System and Method for Adaptive Neural Stimulation

Abstract

Methods and systems for providing electrical stimulation to a patient's spinal cord using electrode leads implanted in the patient's spinal column are described. Embodiments involve cycling between durations during which stimulation is actively applied and durations when no stimulation is applied. The stimulation can be configured such that pain relief washes in during the active stimulation duration and continues for some part of the duration when no stimulation is being applied. Eventually the pain relief may wash out. The washout time may be modeled, so that stimulation may be resumed before the pain relief washes out. The stimulation may be below the patient's perception threshold.

Claims

1. A method of providing electrical stimulation to a patient's spinal cord to treat pain in the patient using one or more electrode leads implantable in the patient's spinal column, wherein each electrode lead comprises a plurality of electrodes, the method comprising: providing a patient-specific model that models washout times as a function of stimulation durations, using the patient-specific model to determine an on-time during which stimulation is to be applied and a corresponding off-time during which stimulation is not to be applied, and providing stimulation to the patient according to the determined on- and off-times.

2. The method of claim 1, wherein the determined on time corresponds to a first stimulation duration of the patient specific model.

3. The method of claim 2, wherein the determined off time is less than a washout time modeled for the first duration by the patient-specific model.

4. The method of claim 3, wherein the electrical stimulation is configured to be below the patient's perception threshold and to provide pain relief to the patient.

5. The method of claim 4, wherein the electrical stimulation is configured so that the pain relief washes in in a period of ten seconds or less.

6. The method of claim 3, wherein determining the on and off times further comprises using one or more stimulation parameters to select the on and off times.

7. The method of claim 6, wherein the one or more stimulation parameters comprise one or more of stimulation intensity, stimulation amplitude, and/or stimulation dose.

8. The method of claim 3, wherein determining the on and off times further comprises using one or more parameters determined using one or more sensors selected from the group consisting of accelerometers, motion detectors, heart rate monitors, sleep sensors.

9. The method of claim 3, wherein determining the on and off times further comprises using one or more features of electrical signals recorded at one or more of the electrodes.

10. The method of claim 3, further comprising deriving the model by: (a) providing test stimulation to the patient for a test first duration, (b) turning off the test stimulation at the end of the test first duration, (c) after a test second duration following the turning off the test stimulation, receiving an indication that the patient has turned on the stimulation, (d) increasing the length of the test first duration, (e) repeating steps (a)-(d) a plurality of times to collect a plurality test first durations and test second durations, and (f) modeling a relationship between test first durations and test second durations.

11. A system for providing electrical stimulation to a patient's spinal cord to treat pain in the patient using one or more electrode leads implantable in the patient's spinal column, wherein each electrode lead comprises a plurality of electrodes, the system comprising: a memory programmed with a patient-specific model that models washout times as a function of stimulation durations, and control circuitry configured to use the patient-specific model to determine an on-time during which stimulation is to be applied and a corresponding off-time during which stimulation is not to be applied, and to provide stimulation to the patient according to the determined on- and off-times.

12. The system of claim 11, wherein the determined on time corresponds to a first stimulation duration of the patient specific model.

13. The system of claim 12, wherein the determined off time is less than a washout time modeled for the first duration by the patient-specific model.

14. The system of claim 13, wherein the electrical stimulation is configured to be below the patient's perception threshold and to provide pain relief to the patient.

15. The system of claim 14, wherein the electrical stimulation is configured so that the pain relief washes in in a period of ten seconds or less.

16. The system of claim 13, wherein determining the on and off times further comprises using one or more stimulation parameters to select the on and off times.

17. The system of claim 16, wherein the one or more stimulation parameters comprise one or more of stimulation intensity, stimulation amplitude, and/or stimulation dose.

18. The system of claim 13, wherein determining the on and off times further comprises using one or more parameters determined using one or more sensors selected from the group consisting of accelerometers, motion detectors, heart rate monitors, sleep sensors.

19. The system of claim 13, wherein determining the on and off times further comprises using one or more features of electrical signals recorded at one or more of the electrodes.

