Reducing stimulation artefact

Abstract

One example discloses a method of stimulating a target neural structure which comprises delivering a pulse train to an electrode implanted in or near the target neural structure. The pulse train comprises one or more compensation pulses which are not charge balanced and impart a charge so as to reduce or prevent a stimulation artefact after the pulse train, and one or more counter balancing pulses which are not charge balanced and impart a charge opposite to the charge imparted by the one or more compensation pulses. The counter balancing pulses substantially charge balance with the compensation pulses. Another example discloses a method of monitoring neural activity, comprising delivering a calibration pulse to an electrode implanted in or near a target neural structure, determining a compensation current or voltage that is to reduce a stimulation artefact based on waveform characteristics extracted from a response to the calibration pulse and delivering a pulse train comprising the determined compensation current or voltage and detecting neural activity of the target neural structure after delivering the pulse train. Apparatus and a storage medium storing instructions for carrying out the method are also disclosed.

Description

REDUCING STIMULATION ARTEFACTTECHNICAL FIELD

[0001] The present application relates to a method and system for delivering a neural stimulus to a target neural structure. In some examples the method and system detect neural activity of the target neural structure after delivering the neural stimulus. In some examples, the techniques described herein may be applied to a system and method for deep brain stimulation (DBS) and methods and systems of monitoring neural activity responsive to DBS.BACKGROUND

[0002] Electrical neuromodulation or electrical stimulation of nerves may be used for various purposes including but not limited to pain management, treatment of Parkinson's disease and movement disorders, epilepsy, psychiatric disorders, urinary and fecal incontinence and gastrointestinal disorders. Deep brain stimulation (DBS) is one type of electrical neuromodulation which is used for treatment of movement disorders as well as other neurological disorders, including epilepsy, obsessive compulsive disorder, and depression.

[0003] Typically a signal generator is connected to one or more electrodes which are implanted in or near the target neural structure and electrical pulses are then delivered to the target neural structure. In the case of DBS the target neural structure may for example be in the subthalamic nucleus (STN), the globus pallidus interna (GPi), and / or the thalamus. Electrodes are connected to a neurostimulator usually implanted within the body and configured to deliver electrical pulses into target areas. It is believed that this electrical stimulation disrupts abnormal brain activity causally linked to a patient’s symptoms. Stimulation parameters can be adjusted using a controller external to the body, remotely connected to the neurostimulator.

[0004] However, it can be difficult to place the electrodes correctly. In the context of DBS, established techniques for intraoperative testing of DBS electrodes to ensure correct positioning in the brain, include x-ray imaging, microelectrode recordings, and clinical assessment. However, these can be inaccurate and consequently, electrodes are often implanted in suboptimal locations, resulting in diminished therapeutic outcomes and unwanted side-effects.

[0005] In DBS and other applications, it would be desirable to accurately record neural activity after delivering a stimulus. For example, recording the neural activity could help tobetter understand the effects of the stimulation and / or to gather data that be used to inform positioning or re-positioning of the electrodes and for other purposes. However, the electrical stimulus may generate a stimulation artefact due to polarization of the nearby neural tissue. This can make it difficult to accurately measure the neural activity after delivering the stimulus, as the amplitude of the stimulation artefact may in some cases be similar to or many times greater than the amplitude of the neural response. This is particularly the case when the neural stimulus and neural activity are measured at the same or nearby electrodes and / or when the neural activity is measured at the same time or close in time to the delivery of the neural stimulus. The stimulation artefact may saturate an amplifier, especially where the amplifier has low power and / or a limited range, which is often the case for bioamplifiers.

[0006] A bioamplifer is an amplifier used to measure electrical signals originating from biological tissue. Such amplifiers need to be able to detect small signals, but often operate on low power so that they may for instance be battery powered. Bioamplifier saturation can arise due to a stimulation artefact and / or residual voltage in capacitive loads.

[0007] Any discussion of documents, acts, materials, devices, articles or the like which has been included in the present specification 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 disclosure as it existed before the priority date of each of the appended claims.BRIEF SUMMARY

[0008] A first aspect of the present disclosure provides a method of delivering a stimulus to a target neural structure. The method comprises delivering a pulse train to an electrode implanted in or near the target neural structure. The pulse train comprises one or more compensation pulses which are not charge balanced and impart a charge so as to reduce or prevent a stimulation artefact after the pulse train. The pulse train further comprises one or more counter balancing pulses which are not charge balanced and impart a charge opposite to the charge imparted by the one or more compensation pulses. The one or more counter balancing pulses substantially charge balance with the one or more compensation pulses so that the pulse train as a whole is substantially charge balanced.

[0009] A third aspect of the present disclosure provides an apparatus comprising a signal generator, a measurement circuit and a processing unit configured to control the signal generator to perform the method of the first aspect of the present disclosure.

[0010] A third aspect of the present disclosure provides a method of monitoring neural activity. The method comprises delivering at least one calibration pulse to one or more electrodes implanted in or near a target neural structure; detecting a resultant waveform resulting from the at least one calibration pulse; extracting one or more waveform characteristics from the resultant waveform; based on the one or more waveform characteristics determining a compensation current, voltage or charge that is to reduce a stimulation artefact; delivering into the target neural structure a pulse train comprising a plurality of pulses together with the determined compensation current, voltage or charge so as to reduce a stimulation artefact of the pulse train; and detecting neural activity of the target neural structure after delivering the pulse train.

[0011] A fourth aspect of the present disclosure provides an apparatus comprising a signal generator, a measurement circuit; and a processing unit configured to control the signal generator and the measurement circuit to perform the method of the third aspect of the present disclosure.

[0012] A fifth aspect of the present disclosure provides a non-transitory computer readable storage medium storing instructions which are executable by a processor to perform the method of the third aspect of the present disclosure.

[0013] Further aspects, features and examples are provided in the appended claims and the description below.BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0014] To easily identify the discussion of any particular element or act, the most significant digit or digits in a reference number refer to the figure number in which that element is first introduced.

[0015] FIG. 1 illustrates an aspect of the subject matter in accordance with one embodiment.

[0016] FIG. 2 illustrates an aspect of the subject matter in accordance with one embodiment.

[0017] FIG. 3A illustrates an aspect of the subject matter in accordance with one embodiment.

[0018] FIG. 3B illustrates an aspect of the subject matter in accordance with one embodiment.

[0019] FIG. 4 illustrates an aspect of the subject matter in accordance with one embodiment.

[0020] FIG. 5 illustrates an aspect of the subject matter in accordance with one embodiment.

[0021] FIG. 6 illustrates an aspect of the subject matter in accordance with one embodiment.

[0022] FIG. 7A illustrates an aspect of the subject matter in accordance with one embodiment.

[0023] FIG. 7B illustrates an aspect of the subject matter in accordance with one embodiment.

[0024] FIG. 8 A illustrates an aspect of the subject matter in accordance with one embodiment.

[0025] FIG. 8B illustrates an aspect of the subject matter in accordance with one embodiment.

[0026] FIG. 9 illustrates an aspect of the subject matter in accordance with one embodiment.

[0027] FIG. 10 illustrates a method 1000 in accordance with one embodiment.

[0028] FIG. 11 A illustrates an aspect of the subject matter in accordance with one embodiment.

[0029] FIG. 1 IB illustrates an aspect of the subject matter in accordance with one embodiment.

[0030] FIG. 12 illustrates an aspect of the subject matter in accordance with one embodiment.

[0031] FIG. 13 A illustrates an aspect of the subject matter in accordance with one embodiment.

[0032] FIG. 13B illustrates an aspect of the subject matter in accordance with one embodiment.

[0033] FIG. 13C illustrates an aspect of the subject matter in accordance with one embodiment.

[0034] FIG. 14A illustrates an aspect of the subject matter in accordance with one embodiment.

[0035] FIG. 14B illustrates an aspect of the subject matter in accordance with one embodiment.

[0036] FIG. 15 illustrates an aspect of the subject matter in accordance with one embodiment.

[0037] FIG. 16 illustrates an aspect of the subject matter in accordance with one embodiment.

[0038] FIG. 17 illustrates an aspect of the subject matter in accordance with one embodiment.

[0039] FIG. 18 illustrates an aspect of the subject matter in accordance with one embodiment.

[0040] FIG. 19A illustrates an aspect of the subject matter in accordance with one embodiment.

[0041] FIG. 19B illustrates an aspect of the subject matter in accordance with one embodiment.

[0042] FIG. 20 illustrates an aspect of the subject matter in accordance with one embodiment.

[0043] FIG. 21 illustrates an aspect of the subject matter in accordance with one embodiment.

[0044] FIG. 22 illustrates an aspect of the subject matter in accordance with one embodiment.

[0045] FIG. 23 illustrates an aspect of the subject matter in accordance with one embodiment.

[0046] FIG. 24 illustrates an aspect of the subject matter in accordance with one embodiment.

[0047] FIG. 25A illustrates an aspect of the subject matter in accordance with one embodiment.

[0048] FIG. 25B illustrates an aspect of the subject matter in accordance with one embodiment.

[0049] FIG. 26 A illustrates an aspect of the subject matter in accordance with one embodiment.

[0050] FIG. 26B illustrates an aspect of the subject matter in accordance with one embodiment.

[0051] FIG. 27 illustrates an aspect of the subject matter in accordance with one embodiment.

[0052] FIG. 28 illustrates an aspect of the subject matter in accordance with one embodiment.DETAILED DESCRIPTION

[0053] 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.

