For example, bladder data 69 may include threshold values or baseline values for at least one of bladder impedance, bladder pressure, sacral or pudendal afferent nerve signals, bladder contraction frequency, or external urinary sphincter EMG templates for use as physiological markers for an associated physiological cycle. Bladder data

69 may also include timing information and physiological markers associated with physiological events, such as a voiding event.

[0092] Memory 56 may also store machine learning algorithm 68. For example, machine learning algorithm 68 could be trained based upon patient indicated voids (through external programmer 24, for example) and sensed physiological markers that are contemporaneous to the indicated voids. For example, machine learning algorithm 68 may determine that a certain amount of change in bladder pressure, rate of change in bladder pressure or duration of change in bladder pressure for the particular patient is indicative of a void.

[0093] Information related to sensed bladder contractions, bladder impedance and/or posture of patient 14 may be recorded for long-term storage and retrieval by a user, and/or used by processor circuitry 53 for adjustment of stimulation parameters fe.g., amplitude, pulse width, and pulse rate) or for use as a physiological marker. In some examples, memory 56 includes separate memories for storing instructions, electrical signal information, stimulation programs 66, machine learning algorithm 68, bladder data 69, classifier algorithm 34 and control policy 36.

In some examples, processor circuitry 53 selects new stimulation parameters for a stimulation program 66 or new stimulation program from stimulation programs 66 to use m the delivery of the electrical stimulation based on patient input and/or monitored physiological markers after termination of the electrical stimulation.

[0094] Generally, therapy delivery circuitry 52 generates and delivers electrical stimulation under the control of processor circuitry 53. As used herein, controlling the deliver _} ' of electrical stimulation may also include controlling the termination of stimulation to achieve the different stimulation and non-stimulation phases of the physiological cycle. In some examples, processor circuitry 53 controls therapy delivery circuitry 52 by accessing memory' 56 to selectively access and load at least one of stimulation programs 66 to therapy delivery circuitry 52. For example, in operation, processor circuitry 53 may access memory 56 to load one of stimulation programs 66 to therapy delivery circuitry 52, In other examples, therapy delivery circuitry 52 may access memory 56 and load one of the stimulation programs 66.

[0095] By way of example, processor circuitry 53 may access memory 56 to load one of stimulation programs 66 to therapy delivery circuitry 52 for delivering the electrical stimulation to patient 14. A clinician or patient 14 may select a particular one of stimulation programs 66 from a list using a programming device, such as external programmer 24 or a clinician programmer. Processor circuitry 53 may receive the selection via telemetry circuitry 58

Therapy delivery circuitry 52 delivers the electrical stimulation to patient 14 according to the selected program for an extended period of time, such as minutes, hours, days, weeks, or until patient 14 or a clinician manually stops or changes the program.

[0096] Therapy delivery circuitry 52 delivers electrical stimulation according to stimulation parameters. In some examples, therapy delivery circuitr 52 delivers electrical stimulation in the form of electrical pulses. In such examples, relevant stimulation parameters may include a voltage amplitude, a current amplitude, a pulse rate, a pulse width, a duty cycle, or the combination of electrodes 29 that therapy deliver _} circuitry 52 uses to deliver the stimulation signal. In other examples, therapy delivery circuitry 52 delivers electrical stimulation in the form of continuous waveforms. In such examples, relevant stimulation parameters may include a voltage or current amplitude, a frequency, a shape of the stimulation signal, a duty cycle of the stimulation signal, or the combination of electrodes 29 therapy delivery circuitry 52 uses to deliver the stimulation signal.

[0097] In some examples, the stimulation parameters for the stimulation programs 66 may be selected to relax bladder 12, e.g., to reduce a frequency of contractions of bladder 12, after termination of the electrical stimulation. An example range of stimulation parameters for the electrical stimulation that are likely to be effective m treating bladder dysfunction, e.g., upon application to the spinal, sacral, pudendal, tibial, dorsal genital, inferior rectal, or perineal nerves, are as follows:

1 Frequency or pulse rate: between about 0.5 Hz and about 500 Hz, such as between about 1 Hz and about 250 Hz, between about 1 Hz and about 20 Hz, or about 10 Hz.

2 Amplitude: between about 0.1 volts and about 50 volts, such as between about 0.5 volts and about 20 volts, or between about 1 volt and about 10 volts. Alternatively, the amplitude may be between about 0.1 milliamps (mA) and about 50 m A, such as between about 0.5 rnA and about 20 mA, or between about 1 mA and about 10 mA.

3 Pulse Width: between about 10 microseconds (gs) and about 5000 gs, such as between about 100 gs and about 1000 gs, or between about 100 gs and about 200 gs.

[0098] W ⁷ hen IMD 16 is monitoring the fill level of the bladder to determine the status of the bladder fill cycle, processor circuitry 53 may monitor impedance of bladder 12 for a

predetermined duration of time to detect contractions of bladder 12, and determine the baseline contraction frequency of bladder 12 by determining a number of contractions of bladder 12 in the predetermined duration of time in other examples, electrodes 19 or 21 may be used to detect an EMG of the detrusor muscle to identify bladder contraction frequencies. Alternatively, a strain gauge sensor signal output or other measure of bladder contraction change may be used to detect the physiological marker of bladder 12, Each of these alternative methods of monitoring the fill level and/or voiding event of bladder 12 may be used m some examples. [0099] In the example illustrated in FIG 2A, impedance circuitry 54 includes voltage measurement circuitry 62 and current source 64, and may include an oscillator (not shown) or the like for producing an alternating signal. In some examples, as described above with respect to FIG 1 , impedance circuitry 54 may use a four- wire, or Kelvin, arrangement. As an example, processor circuitry 53 may periodically control current source 64 to, for example, source an electrical current signal through electrode 19A and sink the electrical current signal through electrode 21 A. In some examples, for collection of impedance measurements, current source 64 may deliver electrical current signals that do not deliver stimulation therapy to bladder 12, e.g., sub-threshold signals, due to, for example, the amplitudes or widths of such signals and/or the timing of delivery of such signals. Impedance circuitry 54 may also include a switching circuitry (not shown) for selectively coupling electrodes 19A, 19B, 21 A, and 21B to current source 64 and voltage measurement circuitry' 62. Voltage measurement circuitry' 62 may measure the voltage between electrodes 19B and 21B. Voltage measurement circuitry' 62 may include sample and hold circuitry' or other suitable circuitry' for measuring voltage amplitudes. Processor circuitry 53 determines an impedance value from the measure voltage values received from voltage measurement circuitry' 52.

[0100] In other examples, processor circuitry 53 may monitor signals received from sensor 22 to detect contraction of bladder 12 and determine the baseline contraction frequency. In some examples, sensor 22 may be a pressure sensor for detecting changes in pressure of bladder 12, which processor circuitry 53 may correlate to contractions of bladder 12. Processor circuitry 53 may determine a pressure value based on signals received from sensor 22 and compare the determined pressure value to a threshold value stored in bladder data 69 to determine whether the signal is indicative of a contraction of bladder 12. In some implementations, processor circuitry 53 monitors pressure of bladder 12 to detect contractions of bladder 12 for a predetermined duration of tune, and determines a contraction frequency of bladder 12 by calculating a number of contractions of bladder 12 in the predetermined time period.

[0101] In some examples, processor circuitry 53 may cause contraction frequency information to be stored as bladder data 69 in memory 56, and may utilize the changes to contraction frequency to track the fill level of the bladder fill cycle or otherwise track the phase of the fill cycle. In some implementations, processor circuitry' 53 may, automatically or under control of a user, determine the contraction frequency over the fill cycle. Processor circuitry 53 may determine that an increase in contraction frequency indicates a later phase of the fill cycle.

In some examples, processor circuitry 53 may track bladder contractions using EMG signals of patient 14. In some implementations, sensor 22 may include an EMG sensor, and processor circuitry 53 may generate an EMG from the received signals generated by sensor 22. Sensor 22 may be implanted proximate to a muscle which is active when bladder 12 is contracting, such as a detrusor muscle. Processor circuitry 53 may compare an EMG collected during the second tune period to EMG templates stored as bladder data 69 (e.g., a short-term running average) to determine whether the contractions of bladder 12 are indicative of particular phases of the bladder fill cycle.