20. The system of claim 13, wherein the control circuitry is further configured to derive the model by: (a) providing test stimulation to the patient for a test first duration, (b) turning off the test stimulation at the end of the test first duration, (c) after a test second duration following the turning off the test stimulation, receiving an indication that the patient has turned on the stimulation, (d) increasing the length of the test first duration, (e) repeating steps (a)-(d) a plurality of times to collect a plurality test first durations and test second durations, and (f) modeling a relationship between test first durations and test second durations.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0033] FIG. 1 shows an Implantable Pulse Generator (IPG) useable for Spinal Cord Stimulation (SCS), in accordance with the prior art.

[0034] FIGS. 2A and 2B show examples of stimulation pulses producible by the IPG employing active charge recovery and passive charge recovery respectively, in accordance with the prior art.

[0035] FIG. 3 shows stimulation circuitry used in the IPG to provide stimulation pulses, in accordance with the prior art.

[0036] FIG. 4 shows an External Trial Stimulator (ETS) useable to provide stimulation before implantation of an IPG, in accordance with the prior art.

[0037] FIG. 5 shows various external devices capable of communicating with and programming stimulation in an IPG and ETS, in accordance with the prior art.

[0038] FIG. 6 shows a Graphical User Interface (GUI) of a clinician programmer external device for setting or adjusting stimulation parameters, in accordance with the prior art.

[0039] FIG. 7 shows an embodiment of using a regulating function to modulate stimulation intensity.

[0040] FIG. 8 shows an integrated system for managing aspects of a patient's SCS therapy.

[0041] FIG. 9 shows qualitative relationships between stimulation intensity and stimulation duration.

[0042] FIG. 10 illustrates a patient's pain response to stimulation cycling on and off.

[0043] FIG. 11 illustrates relationships between washout rate and stimulation intensity.

[0044] FIG. 12 shows relationships between washout rate and stimulation duration.

[0045] FIG. 13 shows an algorithm for determining washout time as a function of stimulation duration.

[0046] FIG. 14 shows an embodiment of an algorithm for determining stimulation cycling options based on predictions of washout times informed using a model of washout time as a function of stimulation duration.

DETAILED DESCRIPTION

[0047] Aspects of this disclosure relate to systems and methods that provide and support the delivery of the patient's SCS therapy. One of the goals of such systems is to maintain and adjust the patient's therapy so as to provide effective pain relief and minimal side effects. Another goal may be to minimize energy usage (i.e., minimize battery drain). While SCS therapy can be an effective means of alleviating a patient's pain, such stimulation can also cause paresthesia. Paresthesia is a sensation such as tingling, prickling, heat, cold, etc. that can accompany SCS therapy. Generally, the effects of paresthesia are mild, or at least are not overly concerning to a patient. Moreover, paresthesia is generally a reasonable tradeoff for a patient whose chronic pain has now been brought under control by SCS therapy. Some patients even find paresthesia comfortable and soothing. SCS therapy that causes paresthesia may be referred to as supra-perception therapy. Nonetheless, at least for some patients, SCS therapy may ideally provide complete pain relief without paresthesia-what is often referred to as “sub-perception” or sub-threshold therapy, i.e., therapy that a patient cannot feel. U.S. Patent Publication No. 2021/0299448 (“the '448 Publication”) and the priority applications cited therein describe methods and systems for providing sub-perception stimulation. The contents of the '448 Publication are hereby incorporated herein by reference. According to some embodiments, the sub-perception stimulation is configured to provide pain relief to the patient, wherein the pain relief washes in very quickly, for example, in less than an hour or less than ten minutes, or less than a minute, or less than ten seconds after beginning stimulation. The methods and systems described herein may be used with both supra- and sub-perception therapy.

[0048] When a patient is first implanted with spinal electrode leads, they will typically undergo a “fitting” process with their clinician, whereby the stimulation is calibrated to treat the individual patient's specific pain. The goal of the fitting process is to determine one or more stimulation programs that best treat the patient. As is known in the art, the stimulation programs may define parameters of the stimulation, such as frequency (v), amplitude (A) and pulse-width (PW), waveform shape, etc. Another aspect the fitting procedure is determining the best location along the lead for providing stimulation. By selecting which electrodes are active for supplying (or sinking) current, and how current is fractionated among those electrodes, the center-point of stimulation, as well as the focus of the resulting electric fields in the tissue, can be adjusted to best recruit neural elements in the spinal cord that are correlated to the location (i.e., the dermatome(s)) on the patient's body where the patient feels pain. This process is referred to as “sweet spot searching.” Aspects of sweet spot searching are discussed in U.S. Patent Publication No. 2019/0366104, for example. As discussed in more detail below, the pulse width of the stimulation may also have an impact on the location of the patient's body where the patient will feel paresthesia. This is because the pulse width impacts which fibers are depolarized by the stimulation. For example, when other aspects of the stimulation are equal, increasing the stimulation pulse width tends to depolarize fibers that are deeper within the spinal cord. This can cause the location of paresthesia (i.e., the impacted dermatome) to shift caudally (downwardly on the patient's body).