[0054] The present application discloses a system and method for delivering a neural stimulus to a target neural structure. Neural activity of the target neural structure may be measured or detected after delivering the stimulus. In the context of this application the neural stimulus is an electrical stimulus and the neural activity is an electrical signal from from the target neural structure. In some examples the neural activity is an evoked neural response which is generated directly in response to the neural stimulus. For example, the neural activity may be an evoked resonant neural response (ERNA) which is a resonant response of the target neural structure to the neural stimulus. In order to generate an evoked resonant neural response, it is usually necessary for the stimulus to have certain characteristics such as a particular amplitude and / or frequency which causes the target neural structure to generate the resonant response. In other examples the neural activity may be underlying spontaneous neural activity. The underlying spontaneous neural activity may be enhanced or otherwise affected by the neural stimulus. In this application, for ease of reference, the neural activity is sometimes referred to as a “neural response” regardless of whether it is an evoked neural response or spontaneous neural activity.

[0055] In some examples the neural stimulus may be applied through one or more electrodes embedded in or near the target neural structure. The neural stimulus may be a pulse train. A resonant neural response may be achieved by injecting a pulse train through one or more electrodes in or near the target neural structure, wherein the pulse train has a frequency which resonates with the target neural structure.

[0056] In some examples the neural stimulus is therapeutic and in other examples the neural stimulus may be non-therapeutic. For example the energy delivered by the neural stimulus may be below a therapeutic threshold. The neural stimulus may have particular characteristics whichreduce or prevent a stimulation artefact thereby enabling more accurate measurement of the neural response. Measuring the neural response provides useful information about the state of the patient, the location of the electrodes relative to the target neural structure. In some examples, the position of one or more electrodes embedded in the target neural structure and / or waveform characteristics of the neural stimulus such as frequency, amplitude and waveform shape may be modified based upon the detected neural response.

[0057] In some examples, such as for deep brain stimulation (DBS), the one or more electrodes may be implanted in or near a target neural structure in a brain. In some examples the neural response may be an evoked neural response that is generated directly by the neural stimulus. In other examples, the neural response may be, spontaneous neural activity , which occurs where the neural stimulus resonates with and enhances pre-existing or spontaneous neural activity. Such a response is sometimes referred to as high frequency oscillations (HFO).

[0058] In some examples, the neural response which is to be detected may have one or more of the following characteristics which are typical of a neural response in DBS: a sinusoidal signal that decays in amplitude with each cycle or number of cycles, a peak-to-peak voltage amplitude of at least 50 pV peak-to-peak, typically at least 500 pV peak-to-peak, at its maximum; a duration of approximately 20 to 50 ms before the peak-to-peak amplitude decays to less than 50 pV peak-to-peak; and a fundamental frequency of around 300 to 400 Hz.

[0059] While examples of the present disclosure will be discussed with reference to DBS and a resonant neural response, the methods and systems described herein are not limited to DBS or resonant responses and may be used in other applications where it is desired to be measure neural activity after delivery of a therapeutic or non-therapeutic neural stimulus. For example, the techniques described herein may be applied to other types of neuromodulation, including types of nerve tissue other than brain tissue and to non-resonant responses.

[0060] FIG. 1 shows a system 100 for stimulating a target neural structure. The neural stimulus may be a stimulus for evoking a neural response. The system 100 comprises a processing unit 102, a signal generator 104 and a measurement circuit 106. The processing unit 102 is arranged to control the signal generator 104 to generate a neural stimulus in the form of a pulse train 108. The pulse train is to be delivered into a target neural structure 110 by one or more electrodes 112 which are embedded in or near the target neural structure. The measurement circuit 106 is arranged to receive and detect a neural response 114 of the targetneural structure 110 to the neural stimulus. The neural response may for example be an evoked neural response or underlying spontaneous neural activity.

[0061] The neural response of a target neural structure may be relatively weak and in some cases may have an amplitude which is hundreds of times less than an amplitude of the neural stimulus. Therefore, the measurement circuit 106 may include an amplifier 116 to amplify the response of the target neural structure to the neural stimulus. The amplifier may be implanted in the body of the patient (e.g. the chest) or located externally. The amplifier may be a low power amplifier with a limited range.

[0062] Applying the neural stimulus may create a stimulation artefact which makes it difficult to accurately detect the neural response. For example, the neural stimulus may create a stimulation artefact which has an amplitude having a similar order of magnitude, or even many times greater order of magnitude, than the amplitude of the neural response . The stimulation artefact is a product of the compound evoked action potentials generated by the neural stimulus in the target neural structure or nearby tissue. While the stimulation artefact decays over time, in many cases it may be desirable to measure the neural response of the target neural structure shortly after the neural stimulus in order to obtain an accurate reading. In this case, at the time at which the neural response is measured, the stimulation artefact may have an amplitude which is comparable to, or even many times greater than, the amplitude of the neural response which is to be measured.

[0063] For example, the neural stimulus may have an amplitude of several volts and create a stimulation artefact having an amplitude of several millivolts. Meanwhile, a typical amplitude of the neural response may be only a few microvolts. In many cases it may be desired to record the evoked neural response within a hundred microseconds of the end of the neural stimulus (e.g. after the end of the stimulating pulse train). There is a risk the amplifier may be saturated by the stimulation artefact and unable to accurately detect the neural response.

[0064] Accordingly, the present application discloses methods and systems which may be used to reduce the stimulation artefact. By generating a pulse train having a compensation pulse or compensation current or voltage it is possible to reduce or even cancel out the stimulation artefact. This will be discussed further below with reference to FIG. 4 to FIG. 8B, after describing a specific example of a stimulation artefact, electrode array and target neural structure.

[0065] FIG. 2 shows an example of a recorded signal when a neural stimulus (pulse train 202) is delivered to a target neural structure in the brain. Specifically FIG. 2 shows a neural response to a 130 Hz pulse train 202 delivered from a neurostimulator via an electrode lead, such as the 3387 electrode lead manufactured by Medtronic (Registered TM), implanted in the subthalamic nucleus (STN) of a Parkinson’s disease (PD) patient. The pulse train 202 comprises a plurality of pulses 204. Each pulse 204 of the pulse train evokes a stimulation artefact 206 in the form of an evoked compound action potential (ECAP) and a neural response in the form of evoked resonant neural activity (ERNA) occurring after the ECAP. The ECAP typically occurs within 1-2 milliseconds of the stimulus pulse. The graph shows the response to the last three consecutive pulses 204 of a 60 second period of continuous stimulation followed by a period of no stimulation. It can be seen that the evoked resonant neural response to each of the first two stimulus pulses shown in FIG. 2 is cut short by the onset of the next stimulus pulse, such that only a single secondary peak is detected. However, the evoked neural resonant response 208 to the third (and final) pulse is able to resonate for longer and so can be clearly seen in the form of a decaying oscillation with at least seven peaks for a post-stimulus period of about 30 milliseconds.

[0066] It has been found that by controlling DBS parameters in certain ways, a non-therapeutic stimulus 202 can be administered which produces an evoked resonant neural response 208 in a patient without having any therapeutic impact or causing undesirable side effects. Such non-therapeutic stimuli can be used to reliably measure ERNA without causing sustained changes to the resonant neural circuit (target neural structure) or the patient’s symptomatic state. Non-therapeutic stimulation is preferably achieved by administering a stimulus comprising a short burst of pulses followed by a period of no stimulation, and the ERNA is measured during this period of no stimulation. By doing so, the total charge or energy provided to the patient is below a therapeutic threshold, and the measured ERNA provides information concerning the patient’s natural state (without therapy). In an alternative embodiment, the overall charge or energy provided to the patient may be reduced by reducing the amplitude of the stimulation signal below a therapeutic threshold. However, doing so may also reduce the amplitude of peaks in the ERNA making it more difficult to observe.

[0067] Patterned stimulation can also be used to monitor and analyse evoked resonant neural activity during therapeutic stimulation of a patient. By patterning the stimulation signal, therapeutic stimulation can be maintained whilst providing time windows in which to monitorresonant responses past that of the first resonant peak or more preferably past two or more resonant peaks. For example, nontherapeutic stimulation may comprise delivering a short burst of pulses (e.g. a pulse train comprising a small number of pulses, followed by a much longer period of no stimulation before delivering a second short burst of pulses). In this way the energy delivered within a unit tim may be kept below a therapeutic threshold. In one example, a non-therapeutic stimulus may comprise bursts (pulse trains) of 10 pulses delivered at a frequency of 130 Hz over a l - second time period with the remaining 120 pulses (which would be present during continuous stimulation) skipped.

[0068] FIG. 3A shows an example of an electrode array in the form of an electrode lead tip 300 comprising a first electrode 302, a second electrode 304, a third electrode 306, a fourth electrode 308, a fifth electrode 310, and a sixth electrode 312. Once implanted into the brain, each of the electrodes 302, 304, 306, 308, 310 and 312 may be used to apply a stimulus to one or more target neural structures and / or to monitor and optionally record the evoked neural response (e.g. ERNA) from neural circuits to the stimulus. In other examples, leads with more electrodes or electrodes with different sizes or topologies may be used. In addition, one or more reference electrodes may be located at a remote site and used to complete the electrical circuit when one or more electrodes on the DBS lead are activated for stimulation or used for signal monitoring.

[0069] FIG. 3B shows an example in which the electrode lead tip 300 of FIG. 3B is implanted into a brain at a target structure, in this case the subthalamic nucleus (STN) 316. It will be appreciated that intersecting the electrode tip 300 with the subthalamic nucleus (STN) 316, which has a typical diameter of 5 to 6 mm, can be a very difficult surgical task. Techniques such as stereotactic imaging, microelectrode recordings, intraoperative x-ray imaging, and applying therapeutic stimulation whilst monitoring patient symptoms, are currently used to localise the electrode tip 316. However, these methods can lack accuracy. Additionally, these methods usually require the patient to be awake for the procedure, since voluntary responses from the patient can be used to confirm that the electrode is at a suitable location relative to a target structure in the brain. For this reason, many potential recipients of DBS therapy turn down the option because they are not comfortable with having to be awake during the surgical procedure.

