Method and apparatus for estimating neural recruitment

Abstract

An implantable device for estimating neural recruitment arising from a stimulus, has a plurality of electrodes. A stimulus source provides stimuli to be delivered from the electrodes to neural tissue. Measurement circuitry obtains a measurement of a neural signal sensed at the electrodes. A control unit is configured to control application of a selected stimulus to neural tissue using the stimulus electrodes; and after the selected neural stimulus, apply a probe stimulus having a short pulse width. A remnant neural response evoked by the probe stimulus is measured; and the control unit estimates from the remnant neural response a neural recruitment caused by the selected neural stimulus.

Description

TECHNICAL FIELD

The present invention relates to measuring a neural response to a stimulus, and in particular relates to measurement of a compound action potential by using one or more electrodes implanted proximal to the neural pathway, in order to estimate neural recruitment resulting from an applied stimuli.

BACKGROUND OF THE INVENTION

There are a range of situations in which it is desirable to apply neural stimuli in order to give rise to a compound action potential (CAP). For example, neuromodulation is used to treat a variety of disorders including chronic pain, Parkinson's disease, and migraine. A neuromodulation system applies an electrical pulse to tissue in order to generate a therapeutic effect. When used to relieve chronic pain, the electrical pulse is applied to the dorsal column (DC) of the spinal cord. Such a system typically comprises an implanted electrical pulse generator, and a power source such as a battery that may be rechargeable by transcutaneous inductive transfer. An electrode array is connected to the pulse generator, and is positioned in the dorsal epidural space above the dorsal column. An electrical pulse applied to the dorsal column by an electrode causes the depolarisation of neurons, and generation of propagating action potentials. The fibres being stimulated in this way inhibit the transmission of pain from that segment in the spinal cord to the brain. To sustain the pain relief effects, stimuli are applied substantially continuously, for example at 100 Hz.

While the clinical effect of spinal cord stimulation (SCS) is well established, the precise mechanisms involved are poorly understood. The DC is the target of the electrical stimulation, as it contains the afferent Aβ fibres of interest. Aβ fibres mediate sensations of touch, vibration and pressure from the skin, and are thickly myelinated mechanoreceptors that respond to non-noxious stimuli. The prevailing view is that SCS stimulates only a small number of Aβ fibres in the DC. The pain relief mechanisms of SCS are thought to include evoked antidromic activity of Aβ fibres having an inhibitory effect, and evoked orthodromic activity of Aβ fibres playing a role in pain suppression. It is also thought that SCS recruits Aβ nerve fibres primarily in the DC, with antidromic propagation of the evoked response from the DC into the dorsal horn thought to synapse to wide dynamic range neurons in an inhibitory manner.

Neuromodulation may also be used to stimulate efferent fibres, for example to induce motor functions. In general, the electrical stimulus generated in a neuromodulation system triggers a neural action potential which then has either an inhibitory or excitatory effect. Inhibitory effects can be used to modulate an undesired process such as the transmission of pain, or to cause a desired effect such as the contraction of a muscle.

The action potentials generated among a large number of fibres sum to form a compound action potential (CAP). The CAP is the sum of responses from a large number of single fibre action potentials. The CAP recorded is the result of a large number of different fibres depolarising.

The propagation velocity is determined largely by the fibre diameter and for large myelinated fibres as found in the dorsal root entry zone (DREZ) and nearby dorsal column the velocity can be over 60 ms⁻¹. The CAP generated from the firing of a group of similar fibres is measured as a positive peak potential P1, then a negative peak N1, followed by a second positive peak P2. This is caused by the region of activation passing the recording electrode as the action potentials propagate along the individual fibres. An observed CAP signal will typically have a maximum amplitude in the range of microvolts, whereas a stimulus applied to evoke the CAP is typically several volts.

To resolve a 10 μV SCP with 1 μV resolution in the presence of an input 5V stimulus, for example, requires an amplifier with a dynamic range of 134 dB, which is impractical in implant systems. As the neural response can be contemporaneous with the stimulus and/or the stimulus artefact, CAP measurements are difficult to obtain. This is particularly so for pain relief where patients typically obtain best effects with a pulse width in the range of 100-500 μs which ensures much of the neural response occurs while the stimulus is still ongoing, making measurement of the neural response effectively impossible.

For effective and comfortable operation, it is necessary to maintain stimuli amplitude or delivered charge above a recruitment threshold, below which a stimulus will fail to recruit any neural response. It is also necessary to apply stimuli which are below a comfort threshold, above which uncomfortable or painful percepts arise due to increasing recruitment of Aδ fibres which are thinly myelinated sensory nerve fibres associated with acute pain, cold and pressure sensation. In almost all neuromodulation applications, a single class of fibre response is desired, but the stimulus waveforms employed can recruit other classes of fibres which cause unwanted side effects, such as muscle contraction if motor fibres are recruited. The task of maintaining appropriate stimulus amplitude is made more difficult by electrode migration and/or postural changes of the implant recipient, either of which can significantly alter the neural recruitment arising from a given stimulus, depending on whether the stimulus is applied before or after the change in electrode position or user posture. Postural changes alone can cause a comfortable and effective stimulus regime to become either ineffectual or painful.

Another control problem, faced by neuromodulation systems of all types, is achieving neural recruitment at a sufficient level required for therapeutic effect, but at minimal expenditure of energy. The power consumption of the stimulation paradigm has a direct effect on battery requirements which in turn affects the device's physical size and lifetime. For rechargeable systems, increased power consumption results in more frequent charging and, given that batteries only permit a limited number of charging cycles, ultimately this reduces the lifetime of the device.

Any discussion of documents, acts, materials, devices, articles or the like which has been included in the present specification is solely for the purpose of providing a context for the present invention. It is not to be taken as an admission that any or all of these matters form part of the prior art base or were common general knowledge in the field relevant to the present invention as it existed before the priority date of each claim of this application. Throughout this specification the word “comprise”, or variations such as “comprises” or “comprising”, will be understood to imply the inclusion of a stated element, integer or step, or group of elements, integers or steps, but not the exclusion of any other element, integer or step, or group of elements, integers or steps.

SUMMARY OF THE INVENTION

According to a first aspect the present invention provides a method of estimating neural recruitment arising from a selected neural stimulus, the method comprising:

• • applying the selected neural stimulus;
  • after the selected neural stimulus, applying a probe stimulus having a short pulse width;
  • measuring a remnant neural response evoked by the probe stimulus; and
  • from the remnant neural response, estimating neural recruitment caused by the selected neural stimulus.

According to a second aspect the present invention provides an implantable device for estimating neural recruitment arising from a selected neural stimulus, the device comprising:

• • a plurality of electrodes including one or more nominal stimulus electrodes and one or more nominal sense electrodes;
  • a stimulus source for providing stimuli to be delivered from the one or more stimulus electrodes to neural tissue;
  • measurement circuitry for obtaining a measurement of a neural signal sensed at the one or more sense electrodes; and
  • a control unit configured to control application of a selected stimulus to neural tissue using the stimulus electrodes; the control unit further configured to, after the selected neural stimulus, apply a probe stimulus having a short pulse width; the control unit further configured to measure a remnant neural response evoked by the probe stimulus; and the control unit further configured to estimate from the remnant neural response a neural recruitment caused by the selected neural stimulus.