[0049] Once the patient has undergone fitting, their IPG (of ETS) will be equipped with one or more stimulation programs that have been optimized for their pain. As mentioned above, the programs may specify the stimulation waveforms, pulse patterns, and parameters such as frequency (v), amplitude (A) and pulse-width (PW), etc. However, after the fitting process and during ongoing chronic therapy, the patient may wish to adjust the intensity of their therapy from time to time. Also, as explained further below, it may be beneficial to have periods of time when no stimulation is provided and/or when the stimulation is less than maximum intensity.

[0050] Accordingly, aspects of this disclosure relate to methods and systems for determining how to regulate the intensity of stimulation to achieve certain goals. For example, FIG. 7 illustrates a sequence 702 of four stimulation programs (programs 1-4). According to some embodiments, the clinician and/or the user may define schedule/order of the programs a priori. According to other embodiments, the user may select the schedule/order of the programs using their RC. According to some embodiments, the program sequence and scheduling may be informed based on data recorded using one or more internal or external sensors, as well as other input from the patient. Such sensor data and other input may be components of an integrated system for adaptively tailoring the patient's therapy, such as described in more detail below.

[0051] One aspect of adaptively regulating the patient's therapy may be to use a regulating/modulating function 704 that is combined (e.g., multiplied) with the stimulation program or sequence of programs to regulate/modulate the intensity of the stimulation, as a function of time. The illustrated regulating/modulating function 704 defines an intensity that ramps quickly over a first duration of t1 to t2, plateaus for a second duration of t2 to t3, and then ramps down over a third duration from t3 to t4. Of course, other regulating function shapes may be defined. For example, the regulating function could be a simple square wave, such that the stimulation is on for a certain duration of time and then is turned off for a second duration. The regulating function may be defined by the user and/or the clinician. As with the program schedule/order described above, the regulating function may be defined by the user and/or the clinician, and/or may be informed based on data recorded using one or more internal or external sensors, as well as other input from the patient.

[0052] The patient's therapy 706 at a given time is defined as a mathematical combination of the stimulation program running at that time and the regulating/modulating function. Specifically, the regulating/modulating function mathematically regulates the intensity of stimulation being delivered at a given time. For example, the stimulation program may be multiplied by the regulating/modulating function. Other mathematical combinations of the regulating function and the therapy program can be used, such as convolution, cross-correlation, and more complex time-variant operations in the space and/or time domain. U.S. Patent Publication No. 2020/0346019 describes methods and systems for modulating a stimulation program or sequence of programs using a modulating function.

[0053] FIG. 8 illustrates an embodiment of an integrated system 800 that includes aspects that may be used to derive or inform the scheduling, sequencing, and/or the regulating of the stimulation programs. The system may use various information to build a model 802 to inform how the patient's therapy can best be managed. The model may be built based (in part) on aspects of the patient's therapy that are related to, or measured by, the patient's IPG 10 and/or their RC 45 (which, as explained above, are in communication with each other). For example, the model may consider aspects of the geometry of the electric field (stimulation field), including the center-point of stimulation (CPS), that is created in the patient's tissue as a result of providing stimulation defined by the stimulation program(s). Information about the field geometry may typically be determined during the fitting process, as described above.

[0054] The model may also consider data recorded using the IPG 10. According to some embodiments, the IPG 10 may be configured to sense electrical signals that are present in the patient's tissue. See U.S. Patent Application Publication 2021/0236829 (“the '829 Application) for an example of an IPG comprising circuitry configured to sense such electrical signals. Some electrical signals may include neural signals such as Evoked Compound Action Potentials (ECAPs) that the stimulation evokes in the tissue. An ECAP comprises a cumulative response provided by neural fibers that are recruited by the stimulation, and essentially comprises the sum of the action potentials of recruited neural elements (ganglia or fibers) when they “fire.” Various parameters relating to recorded ECAP signals can be determined, such as parameters relating to the amplitude, frequency, latency, etc. of the ECAP, as explained in the '829 Application. Another electrical signal that may be measured is the stimulation artifact, which comprises a voltage that is formed in the tissue as a result of the stimulation, i.e., as a result of the electric field that the stimulation creates in the tissue. Another electrical signal that may be measured is the electrical impedances that occur at the electrode/tissue interfaces.