[0070] Instead of using the above methods to locate the electrodes, an alternative method of locating the one or more electrodes in relation to the target neural structure is to use a series ofpatterned stimulations (e.g. a series of pulse trains) to generate and measure an evoked resonant response from a neural target. Such techniques are disclosed in WO 2021 / 022332 which is disclosed herein by reference and can obviate the need for the patient to be awake during the implantation procedure, since an electrode can be located much more accurately at the correct location within the brain and relative to a target neural structure. Some of the example methods and systems disclosed in the present disclosure can further increase the accuracy of locating the electrode(s) by reducing or eliminating the stimulation artefact, thereby enabling the neural response to be measured more accurately.

[0071] The neural stimulus may take the form of a pulse train. The pulse train may comprise a compensation pulse to reduce the stimulation artefact. FIG. 4 shows an example of how a pulse train 410 according to the present disclosure, which includes a compensation pulse. The pulse train 410 may be constructed by combining a plain pulse train 402 with a compensation pulse 404 for reducing or preventing a stimulation artefact after the pulse train and a counter balancing pulse 406.

[0072] The plain pulse train 402 is substantially charge balanced. Charge balanced means that the pulse train 402 as a whole neither imparts a positive charge or a negative charge. The compensation pulse 404 is not charge balanced, but imparts a charge which acts to reduce any stimulation artefact which would otherwise occur after the pulse train has finished. As the compensation pulse 404 is not charge balanced, a counter balancing pulse 406 may be added to counter balance the compensation pulse so that the resultant pulse train 410 is charge balanced as a whole. For instance, the counter balancing pulse 406 is not charged balanced and may impart a charge opposite to the charge imparted by the compensation pulse 404. In this way as the charge of the counter balancing pulse balances with the compensation pulse.

[0073] The process of generating the pulse train 410 may be thought of as adding a compensation pulse 404 and the counter balancing pulse 406 to the plain pulse train 402.While it is possible for the plain pulse train 402, compensation pulse 404 and counter balancing pulse 406 to be generated separately and superimposed on each other, in practice the pulse train 410 may be generated directly by programming or controlling a signal generator.

[0074] FIG. 5 shows an example of the pulse train 502 in more detail. The pulse train 502 includes a plurality of charge balanced multiphase pulses 506, a counter balancing pulse 508 and a compensation pulse 510. The method of stimulating a target neural structure according tothe present disclosure may include delivering a pulse train 502 such as the one shown in FIG. 5 to one or more electrodes implanted in or near a target neural structure.

[0075] The pulse train 502 includes a plurality of charge balanced multiphase pulses 506. Each charge balanced multiphase pulse 506 includes a first phase 506A having a first polarity and a second phase 506B having a second polarity. In some examples the first phase may be a stimulating phase (e.g. a response evoking phase) and the second phase may be a charge recovery phase. In FIG. 5 the charge balanced multiphase pulses are biphasic, but in other examples the charge balanced multiphase pulses may have more than two phases, as long as each pulse is charge balanced.

[0076] The pulse train 502 further comprises a compensation pulse 510. The compensation pulse 510 is not charge balanced and imparts a charge so as to reduce or prevent a stimulation artefact after the pulse train.

[0077] The pulse train 502 further comprises a counter balancing pulse 508. The counter balancing pulses 508 is not charge balanced and imparts a charge opposite to the charge imparted by the compensation pulse. The counter balancing pulse 502 substantially charge balances with the compensation pulse 510 so that the pulse train 502 as a whole is substantially charge balanced.

[0078] The compensation pulse 510 prevents or reduces a stimulation artefact occurring in the period immediately after the pulse train. In this way the pulse train 502 is able to deliver a stimulus which evokes a neural response, but which reduces any stimulation artefact occurring after the pulse train. However, as there is a counter balancing pulse 508, the pulse train 502 as a whole is substantially charge balanced. Over the period of the pulse train the net charged imparted to the target neural structure may be approximately zero. As the pulse train is charge balanced, this enhances the safety of the method and reduces the risk of harming the patient.

[0079] In the example of FIG. 5, the charge balanced multiphase pulses 506 are biphasic. However, in other examples the charge balanced pulses 506 may have more than two phases. In some, examples, there may be a phase gap 528 between adjacent phases 506A, 506B of a pulse as shown in FIG. 5. However the phase gap 528 is shorter in duration than a pulse gap 522 between adjacent pulses of the pulse train.

[0080] In the examples of FIG. 5, the charge balanced multiphase pulses 506 comprise a plurality of biphasic pulses having a first phase 506A of the first (e.g. stimulating) polarity followed by a second charge recovery phase 506B of the second (e.g. charge recovering)polarity. While this order of phases for the charge balanced pulse is most commonly used, in other examples the order could be reversed.

[0081] The charge balanced multiphase pulses 506 may include an inverted pulse (in the illustrated example inverted biphasic pulse 506C). This is referred to as an inverted pulse, as the order of its phases is reversed compared to the other charge balanced multiphase pulses 506. This helps to reduce the stimulation artefact by balancing the distribution of charge over the pulse train. The inverted pulse may be in the second half of the pulse train. In the illustrated example the inverted pulse is the penultimate pulse of the pulse train as this helps to further enhance the reduction in the stimulation artefact by providing a more balanced distribution of charge in the final part of the pulse train.

[0082] While the examples of FIG. 4 and FIG. 5 show a pulse train with only one compensation pulse and one counter balancing pulse, in other examples there may be more than one compensation pulse and / or more than one counter balancing pulse. In that case the one or more counter balancing pulses should charge balance with the one or more compensation pulses. In the example of FIG. 4 and FIG. 5 the compensation pulse 510 and the counter balancing pulse 508 are both triphasic. However, in other examples the compensation pulse(s) and / or the counter balancing pulse(s) may have a different number of phases, for instance a single phase, two phases (biphasic) or have more than three phases etc.

[0083] In the example of FIG. 4 and FIG. 5, the one or more counter balancing pulses 508 impart a charge having the first polarity (the same polarity as the first phase 506A of the charge balanced pulses) and the one or more compensation pulses impart a charge having the second polarity (the same polarity as the second phase 506B of the charge balanced pulses). However, in other examples the polarity of compensation pulses and counter balancing pulses could be reversed.

[0084] In some examples, at least one of the one or more compensation pulses are delivered later in the pulse train than all of the one or more counter balancing pulses 508. This helps to ensure that the compensation pulse reduces the stimulation artefact in the period immediately following the pulse train. In some examples the majority of the one or more compensation pulses 510 are delivered later in the pulse train than the majority of the one or more counter balancing pulses. In some examples all of the one or more compensation pulses 510 are delivered later in the pulse train than all of the one or more counter balancing pulses. Again,this helps to ensure that the compensation pulse reduces the stimulation artefact in the period immediately following the pulse train.

[0085] FIG. 5 shows an example in which the compensation pulse 510 and counter balancing pulse 508 are triphasic. Thus the compensation pulse 510 has a first phase 510A, a second phase 510B and a third phase 5 IOC. There may be a phase gap between some or all of the phases of the compensation pulse. For example in FIG. 5 there is a phase gap 510D between the second phase and the third phase. The phase gap 510D is less than a pulse gap 522 between adjacent pulses of the pulse train 502. Likewise, the counter balancing pulse 508 is triphasic, having a first phase 508A, a second phase 508B and third phase 508C and may have a phase gap 508D between one or more phases of the pulse.

[0086] In the example of FIG. 5 the second phase 510B and third phase 510C of the compensation pulse 510 charge balance with each other. Further, in the example of FIG. 5 the second and third phases of the compensation pulse have the same amplitude as the first and second phases 506A, 506B of the charge balanced pulses. Meanwhile, the first phase 510A of the compensation pulse imparts a charge which acts to reduce the stimulation artefact. The counter balancing pulse 508 has a similar structure with charge balanced second 508B and third phases 508C and a first phase 508A which counter balances the charge imparted by the compensation pulses 510.

[0087] The structure of the compensation pulse and counter balancing pulse in FIG. 5 is just an example and other structure are possible. In some examples the second and third phases of the compensation pulse 510 (and / or counter balancing pulse 508) may not be charge balanced and / or may have a different amplitude to the charge balanced pulses 506. In some examples the order of the phases may be swapped around, for instance with the phase imparting the charge being moved to the second or third phase instead of the first phase. In still other examples, several phases of the compensation (or counter balancing pulse) may impart a charge. The overarching principle is that the one or more compensation pulses charge balance with the one or more counter balancing pulses.

[0088] Likewise, the pulse train shown in FIG. 5 is just an example and in other examples the structure of the pulse train may be modified. For example the amplitude or period of some of the pulses may be changed, or the amplitude, period or ordering of the pulse phases may be changed, while still keeping the pulse train charge balanced as a whole. The overarching principle is that the pulse train is substantially charge balanced as a whole. The pulse train maythus include some charge balanced multiphase pulses, at least one compensation pulse and at least one counter balancing pulse.

[0089] There may be a pause or gap (referred to as a period of no stimulation) after the pulse train 502, during which the neural response can be measured. The pause or gap may be followed by another similar pulse train. Thus the stimulus may be delivered in bursts of pulses, each burst of pulses comprising a pulse train as described above.

[0090] FIG. 6 shows an example of the end of the pulse train 502, including the final pulse, which in this example is a compensation pulse 510, followed by a post pulse train shorting period 604 in which the amplifier or measurement circuit is disconnected from the electrode and a response recording period response recording 602 in which the amplifier is reconnected and the neural response detected. In this example, the post pulse shorting period lasts for 1000 ps and the recording period lasts for 100ms. However in other examples these periods could have different lengths of time.