The present invention thus provides for probing of an un-recruited fibre population which was not recruited by the selected stimulus, by reference to which an understanding of the population recruited by the selected stimulus can be obtained.

Embodiments of the invention may be particularly beneficial in providing for estimation of neural recruitment effected by a selected stimulus having a long pulse width, for example in the range of 100-500 μs, in relation to which it is not possible to directly measure a neural response due to temporal overlap of the stimulus and response.

In preferred embodiments, the probe stimulus is applied quickly after the selected stimulus, within the refractory period of the fibres recruited by the selected stimulus.

In some embodiments, a second probe stimulus is applied after the refractory period of fibres recruited by either the selected stimulus or the probe stimulus, and a second measure of evoked neural response is obtained as caused by the second probe stimulus. In such embodiments, the neural recruitment arising from the selected neural stimulus may be estimated by comparing the remnant neural response to the second measure.

Additionally or alternatively, some embodiments may comprise:

• • a) applying the probe stimulus at a time t after conclusion of the selected stimulus;
  • b) obtaining a measure of a remnant neural response arising from the probe stimulus;
  • c) changing t; and
  • d) repeating (a), (b) and (c) to determine variations in the remnant neural response measure, with varying t.

For example, with increasing t an increase in the remnant neural response may indicate the refractory period of the fibre population recruited by the selected stimulus.

In embodiments of the invention in which an estimate of refractory period is obtained, the refractory period may be monitored over time in order to diagnose onset or progression of a disease.

According to another aspect the present invention provides a computer program product comprising computer program code means to make a computer execute a procedure for estimating neural recruitment arising from a selected neural stimulus, the computer program product comprising computer program code means for carrying out the method of the first aspect.

According to a first aspect the present invention provides a method for measuring a neural response to a stimulus, the method comprising:

• • settling measurement circuitry prior to a stimulus, by connecting a sense electrode to the measurement circuitry to allow the measurement circuitry to settle towards a bio-electrically defined steady state;
  • recovering charge on stimulus electrodes by short circuiting the stimulus electrodes to each other;
  • applying an electrical stimulus from the stimulus electrodes to neural tissue, while keeping the sense electrode disconnected from the measurement circuitry;
  • imposing a delay during which the stimulus electrodes are open circuited and the sense electrode is disconnected from the measurement circuitry and from the stimulus electrodes; and
  • after the delay, measuring a neural response signal present at the sense electrode by connecting the sense electrode to the measurement circuitry.

According to a second aspect the present invention provides an implantable device for measuring a neural response to a stimulus, the device comprising:

• • a plurality of electrodes including one or more nominal stimulus electrodes and one or more nominal sense electrodes;
  • a stimulus source for providing a stimulus to be delivered from the one or more stimulus electrodes to neural tissue;
  • measurement circuitry for amplifying a neural signal sensed at the one or more sense electrodes; and
  • a control unit configured to control application of a stimulus to the neural tissue and measurement of an evoked neural response, the control unit configured to settle the measurement circuitry prior to a stimulus by connecting the or each sense electrode to the measurement circuitry to allow the measurement circuitry to settle towards a bioelectrically defined steady state, the control unit further configured to recover charge on the stimulus electrodes by short circuiting the stimulus electrodes to each other, the control unit further configured to cause the stimulus source to apply an electrical stimulus from the stimulus electrodes to neural tissue while keeping the or each sense electrode disconnected from the measurement circuitry, the control unit further configured to impose a delay during which the stimulus electrodes are open circuited and the sense electrode is disconnected from the measurement circuitry and from the stimulus electrodes, and the control unit further configured to measure a neural response signal present at the sense electrode by connecting the or each sense electrode to the measurement circuitry after the delay.

It is to be understood herein that open circuiting of an electrode involves ensuring that the electrode is disconnected from other electrodes, the stimulus source, the measurement circuitry and from voltage rails. Ensuring that the sense electrode is disconnected from the stimulus electrodes during the delay period avoids charge transfer onto the sense electrode(s) and associated artefact. The present invention recognizes that connecting the sense electrodes to the stimulus electrodes during a post-stimulus delay period can undesirably give rise to such charge transfer and associated artefact, particularly if the delay is short relative to the time constant of the stimulus electrodes, the latter typically being around 100 μs. The sense electrode is preferably open circuited during the post-stimulus delay so as to be disconnected from all other electrodes of the array, to prevent such charge transfer to the sense electrode from other non-stimulus electrodes.

The present invention recognizes that it is beneficial to provide for pre-stimulus settling of the measurement circuitry towards a bio-electrically defined steady state. This ensures that charge recovery occurs in the settling stage prior to the stimulus and not during or immediately after the stimulus and thus does not give rise to artefact during or immediately after the stimulus. Where repeated measurement cycles are undertaken, the present invention further permits the measurement amplifier to accumulate a bias point over multiple cycles rather than re-setting the bias point each cycle. The settle period is preferably sufficiently long to permit the electrodes and circuitry to reach an equilibrium, and for example the settle period may be around 1 ms or greater, as permitted by a stimulus rate. For example if therapeutic stimuli are applied to a dorsal column at about 100 Hz and do not give rise to a slow neural response, then after the approximately 2 ms duration of an evoked fast response up to about 8 ms would be available for the settling period. However, this is generally longer than required and the settling period may be substantially less than 8 ms.

The delay may be in the range of substantially zero to 1 ms, and for example may be about 0.3 ms. Such embodiments permit onset of the neural response to be observed, this typically occurring about 0.3 ms after the stimulus for an electrode 3 cm away from the stimulus site. In embodiments in which an amplifier of the measurement circuitry has a very high dynamic range, the delay may be set to a smaller value. The delay is preferably set to a value which ensures the measurement amplifier is not saturated and therefore performs linearly at all times when connected without experiencing clipping, and for example a feedback loop may be implemented to determine a suitable delay which avoids amplifier saturation for a given stimulus.

In preferred embodiments of the invention, the signal from the or each sense electrode is passed to a sample-and-hold circuit at the input of a measurement amplifier. In such embodiments measurements of a single evoked response may be obtained from a plurality of sense electrodes, even if the measurement circuitry of each electrode is connected to the control unit only by a two wire bus or the like, as is commonly required in implanted electrode arrays.

Additionally or alternatively, a buffer or follower amplifier is preferably provided in some embodiments, between the sense electrode and the measurement amplifier. The buffer is preferably connected to the sense electrode without interposed switches, so that the high reverse impedance of the buffer effectively prevents switching transients from being conveyed to the sense electrode, thereby avoiding artefact which may arise upon the sense electrode if subjected to such transients. The buffer amplifier is also preferably configured to give current gain to drive a storage capacitor of a sample and hold circuit. A series capacitor may be interposed between the sense electrode and the buffer to avoid DC transfer with the tissue.