[0102] In other examples, sensor 22 may be a pressure sensor and processor circuitry 53 may monitor signals received from sensor 22 during at least a portion of the second time period to detect contraction of bladder 12. In some examples, processor circuitry 53 substantially continuously monitors pressure of bladder 12, at least during the second time periods, to detect contraction of bladder 12, and determines a contraction frequency of bladder 12 by determining a number of contractions of bladder 12 in a specified time period. Sensor 22 may also provide longer-term changes in pressure to track the bladder fill status (e.g., increased bladder volume may correspond to increased bladder pressure).

[01Q3] In the example of FIG. 2A, therapy delivery circuitry 52 drives electrodes on a single lead 28. Specifically, therapy deliver _} ' circuitry 52 delivers electrical stimulation to tissue of patient 14 via selected electrodes 29A-29D carried by lead 28. A proximal end of lead 28 extends from the housing of IMD 16 and a distal end of lead 28 extends to a target therapy site, such as a spinal nerve (e.g., an S3 nerve), or a therapy site within the pelvic floor, such as tissue sites proximate a sacral nerve, a pudendal nerve, a tibial nerve, a dorsal genital nerve, an inferior rectal nerve, a perineal nerve, a hypogastric nerve, a urinary sphincter, or any combination thereof. In other examples, therapy delivery circuitry 52 may deliver electrical stimulation with electrodes on more than one lead and each of the leads may carry one or more electrodes. The leads may be configured as an axial lead with ring electrodes or segmented electrodes and/or paddle leads with electrode pads arranged in a two-dimensional array. The electrodes may operate in a bipolar or multi-polar configuration with other electrodes, or may operate in a unipolar configuration referenced to an electrode carried by the device housing or“can” of IMD 16 [0104] As previously described, sensor 22 may comprise a pressure sensor configured to detect changes m bladder pressure, electrodes for sensing pudendal or sacral afferent nerve signals, or electrodes for sensing external urinary sphincter EMG signals (or anal sphincter signals in examples in which IMD 16 provides fecal urgency or fecal incontinence therapy), or any combination thereof Additionally, or alternatively, sensor 22 may comprise a motion sensor, such as a two-axis accelerometer, three-axis accelerometer, one or more gyroscopes, pressure transducers, piezoelectric crystals, or other sensors that generate a signal that changes as patient activity level or posture state changes. Processor circuitry 53 may detect a physiological marker indicative of point during a bladder fill cycle. Sensor 22 may also be a motion sensor that is responsive to tapping (e.g., by patient 14) on skin superior to IMD 16. Processor circuitry 53 may be configured to log patient input using this tapping method (e.g., tapping may indicate that a voiding event is occurring). Alternatively, or in addition, processor circuitry 53 may control therapy circuitry 52 to deliver or terminate electrical stimulation delivery in response to the tapping or certain pattern of tapping.

[0105] In examples in which sensor 22 includes a motion sensor, processor circuitry 53 may determine a patient activity level or posture state based on a signal generated by sensor 22. This patient activity level may be, for example, sitting, exercising, working, running, walking, or any other activity of patient 14. For example, processor circuitry 53 may determine a patient activity level by sampling the signal from sensor 22 and determining a number of activity counts during a sample period, where each activity level of a plurality of activity levels is associated with respective activity counts. In one example, processor circuitry 53 compares the signal generated by sensor 22 to one or more amplitude thresholds stored within memory 56, and identifies each threshold crossing as an activity count. The physical activity may be indicative of a fill level, a voiding event, or any other physiological marker related to the bladder fill cycle.

[0106] In some examples, processor circuitry 53 may control therapy delivery circuitry 52 to deliver or terminate the electrical stimulation based on patient input received via telemetry- circuitry 58. Telemetry circuitry 58 includes any suitable hardware, firmware, software or any combination thereof for communicating with another device, such as external programmer 24 (FIG 1). Under the control of processor circuitry 53, telemetry circuitry 58 may receive downlink telemetry, e.g., patient input, from and send uplink telemetry, e.g., an alert, to external programmer 24 with the aid of an antenna, which may be internal and/or external. Processor circuitry 53 may provide the data to be uplinked to external programmer 24 and the control signals for the telemetry circuit within telemetry circuitry 58, and receive data from telemetry circuitry 58.

[0107] Generally, processor circuitry 53 may control telemetry circuitry 58 to exchange information with external programmer 24 and/or another device external to IMD 16. Processor circuitry 53 may transmit operational information and receive stimulation programs or stimulation parameter adjustments via telemetry circuitry 58. Also, in some examples, IMD 16 may communicate with other implanted devices, such as stimulators, control devices, or sensors, via telemetry circuitry 58.

[0108] Power source 60 delivers operating power to the components of IMD 16. Power source 60 may include a battery' and a power generation circuit to produce the operating power. In some examples, the battery may be rechargeable to allow' extended operation. Recharging may be accomplished through proximal inductive interaction between an external charger and an inductive charging coil within IMD 16. In other examples, an external inductive power supply may transcutaneously power IMD 16 whenever electrical stimulation is to occur.

[0109] As shown in FIG. 2B, IMD 70 is similar to IMD 16 of FIG. 2A, but IMD 70 delivers neurostimulation to patient 14 in the form of drugs instead of electrical stimulation. IMD 70 includes processor circuitry' 73 (e.g., similar to processor circuitry' 53), therapy delivery' module 74 coupled to catheter 75, sensor 76 (e.g., a pressure sensor similar to sensor 22 of FIG. 2 A), telemetry circuitry 78 (e.g., similar to telemetry circuitry 58), memory 80 (e.g., similar to memory 56), and power source 86 (e.g., similar to power source 60. Although IMD 70 does not include impedance circuitry 54, this or other circuitry may be provided in some examples.

[0110] Therapy delivery module 74 may include a drug reservoir and drug pump that moves the drug from the reservoir, through catheter 75, and out to patient 14 In some examples, IMD 70 may include both a drug pump and electrical stimulation generator. Memory 80 may include therapy programs 82, machine learning algorithm 83, bladder data 84, classifier algorithm 85. Therapy programs 82 may include instructions for drug delivery that may be based on one or more physiological markers stored as bladder data 84. Processor circuitry 73 may predict when to deliver a bolus of drug to patient 14 based on a phase of a physiological cycle such as the bladder fill cycle, for example, in a manner similar to that of processor circuitry 53 of FIG. 2A with respect to the delivery of stimulation. |Ό111] FIG 2C is a block diagram illustrating an IMD 59 which is similar to IMD 16 of FIG 2A. In this example, classifier device 35 may function in a similar fashion to processor circuitry 53 executing classifier algorithm 34 of FIG 2A. For example, classifier device 35 may accept a sensor signal from sensor 22 and classify a physiological marker, such as a bladder void, of patient 14 based upon the sensor signal. Control policy device 37 may function m a similar fashion to processor circuitry 53 executing control policy 53 of FIG. 2A. Classifier device 35 may signal control policy device 37 with the classification of the physiological marker and control policy device 37 may control therapy delivery circuitry 52 based on the classification.

For example, control policy device 37 may control therapy delivery circuitry' 52 to turn off stimulation after a void event. In the example of FIG 2C, processor circuitry 57 may execute machine learning algorithm 68. Machine learning algorithm may learn to associate information m sensor signals with physiological states of patient 14 and may adapt one or more of classifier device 35, control policy device 37, or therapy programs 66 so as to change the timing of the of the delivery of stimulation and the withholding of the delivery _' of stimulation. In some examples, classifier device 35 and control policy device 37 may be a single device. In some examples, elements of FIGS 2A, 2B and 2C may he combined in various manners. For example, an IMD may contain processing circuitry that may be configured to execute a classification algorithm and signal a classification to a control policy device.

[0112] FIG. 3 is a block diagram illustrating an example configuration of an external programmer 24. While external programmer 24 may generally be described as a hand-held computing device, external programmer 24 may be a notebook computer, a smart phone, or a workstation, for example. As illustrated in FIG 3, external programmer 24 may include a processor circuitry 90, memory 92, user interface 94, telemetry circuitry 96, and power source 98. Memory 92 may store program instructions that, when executed by processor circuitry 90, cause processor circuitry 90 and external programmer 24 to provide the functionality ascribed to external programmer 24 throughout this disclosure.