[0091] FIG. 7A shows an example of one variation which is similar to FIG. 5, except there is no inverted charge balanced pulse, and the amplitude of the second phase 510B and third phase 510C of the compensation pulse 510 is greater than the amplitude of the charge balanced multiphase pulses 506.

[0092] In the example of FIG. 7A, the first phase 508A of the counter balancing pulse 508 imparts a negative charge of 5nC, while the second phase 508B and third phase 508C impart charges of negative lOnC and positive lOnC respectively. Therefore, the charge imparted by the counter balancing pulse 508 is negative 5nC. Meanwhile the charge balanced multiphase pulses 506 impart substantially no charge as the first phase 506A of each pulse has a negative charge of lOnC, while the second pulse 506B has an equal but opposite charge of positive lOnC. The compensation pulse 510 has a first phase 510A which imparts a charge of positive 5nC, while the second phase 510B and third phase 510B impart equal but opposite charges of negative 15nC and positive 15nC respectively. Accordingly the pulse train is substantially charge balanced as the counter balancing pulse 508 substantially charge balances with the compensation pulse 510.

[0093] It is to be understood that these charges are by way of example only and different values could be used in other implementations. Likewise the number of pulses of each type could be varied in other implementations.

[0094] While the example of FIG. 5 shows a triphasic compensation pulse and a triphasic counter balancing pulse, in some examples the compensation pulse(s) and / or the counter balancing pulse(s) need not be triphasic and may each have one phase, two phases or more than 3 phases.

[0095] FIG. 7B shows an example of a pulse train in which the compensation pulse 510 and the counter balancing pulse 508 are biphasic. In this case the two phases 510A, 510B of the compensation pulse 510 are not charge balanced with each other and the two phases 508 A, 508B of the counter balancing pulse 508 are not charge balanced with each other. However, the compensation pulse is charge balanced with the counter balancing pulse so that the overall charge imparted by the pulse train is negligible.

[0096] In the example of FIG. 7B the compensation pulse (and the counter balancing pulse) has a first phase 510A having a first (e.g. charge delivery) polarity and a second phase 510B having a second (e.g. charge recovery) polarity. This order of phases works especially well as the second phase may then act to reduce the stimulation artefact caused by the first polarity of the first phase. However, in other examples, the order of the phases could be reversed.

[0097] The charge imparted by each pulse is determined by the sum of the charges imparted by each phase of the pulse. The charge imparted by each phase is equal to the duration or period of the phase multiplied by the amplitude of the phase. Thus the charge imparted by a phase can be modified by increasing or decreasing the amplitude or duration of the phase.

[0098] In the example of FIG. 7B the first and second phases of the counter balancing pulse have the same period. However, the first phase 508A of the counter balancing pulse has a larger amplitude than the second phase 508B of the counter balancing pulse. Similarly, the first and second phases of the compensation pulse have the same period. However, the second phase 510B of the compensation pulse has a larger amplitude than the first phase 510A of the counter balancing pulse.

[0099] In the example of FIG. 7B the second phase of the counter balancing pulse and the first phase of the compensation pulse have the same period and same amplitude as the phases of the charge balanced multiphase pulses. However, in other examples the duration or amplitude of the phases of the compensation pulse and counter balancing pulse may differ.

[0100] It will be noted that there is no inverted charge balanced pulse in the example of FIG.7B. However, in alternative implementations, an inverted charge balanced pulse could be added. Likewise, while the examples shows two charge balanced multiphase pulses, onecompensation pulse and one counter balancing pulse, in other examples the number of each type of pulse may differ. Furthermore, while in FIG. 7B the counter balancing pulse is the first pulse in the pulse train and the counter balancing pulse is the last pulse in the pulse train, in other examples the position of the pulses may be varied and need not be at the beginning or end of the pulse train. While FIG. 7B shows a pulse train with one compensation pulse, one counter balancing pulse and two charge balanced multiphase pulses, it is to be understood that this is an example only and the number of pulses of each type could be varied in other implementations.

[0101] In some examples there may be more than one compensation pulse. FIG. 8A shows an example pulse train 800 in which there are three compensation pulses 820, 822 and 824 at the end of the pulse train. The dotted lines 830 indicate parts of the pulse train including charge balanced multiphase pulses which are not shown in FIG. 8A.

[0102] Having more than one compensation pulse helps to spread out the compensation charge which may help to avoid overcompensation. In many examples, including this one, the one or more compensation pulses may be in the last half of the pulse train, as having the compensation pulses later in the pulse train is thought to enhance their effectiveness in reducing the stimulation artefact after the pulse train. In some examples, the one or more compensation pulses may be in the last third of the pulse train. For instance, in the example of FIG. 8 A, if the pulse train is 10 pulses long, and there are 3 compensation pulses at the end of the pulse train, then the compensation charge is spread out over approximately the last 33% of the pulse train. In some examples the one or more compensation pulses may be in the last 25% of the pulse train.

[0103] The charge imparted by the compensation pulses is spread out over the compensation pulses. In this example, each compensation pulse 820, 822, 824 imparts an equal charge, so the compensation charge is split equally between the compensation pulses. In this case, each compensation pulse imparts approximately 33% of the compensation charge. In other examples, the compensation pulses may have different durations or amplitudes and so contribute different portions of the compensation charge. In this example each compensation pulse is triphasic and has balanced second and third phases 820B, 820C, 822B, 822C, 824B, 824C, and unbalanced first phase 820A, 822A, 824A. However in other examples the structure of the compensation pulses could be different.

[0104] In any event, the plurality of compensation pulses charge balance with the one (or more) counter balancing pulses 810. In this example there is one counter balancing pulse 810and three compensation pulses 820, 822, 824, so the charge imparted by the counter balancing pulse is of opposite polarity and equal magnitude to the total charge imparted by the 3 compensation pulses. In this example, the pulses have equal periods, and the charge recovery phases 820A, 822A, 824A of the compensation pulses have equal amplitudes, so the counter balancing pulse 810 has a first (e.g. charge delivery) phase 810A with an amplitude X which is three times the amplitude X / 3 of the second (e.g. charge recovery) phases 820A, 822A, 824A of the compensation pulses. In this example the counter balancing pulse 810 is triphasic and has an unbalanced first phase 810A and balanced second and third phases 81 OB, 810C. However in other examples the structure of the counter balancing pulse could be different.

[0105] FIG. 8B shows a further example, which is similar to FIG. 8A and in which like reference numerals denote like parts. However, in contrast to FIG. 8A, in FIG. 8B the compensation pulses 820, 824 and counter balancing pulses 810, 812 are interspersed.Whereas, in the previous examples, the counter balancing pulses were in a first part of the pulse train and the compensation pulses in a second part of the pulse train, in this example one or more of the compensation pulses (e.g. 820) is prior to one or more of the counter balancing pulses (e.g. 812). In most cases, the last pulse of the pulse train (here 824) will be a compensation pulse, as this helps to better reduce the stimulation artefact in the period immediately following the pulse train.

[0106] In some examples, a compensation pulse may be a multiphase pulse and the last phase 820C, 824C of the compensation pulse may have the second (charge recovery) polarity. This approach helps to avoid over compensation, as the compensation pulse has multiple phases, while also enhancing the compensation effect which reduces the stimulation artefact, as the last phase of the compensation pulse has the second polarity (which is opposite to the first polarity).

[0107] In some examples according to the present disclosure, the pulse train has a frequency of at least 5Hz. This can help to evoke a resonant response. In some examples the pulse train has a frequency of 60-200 Hz, which is especially suitable for deep brain stimulation. In some examples the pulse train has a frequency of around 130 Hz. The frequency of the pulse train may be measured as the inverse of the average time from the start of one pulse to the start of the next pulse; for example, from the onset of the first phase of one pulse to the onset of the first phase of the next pulse. The pulse train may have a substantially constant frequency meaning that the pulses are substantially equally spaced in time.

[0108] Pulse trains as described above with one or more compensation pulses may reduce residual offsets and allow more reliable recordings to be obtained on the stimulating channel without amplifier saturation. The pulse train may be suitable for generation by implantable pulse generator (IPG) devices capable of recording pulse voltage waveforms, even with limited dynamic range.

[0109] The method according to the present disclosure may further comprise, after the pulse train has been delivered, measuring the evoked neural response of the neural structure to the pulse train. The measured evoked neural response may be used to determine whether one or more of the electrodes is in or near a target neural structure, to assess the response of the patient to the neural stimulation, or for other purposes. For example, the method may use a measurement circuit associated with an electrode to detect a neural response of the target neural structure to the neural stimulus. The measurement circuit may comprise using an amplifier to amplify the neural response, for instance as shown in FIG. 1. In some examples, apparatus may be configured to disconnect the measurement circuit and / or amplifier from the electrode prior to delivery of the pulse train and reconnected to the measurement circuit and / or amplifier after delivery of the pulse train.

[0110] FIG. 9 shows an example of a system 900 according to an embodiment of the present disclosure. The system may be used to implement any of the methods described above.

[0111] The system 900 comprises a lead tip 910 comprising a plurality of integrated electrodes 912, 914, 916, 918 together with a 920, a signal generator 930, a measurement circuit 940 and an optional multiplexer 950. The 920 comprises a processor 922, e.g. a central processing unit (CPU), a memory 924, and an input / output (I / O) bus 926 communicatively coupled with an input device 928 and output device 929 and one or more of the processor 922 and the memory 924.

[0112] In some embodiments, the multiplexer 950 is provided to control whether the electrodes 912, 914, 916, 918 are connected to the signal generator 930 and / or to the measurement circuit 940. In other embodiments the multiplexer may not be present. For example, the electrodes 912, 914, 916, 918 may instead be connected directly to both the signal generator 930 and the measurement circuit 940. Although in FIG. 9 all of the electrodes 912, 914, 916, 918 are connected to the multiplexer 950, in other embodiments, only one or some of the electrodes 912, 914, 916, 918 may be connected. In some examples the multiplexer 950 may be used to disconnect the measurement circuit 940 from one or more of the electrodes 912-918 prior to delivery of the pulse train and reconnected them to the measurement circuit and / or amplifier after delivery of the pulse train.