In preferred embodiments of the invention, the stimulus and sense electrodes are selected from an implanted electrode array. The electrode array may for example comprise a linear array of electrodes arranged in a single column along the array. Alternatively the electrode array may comprise a two dimensional array having two or more columns of electrodes arranged along the array. Preferably, each electrode of the electrode array is provided with an associated measurement amplifier, to avoid the need to switch the sense electrode(s) to a shared measurement amplifier, as such switching can add to measurement artefact. Providing a dedicated measurement amplifier for each sense electrode is further advantageous in permitting recordings to be obtained from multiple sense electrodes simultaneously.

The measurement may be a single-ended measurement obtained by passing a signal from a single sense electrode to a single-ended amplifier. Alternatively, the measurement may be a differential measurement obtained by passing signals from two sense electrodes to a differential amplifier.

While recovering charge by short circuiting the stimulus electrodes together, it may in some embodiments be advantageous to disconnect the sense electrode from the measurement circuitry, for example by setting a sample-and-hold circuit to “hold”.

Embodiments of the invention may prove beneficial in obtaining a CAP measurement which has lower dynamic range and simpler morphology as compared to systems more susceptible to artefact. Such embodiments of the present invention may thus reduce the dynamic range requirements of implanted amplifiers, and may avoid or reduce the complexity of signal processing systems for feature extraction, simplifying and miniaturizing an implanted integrated circuit. Such embodiments may thus be particularly applicable for an automated implanted evoked response feedback system for stimulus control. Thus, in a further aspect, the present invention provides a method for feedback control of a neural stimulus, the method comprising an implanted control unit obtaining a CAP measurement in accordance with the method of the first aspect, and the implanted control unit using the obtained CAP measurement to control the delivery of subsequent neural stimuli by the implant.

In some embodiments of the invention, an averaged CAP measurement may be obtained by (i) delivering a first biphasic stimulus which starts with a pulse of a first polarity and then delivers a pulse of a second polarity opposite to the first polarity, and obtaining a first measurement of a CAP evoked by the first stimulus; (ii) delivering a second biphasic stimulus which starts with a pulse of the second polarity and then delivers a pulse of the first polarity, and obtaining a second measurement of a CAP evoked by the second stimulus; and (iii) taking an average of the first measurement and the second measurement to obtain an averaged measurement. Such embodiments exploit the observation that artefact polarity usually reflects the stimulus polarity, whereas the CAP polarity is independent of the stimulus polarity and is instead determined by the anatomy and physiology of the spinal cord membrane, so that averaging the first and second measurements will tend to selectively cancel out artefact. Further noting that an “anodic first” biphasic stimulus usually has a lower stimulus threshold for neural recruitment than a “cathodic first” biphasic stimulus, the averaged measurement may have a morphology of either (i) a typical CAP of half amplitude if only the anodic-first stimulus exceeds the stimulus threshold; (ii) the average of two CAPs of different amplitude if both stimuli exceed the stimulus threshold but the cathodic first stimulus does not cause saturation recruitment; or (iii) a typical CAP if both stimuli exceed saturation recruitment. Some embodiments may therefore obtain a curve of the averaged measurement vs. stimulus amplitude in order to obtain information regarding the recruitment effected by each stimulus, and such information may be used for feedback control by the implant.

In some embodiments, the method of the present invention may be applied contemporaneously with administration of a drug, in order to gauge efficacy of drug delivery. For example, the implant may comprise or be operatively connected to a drug reservoir and drug delivery pump, with the pump being controlled by feedback based on CAP measurements.

According to another aspect the present invention provides a computer program product comprising computer program code means to make an implanted processor execute a procedure for measuring a neural response to a stimulus, the computer program product comprising computer program code means for carrying out the method of the first aspect.

The present invention recognises that when considering spinal cord stimulation, obtaining information about the activity within the spinal segment where stimulation is occurring is highly desirable. Observing the activity and extent of propagation both above (rostrally of) and below (caudally of) the level of stimulation is also highly desirable. The present invention recognises that in order to record the evoked activity within the same spinal segment as the stimulus requires an evoked potential recording system which is capable of recording an SCP within approximately 3 cm of its source, i.e. within approximately 0.3 ms of the stimulus.

In preferred embodiments the stimulus comprises a bi-phasic pulse, and the stimulus electrodes have no capacitors. In contrast to a monophasic pulse and capacitor arrangement, such embodiments permit the stimulus electrode current to be interrupted, or forced to zero, at those times where it would interfere with measurement. Omitting capacitors is also desirable in order to minimise the size of the implanted device.

BRIEF DESCRIPTION OF THE DRAWINGS

An example of the invention will now be described with reference to the accompanying drawings, in which:

FIG. 1 illustrates an implantable device suitable for implementing the present invention;

FIG. 2 is a schematic of a feedback controller which refines future stimuli based on estimated recruitment of neurons by past stimuli;

FIG. 3 illustrates the masked to unmasked stimulation paradigm provided by the present embodiment of the invention;

FIG. 4 illustrates recordings of actual evoked responses in accordance with the method of one embodiment of the present invention; and

FIG. 5 is a plot of the (P1-N1) amplitude of measurements of responses evoked by two pulses, for varying inter-stimulus interval.

FIG. 6 illustrates currents and voltages which can contribute to SCP measurements;

FIG. 7 illustrates the circuitry of one embodiment of the present invention, throughout five phases of a measurement cycle;

FIG. 8 illustrates idealised waveforms arising in the circuit of FIG. 2 during each phase of the measurement cycle;

FIG. 9 illustrates SCP measurements made using the embodiment of FIG. 2;

FIG. 10 illustrates the circuitry of an alternative embodiment of the invention implementing differential CAP measurements;

FIG. 11 illustrates delayed activation of a measurement amplifier to avoid clipping;

FIG. 12 illustrates an embodiment in which alternate phased stimuli are used to obtain an averaged CAP measurement;

FIG. 13a illustrates the “anodic first” and “cathodic first” CAP responses induced by the method of FIG. 12, while FIG. 13b illustrates the averaged measurement obtained therefrom;

FIG. 14 illustrates the CAP response to anodic-first and cathodic-first stimuli, respectively, with increasing stimulus amplitude;

FIG. 15 illustrates the nature of differential CAP measurements in the spinal cord;

FIG. 16 illustrates yet another embodiment of the invention, in which a single electrode serves both as a stimulus electrode and also as the sense electrode in the same measurement cycle;

FIG. 17 illustrates a model of a metal electrode in a conductive solution;

FIG. 18 illustrates segmented electrodes which may be used to reduce artefact without sacrificing noise, impedance or current carrying capacity; and

FIGS. 19a and 19b illustrate the effect of epidural administration of Lignocaine on suppression of the spinal evoked responses.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

FIG. 1 illustrates an implantable device 100 suitable for implementing the present invention. Device 100 comprises an implanted control unit 110, which controls application of neural stimuli, and controls a measurement process for obtaining a measurement of a neural response evoked by the stimuli from each of a plurality of electrodes. Device 100 further comprises an electrode array 120 consisting of a three by eight array of electrodes 122, each of which may be selectively used as either the stimulus electrode or sense electrode, or both.