[0055] The model may also consider usage logs, which may be provided from the patient's RC 45. The usage logs can provide analytics related to how the patient tends to adjust their therapy. For example, if the patient tends to repeatedly increase their stimulation at a certain point during the day, that may indicate that something characteristic about that time of day that triggers or exacerbates the patient's pain. Usage logs may also indicate which stimulation programs the patient like and dislikes. According to some embodiments, the patient may periodically use the RC to provide rankings of their pain and/or their therapy, for example, using questionnaires.

[0056] The system 800 may also include a clinical database 804, which may comprise a cloud-based data center/server. The clinical database 804 may be configured to exchange information with the patient RC 45, which may be internet connectable. The clinical database 804 may thereby receive RC- and IPG-related data from the RC. According to some embodiments, the model 802 (or aspects thereof) may be kept in the remote center/server. According to other embodiments, the model (or aspects thereof) may be resident in the patient's RC (or IPG).

[0057] The model 802 can also receive data from one or more external sensors as input. Examples of external sensors may include accelerometers (or other motion/activity sensors), heartrate monitors, sleep sensors, and the like. The model may use such data to correlate the patient's therapy to the patient's activity, for example.

[0058] One aspect of the patient's therapy that the model 802 may seek to control/adjust is the dosage and/or the intensity of the patient's therapy. The dosage may be expressed as the amount of charge Q (e.g., expressed in Coulombs) that a patient receives. The stimulation dose (i.e., the intensity) may be expressed as the charge per unit time, (e.g., charge per second (Q/s)). Embodiments of the disclosed methods and system concern determining appropriate dosages and durations for the stimulation provided to the patient.

[0059] FIG. 9 shows a graph 900 illustrating a qualitive relationship between stimulation intensity and stimulation duration. The graph 900 is divided into zones corresponding to different effects that certain combinations of intensity and duration may have on the patient. For example, if the duration of the stimulation is too short (zone 902), then even moderately intense therapy will fail to treat the patient's symptoms. Likewise, if the intensity of therapy is too low (zone 904), the therapy will be ineffective even if the duration is long. Effective therapy can be achieved when both the stimulation intensity and the stimulation duration are sufficient (zone 906). If the stimulation intensity is too high, then overstimulation can occur (zone 908). The stimulation may cause pain or other side effects for the patient.

[0060] Zone 910 is referred to as the bolus area. As used herein, “bolus stimulation” refers to a relatively high stimulation intensity applied for a relatively short duration (typically about 15-30 minutes), after which the stimulation is typically turned off for some length of time. The inventors have found that using one to several boluses of stimulation applied a few times throughout the day (e.g., 4-6 times) may treat the patients' pain even though stimulation is not applied continuously throughout that time.

[0061] FIG. 10 illustrates an example of how a patient's pain may respond to intermittent or bolus stimulation. Curve 1002 illustrates the patient's pain level and curve 1004 illustrates the therapy. Assume that at time t(0) the patient's pain is intense and the stimulation is off. At time t(1) the stimulation is applied, and the patient's pain level quickly drops (hopefully, to zero or close to zero). In other words, the therapy washes in very quickly. In the illustrated embodiment, the stimulation intensity is ramped down over a duration from t(2) to t(4). At some time as the stimulation is ramped down (or following the ramp down), the patient's pain begins returning. In the illustration, the patient's pain begins to return at time t(3). The process of the pain slowly returning once the stimulation is reduced or eliminated is referred to as “washout.” In the illustration the therapy resumes at time t(5) and the patient's pain again drops very quickly.