[0113] In general, external programmer 24 comprises any suitable arrangement of hardware, alone or in combination with software and/or firmware, to perform the techniques attributed to external programmer 24, and processor circuitry 90, user interface 94, and telemetr circuitry' 96 of external programmer 24. In various examples, external programmer 24 may include one or more processors, such as one or more microprocessors, DSPs, ASICs, FPGAs, or any other equivalent integrated or discrete logic circuitry, as well as any combinations of such components. External programmer 24 also, in various examples, may include a memory 92, such as RAM, ROM, PROM, EPROM, EEPROM, flash memory , a hard disk, a CD-ROM, comprising executable instructions for causing the one or more processors to perform the actions attributed to them. Moreover, although processor circuitry 90 and telemetry circuitry 96 are described as separate circuitry, in some examples, processor circuitry 90 and telemetry circuitry 96 are functionally integrated. In some examples, processor circuitry 90 and telemetry circuitry 96 and telemetry circuitry 58 correspond to individual hardware units, such as microprocessors, ASICs, DSPs, FPGAs, or other hardware units. In other examples, any of processor circuitry 90 and telemetry circuitry 96 and telemetry circuitry 58 may correspond to multiple individual hardware units, such as microprocessors, ASICs, DSPs, FPGAs, or other hardware units.

[0114] Memory 92 may store program instructions that, when executed by processor circuitry ⁷ 90, cause processor circuitry ⁷ 90 and external programmer 24 to provide the

functionality ascribed to external programmer 24 throughout this disclosure. In some examples, memory 92 may further include program information, e.g., stimulation programs defining the neurostimulation, similar to those stored in memory ⁷ 56 of IMD 16. The stimulation programs stored in memory 92 may be downloaded into memory ⁷ 56 of IMD 16.

[0115] In certain examples, the system includes a user interface 94 that allows the patient to provide input. IMD 16 may respond to patient-supplied data from the user interface by altering therapy. For example, a patient may use external programmer 24 (e.g., a handheld device) to record (by pushing a button) a physiological e\ ent of interest. Processor circuitry 53 of IMD 16 may respond by turning the therapy on or off, or by adjusting the therapy (e.g., the stimulation strength) or by changing the therapy program. With reference to the urological applications discussed herein, the patient could push a button on external programmer 24 (e.g., their smartphone) when the bladder is voided. This sends a signal to IMD 16 to turn off for a period of time that is either pre-programmed based on the voiding characteristics of the patient, or (heuristiealiy) determined by an algorithm that interprets the frequency with winch the patient sends the“voiding” signal, such as machine learning algorithm 68. In such examples, the patient input is used by the distributed“closed loop” system as a source of data about the physiological event of interest that is impacted by the therapy. [0116] Consistent with various examples, the patient-supplied data may be used to configure the classifier algorithm, such as classifier algorithm 34. For example, machine learning algorithm 68 may be trained based upon patient indicated voids and senses physiological markers that are contemporaneous to the indicated voids. For example, machine learning algorithm 68 may determine that a certain amount of change m bladder pressure, rate of change m bladder pressure or duration of change in bladder pressure for the particular patient is indicative of a void.

[0117] User interface 94 may include a button or keypad, lights, a speaker for voice commands, a display, such as a liquid crystal (LCD), light-emitting diode (LED), or cathode ray tube (CRT). In some examples the display may be a touch screen. As discussed in this disclosure, processor circuitry 90 may present and receive information relating to electrical stimulation and resulting therapeutic effects via user interface 94. For example, processor circuitry 90 may receive patient input via user interface 94. The input may be, for example, in the form of pressing a button on a keypad or selecting an icon from a touch screen.

[0118] Processor circuitry 90 may also present information to the patient in the form of alerts related to deliver} of the electrical stimulation to patient 14 or a caregiver, as described in more detail below, via user interface 94. Although not shown, external programmer 24 may additionally or alternatively include a data or network interface to another computing device, to facilitate communication with the other device, and presentation of information relating to the electrical stimulation and therapeutic effects after termination of the electrical stimulation via the other device.

[0119] Telemetry circuitry 96 supports wireless communication between IMD 16 and external programmer 24 under the control of processor circuitry 90. Telemetry circuitry 96 may also be configured to communicate with another computing device via wireless communication techniques, or direct communication through a wired connection. In some examples, telemetry circuitry 96 may be substantially similar to telemetry circuitry 58 of IMD 16 described above, providing wireless communication via an RF or proximal inductive medium. In some examples, telemetry circuitry 96 may include an antenna, which may take on a variety of forms, such as an internal or external antenna.

[0120] Examples of local wireless communication techniques that may be employed to facilitate communication between external programmer 24 and another computing device include RF communication according to the 802.11 or Bluetooth specification sets, infrared communication, e.g., according to the lrDA standard, or other standard or proprietary telemetry protocols. In tins manner, other external devices may be capable of communicating with programmer 24 without needing to establish a secure wireless connection.

[0121] Power source 98 delivers operating power to the components of programmer 24. Power source 98 may include a battery and a power generation circuit to produce the operating power in some examples, the battery may be rechargeable to allow extended operation.

[0122] FIG. 4 is a flow diagram that illustrates an example technique for determining a timing for therapy delivery based on a physiological marker. The technique of FIG. 4 will be described with respect to processor circuitry 53 of IMD 16 for purposes of illustration.

However, in other examples, processor circuitry 73 of IMD 70 or processor circuitry 90 of external programmer 24 may perform similar functions or employ a distributive functionality with other devices (e.g., functionality split between external programmer 24 and IMD 16).

[0123] As shown m FIG. 4, processor circuitry 53 may monitor patient 14 for a physiological marker (100). The physiological marker may be associated with a dysfunctional state of patient 14. If processor circuitry' 53 does not detect the physiological marker (“NO” branch of block 102), processor circuitry 53 may continue to monitor for the physiological marker (100). If processor circuitry' 53 detects the physiological marker (“YES” branch of block 102), processor circuitry 53 may determine timing for delivery of therapy after the physiological marker (104). For example, processor circuitry 53 may determine the duration of time between the

physiological marker and delivery of neurostimulation and/or the duration of the phase during which neurostimulation will be delivered. As discussed herein, processor circuitry 53 may determine the timing of therapy delivery based on previously detected or recorded dysfunctional states and whether the neurostimulation should be delivered prior to and/or during the dysfunctional state.

[0124] Processor circuitry 53 may then deliver the neurostimulation therapy according to the determined timing (106). If processor circuitry' 53 is to continue delivering therapy to treat another predicted dysfunctional state (“YES” branch of block 108), processor circuitry 53 may continue to monitor patient 14 for the physiological marker (100). If processor circuitry 53 is to stop delivering therapy (“NO” branch of block 108), processor circuitry' 53 terminates the therapy deli very program (110). The process of FIG 4 may be applicable to physiological cycles that include a dysfunctional state such as a bladder fill cycle for patient 14 suffering from incontinence. However, processor circuitry 53 may track other physiological cycles or recurring events to predict when to deliver neurostimulation to treat an upcoming dysfunctional state. The process of FIG. 4, similar to the process of FIG. 5, may be suitable for treating dysfunction associated with organs other than the bladder, such as the large and small bowel, stomach and/or intestines, liver, or spleen, as some examples.

[0125] In some examples, processor circuitry 53 may withhold neurostimulation delivery during the first phase after detection of the physiological marker. In other examples, processor circuitry ⁷ 53 may control the medical device to deliver a first neurostimulation therapy (or multiple different therapies) between the physiological marker and the neurostimulation timed to the physiological marker. Processor circuitiy 53 may transition from the first neurostimulation therapy to a second neurostimulation timed to the physiological marker by changing one or more parameters of the first neurostimulation therapy according to the parameters of the second neurostimulation therapy. Processor circuitiy 53 may change parameters such as at least one of a voltage amplitude, a current amplitude, a pulse frequency, a stimulation waveform frequency, a pulse width, or an electrode combination. In this manner, the physiological marker may signal the timing of a modification to neurostimulation.

[0126] FIG. 5 is a flow diagram that illustrates an example technique for determining phases to withhold and delivery ⁷ neurostimulation to manage bladder dysfunction. The technique of FIG. 5 will be described with respect to processor circuitry 53 of IMD 16 for purposes of illustration. However, in other examples, processor circuitry 73 of IMD 70 or processor circuitry 90 of external programmer 24 may perform similar functions or employ a distributive functionality to with other devices (e.g., functionality split between external programmer 24 and IMD 16).