FIG. 2 is a schematic of a feedback controller which refines future stimuli based on estimated recruitment of neurons by past stimuli. The present embodiment provides for the recruitment estimator in FIG. 2 to obtain a measurement of a masked neural response arising in response to a probe stimuli applied during a refractory period of a therapeutic stimulus, and also provides for measurement of an unmasked neural response arising in response to a probe stimuli applied after a refractory period of the same or equivalent subsequent therapeutic stimulus. Comparing the ratio or difference between the masked and unmasked neural responses indicates a level of recruitment achieved by the therapeutic stimulus.

In this embodiment the evoked CAP measurements are made by use of the neural response measurement techniques set out in the Australian provisional patent application No. 2011901817 in the name of National ICT Australia Ltd entitled “Method and apparatus for measurement of neural response” from which the present application claims priority.

Long pulse widths on the order of 400 μs, as used in many commercially available stimulators, cause problems for the measurement of evoked response, as much of the neural response passes the recording electrodes during the stimulus period. That is, in such a biphasic pulse, at least 0.8 ms passes from stimulus onset before measurement is possible. As shown in FIG. 3a, the therapeutic stimulus 302 continues for a sufficiently long period of time that it substantially temporally overlaps the evoked neural response 304. The signal amplitudes in FIG. 3 are not to scale, and the therapeutic stimulus is of the order of volts while the neural response measurement is of the order of tens of microvolts, so that in the case shown in FIG. 3a the evoked response is practically impossible to measure directly. Nevertheless, for many reasons it is desirable to measure or estimate the amplitude of the response R_t induced by stimulus 302.

FIGS. 3b and 3c illustrate the masked to unmasked stimulation paradigm provided by the present embodiment of the invention. In order to estimate how many fibres are recruited in the neural response 304 arising from the long therapeutic pulse 302, a shorter probe pulse 306 is delivered shortly after the therapeutic stimulus 302. The neural response 308 caused by probe pulse 306 is not contemporaneous with any stimulus, and is therefore able to be measured without being swamped by large stimulus voltages. Notably, by delivering the probe pulse 306 during the refractory period of the fibres triggered in response 304, the response 308 has an amplitude R_I which is proportional to the number of fibres which were not triggered by the long pulse 302.

After a time delay of sufficient length to allow all fibres triggered as part of either response 304 or response 308 to exit their refractory states, another short probe pulse 310 is delivered as shown in FIG. 3c. Probe pulse 310 preferably has the same parameters as probe pulse 306.

Obtaining a measure of response 312 provides an unmasked response amplitude measurement R_S, with R_S>R_I, against which the first, masked response 308 can be compared. This masked/unmasked ratio (R_I. R_S) can be used to estimate what proportion of the accessible fibre population was stimulated in response 304 by therapeutic stimulus 302, thereby allowing R_t to be estimated. Notably, when performed sufficiently quickly that a fibre-to-electrode distance will remain substantially constant, this technique is not susceptible to the problem of unknown fibre-to-electrode distance as the ratios cancel the effect of variable electrode-to-fibre distance.

In addition to determining recruitment of long pulse width stimuli, it can be useful to measure physiological parameters such as refractory periods in order to give a diagnosis of various conditions or diseases. Thus, in another embodiment of the invention the refractory period is estimated by first obtaining a measure R_S of the unmasked neural response to a given probe stimulus. Then, two stimuli are applied close together separated by a variable delay t_d (FIG. 3b). With increasing t_d, the amplitude R_I can be expected to markedly increase when the onset of pulse 306 is delayed sufficiently to allow the average refractory period of the neural population recruited in response 304 to conclude, so that observing such an increase in R_I allows that population's refractory period to be estimated. There are a number of neurological conditions and non-neurological conditions which can affect the refractory period. This measurement technique may thus serve as a useful diagnostic indicator.

FIG. 4 illustrates recordings of actual evoked responses in accordance with the embodiment of FIG. 3. The recordings of a response pair were made on 8 spaced apart electrodes along the spinal column as the evoked responses 404, 408 travelled along the spinal column adjacent to the array. As can be seen, an initial response 404 is evoked by a first stimulus, and then a second response 408 is evoked immediately afterwards in the refractory period of the neural population recruited as part of response 404. Response 408 is of reduced, but non-zero, amplitude. The relative ratios of the amplitudes of the measurements of the two responses thus permit the above-described information to be elicited.

FIG. 5 is a plot of the (P1-N1) amplitude of measurements of responses 502, 506 respectively evoked by a first pulse 302 and a second pulse 306 of equal amplitude and pulse width, for varying inter-stimulus interval t_d. As can be seen at 502, the first pulse 302 produces the same recruitment and response amplitudes irrespective of t_d. However, the recruitment effected by the second pulse 306 varies considerably with t_d, as shown by 506. Two fibre population characteristics are evident in this plot, either or both of which may be investigated in accordance with the present invention in order to determine suitable stimulus parameters and/or physiological state or change. First, pulse 302 will depolarise some fibres close to threshold, but without activating them. This partial depolarisation means that for small t_d, in the range (512) of about 0 to 200 μs, where pulse 306 is sufficiently close in time to pulse 302, some fibres that had not been activated by 302 may be activated by 306 more easily than is the case for the remainder of the refractory period for t_d>200 s. This depolarisation will decay with time, usually to resting levels before the end of the absolute refractory period for the fibres that were activated by 302. This means for short inter-stimulus intervals (e.g. <200 us), there will be a response 308 from fibres which had residual depolarisation from 302. Second, for t_d greater than about 400 μs, a relative refractory period 514 commences, during which fibres activated by 302 gradually become able to be activated again. Between the remnant depolarisation period 512 and the relative refractory period 514, the absolute refractory period dominates and the second pulse 306 is almost entirely unable to recruit any response (it is noted that curve 506 is at levels around 5 μV in this period which may be noise and does not necessarily indicate any response has been evoked). Thus assessing curve 506 instantaneously permits a current state of both (a) the residual depolarisation decay 512, and (b) onset of the relative refractory period 514 to be assessed. Monitoring curve 506 over time permits changes in these characteristics to be determined, for example to be used for feedback to optimise therapeutic stimuli or in order to diagnose or monitor an underlying disease.

While FIG. 3b shows the probe pulse 306 as having the same amplitude as therapeutic pulse 302, alternative embodiments may advantageously use probe pulses 306 and/or 310 which are of a different amplitude to therapeutic pulse 302. For example, therapeutic pulse 302 is usually set to a comfortable level for the patient, and at such a level not all fibres are usually recruited by pulse 302. Pulse 306 may therefore be set to have a greater amplitude and/or a greater total charge than therapeutic stimulus 302 in order to ensure that the probe pulse 306 will recruit at least some fibres even when applied during the refractory period of fibres recruited as part of response 304.

In another embodiment the probe stimulus 306 may be configured to have reduced recruitment capability as compared to pulse 302, so that if pulse 306 is applied during the absolute refractory period of fibres recruited as part of response 304 then pulse 306 will recruit no additional response. In such embodiments, when the relative delay t_d is such that probe stimulus 306 occurs in the relative refractory period of response 304, being the period in which some fibres recruited as part of response 304 have concluded their refractory period but some have not, then the probe stimulus response 308 will begin to recruit fibres. Determining the value of t_d at which a threshold exists for response 308 starting to arise provides useful information regarding the refractory period of response 304.