[0113] The measurement circuit 940 may include one or more amplifiers and digital signal processing circuitry including but not limited to sampling circuits for measuring neural responses to stimulation, including ERNA. In some embodiments the measurement circuit 940 may also be configured to extract other information from received signals, including local field potentials for measurement of High Frequency Oscillations (HFOs) and the like. The measurement circuit 940 may also be used in conjunction with the signal generator 930 to measure electrode impedances. The measurement circuit 940 may be external to or integrated within the 920. Communication between the measurement circuits 940 and / or the 930 on the one hand and the I / O port on the other may be wired or may be via a wireless link, such as over inductive coupling, WiFi (Registered TM), Bluetooth (Registered TM) or the like. Power may be supplied to the system 900 via at least one power source 960. The power source 960 may comprise a battery such that elements of the system 900 can maintain power when implanted into a patient.

[0114] The signal generator 930 is coupled via the multiplexer 950 to one or more of the electrodes 912, 914, 916, 918 and is operable to deliver electrical stimuli to respective electrodes based on signals received from the 920. To this end, the signal generator 830, the multiplexer 950 and the 920 are also communicatively coupled such that information can be transferred therebetween.

[0115] Whilst the signal generator 930, multiplexer 950, and the 920 in FIG. 9 are shown as separate units, in other embodiments the signal generator 930and multiplexer 950 may be integrated into the 920. Furthermore, either unit may be implanted or located outside the patient’s body.

[0116] The system 900 may further comprise one or more input devices 928 and one or more output devices 829. Input 928 may include but are not limited to one or more of a keyboard, mouse, touchpad and touchscreen. Examples of output devices include displays, touchscreens, light indicators (LEDs), sound generators and haptic generators. Input and / or output devices 108, 110 may be configured to provide feedback (e.g. visual, auditory or haptic feedback) to a user related, for example, to characteristics of ERNA or subsequently derived indicators (such as proximity of the lead tip and one or more electrodes 910 relative to target neural structures. To this end, one or more of the input devices 928 may also be an output device 929 , e.g. atouchscreen or haptic joystick. Input and output 928, 929 may also be wired or wirelessly connected to the 920. Input and output devices 928, 929 may be configured to provide the patient with control of the device (i.e. a patient controller) or to allow clinicians to program stimulation settings, and receive feedback of the effects of stimulation parameters on ERNA and / or HFO characteristics.

[0117] One or more elements of the system 900 may be portable. One or more elements may be implantable into the patient. In some embodiments, for example, the signal generator 930 and lead 910 may be implantable into the patient and the 920 may be external to the patient’s skin and may be configured for wireless communication with the signal generator via RF transmission (e.g. induction, Bluetooth (Registered TM), Wifi etc.). In other embodiments, the 920 , signal generator 930 and lead 910 may all be implanted within the patient’s body. In any case, the signal generator 930 and / or the 920 may be configured to wirelessly communicate with a controller (not shown) located external to the patient’s body.

[0118] In some examples the memory or another non-transitory computer readable storage medium may store instructions which are executable by the processor to control the system to carry out any of the methods described herein.

[0119] One embodiment of the present disclosure provides a system and method for localising the lead tip 900 within a target structure of the brain using measured evoked or stimulated activity such as ERNA and / or HFO activity. During an operation for implantation of the lead tip 900 into the brain, instead of relying on low accuracy positioning techniques as described above to estimate the location of electrodes relative to neural structures within the brain, the system 900 may be used to provide real-time feedback to the surgeon based on characteristics such as the strength and quality of evoked response signals received from one or more electrodes of the lead tip 910. This feedback may be used to estimate position within the target structure in three dimensions and to inform the decision of whether to reposition the electrodes or remove and reimplant the electrodes along a different trajectory.

[0120] The above examples describe a method and apparatus for generating a pulse train and detecting a neural response, while reducing or preventing a stimulation artefact. The above described method and apparatus achieve this by including one or more compensation pulses in the pulse train.

[0121] The method may further comprise setting a charge to be delivered by the one or more compensation pulses based on a measurement arising from one or more calibration pulsesdelivered prior to the pulse train. The measurement may for instance be a current or voltage measurement of a response to the one or more calibration pulses.

[0122] FIG. 10 shows a method 1000 of monitoring neural activity. In particular the method 1000 determines an appropriate compensation current or voltage for reducing the stimulation artefact of a pulse train.

[0123] At block 1002, the method delivers at least one calibration pulse to one or more electrodes implanted in or near a target neural structure. In some examples the calibration pulse is a current pulse, e.g. an electrical pulse having a specified current amplitude.

[0124] At block 1004, the method detects a resultant waveform resulting from the at least one calibration pulse. In some examples the resultant waveform is a voltage waveform.

[0125] At block 1006, the method extracts one or more waveform characteristics from the resultant waveform. The one or more waveform characteristics extracted from the resultant waveform may for example include a plurality of impedance measurements over time based on the at least one calibration pulse and resultant waveform. In som examples, the waveform characteristics extracted from the resultant waveform include a plurality of amplitude measurements of the waveform over a period of time and / or one or more measurements of a shape and / or slope of the resultant waveform over a period of time.

[0126] At block 1008, based on the one or more extracted waveform characteristics, the method determines a compensation current or voltage that is to reduce a stimulation artefact. That is the method determines an appropriate level of compensation or level of charge that will reduce or minimise the stimulation artefact.

[0127] At block 1010, the method delivers into the target neural structure a pulse train comprising a plurality of pulses together with the determined compensation current or voltage so as to reduce a stimulation artefact of the pulse train.

[0128] At block 1012, the method detects a response of the target neural structure to the pulse train.

[0129] Both the resultant waveform at block 1004 and the response of the target neural structure at block 1012 may be detected by an amplifier or a measurement circuit comprising an amplifier, for example using a system as shown in Fig.l or FIG. 9. The compensation current or voltage may be configured to reduce or avoid saturation of the amplifier in a pre -determined period of time after the pulse train has ended.

[0130] For example, at block 1008 the system may determine the compensation current or voltage at a level so as to avoid saturation of the amplifier. The compensation current or voltage is included in the pulse train in block 1010 and in this way saturation of the amplifier may be avoided leading to increased accuracy for the detection and measurement of the neural response at block 1012.

[0131] FIG. 11A shows an example of a method 1100 not in accordance with the present disclosure. In the example method 1100, a plain pulse train 1102 is delivered by one or more electrodes embedded in or near the target neural structure. The plain pulse train 1102 comprises a plurality of charge balanced bipolar pulses. The one or more electrodes detect a neural response to the plain pulse train 1102 which is amplified by an amplifier associated with a measurement circuit. However, due to the stimulation artefact arising after the plain pulse train 1102, the amplifier is saturated. The neural response 1104 is therefore truncated or may not be detectable at all.

[0132] FIG. 1 IB shows an example of a method 1110 in accordance with the present disclosure. The method 1110 involves determining an appropriate level of compensation current or voltage (or compensation charge) based on the waveform resulting from a calibration pulse, as described above for the method 1000 of FIG. 10. In the example of FIG. 11A two compensation pulses are used.

[0133] At 1112 a first calibration pulse 1112 is delivered to the one or more electrodes , this generates a first resultant waveform which is detected. A second calibration pulse 1116 is then delivered to the one or more electrodes so as to generate a second resultant waveform 1118 which is detected.

[0134] The charge delivered by the second calibration pulse 1116 is greater than the charge delivered by the first calibration pulse 1112 . For instance, as shown in FIG. 1 IB the second calibration pulse may have a greater amplitude than the first calibration pulse. In some examples the amplitude of the second calibration pulse 1116 may be determined based on the amplitude of the first resultant waveform. This may help to generate a second calibration pulse 1116 at a level which evokes a second resultant waveform 1118 close to but not exceeding a dynamic range of the amplifier. The second resultant waveform 1118 may then be analysed to determine information used for setting the compensation current or voltage.

[0135] At block 1130, a compensation current or voltage is determined. For example, this may involve determining an appropriate charge to be delivered to reduce or eliminate the1stimulation artefact or to bring the signal level of a detected waveform back into the dynamic range of the amplifier. Such a charge may be delivered as a compensation current of compensation voltage which is to be added to the pulse train. For instance the compensation current or voltage may take the form of a calibration pulse or a phase of a calibration pulse, as described in the earlier examples in this application with reference to FIG. 1 to FIG. 9. The dynamic range of the amplifier is the range of voltages at the input of the amplifier for which the amplifier gain is constant. Outside this range the gain is reduced and the output of the amplifier may be limited to one extreme voltage. Therefore, the dynamic range is the input range over which the amplifier’s function is linear.

[0136] At block 1128 a pulse train including the compensation current or voltage is delivered via the the one or more electrodes. The compensation current or voltage may take the form of a pulse or a phase of a pulse train (e.g. one or more compensation pulses or phases thereof). The pulse train may also include one or more counter balancing pulses in order that the pulse train as a whole is substantially charge balanced. In this example, due to the compensation current or voltage, the stimulation artefact after the pulse train is reduced or avoided and the amplifier is not saturated. Accordingly, the full neural response can be detected at block 1136.

[0137] The pulse train including the compensation current or voltage may have any of the characteristics of the examples described above with reference to FIG. 1 to FIG. 9.