[0127] As shown in FIG. 5, processor circuitry 53 may monitor the bladder fill cycle of bladder 12 of patient 14 (120) The fill cycle may be defined by beginning immediately after the end of a voiding event and ending at the end of the next voiding event. However, the fill cycle may include the voiding event at the beginning of the cycle in other examples. If processor circuitry 53 does not detect a voiding event signaling the end of the fill cycle (“NO” branch of block 122), processor circuitry 53 may continue to monitor the fill cycle of the bladder (120). If processor circuitry 53 detects the end of the fill cycle (“YES” branch of block 122), processor circuitry 53 may determine to withhold delivery of neurostimulation therapy during the first phase of the fill cycle (124). in some examples, the neurostimulation may be terminated prior to or during the voiding event, so processor circuitry 53 may simply continue not delivering stimulation. However, if IMD 16 is still delivering stimulation at the end of the voiding event, processor circuitry 53 may terminate delivery of the neurostimulation and then withhold further delivery of neurostimulation during the first phase.

[0128] Based on a physiological marker indicative of a phase of the fill cycle, processor circuitry 53 may determine a second phase for the fill cycle (126). The second phase is defined for delivery of neurostimulation at an appropriate time prior to a predicted dysfunctional state of bladder 14 (e.g., overactive bladder contractions). The second phase may have a duration determined to provide therapy when the tissue is receptive to the therapy and a starting point prior to the dysfunction in order to reduce or eliminate the dysfunction. If the second phase has not started yet (“NO” branch of block 128), processor circuitry 53 may continue to wait until the first phase has ended. If processor circuitry 53 determines that the second phase should start (“YES” branch of block 128), processor circuitry 53 starts the second phase and controls therapy- delivery circuitry ⁷ 52 to deliver neurostimulation therapy to patient 14 in order to inhibit or reduce contractions of bladder 12 during the second phase (130).

[0129] The neurostimulation therapy ⁷ delivered during the second phase of the fill cycle may be configured to reduce or eliminate contractions of bladder 12 (e.g., relax the bladder and reduce the likelihood of incontinence). Although only one phase for stimulation is described, processor circuitry 53 may determine two or more phases of the fill cycle during which neurostimulation is delivered. Generally, the second phase, and corresponding neurostimulation, at least begins prior to a dysfunctional state that may occur during that bladder fill cycle.

However, processor circuitry 53 may control therapy delivery ⁷ circuitry 52 to deliver

neurostimulation until the following voiding event. Processor circuitry 53 may control therapy delivery circuitry 52 to deliver a different neurostimulation, e.g., with different stimulation parameters than the stimulation delivered during the second phase, just prior to and/or during the next voiding event to promote voiding. The voiding event may also act as a physiological marker for this voiding promoting stimulation.

[0130] In other examples, processor circuitr 53 may deliver neurostimulation during the first phase of the process of FIG. 5 instead of withholding all neurostimulation delivery. For example, processor circuitry 53 may control therapy delivery circuitry 52 to deliver a first neurostimulation therapy for multiple different therapies) during the first phase and prior to the second phase. Processor circuitry 53 may transition from the first neurostimulation therapy of the first phase to the second neurostimulation of the second phase by changing one or more stimulation parameters defining the first neurostimulation therapy to achieve the parameters of the second neurostimulation therapy for the second phase. Processor circuitr 53 may change parameters such as at least one of a voltage amplitude, a current amplitude, a pulse frequency, a stimulation waveform frequency, a pulse width, or an electrode combination or may change the therapy program. In this manner, the physiological marker may signal the timing of a modification to neurostimulation, not just when to start neurostimulation.

[0131] FIG. 6 is a flow diagram that illustrates an example technique for training a device to automatically associate a bladder void with physiological markers m sensor signals. Processor circuitry ⁷ 53 of IMD 16 may monitor physiological marker(s) in sensor signals, e.g., in a signal from sensor 22 or sensor(s) external to IMD 16 (600). In some examples, the physiological marker(s) may include bladder pressure, the rate of change in bladder pressure, the amount of change in bladder pressure and/or the duration of change in bladder pressure. Patient 14 may alert IMD 16 (e.g., through telemetry circuitry 96 of external programmer 24 and telemetry ⁷ circuitry' 58 of IMD 16) that patient 14 has just voided their bladder (602). IMD 16 may then record the physiological marker(s) in bladder data 69 of memory 56 and associate them with a voiding event (604). Processor circuitry ⁷ 53 and machine learning algorithm 68 of IMD 16 may compare the physiological marker(s) in bladder data 69 with previously stored physiological markers in bladder data 69 associated with a voiding event to determine if any of the

physiological markers are indicative of a voiding event (606). For example, if a certain rate of change in bladder pressure is frequently associated with a voiding event, then that rate of change may be indicative of a voiding event. If none of the physiological markers are indicative of a voiding event, processor circuitry 53 of IMD 16 may continue to monitor physiological markers (600). If any of the physiological markers are indicative of a voiding event, processor circuitry 53 and machine learning algorithm 68 may, for example, alter classifier algorithm 34 to classify a physiological state of patient 14 as a void when processor circuitry 53 receives the

physiological markers indicative of a voiding event in sensor signals, e.g., in a signal from sensor 22 or sensor(s) external to IMD 16 (608), even in the absence of a patient input. [0132] In some examples, the example of FIG. 6 may be used for physiological states other than a void. For example, the example of FIG. 6 may be used for learning which physiological markers may be indicative of bladder leakage. In this example, patient 14 may alert IMD 16 that they have just experienced bladder leakage and machine learning algorithm 68 may determine if any of the physiological markers are indicative of bladder leakage and alter classifier algorithm 34 to classify a physiological state of patient 14 as a bladder leakage event when processor circuitry 53 receives the physiological markers indicative of a bladder leakage event. In other examples, the example of FIG. 6 may be used for learning which physiological markers may be indicative of an upcoming voiding event.

[0133] FIG. 7 is a flow diagram that illustrates an example technique for automatically adapting parameters or timing of therapy. Processor circuitiy 53 of IMD 16 may monitor physiological marker(s) in sensor signals, e.g., in a signal from sensor 22 or sensor(s) external to IMD 16 (700). In some examples, the physiological marker(s) may include bladder pressure, the rate of change in bladder pressure, the amount of change in bladder pressure and/or the duration of change in bladder pressure. IMD 16 may determine whether it is the end of a bladder fill cycle, for example a void event has occurred (702). For example, processor circuitry 53 may determine that a void event has occurred based on physiological markers in the sensor signals. Alternatively, IMD 16 may receive an indication from patient 14 (e.g., through telemetry circuitry ⁷ 96 of external programmer 24 and telemetry ⁷ circuitiy 58 of IMD 16) that patient 14 has just voided their bladder. If IMD 16 does not determine it is the end of a bladder fill cycle, processor circuitry 53 of IMD 16 may continue to monitor the physiological markers (700).

[0134] If IMD 16 does determine that it is the end of a bladder fill cycle, IMD 16 may record the duration of the bladder fill cycle (e.g., the time from the last void to the present void) (704). Processor circuitry 53 and machine learning algorithm 68 of IMD 16 may compare the duration of the bladder fill cycle with durations of past bladder fill cycles stored in bladder data 69 in memory ⁷ 56, for example. Processor circuitry 53 and machine learning algorithm 68 may then determine whether parameters or timing of therapy should change (708). The parameters of therapy may include the pulse width, frequency, and intensity (voltage or current) of the stimulation, and the timing of therapy may include when to deliver therapy and when to withhold delivery of therapy. |Ό135] For example, processor circuitr 53 and machine learning algorithm 68 may determine that there is a trend in the duration of a bladder fill cycle, such as a bladder fill cycle duration is increasing by generally X minutes a week. As additional power savings may be gained by increasing the time period to withhold delivery of stimulation, processor circuitry 53 and machine learning algorithm 68 may determine it should increase the time period to withhold deliver _} of stimulation, e.g., by X/2 minutes. Processor circuitry 53 and machine learning algorithm 68 may then change the timing of therapy (710).

[0136] FIG. 8 is an example timing diagram 140 of a bladder fill cycle and delivery of neurostimulation timed to a phase of the bladder fill cycle. As shown in FIG. 8, timing diagram 140 shows the bladder fill cycle with the fill level 142 (e.g., the percentage of full bladder volume) increasing over time during the bladder fill cycle. Portion 146 shows the bladder filling with urine, and portion 148 of fill level 142 show¾ that bladder 12 is emptying during a voiding event. Stimulation level 144 show¾ whether stimulation is being delivered or not.