Routinely, during assessment of patients for spinal cord stimulation therapy, the patient will undergo a trial stimulation procedure. This is where the patient is implanted with a percutaneous lead with an externalised set of contacts. The lead is attached to an external pulse generator and the patient has use of the device for several days. At the end of the trial period the clinician and patient assess the performance of the system with regard to pain relief and a choice is made whether or not to proceed with a full implantation. The take-home device for trial purposes may consist of both a stimulus generator but also an evoked response measurement system. The ERT responses recorded during the trial period could be used to adjust the stimulus parameters as described above.

The ERT system measures amplitude growth functions etc., collected at time of surgery and during the trial stimulation period, and together with subjective performance measures could be used to develop a correlation between the response parameters and the patient outcomes. For instance, there is considerable variation in threshold response and there may exist a correlation between threshold and outcome where lower thresholds generate better outcomes. There are a large number of neurological parameters that can be collected in performing ERT measures, including refractory periods. Systematic collection of this data across a number of patients will allow analysis for correlation with outcome.

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. For example the neural response measurement may be conducted in accordance with any suitable CAP measurement technique. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

FIG. 6 shows the currents and voltages that contribute to SCP measurements. These signals include the stimulus current 102 applied by stimulus electrodes, which is a charge-balanced biphasic pulse to provide low artefact. In the case of spinal cord stimulation, the stimulus currents 102 used to provide paraesthesia and pain relief typically consist of pulses in the range of 3-30 mA amplitude, with pulse width typically in the range of 100-400 μs. The stimuli can comprise monophasic or biphasic pulses.

The stimulus 102 induces a voltage on adjacent electrodes, referred to as stimulus crosstalk 104. Where the stimuli 102 are SCP stimuli they typically induce a voltage 104 in the range of about 1-5 V on a SCP sense electrode.

The stimulus 102 also induces electrode artefact, which is a residual voltage on an electrode resulting from uneven charge distribution on its surface. The electrode artefact is indicated in the voltage waveform 104 after cessation of stimulus crosstalk. The stimulus 102 disturbs the galvanic interface between the sense electrode and the tissue, so that after stimulus crosstalk in voltage 104 concludes, a voltage known as the electrode artefact continues on the electrode, as indicated in waveform 104 in FIG. 6. Electrode artefact is very difficult to measure, and depends on factors such as the stimulation pulse, the geometry of the electrodes and the bio-electrical nature of the tissue surrounding the electrodes. Electrode artefact can have a typical value of 50 μV at a time 50 μs after stimulation ceases. Electrode artefact is difficult to measure because it is indistinguishable from electrical artefact, the latter being caused by the amplifier's exposure to the high stimulation voltages. Further, the causes of electrical artefact can be subtle, and therefore hard to identify and eliminate.

An appropriate stimulus 102 will also induce nerves to fire, and thereby produces an evoked neural response 106. In the spinal cord, the neural response 106 has two major components: a fast response lasting ˜2 ms and a slow response lasting ˜15 ms. The amplitude of the evoked response seen by epidural electrodes is typically no more than hundreds of microvolts, but in some clinical situations can be only tens of microvolts.

In practical implementation a measurement amplifier used to measure the evoked response does not have infinite bandwidth, and will normally have infinite impulse response filter poles, and so the stimulus crosstalk 104 will produce an output 108 during the evoked response 106, this output being referred to as electrical artefact. Electrical artefact can be in the hundreds of millivolts as compared to a SCP of interest in the tens of microvolts. Electrical artefact can however be reduced by suitable choice of a high-pass filter pole frequency.

The measurement amplifier output 110 will therefore contain the sum of these various contributions 102-108. Separating the evoked response of interest (106) from the artefacts 104 and 108 is a major technical challenge. For example, to resolve a 10 μV SCP with 1 μV resolution, and have at the input a 5V stimulus, requires an amplifier with a dynamic range of 134 dB. As the response can overlap the stimulus this represents a difficult challenge of amplifier design.

FIGS. 7a-7e are schematic diagrams of the five phases of operation of a sample and hold (S/H) measurement amplifier in accordance with one embodiment of the present invention. The stimulus and measurement circuitry 200 comprises a buffer amplifier 206 that is always connected to the sense electrode 202 such that there is no switch between the sense electrode 202 and the buffer amplifier 206. The output of the buffer amplifier 206 drives a sample and hold circuit 208, followed by a high gain amplifier 210 with unity gain at DC. The front-end amplifier 206 has sufficiently wide bandwidth that it can follow the voltage induced on the sense electrodes 202 by the stimulus pulse, and settle before the SCP begins. A current source 212 can be selectively connected to stimulus electrodes 204 to deliver a stimulus. The stimulus electrodes 204 and sense electrode 202 are in the same electrode array of a single implanted device.

The stimulus and measurement circuitry 200 operates to obtain a SC measurement using five phases. The first phase shown in FIG. 7a open circuits the stimulus electrodes 204 and connects the sense electrode 202 to the measurement amplifier 210 by setting the sample and hold circuit to “sample”. The first phase shown in FIG. 7a allows the amplifier chain 206, 210 to settle, with no disturbance from the stimulating electrodes 204.

In the second phase shown in FIG. 7b, the stimulus electrodes 204 are short circuited to each other. This allows the stimulating electrodes 204 to recover charge, so as to avoid DC injection to the tissue as is required for electrical implants. During this phase, the sample-and-hold 208 is set to “hold” so that charge transfer on the stimulus electrodes 204 does not disrupt the measurement amplifier 210.

In the third phase shown in FIG. 7c, the stimulation is applied. The stimulus electrodes 204 are switched to the current source 212, and the sample-and-hold 208 is set to “hold” so that the large stimulus crosstalk seen on electrode 202 is not presented to the measurement amplifier 210.

The fourth phase shown in FIG. 7d provides for a post-stimulus delay. In this phase the stimulus electrodes 204 are open circuited, and the sample-and-hold remains in the “hold” position, to allow the electrodes 202, 204 settle towards equilibrium, as defined by bio-electrical conditions.

Finally, in the fifth phase shown in FIG. 7e, the SCP present at sense electrode 202 is measured by switching the sample-hold 208 to “sample”.

When performing repeated measurement cycles in this fashion, it is noted that the switch positions are the same in the phase 1 “settling” and the phase 5 “measuring” states. Thus, the state of phase 5 is maintained, by virtue of a subsequent phase 1, until the electrodes and circuitry are in equilibrium, even after the time that useful SCP data is no longer present or being captured. Such embodiments thus provide a greater length of the “settle” state.

FIG. 8 shows idealised waveforms arising during the SCP measurement process of FIG. 7. FIG. 8 illustrates the current 802 of stimulus electrodes 204, and the output voltage 804 of amplifier 210, during each of the five phases of the measurement cycle. Importantly, it can be seen that phase 1 permits the amplifier bias point to settle to a steady state as defined by bio-electrical conditions at the sense electrode, while phases 2-4 do not disrupt the amplifier 210 bias point.