[0138] For example, the compensation current or voltage may be delivered prior to a final pulse of the pulse train. For instance it may delivered, at least in part during the final pulse of the pulse train but prior to a final phase of the final pulse. The compensation current or voltage may be delivered in a multiphase pulse which is charge unbalanced, and wherein the pulse train as a whole is substantially charge balanced. The pulse train in block 1010 or 1128 may comprise a plurality of bipolar pulses and at least one triphasic pulse and wherein the compensation current or voltage is a phase of the at least one triphasic pulse, in which case the triphasic pulse is not charge balanced. In some examples, the pulse train may include a first triphasic pulse which is not charge balanced, a second triphasic pulse which is not charge balanced, and the combination of the first triphasic pulse and second triphasic pulse may be substantially charge balanced. In some examples, the second triphasic pulse is a last pulse of the pulse train.

[0139] In some examples, the one or more electrodes may be implanted in or near a target neural structure of a brain and in some examples the detected response may be an evoked resonant neural activity (ERNA).

[0140] In some examples the compensation current or voltage may be determined by utilising the extracted waveform characteristics of the resultant waveform in a predetermined model of the one or more electrodes. For example, the extracted waveform characteristics may be input into the predetermined model and the predetermined model may output a suitable compensation current or voltage based on the extracted waveform characteristics.

[0141] FIG. 12 shows an example of such an approach in which at block 1202 one or more waveform characteristic of the resultant waveform (or resultant waveforms) is input into a predetermined model 1210. At block 1204 the predetermined model 1210 determines an effective surface area of an electrode based on the detected waveform characteristics of the resultant waveform detected at that electrode. At block 1206, the predetermined model 1210 calculates a compensation current or voltage based on the determined effective area and one or more of the detected waveform characteristics.

[0142] In other examples, the predetermined model 1210 may use a different method to determine the recommended compensation current or voltage based on the extracted waveform characteristics. For example, the method model may not estimate the effective surface area of the electrode as an intermediate step. The predetermined model may be an empirical model. In some examples the predetermined model may be a neural network trained by experimentally gathered data matching the detected waveform characteristics with an offset needed to reduce the stimulation artefact.

[0143] At block 1208 the predetermined model 1210 outputs a recommended compensation current or voltage.

[0144] The compensation current or voltage may be included in an output pulse train as one or more compensation pulses or phases thereof. For example, the compensation current may be applied as first phase 510A of the compensation pulse 510 of FIG. 5, FIG 7A, FIG. 7B or as one of the phases 820A, 822A, 824A of the compensation pulses 820, 822, 824 of FIG.8 A or one of the phases of 820A, 824A of the compensation pulses 822, 824 of FIG.8 B.

[0145] In some examples the compensation current or voltage may be determined by utilising the extracted waveform characteristics of the resultant waveform in a predetermined model of the one or more electrodes. For example the extracted waveform characteristics may be inputinto the predetermined model and the predetermined model may outputs a suitable compensation current or voltage based on the extracted waveform characteristics.

[0146] FIG. 13A shows an example of such an approach in which a system 1300 comprises a computing device 1310 which includes a computing device 1310, a signal generator 1320, a plurality of electrodes 1340 which may be implanted in or near a target neural structure 1360, a measurement circuit 1330 and one or more return or reference electrodes 1350. The system may include any of the components shown in FIG. 9. For instance the computing device 1310 may have any of the features of the processing unit show in in FIG. 9 and the measurement circuit 1330 may include an amplifier (not shown).

[0147] The computing device 1310 controls the signal generator 1320 to deliver a calibration pulse 1370 to one 1342 of the plurality of electrodes 1340. The measurement circuit 1330 detects a resultant waveform 1375 at the electrode 1342 produced as a result of the calibration pulse and communicates this to the computing device 1310. The resultant waveform may be measured with respect to one or more return or reference electrodes 1350 which complete the circuit. Two or more calibration pulses and resultant waveforms 1375 may be delivered and measured. In the example of FIG. 13 A, the same electrode 1342 is used for both delivering the calibration pulse and detecting the resultant waveform. That is, the same electrode 1342 is both a stimulation electrode to which the calibration pulse is delivered and a recording electrode at which the resultant waveform is detected. The stimulation electrode may sometimes be referred to as the evoking electrode, especially if the electrode is used to produce ERNA.

[0148] FIG. 13B shows the system 1300 of FIG. 13A delivering a pulse train 1380. The pulse train has one or more compensation pulses based on a compensation current or voltage. The compensation current or voltage is determined based on one or more waveform characteristics extracted from the resultant waveform 1375. Again the same electrode 1342 is used for both delivering the pulse train 1380 and recording the neural response 1390.

[0149] While the electrode 1342 which is used for delivering the calibration pulse and recording the resultant waveform in FIG. 13A is the same as the electrode 1342 for delivering the pulse train and detecting the neural response in FIG. 13B, in other examples different electrodes could be used. For example, it may be assumed that nearby electrodes have similar characteristics and will produce a similar stimulation artefact if they are of similar shape and design; or in some cases physical calculations may be used to predict a difference if the nearby electrode has a different shape or design.

[0150] In the example of FIG. 13C, the pulse train 1380 is delivered through the electrode 1342, but the recording electrode 1344 is a nearby electrode (e.g. an electrode on the same lead, or in the illustrated example an adjacent electrode on the same lead).

[0151] Likewise, different electrodes may be used for delivering the compensation pulse and receiving and recording the resultant waveform, as shown in FIG. 14A, where the stimulation electrode is electrode 1342 and the recording electrode is electrode 1344 (in this example an adjacent electrode). In this case, the calibration pulse should be delivered to both the stimulation electrode 1342 and the recording electrode 1344, preferably at the same time.

[0152] In the example of FIG. 14B, the electrode 1342 is a stimulation electrode for delivering the pulse train and the nearby electrode 1344 is a recording electrode for detecting the neural response. In this case, where the stimulation and recording electrodes are different, the compensation pulse may be delivered to the recording electrode rather than the stimulation electrode.

[0153] Under the control of the computing device 1310, the system 1300 may carry out the method of FIG. 10, FIG. 11B, or FIG. 12. For instance, the computing device 1310 may determine an effective surface area of the stimulation and / or recording electrode based on the detected waveform characteristics of the resultant waveform and / or determine a compensation current or voltage based on the detected waveform characteristics.

[0154] EXAMPLE

[0155] One example implementation of the present disclosure for determining a level of compensation current or voltage to be applied to a pulse train to reduce or prevent a stimulation artefact will now be described. The compensation current or voltage is applied to the pulse train in one or more compensation pulses. This pulse train is designed to evoke a neural response (e.g. ERNA) whilst also attempting to keep the amplifier baseline within an acceptable operating range. FIGS. 5 to 8B show examples of such a pulse train with compensation current (in the form of compensation pulses).

[0156] The level of compensation may be calculated by sending one or more calibration pulses into the target neural structure, detecting waveform characteristics of resultant waveform(s) resulting from the one or more calibration pulses and entering the waveform characteristics into a predetermined model.

[0157] In this example, the method of determining the level of compensation includes delivering a first calibration pulse and a second calibration pulse to one or more electrodes anddetecting first and second resultant waveforms evoked by the calibration pulse (e.g. as shown in blocks 1112 to 1118 of FIG. 1 IB). In some examples the calibration pulses may be delivered to each of the electrodes in turn and resultant waveforms detected for each.

[0158] The first and second calibration pulses may be referred to as first and second impedance tests as they enable the impedance of the electrode in-situ to be inferred from the resultant waveform. The first calibration pulse has a relatively lower amplitude than the second calibration pulse. The first calibration pulse should have a peak-to-peak amplitude well within the amplifier dynamic range, while the second calibration pulse may have a peak-to-peak amplitude which approaches the dynamic range of the amplifier.

[0159] The first calibration pulse may be in the form of a constant-current, biphasic symmetric pulse. It may be delivered to a single electrode (the ‘stimulating electrode’) and record the resultant voltage waveform with all non-stimulating electrodes configured to be disconnected from stimulation and recording pathways.

[0160] The first resultant waveform is detected and one or more waveform characteristics extracted. An amplitude for the second calibration peak may be determined based on the one or more extracted waveform characteristics of the first resultant waveform. For example, the peak-to-peak amplitude of the first resultant voltage waveform may be used to determine the peak-to-peak amplitude of the second calibration pulse. In this way, based on the first resultant waveform, an appropriate amplitude for the second calibration pulse may be determined such that second calibration pulse produces a larger resultant waveform without exceeding the amplifier dynamic range.

[0161] In one example, a first calibration pulse having an amplitude of 200 p A was used and the first resultant voltage waveform was found to have peak-to-peak amplitude of between 100 and 500 mV, depending on load. The second stimulation amplitude was around 1 mA into a 1 k load and the second resultant waveform was around 2 V peak-to-peak (which in this case was -80% of the amplifier dynamic range of + / - 1.25 V at lx gain). However, in other examples the second stimulation amplitude may be smaller for larger load impedances to avoid unintended limiting of the current by the stimulator circuit and / or saturation of the amplifier used to measure the voltage waveform.