[0137] Phase Ti is the first phase of the bladder fill cycle that starts at time 150, and phase Tz is the second phase of the bladder fill cycle that starts at time 152 and runs to time 154. During phase Ti, system 10 may withhold stimulation delivery. When the first phase Ti is completed at time 152, system 10 may start the second phase T2 and deliver neurostimulation during this second phase. The timing of the second phase during the fill cycle may be selected to reduce or eliminate dysfunctional states of the bladder as the fill level 142 increases. At the end of phase T2, the patient may need to void at time 154. The beginning of the voiding event during the period T3 is at time 154 and the end of the voiding event is at time 156. Although system 10 terminated delivery of neurostimulation at time 1 54, neurostimulation may continue during the voiding event in other examples.

[0138] Although the second phase T2 is shown as occupying almost the entire second half of the bladder fill cycle, the second phase may occur at other times and with other durations of the fill cycle in other examples. For example, the duration of T2 may only last for a quarter of the fill cycle, and Tz may only occur during the third quartile or the fourth quartile of the bladder fill cycle. In any case, the timing of the first phase and the second phase may be based on a voiding event (e.g., a physiological marker) and historical fill cycles in order to predict when to deliver neurostimulation in order to reduce or eliminate the dysfunctional state (e.g., overactive bladder) during the physiological cycle (e.g., the fill cycle). [0139] FIGS. 9A-12B relate to experimental data and also show conceptual timing of neurostimulation to reduce bladder dysfunction. The experimental data was obtained using a novel sacral neurostimulation (SCS) model in rats with bilateral bipolar electrode stimulation of the L6S1 roots. With this stimulation, together with some other stimulus parameters, bladder areflexia (i.e. complete blockade of bladder-to-bladder reflexes) was produced under conditions of SNS below somatomotor threshold in many animals. These stimulation delivery paradigms were examined to determine whether intermittent application of stimulation timed to certain intervals during the filling portion of the fill cycle (e.g. immediately post-void, timed to mid-cycle or immediately preceding voids) may produce similar effects of increasing bladder capacity shown by continuous stimulation throughout the fill cycle. Battery life may be increased and patient side effects may be decreased by such an approach of intermittent neurostimulation using a timed approach based on voiding events and/or fill cycle levels.

[0140] Electrodes were fashioned from 50 micron diameter stainless steel wire with epoxy coating. The electrodes were paired as poles and joined at the binding post of the stimulator terminals. Rostral poles were positioned rostrally on the nerve bilaterally and the leads were then combined and inserted into the output terminal binding post of the FHC Pu lsar 6bp stimulator, and the caudal poles w ^r ere positioned caudally on the nerve bilaterally and the leads were then combined and inserted into the ground terminal binding post. Parafilm acted to electrically isolate the electrodes. Female Sprague-Dawley rats (250-275 g BW, n=13) were anesthetized with urethane (1.2 g/kg s.c.). Jugular vein catheters were inserted unilaterally for hydration, a transvesical catheter was inserted into a cystotomy in the apex of the bladder dome and ligated in position, and the L6-S1 trunks were isolated as they passed from ventral to dorsal sacrum. Small sheets of Parafilm were inserted between the nerves and the sciatic, inferior iliac vein and other surrounding tissues, isolating the L6-S1 trunks of the spinal cord electrically in that region. Then, two leads were conjoined for co-insertion into the positive pole of the stimulator and placed bilaterally at the rostral aspect of the exposed L6-S1 trunk of the spinal cord. Two additional lead were conjoined for insertion into the negative pole of the stimulation device and placed bilaterally 0.5-1.0 cm caudal to the positive poles. The trunks were covered with mineral oil and the wares were anchored into place on midline tissues using tissue adhesive. The back skm was carefully closed with wound clips. Animals were mounted in Bailman cages to ensure that the bladder catheter was free to move during filling and voiding. A heat lamp was placed nearby to maintain body temperature. The bladder catheters were hooked up to infusion pumps and pressure transducers by 4-way stopcock and infusion was begun for a one hour recovery/accommodation period at a flow rate of 0.1 ml/min. Following controlled and continuous cystometry', the bladder was emptied and single cystometrograms were performed until a stable baseline True Bladder Capacity was established.

[0141] FIGS. 9A and 9B are graphs 160 and 162 showing example bladder volumes for neurostimulation delivered during different phases of a bladder fill cycle for the experimental rats. As shown in graph 160 of FIG. 9A, sacral neurostimulation (SNS) (i.e., a form of neurostimulation delivered to sacral nerves) was delivered at the onset of bladder filling for 25%, 50%, 75%, or 100% of the previous control bladder filling cycle duration (n=10). There were no significant differences in pre-SNS baseline control bladder capacities as shown in graphs 164 and 166 of FIGS. 10A and 10B, respectfully. Graphs 164 and 166 of FIGS. 10A and 10B show that consecutive SNS applications did not significantly alter return to baseline conditions and that the effects seen with SNS were not a reflection of alterations in baseline control values. In other words, subsequent changes in bladder capacity' should be the result of stimulation delivery' at the respective time.

[0142] Graph 160 of FIG. 9 A shows that under conditions of increasing SNS duration, the positive effect of SNS on bladder capacity is maximal when the stimulation was delivered for 75-100% of the fill cycle. Graph 160 may suggest that SNS stimulation should be delivered for at least 75% duration of fill time or that SNS is maximally effective during the final 50% of the filling cycle. Graph 162 of FIG. 9B shows the variation in samples for each stimulation duration that begins when the fill cycle begins (e.g , upon completion of the previous voiding event). In this manner, SNS durations of both 75% and 100% of control fill times resulted in significantly larger bladder capacities than control.

[0143] FIGS. 11 A and 1 IB are graphs showing example bladder volumes for

neurostimulation delivered during different phases of a bladder fill cycle. Graphs 168 and 170 show the results of an experiment in which SNS delivered to discrete and different regions of the fill cycle was evaluated for any effects on the bladder capacity. SNS was applied during the first 25%, the second 25%, the third 25%, the fourth 25% and the first and second 50% of control fill times in random (all periods randomized in order of application) or pseudorandom order (all of the 25% randomized in order, followed by the 50%s randomized in order). No differences in the randomization approaches were found.

[0144] The results of graph 168 indicate that the maximal effect of SNS is achieved when stimulation is applied during the last half or last quarter of the filling cycle. As shown in graph 168, the bladder capacity for the last 75% of the fill cycle (bar 25-4) appears to be the most effective time to deliver stimulation during the fill cycle. The remaining earlier periods of the fill cycle may go unstimulated such that a system may withhold stimulation during the early periods of the fill cycle and still achieve effective treatment for incontinence. Similarly, the last 50% of the fill cycle (bar 50-2) may be an appropriate time for delivery of stimulation with the first 50% remaining unstimulated. Graph 170 shows the variation in bladder capacities due to SNS for all samples in the experiment.

[0145] As shown m FIGS. 12A and 12B, graphs 172 and 174 show ^r example bladder volumes prior to neurostimulation delivery as the baseline control bladder capacities prior to each successive SNS application. No statistical significance was detected by Friedman Test.

Therefore, consecutive SNS applications did not significantly alter return to baseline conditions and that the effects seen with SNS were not a reflection of alterations in baseline control values. The more effective stimulation phases at the end of the fill cycle appear to be the result of stimulation at those times timed during the fill cycle.

[0146] Experimental testing was also carried out on female sheep by delivering closed-loop sacral neuromodulation. An experimental system general-purpose neuromodulation system was used. FIG. 13 is a conceptual diagram of the experimental general-purpose neuromodulation system. The implantable component of the system 200 was a rechargeable stimulator 202 that supported up to four stimulation leads and could deliver programmable independent electrode control across 16 channels. Rechargeable stimulator 202 was also capable of sensing two different signals from the body, including measuring up to four channels of electrical biopotentials (both in time and frequency domain) from the implanted electrodes and measuring the mertial movement of the device within the body with an on-board accelerometer.

[0147] In the experimental testing, the sensed data was streamed to an external PC over distance telemetry using a collection of external instruments in the system. The system 200 included telemetr -based instruments for both communication and recharging the implan ted device (not shown for simplicity). The system 200 included an application programming interface (API) 222 that allows development of custom applications. API 222 may allow a researcher to design experiments that follow _'’ approved protocols. API 222 may also allow application developers to integrate into their experiments third party sensors that may be used during testing to further refine the delivered therapy and provide additional test data.