An advantage of this circuit is that in the phase 2 equilibration, the circuitry around amplifier 210 is a low-pass filter, and is therefore relatively immune to noise and input transients. This also allows the amplifier 210 to accumulate its bias point over successive measurement cycles, as it does not need to be reset for each cycle. Moreover, because of the buffer 206 before the sample/hold 208, the input-referred effect (i.e. the effect upon sense electrode 202) of the charge injection into the sample/hold 208 is lower.

In the embodiment of FIG. 7, the sense electrode 202 is never shorted to the stimulus electrodes 204, recognising that this creates dis-equilibrium in the sense electrodes and adds artefact, rather than having the effect of creating equilibrium as previously thought. In some embodiments, it may be possible to overlap the “settle” (equilibrate) phase of FIG. 7a, and the “charge recovery” phase of FIG. 7b, although it would be expected that the artefact would be higher, and the time taken to reach equilibrium longer.

FIG. 9 is a plot of 22 separate measurements of ovine SCP made using the embodiment of FIG. 7. The measurements were obtained sequentially for differing stimuli, the stimuli comprising biphasic current pulses of 40 μs pulse width and a current amplitude which varied from 0-10 mA. The measurements were then plotted on a single chart to produce FIG. 9.

The waveforms of FIG. 9 have lower dynamic range and simpler morphology than measurements produced by previous approaches, due to the absence of stimulus crosstalk and reduced artefact. When wishing to provide a system built on an implanted integrated circuit, wide dynamic range amplifiers are difficult to design, as are signal processing systems for feature extraction. Beneficially, the nature of the measured waveforms shown in FIG. 9 permits, for example, a circuit for extracting the peak-to-peak SCP amplitude to have fewer components than would be required to operate upon the waveform produced by previous approaches. Thus the techniques of the present invention for artefact reduction greatly assist in building a practical implanted, evoked response feedback system.

Moreover, it is notable that in this case of a 40 μs pulse width the measurement system is settled and ready to record prior to onset of the evoked CAP. The sense electrode was less than 50 mm from the stimulus electrode, and a post-stimulus delay of 50 μs was observed before the measurement amplifier was switched in to obtain the recordings shown in FIG. 9.

FIG. 10 illustrates the circuitry of an alternative embodiment of the invention in which a differential measurement amplifier is used, and charge recovery is via a voltage rail V_dd. As can be seen, in accordance with the present invention the measurement phases are carried out in a corresponding manner despite the use of different hardware.

In the embodiments of either FIG. 7 or FIG. 10, artefact can cause the high-gain measurement amplifier 210 to clip, and the amplifier can subsequently be slow to recover. However, in preferred embodiments the sample point, being the transition from the “stimulate” to “measure” phases, is delayed, allowing clipping to be avoided. FIG. 11 illustrates the manner of determining a suitable delay 602. Such embodiments may permit use of a higher amplifier gain than would otherwise be the case. In particular, a variable delay and increased amplifier gain may be particularly apt in circumstances where high-gain is desired, and parts of the SCP of interest do not immediately follow the stimulation. Thus, delaying the start of measurement will avoid the side effects of clipping.

In another embodiment of the invention shown in FIG. 12, a method to eliminate artefact from an SCP measurement is to alternate the phase of stimulus waveforms and take an average of obtained measurements. FIG. 12 shows the stimulus current for a positive “anodic-first” stimulus 702, and the stimulus current for a negative “cathodic first” stimulus 704. In this embodiment these are applied in succession with respective CAP measurements obtained after each stimulus. The respective measurement electrode voltages 706 and 708 arising from each such stimulus are also shown. It will be observed where indicated in waveforms 706, 708 that the artefact from each of the two stimuli are of identical magnitude, but opposite sign. In most situations it will be found that the artefact polarity depends on the stimulus polarity. An example of this would be electrical artefact caused by the high-pass poles of the front-end amplifier 206. Clearly, either phase could be used for stimulating nervous tissue, though their effects will differ.

In contrast, the positive and negative phase stimuli 702, 704 produce SCPs of differing amplitudes, but approximately similar shape and importantly of similar polarity, as this is determined by the anatomy and physiology of the spinal cord membrane. Thus, when the voltages 706, 708 resulting from the positive and negative phase stimuli 702, 704 are recorded, and averaged, the opposite phase stimulation artefacts substantially cancel, leaving the SCP or a combination of the two SCPs 710. Note that in practical situations, the artefact can have much higher amplitude than the SCP, making it much harder to detect the SCP than is apparent from FIG. 12.

The response of the spinal cord to these two polarities of stimulation are referred to as the “anodic first” and “cathodic first” SCP responses, as referred to the electrode considered to be that closest to the recording electrode. I.e. anodic-first stimulation makes the stimulating electrode closest to the sense electrode anodic in the first phase of stimulus. Usually cathodic-first stimulation has a lower threshold for neural activation than is the case for anodic-first stimulation. Nevertheless, the SCP polarity is independent of whether the stimulus is anodic first 702 or cathodic first 704.

FIG. 13a illustrates spinal cord measurements obtained in response to anodic-first and cathodic-first stimulations, respectively. Note that measurement obtained in response to the anodic-first stimulation lacks the characteristic P1-N1-P2 form, indicating that the anodic-first stimulation did not evoke a neural response in this case. In contrast, the measurement obtained in response to the cathodic-first stimulus exhibits a significant evoked neural response.

FIG. 13b shows an average of the two responses in FIG. 15a. As can be seen, while the characteristic form of the SCP has been altered, the artefact is essentially removed.

This embodiment of the invention further recognises that the averaged waveform of FIG. 13b can be used to obtain a range of information despite the atypical SCP form. In this regard, FIG. 14 illustrates SCP growth curves against stimulus amplitude, for both anodic-first and cathodic-first stimuli. FIG. 14 also shows the growth behaviour of the average SCP against stimulus amplitude. It can be seen from FIG. 14 that the threshold of the average response is identical to the threshold of the more sensitive cathodic-first stimulation.

When the stimulus amplitude is in the range 902 such that only the cathodic-first stimulus produces an SCP, then the averaged SCP waveform would have a normal SCP morphology but would be half the amplitude compared to a true cathodic first SCP due to the averaging. In the region 904 where both the anodic first and cathodic first responses contribute to the averaged SCP, the resultant averaged SCP waveform will have morphology in between the two measurements. It would not directly represent an SCP, but rather the average of two different SCPs. Nevertheless, this waveform could still be valuable for example in implementing an automatic control loop for stimulation adjustment, as it gives a value proportional to neural recruitment.

Some embodiments of the invention, such as the embodiment of FIG. 10, may use differential amplifiers so as to detect the voltage difference between two sense electrodes. Differential amplifiers simplify the task of separating electrode artefact. If they are connected to electrodes with similar area, and separated from the stimulation electrodes in a similar manner, then they receive similar levels of electrode artefact and this will be removed when their difference voltage is obtained. However, in such a system the voltage recorded by the amplifier is the difference between the voltages at two points along a bundle of neurons, and can thus be difficult to interpret. When making SCP measurements, it is preferable to use single-ended amplifiers as they more accurately measure the SCP, and they are more sensitive in measuring the SCP.