[0162] An example of the voltage waveform is shown in FIG. 15. Waveform characteristics for the first and second waveform are extracted (corresponding to block 1006 of FIG. 10). The extracted waveform characteristics, may include (but are not limited to):

[0163] a. Magnitude and Sign of the voltage waveform at various points, which may include for instance (with reference to the sample numbers in FIG. 15):i. the last pre-stimulus sample (Sample 5)ii. the beginning of the first phase (Sample 6)iii. the end of the first phase (Sample 11)iv. the beginning of the inter-phase gap (Sample 12)v. the end of the inter-phase gap (Sample 17)vi. the beginning of the second phase (Sample 18)vii. the end of the second phase (Sample 23)viii. the first post-stimulus sample (Sample 24)

[0164] b. Slew Rate or Slope of the voltage waveform at one or more of :i. the beginning of the first phase (Sample 7 - Sample 6)ii. the end of the first phase (Sample 11 - Sample 10)iii. during the inter-phase gap (Sample 13 - Sample 12; Sample 17 - Sample 16) iv. the beginning of the second phase (Sample 19 - Sample 18)v. the end of the second phase (Sample 23 - Sample 22)

[0165] c. Slew Rate or Slope of the voltage waveform relative to its magnitude and sign at one or more of:i. the end of the first phase (Sample 11 - Sample 10 / Sample 11)ii. the beginning of the second phase (Sample 19 - Sample 18 / Sample 19)iii. the end of the second phase (Sample 23 - Sample 22 / Sample 23)

[0166] d. The relative change in impedance between the first and second impedance tests at one or more of:i. the end of the first phase (Sample 11 (2nd test) / Sample 11 (1st test))ii. the end of the second phase (Sample 23 (2nd test) / Sample 23 (1st test))

[0167] While in the example above, a resultant voltage form was detected, in other examples a resultant current waveform may be detected and waveform characteristics extracted from the resultant current waveform.

[0168] The one or more extracted waveform characteristics may be entered into a predetermined model to determine a recommended compensation current or voltage. The model estimates a compensation current which when added to a pulse train (which may be a pulse train for evoking a neural response such as ERNA) will allow evocation and detection of theneural response, whilst also keeping the amplifier baseline within an acceptable operating range.

[0169] The predetermined model may estimate the effective surface area of the electrode as an intermediate step in determining the compensation current or voltage. The predetermined model may thus include a first model which estimates the effective surface area for a given electrode and a second model which determines the recommended compensation current or voltage based on the estimated effective area and extracted waveform characteristics of the first and / or second resultant waveform.

[0170] Various models are possible. In some examples an empirical approach may be used to model the response of the electrode in the biological tissue (e.g. target neural structure) to various applied electric stimulus and resulting stimulation artefact and predict a suitable compensation current to reduce or cancel the stimulation artefact. One way in which the predetermined model may be constructed is described in the following.

[0171] First the capacitive element of the load (cap) is estimated or derived from the slope of the voltage waveform during the first phase of the second resultant voltage waveform. The resistive element of the load (imp) is estimated or derived from the voltage of the voltage waveform at the start of the first phase of the second resultant voltage waveform. Based on the load (cap) and the load (imp), the maximum achievable dynamic range of the (bio) amplifier is then estimated, expressed as a fraction of its theoretical maximum dynamic range. In one example, the maximum dynamic range in the test conditions was 4920.2 pV, and so the maximum achievable dynamic range was estimated to vary between 0 and 4920.2 pV, depending on load characteristics.

[0172] FIG. 16 shows a resultant voltage waveform arising from a load value with a moderate value of load capacitance (cap). It can be observed that the amplifier is saturated for approximately 20 milliseconds.

[0173] FIG. 17 shows the resultant voltage waveform arising from a load value with a high value of load capacitance (cap). It can be observed that the amplifier is saturated for approximately 30 milliseconds. In general, it was observed that larger values of resistive load impedance (imp) result in a small decrease in the maximum achievable dynamic range, whereas larger values of capacitive load impedance (cap) result in a large decrease in the maximum achievable dynamic range.

[0174] The radius of the curve described by the slope of the first phase of the second resultant voltage waveform was estimated. FIG. 16 shows a resultant voltage waveform arising from a load value with a moderate value of load capacitance (cap). It can be observed that the slope of the first phase is approximately linear. FIG. 17 shows a resultant voltage waveform arising from a load value with a high value of load capacitance (cap). It can be observed that the slope of the first phase includes some curvature. In general, it was observed that for larger values of capacitive load impedance (cap), the curvature of the first and second phases increases.

[0175] Taking a Boston Vercise Cartesia DBS lead as an example. For that type of lead, the actual electrode surface areas are 1.5 mm2for the segmented electrodes and 6 mm2for the tip and ring electrodes. However, the effective surface area may decrease when placed in biological tissue due to the presence of fluids or air bubbles. Also, the effective surface area may be increased in the case of roughening or pitting of the electrode surface, either during manufacturing or following implantation. Therefore, the effective surface area may be estimated based on the characteristics of a resultant waveform resulting from a known stimulation (referred to as a calibration pulse).

[0176] FIG. 18 shows an example of a model for estimating the effective surface area. In the model, : z200w is the first resultant voltage waveform divided by the first calibration pulse amplitude (e.g. 200 pA), zamp is the second calibration pulse amplitude, zw is the second resultant voltage waveform divided by the second calibration pulse amplitude, c is a vector of parameters for the fit, x new **2 is the estimated effective surface area (SA) of the single electrode (SA = x_new^2 ). A description of all terms in the model is given in the table shown in FIG.s 19A to 19B. Optimal coefficients c[-l] to c

[0031] were calculated to fit empirically gathered data and further modelling performed to check the validity of the model. It was found that the model for estimating the effective surface area performed well across all tested conditions in predicting the surface area of the electrode, with an accuracy of approximately ±1 mm2, as shown in FIG. 20.

[0177] A model for the recommended compensation current was empirically derived from observations of achievable Minimum Dynamic Range (MDR) for ranges of injected current in varying test conditions. Minimum Dynamic Range (MDR) was defined as twice the minimum distance from the amplifier baseline to the amplifier operating limit in a time window of 2.5 to 20 milliseconds following the last 1100 ps post-pulse shorting period. For the ERNA signal of interest, the requirement was set to + / -1500 pV , corresponding to approximately twice thevalue of the largest ERNA signal (1539 pV peak-to-peak) that had been observed in a historic dataset. The average amplitude of the ERNA signal has been 500 pV peak-to-peak.

[0178] FIG. 21 shows example recordings of amplifier baseline in two recording configurations. Minimum Dynamic Range (MDR) is defined as twice the minimum distance from the amplifier baseline to the amplifier operating limit. The blue line (which is highest at the left most side if the graph) demonstrates an example recording of amplifier baseline when a bullet electrode (El) is the stimulation and recording electrode. Within the shaded area of interest (2.5 to 20 milliseconds) the baseline is maximally offset from zero at 2.5 ms, with a voltage of approximately 3000 pV. The voltage difference between this offset and the amplifier operating limit is approximately 1920 pV, and is denoted as MDR El / 2. Therefore MDR El is 2 x 1920 pV = 3840 pV. Similarly, the red line demonstrates an example recording of amplifier baseline when a segmented electrode (E4) is the stimulation and recording electrode. Within the shaded area of interest (2.5 to 20 milliseconds) the baseline is maximally offset from zero at around 5 ms, with a voltage of approximately 2000 pV. The voltage difference between this offset and the amplifier operating limit is approximately 2920 pV, and is denoted as MDR E4 / 2. Therefore MDR E4 is 2 x 2920 pV = 5840 pV.

[0179] FIG. 22 shows a series of observations for Minimum Dynamic Range / 2 for a range of injected current for four test conditions, further segregated by the observed load impedance values. Each dataset is superimposed by a fitting of the equation y = c -|m*x - b|. Figure 22 thus collates a series of observations of Minimum Dynamic Range / 2 for a range of injected current for eight test conditions, comprising 2 sizes of electrode (bullet vs. segment), 2 submersion depths (minimally vs. fully), and 2 separation distances (minimal vs. maximum). It can be seen that the data points for each test condition approximate a triangle shape, in the form of y = -abs( m * x - b ) + c , where m is the slope gradient, b is the horizontal shift, and c is the vertical apex ( Minimum Dynamic Range / 2 ), which should be the theoretical maximum dynamic range achievable, e.g. -4920.2 pV for the bioamplifier. The achieved Minimum Dynamic Range / 2 decreases when the injected current is either decreased or increased about -b . The value for Injected Current that produces a maximum Minimum Dynamic Range / 2 varies across test conditions. Additionally, the slope about the maximum Minimum Dynamic Range / 2 varies for each test condition: as can be seen by the larger width of the triangle shape described by the data for a bullet electrode, leads fully submerged, at maximum leadseparation.

[0180] FIG. 23 is derived from the same dataset as for FIG. 22, but with the value of the x-axis being multiplied by the estimated Load Impedance (using the impedance value computed from the sample of phase 1 of the corresponding impedance test voltage waveform (i.e., sample 11) to yield a value of Estimated Voltage Injected. It can be observed that the slopes for each test condition are now approximately equal, suggesting that the differing Load Impedance values across test conditions has a contribution to the slope of the relationship between Minimum Dynamic Range / 2 and Injected Current in FIG. 22.

[0181] FIG. 23 shows a series of observations for Minimum Dynamic Range / 2 for a range of injected voltage for four test conditions. Each dataset is superimposed by a fitting of the equation y = c -|m*x - b|, however, the gradient m is set to be equal amongst all conditions.

[0182] The model for the recommended compensation current solves the relationship MDR = - abs( m * x - xadj ) + c , where MDR is the achievable Minimum Dynamic Range, m is the slope of the relationship in FIG. 22, x is the Estimated Voltage Injected during empirical observations, xadj is the voltage difference from 0 mV, and c is the maximum achievable dynamic range of the bioamplifier, expressed as a fraction of its theoretical maximum dynamic range.

[0183] FIG. 24 shows a series of observations for Minimum Dynamic Range / 2 for a range of values of abscissa adjusted in the solved relationship for MDR. FIG. 24 uses the same dataset as for FIG. 22 and FIG. 23, but with the value of the x-axis being the solved value of (m * x -xadj) in the formula above. It can now be seen that the slope and y -intercept are substantially equivalent for Minimum Dynamic Range / 2 versus the adjusted abscissa across the four test conditions. It is now possible to solve the relationship for a value of (m * x - xadj ) = 0, thereby achieving the maximum value of Minimum Dynamic Range / 2 .