[0062] Aspects of the disclosure relate to methods and systems for modulating the patient's therapy to account for washout. An aspect of this problem involves predicting the rate at which washout will occur, so that the stimulation can be managed to prevent the patient's pain from returning to a significant level. The inventors have found that the washout time depends on several factors, including (1) the stimulation duration; (2) the stimulation intensity; and (3) the passage of time since the patient first began receiving SCS therapy (for example, the time that has passed since the patient was first implanted with an IPG). Herein, the term “therapy history” is used to refer to the amount of time that has passed since the patient first began receiving SCS therapy. It is important to distinguish between the stimulation duration and the therapy history. The stimulation duration is the length of time that stimulation has been applied essentially without interruption. As used herein, stimulation duration is denoted as t-t(0), or Δt. The stimulation durations shown in FIG. 10 are the durations t(4)-t(1) and t(6)-t(5). The stimulation duration typically comprises multiple pulses of stimulation. The stimulation duration is typically at least a second, and may be many seconds, minutes, hours, days, weeks, or months. Likewise, the off time between the stimulation duration (i.e., the washout time) is also typically at least a second and may be seconds, minutes, hours, days, weeks, or months.

[0063] The stimulation duration should not be confused with the therapy history, which refers to the passage of time since the patient first began receiving SCS therapy. The therapy history is denoted herein as T−T(0), where T(0) is the time (or date) when the patient first began receiving SCS therapy. For example, T(0) may be on or near the day when the patient's IPG was implanted. Therapy history may also be denoted ΔT.

[0064] FIG. 11 illustrates a graph relating washout time (W) on the y-axis as a function of stimulation intensity on the x-axis. Each of the separate plots correspond to different stimulation durations. The intensities are normalized with respect to patient's perception threshold (Pth). In other words, the intensity of 100% Pth is the intensity at which the patient can first perceive paresthesia arising from the stimulation. As mentioned above, the stimulation dose can be expressed as the charge per second (Q/s). The intensity is a product of the frequency (ν), pulse width (PW), and amplitude (A) of the stimulation. In other words, (Q/s)=ν(PW)A.

[0065] Generally, larger stimulation doses recruit more neural fibers, depolarize the fibers to a greater extent and/or for longer periods of time, and generate more action potentials. Also, generally, any of v, PW, and/or A can be manipulated to increase or decrease the stimulation dose Q/s. However, there are caveats. The electrical properties of the patient's biological tissue may impact the extent to which the frequency v may be adjusted while maintaining predictable behavior. Briefly, the patient's tissue can act as a low-pass filter since it comprises substantial amounts of water. Accordingly, different behavior and mechanisms of action may occur above frequencies of about 400-500 Hz. Thus, according to some embodiments, the frequency may be maintained below about 300 Hz. According to some embodiments, the stimulation frequency may be about 60 to about 90 Hz. Also, recall that the pulse width PW of the stimulation may be adjusted during the fitting process to preferentially target specific neural elements. Accordingly, the PW for the patient's stimulation program is typically defined during fitting and typically would not be manipulated for the purpose of adjusting the stimulation intensity for dosing (though it could, according to some embodiments). In other words, generally, the PW may be adjusted to modulate neural targeting and the stimulation amplitude may be adjusted to modulate dosing. All of this is to say that adjustments to the stimulation dose Q/s may most easily be accomplished by adjusting the stimulation amplitude (A) (though any one or more of the parameters could be adjusted).

[0066] Referring again to FIG. 11, notice that the washout time is a function of the stimulation amplitude. That is, higher stimulation intensities lead to longer washout times. Also, the washout time is a function of stimulation duration (Δt), with longer durations yielding longer washout times, even with relatively lower amplitude stimulation. Accordingly, both stimulation intensity and stimulation duration may be modulated to regulate the washout time.

[0067] FIG. 12 shows a graph of washout time (W) on the y-axis as a function of stimulation duration (Δt) on the x-axis. Two curves are shown, each corresponding to different therapy histories (ΔT). The lowermost curve 1202 is for a therapy history that is between 0 and 180 days. In other words, the patient may have began receiving SCS therapy (for example, had their IPG implanted) sometime between 0 and 180 days prior to the collection of the data shown in the FIG. 12. The topmost curve 1204 has a therapy history that is between 90 and 360 days. Assume that all of the measurements were recorded using a stimulation that has an optimum stimulation dose Q(opt)/s. In other words, the stimulation uses an optimum pulse-width PW(opt) and an optimum stimulation amplitude I(opt). As mentioned above, the optimum PW(opt) may be determined for targeting the neural elements for blocking the patient's pain. The optimum stimulation amplitude A(opt) may be determined based on self-dosing (i.e., the amplitude that the patient chooses on their RC) and/or based on any of the measurements discussed above with reference to FIG. 8.