Differential amplifiers are often used because they provide a means to reduce electrode artefact, when other means have been insufficient. However, FIG. 15 illustrates a problem of measuring SCPs with differential amplifiers. It shows a spinal cord potential. As this potential travels along the spine at a velocity, which can be as high as 80 m·s-1, it can also be considered as a spatial wave. Given that a peak-to-peak cycle of the fast response of an SCP typically lasts for 1 ms, the wave will travel 8 cm in this time. Using this 1 ms=8 cm scale, a 5 cm electrode array is drawn alongside the SCP in FIG. 15. Connected to this electrode array are two amplifiers configured to make differential SCP measurements from separate pairs of sense electrodes. As can be seen from FIG. 15, the difference between the voltages on the adjacent electrodes will be quite small and significantly smaller than the peak to peak amplitude of the SCP, and thus more susceptible to electrical noise generated by the amplifier. The output of the amplifier will approximate the differential of the SCP, and thus be harder to interpret than a simple measure of the SCP itself. If measuring evoked SCPs with a micropackage stimulator design, for example in a system using a two-wire bus, differential measurements between non-adjacent electrodes are not possible. Further, if wishing to measure the slow response of the SCP, which has a period of about 6 ms and correspondingly reduced signal gradients, differential measurements are even more difficult to effect. Thus it will be appreciated that single-ended measurements are preferable, as long as artefact can be kept at a sufficiently low level.

With the measurement sequence of the present invention, the artefact is reduced so that some embodiments may instead use a single-ended amplifier, even in situations where previously they would have suffered from too much electrode artefact. Moreover, trials to date show that recording can be initiated with an extremely short time interval from cessation of the stimulus, permitting the same electrode array to be used for recording and stimulation, and even permitting recordings to be made on the electrode immediately adjacent to the stimulus electrode in an electrode array with electrode spacings of less than 10 mm.

Single ended amplifiers have the further advantage that they consist of fewer capacitors and amplifier components than differential amplifiers, so will take up less space on a silicon chip, which is a significant benefit when intended for use in an implanted system with many electrodes and where the silicon area for each amplifier is limited.

Preferred embodiments of the invention may comprise a separate amplifier chain (e.g. 206, 208, 210, see FIG. 7) for every electrode, organised in parallel manner, permitting simultaneous recording of a single CAP from multiple sense electrodes in parallel, and also eliminating the switching noise arising in systems which switch the sense electrode to a shared measurement amplifier.

FIG. 16 illustrates another embodiment of the invention in which the sense electrode is one of the stimulus electrodes. The follower/buffer amplifier in this embodiment has sufficient dynamic range to withstand the stimulus phase, however in alternative embodiments a switch may be interposed between the measurement electrode and the measurement circuitry to isolate the follower/buffer amplifier during the stimulus phase.

Further embodiments of the invention may employ divisible electrodes, as discussed below with reference to FIGS. 17 and 18. When considering electrode artefact in particular, the sources of electrode artefact are relatively poorly understood. The surface of a metal electrode can be modelled as an RC network. For an accurate model, an infinite-phase element is required, but for the explanation of artefact a simple RC model will suffice, as shown in FIG. 17a. A conductive solution can be modelled as a mesh of resistors. Where a conductive solution meets a piece of metal of finite dimensions, the metal provides an alternative conduction path to the solution. This charges the electrode-to-tissue capacitances at the “ends” of the electrodes, with opposite polarities. The electrode does not acquire net charge, but it does cease to be in equilibrium. After the external current ceases, then the electrode will pass current through the solution as it re-equilibrates for a short time after the stimulus. This current will affect the potential of another electrode in the solution, and in the case of multielectrode arrays a unique such current will arise at every electrode in response to local conditions experienced at that electrode. The cumulative impact of such reequilibration currents is seen by a sense electrode as electrode artefact.

A similar effect happens when current flows between two electrodes, as shown in FIG. 17b. During application of a stimulus, the current preferentially flows between the parts of the electrodes where they are closest. When the current is interrupted, the charge on the surface of the electrodes must re-equilibrate; this also leads to a residual current and contributes to electrode artefact seen by a sense electrode.

The model of FIG. 17 predicts that using smaller electrodes will reduce artefact. However, smaller electrodes will have higher noise when used as measurement electrodes, and higher resistance and lower current carrying capacity when used as stimulus electrodes. Two means to reduce artefact without sacrificing noise, impedance or current carrying capacity are shown in FIGS. 18a and 18b. The electrode configuration of FIG. 18a reduces artefact induced in a single metallic electrode; the electrode is composed of two or more smaller electrodes that can be disconnected during a stimulation phase, and reconnected during a measurement phase. In the configuration of FIG. 18b, an electrode is segmented, and individual current sources are provided for each segment. This forces the current in the segments to match, and so reduces artefact.

The evoked response telemetry of the present invention may in some embodiments be used to monitor the effect of a delivered compound. The administration of compounds (drugs or other chemical therapeutics) to effect a change in the nervous system is common for treatment of a wide number of diseases and disorders. Anaesthetics of various types are administered to the spinal cord for the relief of pain. Perhaps the most common form is administration of anaesthetics in the epidural space for pain relief during child birth.

In such embodiments, a catheter comprising a drug delivery tube may be fitted with electrode elements and configured to obtain neural response measurements in accordance with the present invention in order to monitor drug-induced effects on the neural response. Alternatively an electrode array may be temporarily or permanently implanted and used to apply neural stimuli and monitor the neural response. The neural response measurements may be obtained repeatedly during administration of the drug in order to directly measure the effect of the administered drug and control the dosage delivered.

FIGS. 19a and 19b illustrate the effect of administration of anaesthetic to the spinal cord, with a neural response being present prior to administration and largely being absent subsequent to administration. As can be seen, there is a direct correlation between the measured evoked response and the dosage of the anaesthetic. A “partial block” may be effected by ceasing administration of the anaesthetic once the neural response amplitude reduces to a desired level. The technology described herein is suitable for full implantation within the body of a subject and as a result the evoked potential monitoring could be used in the administration of an active compound to produce a therapeutic benefit. The system could be integrated with in an implantable pump to control the administration of the compound.

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

Claims

1. A method of applying stimulus to a tissue, the method comprising: delivering, by a plurality of stimulus electrodes, a therapeutic pulse having a first pulse width,delivering a probe pulse having a second pulse width after the therapeutic pulse, wherein the second pulse width is less than the first pulse width, and wherein the probe pulse evokes a neural response,measuring, by one or more sense electrodes, the neural response evoked by the probe pulse;determining a neural recruitment caused by the probe pulse based on the neural response; anddetermining a parameter for a subsequent therapeutic pulse based on the neural recruitment.

2. The method of claim 1, wherein the first pulse width is in a range of about 100-500 μs.

3. The method of claim 1, wherein the determining the parameter comprises: comparing the neural recruitment with a target neural recruitment;deriving, using a controller, a command from the comparison; andselecting, based on the command, the parameter for the subsequent therapeutic pulse.