[0148] As the sensing site for bladder pressure was distant from the site of stimulation in the testing, the physiological sensor data was provided from a separate physiological sensor 204 implanted at a different location. The physiological sensor 204 was designed to wirelessly transmit bladder pressure data to an external classification system on the external PC which included a configurable classifier device 210 (similar to processor circuitry 53 executing classifier algorithm 34) that could be used to detect when a void had occurred. Classifier device 210 then could transmit that a void occurred to control policy device 220 (similar to processor circuitry 53 executing control policy 36) which then turned off stimulation for a pre-defined time. In the testing, the predetermined time w¾s half of the bladder filling cycle. After the pre defined time elapses, stimulation w¾s turned back on. The overall result w¾s that data from the physiological sensor was used to control rechargeable stimulator 202 by turning therapy on and off, or adjusting stimulation parameters (e.g , amplitude, frequency or pulse width).

[0149] While system 200 makes use of an external communication link for data processing, various examples are directed toward two implanted devices (e.g., IMD 16 and a sensor external to IMD 16, but implanted in patient 14) communicating directly with each other directly through either a wired or wireless data communication channel. This may reduce or eliminate the need for external components and may be more practical for human translation.

[0150] Various examples are directed toward sensing implants that are configured for use with multiple systems and therapy solutions. For instance, the sensing implants may be designed with standardized communication protocols that are used across several different systems.

[0151] The experimental testing was carried out for bladder function, but it is recognized that similar solutions and systems may be applied to various urinary, bowel, or pelvic floor disorder for which a relevant physiological event may be sensed. A few examples of physiological signals include pressure, volume, EMG, EKG, nerve/electrical activity, movement/motion, impedance, body temperature, blood pressure, blood or urine flow _' -, or location. A system may perform acute or chronic data collection from an appropriate sensor which then would communicate using distributed methods to a standardized stimulator providing therapy at potentially a variety of distant neuromodulation targets, for example IMD 16.

[0152] FIG. 14 is a graph showing example bladder volumes changes based on

neurostimulation delivered during different phases of a bladder fill cycle in the experiment involving female sheep. Graph 176 shows the results of an experiment in which SNS delivered to discrete and different regions of the bladder fill cycle was evaluated for any effects on the bladder capacity, or the amount of urine that could be stored by the bladder during a single balder fill cycle. The results of graph 176 w/ere derived from eleven trials with four fully conscious, female sheep implanted with an IMD. In each trial, fluid was added to the bladder at a rate of 15 milliliters (mL) until the animal voided the contents of the bladder. The total amount of fluid that had been added prior to voiding was recorded for each condition. Three baseline bladder fill cycles were conducted to establish a mean bladder filling time, at which time subsequent and separate bladder fill cycles were performed m which neuromodulation was delivered for the first half of the bladder fill cycle, the second half of the bladder fill cycle, and the full period of the bladder fill cycle, respectively. Each bar in graph 176 is shown with respective error parts.

[0153] As shown in graph 176, baseline bladder fill cycles resulted in approximately 40 mL of volume held prior to voiding when no neurostimulation was delivered during the bladder cycle. When neurostimulation was delivered during only the first half of the bladder fill cycle (i.e , no neurostimulation was delivered during the second half of the bladder fill cycle), the fill volume increased to approximately 70 mL. However, this volume increase was not statistically significant compared to the baseline. When neurostimulation was delivered only during the second half of the bladder fill cycle (i.e., no neurostimulation during the first half), the total fill volume was increased a statistically significant amount over the baseline volume to

approximately 100 mL. Similar increases to bladder volume over baseline were observed during neurostimulation delivery during the entire, or full, bladder fill cycle. Therefore, these results indicate that neurostimulation may be delivered during the second half, or a second phase, of the bladder fill cycle instead of the first half, or first phase of the bladder fill cycle. Withholding stimulation during the first phase to avoid continual stimulation delivery may increase long term efficacy and prevent accommodation and/or muscle fatigue while maintaining effective therapy. [0154] FIGS. 15A and 15B are graphs showing example physiological events during a physiological cycle and predictive delivery of neurostimulation to avoid a dysfunctional state of the physiological cycle. Instead of electrical and pharmaceutical based neuromodulation therapies being delivered continuously, these therapies may be delivered in response to detection of a dysfunctional state. As shown in FIG 15 A, event line 190 may show bladder contractions, as one example, during two bladder fill cycles wherein the bladder contractions occur during tune periods B1 and Cl . The bladder contractions exceed a therapy threshold to arrive at a dysfunctional state (e.g., overactive bladder that may result in incontinence) that may benefit from stimulation during periods B1 and Cl. However, this stimulation is too late in time to prevent the dysfunctional state.

[0155] Instead of waiting to detect the dysfunction before delivering stimulation, detection of a physiological marker A prior to the dysfunction may allow a system to predict when stimulation should be delivered prior to the dysfunction occurring. Marker A may be a detectable physiological event, such as a voiding event, a bladder fill level, or minor detrusor contractions for the case of incontinence. As shown in FIG 15B, event line 192 also includes marker A. However, upon detecting the physiological marker A, the system may track a first phase 194 A during which stimulation therapy is withheld from the patient. Upon expiration of the first phase 194A, the system may deliver neurostimulation during the second phase B2 in anticipation of the dysfuncti onal state. Neurostimulation during C2 of the next fill cycle may have a similar effect. In this manner, the system may refrain from delivering stimulation when it is unnecessary for therapy and possibly prevent the dysfunctional state from occurring by pre treating the organ or related tissue. In addition, preemptively delivering stimulation may reduce the time required for stimulation delivery because the stimulation does not need to counteract the dysfunction that has already occurred. Timing delivery of neurostimulation to one or more physiological markers, instead of waiting to detect a dysfunction, may improve efficacy, reduce side effects from shorting stimulation durations, reduce accommodation, and improve battery life or drug fill intervals.

[0156] In the example of urinary bladder applications, urinary physiology is governed by- phases of bladder filling and urine storage (e.g., a filling phase) coupled with phases of bladder contractions associated with urine release (e.g., a voiding phase). Additionally, these two broad phases may also be subdivided into at least three discrete subphases, as the voiding phase is composed of void initiation, void maintenance and void termination phases. Filling, in turn, is composed of at least three subphases including fill initiation, fill maintenance, and fill termination. Each sub phase is associated with specific neural sensory inputs and motor outputs that establish a complete functional network. By timing neurostimulation onto these specific subphases of function, the system and method may achieve an improved network function according to relative physiological and temporal signals within and associated with these subphases. Similar functions for other organ systems may also be described according to physiological markers and temporal relationships with those markers with phases and subphases of activity or states.

[0157] In some examples, one or more non- voiding (e.g., no urination expulsion) bladder contractions associated with phases of bladder filling and non- voiding bladder contractions may be used as physiological markers indicative of sensations of urgency. Three types of non- voiding bladder contractions may include type I, type II, and type III contractions. Type I contractions include base-to-dome in propagation, small magnitude, contractions that typically lead to bladder volume accommodation. Type I contractions are associated with normal filling and accommodation of the bladder (increases in volume) as the bladder fills with urine. Type II contractions are dome-to-base in propagation, larger magnitude, reverse contractions that are observed during high bladder volumes and pressures or during certain types of bladder irritation (e.g., disease state). Type II contractions (or dysfunctions of these) are associated with fullness (urge) and potentially bladder or pelvic pain. The presence of Type II contractions at lower bladder volumes may be associated with a variety of idiopathic urge frequency and/or urge incontinence conditions. It also possible that bladder pain syndrome could include an increase in Type II contractions. Type III contractions are similar to Type I contractions in originating from base-to-dome, but Type III contractions invade into the dome and return towards base. These contractions are larger in magnitude and have been primarily observed in older rats, for example. These contractions (or dysfunctions of these) might be associated with the increase in urge and frequency symptoms common in older patients.

[0158] These types of bladder contractions could be identified using either chronic or acute measurement techniques. Acutely, the contractions could be measured in-office via techniques with multi-channel pressure catheters, or a variety of imaging techniques from endoscopic video recording, ultrasound and/or functional M Rf Filling cystometry could be used to help quickly identify different types of contractions present in a patient for diagnostic purposes as well as therapy selection for an individual patient. Chronically, techniques for recording non-voiding contractions could include focal EMG, chemical sensors, or local pressure or mechanical sensors within the bladder or outside the bladder wall. These long-term recording methods could help identify different contraction types and changes in frequency or amplitude after a therapy has been applied.