[0068] Notice that the washout time increases as a function of the stimulation duration (Δt). Also notice that the washout times are longer for longer therapy histories (ΔT). The inventors have discovered that, generally, the washout time W is proportional to the stimulation duration (Δt) multiplied by a coefficient a, which is a function of the stimulation dose (Q/s) and the therapy history (ΔT). That is: W∝Δt a(Q/s, ΔT). The fact that the washout time is a function of the therapy history suggests that as the patient undergoes SCS therapy, some amount of neuroplasticity takes place in the patient's spine and/or in their brain. As a result, a patient with a longer therapy history needs less stimulation and/or stimulation less often than they did at earlier therapy history times to achieve the same washout times.

[0069] It will be appreciated that the curves 1202 and 1204 associate a first duration (i.e., the stimulation duration, during which stimulation is provided) with a second duration (i.e., the washout time). Information such as the information contained in the curves 1202 and 1204 may be used to determine aspects of a patient's stimulation therapy, such as determining a regulating function (e.g., function 704, FIG. 7). Since the curves indicate how washout time changes as a function of stimulation duration, they can inform how to cycle stimulation on and off so as to avoid washout (i.e., the return of the patient's pain) before stimulation is resumed. For example, according to some embodiments, the curves can be used to minimize stimulation duration (e.g., to save energy usage) while maximizing washout time. Accordingly, an aspect of the disclosed methods and systems comprises deriving curves (i.e., mathematical models) such as 1202/1204 for the patient, and selecting a stimulation duration using the curves.

[0070] Different locations on the curve can be used to inform the cycle times, depending on various operational goals. According to some embodiments, the stimulation duration (the time when stimulation is active) may be set at a value that corresponds to the maximum inclination of the curve (point 1206) (Note that the values discussed here are only shown on plot 1202. But similar values could be calculated for plot 1204, or whichever plot is appropriate for the patient's therapy history (ΔT)). The maximum inclination can be determined analytically as the point where the slope of the curve 1202 is at a maximum (that is, where the first derivative of the curve is maximum, the second derivative of the curve is 0, and the third derivative is less than 0). Selecting the stimulation duration at the point of maximum inclination of the curve provides the optimum efficiency for the stimulation, i.e., it maximizes the washout time with the minimum stimulation duration. According to some embodiments, if the on time for the stimulation is set for a stimulation duration corresponding to the maximum inclination, then the off time may be set at some fraction of the predicted washout time W(t), to allow a margin of safety. For example, the off time may be set at 75-80% W(t).

[0071] According to some embodiments, the stimulation duration may be set at the value on the curve corresponding to the minimum slope 1208. That duration would require longer duty cycles (perhaps days or weeks) and therefore increased energy usage. But such stimulation durations would provide less variability in washout times and might, therefore, be perceived as safer for the patient. According to some embodiments, if the stimulation duration is set for the minimal slope, then the off time may be set at or very near the time predicted by W(t). Generally, once a curve (such as 1202 or 1240) is determined for a patient, stimulation according to any predetermined point on the curve may be targeted, depending on the therapeutic goals.

[0072] FIG. 13 illustrates and embodiment of an algorithm 1300 for determining a model of washout time as a function of the stimulation duration for a patient. At step 1302 assume that the patient has been using one or more stimulation programs continuously for some time. Assume that the patient is satisfied with the stimulation dose of the programs, denoted here as the optimum stimulation dose (Q(opt)/s). As described above, a clinician may have determined the optimum pulse width PW(opt) and frequency v during a fitting process, so as to target specific neural elements. The patient may be using an optimum stimulation amplitude A(opt) that they may have arrived at by adjusting the stimulation to a comfortable and effective level. The calibration routine begins at step 1304. When the calibration routine begins the optimum stimulation is continued for a first duration−1 second in this case (step 1306). At the end of the duration the system automatically turns off the stimulation (step 1308). According to some embodiments, the system may not inform the patient that the stimulation has ceased. According to some embodiments, the stimulation may only be turned off during the day when the patient is awake. The patient will begin to feel pain and turn their stimulation back on (step 1310) at some point. At step 1312 the system records the duration between when the stimulation was turned off and when the patient turned it back on (i.e., the washout time). At step 1314 the test stimulation duration is increased, and the steps 1308-1314 are repeated until the algorithm reaches a maximum stimulation duration to be tested. The stimulation duration may be increased in any manner. According to some embodiments, the stimulation duration is increases by doubling the previous stimulation duration (i.e., N=1, 2, 4, 8, 16, . . . ). The algorithm 1300 may test a range of stimulation durations up to a maximum stimulation duration. For example, the maximum stimulation may be about a week (about 10.sup.6 seconds). The algorithm interpolates the recorded stimulation durations and washout times and builds the model (step 1316) once the maximum stimulation duration to be tested is reached.