4. The method of claim 1, wherein measuring the neural response comprises: settling measurement circuitry prior to the probe pulse, by connecting the one or more sense electrodes to the measurement circuitry to allow the measurement circuitry to settle towards a bio-electrically defined steady state;recovering charge on the plurality of stimulus electrodes by short circuiting the plurality of stimulus electrodes to each other;keeping the one or more sense electrodes disconnected from the measurement circuitry while delivering the probe pulse;imposing a delay during which the plurality of stimulus electrodes are open circuited and the one or more sense electrodes are disconnected from the measurement circuitry and from the plurality of stimulus electrodes; andmeasuring, after the delay, the neural response by connecting the one or more sense electrodes to the measurement circuitry.

5. The method of claim 1, wherein the probe pulse is delivered after a refractory period of fibres recruited by the therapeutic pulse.

6. The method of claim 3, wherein the selecting the parameter comprises refining a parameter for a previously delivered therapeutic pulse.

7. The method of claim 3, wherein the target neural recruitment is a predetermined neural recruitment.

8. The method of claim 1, wherein the probe pulse is delivered at least 1 ms after the end of the therapeutic pulse.

9. An implantable device for applying neural stimulus to neural tissue, the device comprising: a plurality of electrodes including one or more nominal stimulus electrodes and one or more nominal sense electrodes;a stimulus source for delivering stimulus pulses via the one or more stimulus electrodes to neural tissue;measurement circuitry for obtaining a measurement of a neural signal sensed at the one or more sense electrodes; anda control unit configured to: control the stimulus source to deliver a therapeutic pulse to neural tissue via the one or more stimulus electrodes, the therapeutic pulse having a first pulse width;control the stimulus source to deliver a probe pulse to the neural tissue having a second pulse width after the therapeutic pulse, wherein the second pulse width is less than the first pulse width, and wherein the probe pulse evokes a neural response;control the measurement circuitry to measure the neural response evoked by the probe pulse;determine a neural recruitment caused by the probe pulse based on the neural response; anddetermine a parameter for a subsequent therapeutic pulse based on the neural recruitment.

10. The implantable device of claim 9, wherein the first pulse width is in a range of about 100-500 μs.

11. The implantable device of claim 9, wherein the control unit is configured to determine the parameter for the subsequent therapeutic pulse by: comparing the neural recruitment with a target neural recruitment;deriving a command from the comparison; andselecting, based on the command, the parameter for the subsequent therapeutic pulse.

12. The implantable device of claim 9, wherein the control unit is configured to control the stimulus source to deliver the probe pulse after a refractory period of fibres recruited by the therapeutic pulse.

13. The implantable device of claim 11, wherein the control unit is configured to select the parameter by refining a parameter for a previously delivered therapeutic pulse.

14. The implantable device of claim 11, wherein the target neural recruitment is a predetermined neural recruitment.

15. The implantable device of claim 9, wherein the control unit is configured to control the stimulus source to deliver the probe pulse at least 1 ms after the end of the therapeutic pulse.

16. The implantable device of claim 9, wherein the control unit is configured to control the measurement circuitry to measure the neural response by: settling the measurement circuitry prior to the probe pulse by connecting one or more sense electrodes to the measurement circuitry to allow the measurement circuitry to settle towards a bio-electrically defined steady state,recovering charge on the one or more stimulus electrodes by short circuiting the one or more stimulus electrodes to each other;keeping one or more sense electrodes disconnected from the measurement circuitry while delivering the probe pulse,imposing a delay during which the one or more stimulus electrodes are open circuited and one or more sense electrodes are disconnected from the measurement circuitry and from the stimulus electrodes; andmeasuring, after the delay, the neural response by connecting one or more sense electrodes to the measurement circuitry.

17. A method of applying stimulus to a tissue, the method comprising: delivering, by a plurality of stimulus electrodes, a plurality of therapeutic pulses having a first pulse width,delivering a plurality of probe pulses having a second pulse width interleaved with at least some therapeutic pulses of the plurality of therapeutic pulses, wherein the second pulse width is less than the first pulse width, and wherein the plurality of probe pulses evoke neural responses,measuring, after a probe pulse of the plurality of probe pulses and prior to a subsequent therapeutic pulse of the plurality of therapeutic pulses, a neural response evoked by the probe pulse of the plurality of probe pulses via one or more sense electrodes;determining a neural recruitment caused by the probe pulse of the plurality of probe pulses based on the neural response; anddetermining a parameter for the subsequent therapeutic pulse based on the neural recruitment.

18. The method of claim 17, wherein the first pulse width is in a range of about 100-500 μs.

19. The method of claim 17, wherein the determining the parameter comprises: comparing the neural recruitment with a target neural recruitment;deriving, using a controller, a command from the comparison; andselecting, based on the command, the parameter for the subsequent therapeutic pulse.

20. The method of claim 19, wherein the selecting the parameter comprises refining a parameter for a previously delivered therapeutic pulse.

21. The method of claim 19, wherein the target neural recruitment is a predetermined neural recruitment.

22. An implantable device for applying neural stimulus to neural tissue, the device comprising: a plurality of electrodes including one or more nominal stimulus electrodes and one or more nominal sense electrodes;a stimulus source for delivering stimulus pulses via the one or more stimulus electrodes to neural tissue;measurement circuitry for obtaining a measurement of a neural signal sensed at the one or more sense electrodes; anda control unit configured to: control the stimulus source to deliver a plurality of therapeutic pulses to neural tissue via the one or more stimulus electrodes, the plurality of therapeutic pulses having a first pulse width;control the stimulus source to deliver a plurality of probe pulses to the neural tissue having a second pulse width interleaved with at least some therapeutic pulses of the plurality of therapeutic pulses, wherein the second pulse width is less than the first pulse width, and wherein the plurality of probe pulses evoke neural responses;control the measurement circuitry to measure, after a probe pulse of the plurality of probe pulses and prior to a subsequent therapeutic pulse of the plurality of therapeutic pulses, a neural response evoked by the probe pulse of the plurality of probe pulses;determine a neural recruitment caused by the probe pulse of the plurality of probe pulses based on the neural response; anddetermining a parameter for the subsequent therapeutic pulse based on the neural recruitment.

23. The implantable device of claim 22, wherein the first pulse width is in a range of about 100-500 μs.

24. The implantable device of claim 23, wherein the control unit is configured to determine the parameter for the subsequent therapeutic pulse by: comparing the neural recruitment with a target neural recruitment;deriving a command from the comparison; andselecting, based on the command, the parameter for the subsequent therapeutic pulse.

25. The implantable device of claim 24, wherein the control unit is configured to select the parameter by refining a parameter for a previously delivered therapeutic pulse.

26. The implantable device of claim 24, wherein the target neural recruitment is a predetermined neural recruitment.

27. The implantable device of claim 22, wherein the control unit is configured to control the stimulus source to deliver each probe pulse of the plurality of probe pulses after a refractory period of fibres recruited by a previously delivered therapeutic pulse.

28. The implantable device of claim 22, wherein the control unit is configured to control the stimulus source to deliver each probe pulse of the plurality of probe pulses at least 1 ms after the end of a previously delivered therapeutic pulse.