[0159] Distinguishing these different types of bladder contractions (especially Types II &

III) by various methods could allow for differential diagnosis and subsequent efficacious respective therapies based upon these specific dysfunctions. For example, detection of Type II and/or Type III contractions would also serve as a physiological marker for idiopathic overactive bladder and potentially differentiate bladder specific (e.g., bladder muscle or neural input) disease from generalized disorders (e.g., anxiety, depression) that may negatively impact urinary behavior. In this manner the system may use the different contractions as a marker to time subsequent therapy and reduce dysfunctional events such as overactive bladder.

[0160] Specific urinary incontinence features may include using patient sensation of urination, patient sensation of emptiness, and/or of pam or bladder fullness as a signal relating to bladder filling phase. Patient reports (e.g., user input) of urination or reports of urine leakage may also be used to detect voiding phase or abnormal voiding, respectively. This user input includes one or more signals from which physiological markers may be identified. In this manner, leakage, or a leakage event, is different from a voiding event. In the ease of the bladder, leakage refers to an amount of urine that exits the bladder that is less than a full emptying of the bladder, whereas full emptying (e.g., emptying until the bladder is empty or substantially empty where urine no longer exits the bladder) is considered a voiding event or void event. For example, leakage may refer to a small amount of urine that exits the bladder and then stops such that the remainder of urine is still retained within the bladder. Leakage events may result in abnormal voiding and an unstable or abnormal fill cycle. In some examples, the system may ignore identified leakage events when determining a voiding event that stops and starts a fill cycle. In other examples, the system may characterize the fill cycle as an unstable fill cycle in response to identifying a leakage event and treat the fill cycle differently than a fill cycle that is normal and without any identified leakage events. [0161] Patient actuation (e.g., user input) of specific sensations or events recorded to a device may be used to indicate specific phases within filling or voiding. Therapy timing systems may be linked to the patient-indicated signals so that therapy delivery may be partitioned and delivered appropriately relative to the indicated signals. In addition, automated and objective sensing of bladder physiological markers may be utilized to link with urinary phase, such as pressure signals and large rapid changes in bladder pressure that indicate urination events or linked with urge (i.e., a patient sensation of a need to void). Bladder movement (detected by accelerometers, piezoelectric sensors, or similar sensors) may be used as physiological markers of voiding or instability of bladder filling. Pressure spectra of the bladder may be used to identify dysfunction of normal filling. The system may utilize bladder or urethral pressure, patterns or pressure spectra, or external bladder pressures detected near the bladder. Other physiological signals include nerve activity, urine flow and EMG activities from the internal or external urethral sphincter or detrusor muscle.

[0162] A variety of physiological markers are described herein and may be used to determine the start, end, and/or progression point within a physiological cycle. Physiological markers may be indicated by an identified event such as voiding, leaking, muscle activation, or other events related to urine or fecal voiding. These events may be detected automatically by implanted or external sensors. For example, a wetness sensor may detect voi ding or leakage external of the patient, a pressure sensor may detect bladder pressure and/or sphincter pressure via an implanted or external position, or electrodes may generate an electrogram indicative of pelvic floor muscle activity. In some examples, external and/or implanted electrodes may generate an

electroencephalogram (EEG) from which a physiological marker indicative of pelvic muscle contraction related to one or more events of the physiological cycle. Other physiological markers may be derived from external monitoring of patient behavior, patient activity, or even patient location. For example, individual events or activities related to a bladder fill cycle (e.g., pacing, fidgeting, swaying, or other activities indicative of impending voiding) may be used to identify voiding or impending voiding. As another example, a pattern of events or activities may be used to identify a point in the physiological cycle, such as a period of non-motion

immediately following pacing or fidgeting, which may indicate that voiding is occurring during the non-motion period, or detection of increasing use of legs during pacing, contraction of buttocks or other muscle activity, which may indicate that the bladder fill cycle is approaching the end. In some examples, a sensor (e.g., proximity sensor and/or location sensor) may indicate when the patient is within an area identified as a restroom where the patient typically voids and interpret the presence within this area as the patient is voiding. The programmer or IMD may directly detect these areas or locations (e.g., using global position system (GPS) sensors and/or one or more proximity sensors) or receive indications of the patient location from another sensor or device (e.g., from a smartphone or dedicated presence-sensitive sensor such as a pressure sensor on the seat of a toilet, trigger on the toilet flushing mechanism, or any other such sensor). In other examples, a physiological marker may be detected based on input from a patient such as an indication that the patient has voided, needs to void, or is requesting additional or alternative therapy in order to prevent a voiding event.

[0163] In addition, a single physiological marker may be identified from two or more signals. These signals may be synchronized in time such that the physiological marker is identified when an aspect of each synchronized signal matches a predetermined value or exceeds a predetermined threshold. For example, the physiological marker for a voiding event may be the detection of bladder contraction and wetness from a wetness sensor. As another example, a voiding event may be detected when the patient is detected to be within a restroom and the patient is not moving for a at least a certain period of time that indicates the pati ent has voided.

In this manner, the system may detect a physiological marker by analyzing two or more signals.

[0164] Phases of bladder function may be detected and therapy delivery may be linked to these critical phases. If the normal phases of function are not present, or dysfunctional events are detected, then therapies may be delivered relative to a desired normal function. Timing may be used to appropriately delay stimulation relative to major physiological events or identifiable bladder stages. For example, the system may detect voiding or unstable detrusor contractions and link such events to neurostimulation delivery or withholding. Voiding, whether recorded by patient actuation or physiological recording, may be used by a system to initiate a timer that allows the therapy to be discontinued, or withheld, for a duration of a phase following this phase and then therapy may be reinitiated following this phase. Initiation of therapies may therefore be applied in shorter durations linked to voiding events, with therapy delivery being needed in the latter phase of filling instead of the early phase after a void event, for example.

[0165] It should be noted that system 10, and the techniques described herein, may not be limited to treatment or monitoring of a human patient. In alternative examples, system 10 may be implemented in non-human patients, e.g., primates, canines, equines, pigs, and felines. These other animals may undergo clinical or research therapies that my benefit from the subject matter of this disclosure.

[0166] The techniques of this disclosure may be implemented in a wide variety of computing devices, medical devices, or any combination thereof. Any of the described units, circuitry or components may be implemented together or separately as discrete but interoperable logic devices. Depiction of different features as circuitry or units is intended to highlight different functional aspects and does not necessarily imply that such circuitry or units must be realized by separate hardware or software components. Rather, functionality associated with one or more circuitry or units may be performed by separate hardware or software components, or integrated within common or separate hardware or software components.

[0167] The disclosure contemplates computer-readable storage media comprising

instructions to cause a processor to perform any of the functions and techniques d escribed herein. The computer-readable storage media may take the example form of any volatile, non-volatile, magnetic, optical, or electrical media, such as a RAM, ROM, NVRAM, EEPROM, or flash memory that is tangible. The computer-readable storage media may be referred to as non- transitory. A server, client computing device, or any other computing device may also contain a more portable removable memory type to enable easy data transfer or offline data analysis.

[0168] The techniques described in this disclosure, including those attributed to various circuitry and various constituent components, may be implemented, at least in part, in hardware, software, firmware or any combination thereof. For example, various aspects of the techniques may be implemented within one or more processors, including one or more microprocessors, DSPs, ASICs, FPGAs, or any other equivalent integrated, discrete logic circuitry, or other processing circuitry, as well as any combinations of such components, remote servers, remote client devices, or other devices. The term“processor circuitry” or“processing circuitry” may generally refer to any of the foregoing logic circuitry, alone or in combination with other logic circuitry, or any other equivalent circuitry.

[0169] Such hardware, software, firmware may be implemented within the same device or withm separate devices to support the various operations and functions described in this disclosure. In addition, any of the described units, circuitry or components may be implemented together or separately as discrete but interoperable logic devices. Depiction of different features as circuitry or units is intended to highlight different functional aspects and does not necessarily imply that such circuitry or units must be realized by separate hardware or software components. Rather, functionality associated with one or more circuitry or units may be performed by separate hardware or software components, or integrated within common or separate hardware or software components. For example, any circuitry described herein may include electrical circuitry configured to perform the features attributed to that particular circuitry, such as fixed function processing circuitry, programmable processing circuitry, or combinations thereof.