[0073] Note that other methods of determining a relationship between the washout time and the stimulation duration and/or the stimulation intensity may be used. For example, an algorithm could receive a desired washout time as input and then modulate different stimulation durations until the desired washout time is achieved. Alternatively, the algorithm may determine a relationship between stimulation intensity and washout time by modulating one of the values and interpolating its impact on the other. Note that the curves 1202 and 1204 shown in FIG. 12 depend on both the patient's therapy history and on the intensity of the stimulation. Accordingly, for a given therapy history the model may include multiple curves, each corresponding to different stimulation dose. The models may be embodied as curves, estimated functions, analytically derived functions, look-up tables, databases, and the like, which relate washout time (and, accordingly, a time at which the stimulation should be resumed) to the stimulation duration and/or the stimulation dose. The models may be resident in the patient's RC 45, IPG 10, or in a remote data site, such as the clinical database 804.

[0074] FIG. 14 illustrates an embodiment of an algorithm 1400 for using the washout v. stimulation duration model. Assume that at step 1402 the patient is using an optimal therapy program on a continuous basis. At step 1404 the system can automatically characterize a washout v. stimulation duration model (and/or a washout v. stimulation intensity model), for example, using the algorithm 1300 (FIG. 13). Once derived, the model(s) can be used to predict washout times under different conditions (step 1406) and to automatically define cycling options (1408) for the patient's stimulation therapy. Specifically, the models can be used to define when the stimulation should be cycled on to avoid pain arising. For example, the model can be used to define a program sequence 702 and/or a regulating function (704) that modulates/adjusts the stimulation intensity and on/off time (see FIG. 7) to achieve therapeutic goals while optimizing energy consumption. According to some embodiments, in addition to providing an association between the stimulation duration and the washout time, the models can also use additional information and analytics to inform the scheduling/dosing. For example, in the context of a system such as the system 800 (FIG. 8), the model may use information from internal and/or external sensors. For example, when internal and/or external information suggests that the patient is active or otherwise likely to be more susceptible to pain upon washout, the algorithm may weight the stimulation duration more conservatively to prevent the therapy from washing out sooner than expected. For example, the algorithm may select a stimulation duration that is longer and/or that is located on the minimum slope part of the washout v. stimulation duration curve (see point 1208, FIG. 12), as discussed above. Alternatively, when the patient is less likely to be susceptible, the algorithm may configure the stimulation duration to coincide with the maximum inclination part of the washout v. stimulation duration curve (see point 1206, FIG. 12) to provide the most efficient use of energy. According to other embodiments, the algorithm may receive information related to an amount of charge remaining on the IPG's battery and may adjust cycling accordingly. According to some embodiments, the graphs, mathematical functions, look-up tables, databases, or the like, which associate the stimulation duration with an associated washout time may also be indexed according to the further information that the model may use to pick on/off times for cycling the patient's stimulation, such as information values derived from external sensors, internal measurements/recordings, therapy history, and/or stimulation intensity.

[0075] According to some embodiments, the stimulation may be cycled on and off based on the models and algorithms without input from the patient. According to some embodiments, the stimulation programs may provide sub-perception stimulation and the patient may not be aware that their stimulation is cycling on and off. According to other embodiments, the patient may select to cycle their stimulation on and off and/or select a stimulation duration. In such embodiments, the algorithms may determine when to automatically cycle the stimulation on based on the calculated washout time.

[0076] Although particular embodiments of the present invention have been shown and described, it should be understood that the above discussion is not intended to limit the present invention to these embodiments. It will be obvious to those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the present invention. Thus, the present invention is intended to cover alternatives, modifications, and equivalents that may fall within the spirit and scope of the present invention as defined by the claims.