[0170] The techniques described in this disclosure may also be embodied or encoded in an article of manufacture including a computer-readable storage medium encoded with instructions. Instructions embedded or encoded m an article of manufacture including a computer-readable storage medium encoded, may cause one or more programmable processors, or other processors, to implement one or more of the techniques described herein, such as when instructions included or encoded m the computer-readable storage medium are executed by the one or more processors. Example computer-readable storage media may include random access memory (RAM), read only memory (RDM), programmable read only memory (PROM), erasable programmable read only memory (EPROM), electronically erasable programmable read only memory' (EEPROM), flash memory', a hard disk, a compact disc ROM (CD-ROM), a floppy disk, a cassette, magnetic media, optical media, or any other computer readable storage devices or tangible computer readable media. The computer-readable storage medium may also be referred to as storage devices.

[0171] In some examples, a computer-readable storage medium comprises non- transitory medium. The term“non-transitory” may indicate that the storage medium is not embodied in a carrier wave or a propagated signal. In certain examples, a non-transitory storage medium may store data that may, over time, change (e.g., in RAM or cache).

[0172] The following examples are disclosed.

[0173] Example 1 . An adaptive system for providing electrical stimulation to a patient, the system comprising: memory configured to store one or more programs; and processor circuitry coupled to the memory and configured to execute the one or more programs, the processor circuitry being configured to: monitor a sensor signal and classify a physiological marker of the patient based upon the sensor signal, the physiological marker indicative of a phase of a physiological cycle; generate a control signal based on the classified physiological marker of the patient, wherein the control signal controls an implantable stimulation device to provide the electrical stimulation, in accordance with one or more stimulation parameters of a stimulation program, at a target site within the patient to one more of: enable a sphincter to open to allow a voiding event or constrict the sphincter to inhibit the voiding event; adapt one or more of a manner in which the processor circuitry classifies the physiological markers, generates the control signal, or the one or more stimulation parameters of the stimulation program based on the classified physiological markers or patient input so as to automatically adjust one or more of timing of the delivery of the electrical stimulation or the stimulation parameters of the electrical stimulation.

[0174] Example 2. The system of example 1, wherein the one or more programs comprises a classifier program and wherein the processor circuitry is configured to execute the classifier program to monitor the sensor signal and classify the physiological marker of the patient.

[0175] Example 3. The system of any of examples 1 or 2, wherein the one or more programs comprises a control policy and wherein the processor circuitry is configured to execute the control policy to generate the control signal

[0176] Example 4. The system of example 3, wherein the control policy comprises an adaptable baseline control policy, the adaptable baseline control policy initially being configured to control the implantable stimulation device to provide the electrical stimulation continually.

[0177] Example 5. The system of any combination of examples 1-4, wherein the one or more programs comprises a machine learning algorithm and wherein the processor circuitry is configured to execute the machine learning algorithm to adapt one or more of the manner in which the processor circuitry classifies the physiological markers, generates the control signal, or the one or more stimulation parameters of the stimulation program.

[0178] Example 6. The system of any combination of examples 1 -5, wherein the processor circuitry is further configured to withhold electrical stimulation during a first phase of the physiological cycle based upon timing of the electrical stimulation signal.

[0179] Example 7. The system of example 6, wherein the processor circuitry is further configured to generate the control signal so as to provide the electrical stimulation during a second phase of the physiological cycle based upon timing of the electrical stimulation signal.

[0180] Example 8. The system of example 7, wherein the processor circuitry is further configured to: determine the first phase of the physiological cycle and the second phase of the physiological cycle based at least in part upon the physiological marker; and automatically adjust the timing of the electrical stimulation signal so as to withhold electrical stimulation during the first phase of the physiological cycle and provide the electrical stimulation during the second phase of the physiological cycle.

[0181] Example 9. The system of example 8, wherein the processor circuitry is further configured to: determine the first phase of the physiological cycle and the second phase of the physiological cycle based at least m part upon the physiological marker and a duration of one or more previous physiological cycles; and automatically adjust the timing of the electrical stimulation signal so as to withhold electrical stimulation during the first phase of the physiological cycle and provide the electrical stimulation during the second phase of the physiological cycle.

[0182] Example 10. The system of any combination of examples 1-9, wherein the physiological marker is associated with bladder voiding.

[0183] Example 11. The system of any combination of examples 1-10, wherein the sensor signal is indicative of at least one of pressure, contraction, and volume.

[0184] Example 12. The system of example 1 1, wherein the sensor signal is indicative of at least one of amount of change in pressure, duration of change in pressure or rate of change in pressure.

[0185] Example 13. A method comprising: monitoring a sensor signal and classifying a physiological marker of the patient based upon the sensor signal, the physiological marker indicative of a phase of a physiological cycle; generating a control signal based on the classified physiological marker of the patient, wherein the control signal controls an implantable stimulation device to provide the electrical stimulation, m accordance with one or more stimulation parameters of a stimulation program, at a target site within the patient to one more of: enable a sphincter to open to allow a voiding event or constrict the sphincter to inhibit the voiding event; adapting one or more of a manner m which processor circuitry classifies the physiological markers, generates the control signal, or the one or more stimulation parameters of the stimulation program based on the classified physiological markers or patient input so as to automatically adjust one or more of timing of the deliver of the electrical stimulation or the stimulation parameters of the electrical stimulation. |Ό186] Example 14. The method of example 13, wherein the monitoring the sensor signal and classifying the physiological marker is performed by processor circuitry executing a classifier program.

[0187] Example 15. The method any of examples 13 or 14, wherein the generating the control signal is performed by processor circuitry executing a control policy.

[0188] Example 16. The method of example 15, wherein the control policy comprises an adaptable baseline control policy, the adaptable baseline control policy initially being configured to control the implantable stimulation device to provide the electrical stimulation continually.

[0189] Example 17. The method of any combination of examples 13-16, wherein the adapting one or more of the manner in which processor circuitry classifies the physiological markers, generates the control signal, or the one or more stimulation parameters of the stimulation program is performed by processor circuitry executing a machine learning algorithm.

[0190] Example 18. The method of any combination of examples 13-17, further comprising withholding electrical stimulation during a first phase of the physiological cycle based upon timing of the electrical stimulation signal.

[0191] Example 19. The method of example 18, further comprising generating the control signal so as to provide the electrical stimulation during a second phase of the physiological cycle based upon timing of the electrical stimulation signal.

[0192] Example 20. The method of example 19, further comprising: determining the first phase of the physiological cycle and the second phase of the physiological cycle based at least in part upon the physiological marker; and automatically adjusting the timing of the electrical stimulation signal so as to withhold electrical stimulation during the first phase of the physiological cycle and provide the electrical stimulation during the second phase of the physiological cycle.

[0193] Example 21. The method of example 20, wherein the determining the first phase of the physiological cycle and the second phase of the physiological cycle is further based upon a duration of one or more previous physiological cycles; and automatically adjusting the timing of the electrical stimulation signal so as to withhold electrical stimulation during the first phase of the physiological cycle and provide the electrical stimulation during the second phase of the physiological cycle. |Ό194] Example 22. The method of any combination of examples 13-21, wherein the physiological marker is associated with bladder voiding.

[0195] Example 23. The method of any combination of examples 13-22, wherein the sensor signal is indicative of at least one of pressure, contraction, and volume.

[0196] Example 24. The method of example 23, wherein the sensor signal is indicative of at least one of amount of change in pressure, duration of change in pressure or rate of change m pressure.

[0197] Example 25. A non-transitory storage medium containing instructions which when executed by one or more processors, cause the one or more processors to: monitor a sensor signal and classify a physiological marker of a patient based upon the sensor signal, the physiological marker indicative of a phase of a physiological cycle; generate a control signal based on the classified physiological marker of the patient, wherein the control signal controls an implantable stimulation device to provide the electrical stimulation, in accordance with one or more stimulation parameters of a stimulation program, at a target site within the patient to one more of: enable a sphincter to open to allow a voiding event or constrict the sphincter to inhibit the voiding event; adapt one or more of a manner in which the processor circuitry' classifies the physiological markers, generates the control signal, or the one or more stimulation parameters of the stimulation program based on the classified physiological markers or patient input so as to automatically adjust one or more of timing of the delivery of the electrical stimulation or the stimulation parameters of the electrical stimulation.

[0198] Various examples have been described herein. Any combination of the described operations or functions is contemplated. These and other examples are within the scope of the following claims. Based upon the above discussion and illustrations, it is recognized that various modifications and changes may be made to the disclosed examples in a manner that does not require strictly adherence to the examples and applications illustrated and described herein. Such modifications do not depart from the true spirit and scope of various aspects of the disclosure, including aspects set forth in the claims.