[0184] A generalised model that solves for xadj can be expressed as:Xadj = ( linj * -Z Pie ) 0.75 + ( Co * SA ) + ( Cl * Zpie * SA ) + ( C2 / ( C23 * SA )) ... where, xadj is the voltage difference from 0 mV, linj is the Injected Current during empirical observations Zpie is the Voltage at the end of the first phase in the second resultant voltage waveform ( Figure 1 , Sample 11) SA is the estimated Effective Surface Area for the given electrode, and Co ... C23 are parameters to the fit.

[0185] For convenience, the model may be rearranged to solve for recommended compensation current , in the form:Recommended compensation current = - numerator / denominator

[0186] where, numerator = (c[0] * sa) + (c[l] * zw[ll] * sa) + (c[2] / (c

[0023] + sa)) + (c[3] / (c

[0024] + zw[ll])) + (c[4] * zw

[0023] ) + (c[5] * zw[6]) + (c[6] * zw

[0018] ) + (c[7] * zw

[0012] ) + (c[8] * zw

[0017] ) + (c[9] * zw

[0023] * sa) + (c

[0010] * zw[6] * sa) + (c[l 1] * zw

[0018] * sa) + (c

[0012] * zw

[0012] * sa) + (c[l 3] * zw

[0017] * sa) + (c

[0014] * Math.Pow(sa, 2)) (c[l 5] / Math.Pow(r, 2)) + (c

[0016] / zw

[0023] ) + (c

[0017] / zw[6]) + (c[l 8] / zw

[0018] ) + (c

[0019] / zw

[0012] ) + (c

[0020] / zw

[0017] ) + (c

[0021] / Math.Pow(zw[l 1] + zw

[0023] , 2)) + (c

[0022] / Math.Pow(zw[6] + zw

[0018] , 2)) + c[A2];anddenominator = Math.Pow(-zw[l 1], 0.75);

[0187] where zw is the second resultant voltage waveform divided by the amplitude of the compensation pulse (second stimulus amplitude), sa is the estimated Effective Surface Area for the given electrode, r is the radius of the curve described by the slope of the first phase of the second resultant voltage waveform, and c[0] ... c

[0023] are parameters to the fit.

[0188] FIG. 25A shows the numerator expressed as non_current_params, while FIG. 25B shows the denominator expressed as current_params.

[0189] Once both the numerator and denominator have been calculated, the recommended compensation current can be determined from the equation:Recommended compensation current = - numerate r / denominator

[0190] A description of all terms in the model for recommended mpensation current is given in FIGS 26A to 26B.

[0191] FIG. 27 shows verification of the model for a resistive load. Here, the Observed Minimum Dynamic Range is plotted against the Estimated Minimum Dynamic Range. The observation substantially agrees with the estimate, and the goodness -of-fit R 2 value is 0.9071.

[0192] FIG. 28 shows verification of the model for all data gathered during empirical modelling, using a range of resistive and capacitive loads. Here, the Observed Minimum Dynamic Range is plotted against the Estimated Minimum Dynamic Range. The observation somewhat agrees with the estimate, and the goodness -of-fit R 2 value is 0.3492.

[0193] The recommended compensation current was applied to a pulse train comprising a number of multi-phase charge balanced pulses of equal amplitude as shown in FIGs. 4 and 5. The resulting pulse train with the applied compensation current included a triphasic counterbalancing pulse 508, including a first phase 508A, with a negative amplitude equal to the recommended compensation current, a second phase 508B, with a negative amplitude equal to the desired stimulation amplitude, and a third phase 508C, with a positive amplitude equal to the desired stimulation amplitude. The compensation pulse also included a post-pulse shorting period 508D of 500 ps after the second phase.

[0194] The pulse train further included a number of charge balanced multiphase pulses, in this case pulses 2 through 8 which were bi-phasic and comprised a first phase 506A with a negative amplitude equal to the desired stimulation amplitude and a second phase 506B with a positive amplitude equal to the desired stimulation amplitude. There was a post-pulse shorting period of 500 ps after the second phase of each charge balanced biphasic pulse. The pulse train further included an inverted pulse, in this example pulse 9, which was bi-phasic and had phases that were opposite polarity to pulses 2 through 8. The inverted pulse included a first phase, with a positive amplitude equal to the desired stimulation amplitude, a second phase, with a negative amplitude equal to the desired stimulation amplitude and there was a post-pulse shorting period of 500 ps after the second phase.

[0195] The pulse train further comprised a compensation pulse 510, in this example pulse 10 which was tri-phasic and comprised a first phase, 510A with a positive amplitude equal to the recommended compensation current, a second phase 510B, with a negative amplitude equal to the desired stimulation amplitude and third phase 510C with a positive amplitude equal to the desired stimulation amplitude. There was a post-pulse shorting period 604 of 1100 ps after the second phase of the compensation pulse (see FIG. 6).

[0196] The pulse train thus included a triphasic compensation pulse imparting the recommended compensation current in a first phase of the compensation pulse and a triphasic counter balancing pulse imparting a charge equal but opposite to the recommended compensation current in a first phase thereof. However, it is to be understood that the compensation current may be applied in other ways, for example but not limited to the methods described with reference to FIG. 6 to FIG. 8B.

[0197] Prior to delivery of each pulse of the pulse train the amplifier (which may be part of a measurement circuit) may be disconnected from the recording pathway so as to avoid the presence of stimulation artefact entering the amplifier. Following delivery of each pulse of the pulse train, the amplifier may be reconnected to the recording pathway so as to provide a grounding terminal for the otherwise floating amplifier inputs. The amplifier may also bedisconnected from the recording pathway after the final pulse of the pulse train for a post -pulse shorting period 604, e.g. as shown in FIG. 6. Following the completion of a post -pulse shorting period 604 at the end of the pulse train, the amplifier may be reconnected and recording of the neural response initiated.

[0198] While the above disclosure has been described with reference to several examples, it is to be understood that the scope is not limited to the examples and variations are possible while still remaining within the scope of the appended claims.

Claims

CLAIMS1. A method of stimulating a target neural structure, the method comprising:delivering a pulse train to an electrode implanted in or near the target neural structure, the pulse train comprising:one or more compensation pulses which are not charge balanced and impart a charge so as to reduce or prevent a stimulation artefact after the pulse train; andone or more counter balancing pulses which are not charge balanced and impart a charge opposite to the charge imparted by the one or more compensation pulses;wherein the one or more counter balancing pulses substantially charge balance with the one or more compensation pulses so that the pulse train as a whole is substantially charge balanced.

2. The method of claim 1, wherein the pulse train further comprises a plurality of charge balanced multiphase pulses, each charge balanced multiphase pulse including a first (e.g. charge delivery) phase having a first polarity and a second (e.g. charge recovery) phase having a second polarity.

3. The method of claim 2, wherein for the charge balanced multiphase pulses, a phase gap between adjacent phases is shorter in duration than a pulse gap between adjacent pulses of the pulse train.

4. The method of any one of claims 2 or 3 wherein the one or more counter balancing pulses impart a charge having the first polarity and the one or more compensation pulses impart a charge having the second polarity.

5. The method of any one of claims 1 to 4 wherein at least one of the one or more compensation pulses is delivered later in the pulse train than all of the one or more counter balancing pulses.

6. The method of any one of claims 1 to 5, further comprising setting a charge to be delivered by the one or more compensation pulses based on a measurement arising from a response to a calibration pulse delivered prior to the pulse train.

7. The method of any one of claims 1 to 6 wherein the one or more compensation pulses are delivered later in the pulse train than the one or more counter balancing pulses.

8. The method of any one of claims 1 to 7 further comprising, after the pulse train has been delivered, measuring neural activity of the target neural structure.

9. The method of claim 8 wherein the neural activity is a neural response evoked by the pulse train.

10. The method of claim 8 wherein the neural activity is spontaneous neural activity by the target neural structure.

11. The method of any one of claims 1 to 10 wherein the pulse train has a frequency of at least 5Hz.

12. The method of any one of claims 1 to 11 wherein the one or more electrodes are implanted into a target neural structure in a brain.

13. The method of any one of claims 1 to 12 wherein the pulse train comprises a plurality of charge balanced multiphase pulses having a first phase of the first polarity followed by a second phase of the second polarity and an inverted pulse in a second part of the pulse train which has a first phase of the second polarity followed by a second phase of the first polarity.

14. The method of claim 13 wherein the inverted pulse is the penultimate pulse of the pulse train.

15. The method of any one of claims 1 to 14 wherein the one or more compensation pulses are in the last 25% of the pulse train.

16. The method of any one of claims 1 to 15 wherein the last pulse of the pulse train is a compensation pulse.

17. The method of any one of claims 1 to 16 wherein the compensation pulse is a multiphase pulse and the last phase of the compensation pulse has the second polarity.

18. The method of any one of claims 1 to 17 further comprising using a measurement circuit associated with the one or more electrodes to detect neural activity of the target neural structure after delivering the pulse train, wherein the measurement circuit comprises an amplifier for amplifying the neural activity and wherein the amplifier is disconnected from the one or moreelectrodes prior to delivery of the pulse train and reconnected to the one or more electrodes after delivery of the pulse train.

19. An apparatus comprising:a signal generator;a measurement circuit; anda processing unit configured to control the signal generator to perform the method of any one of claims 1 to 18.

20. A non-transitory computer readable storage medium storing instructions which are executable by a processor to perform the method of any one of claims 1 to 18.

21. A method of monitoring neural activity, the method comprising:delivering at least one calibration pulse to one or more electrodes implanted in or near a target neural structure;detecting a resultant waveform resulting from the at least one calibration pulse; extracting one or more waveform characteristics from the resultant waveform; based on the one or more waveform characteristics determining a compensation charge, current or voltage that is to reduce a stimulation artefact;delivering into the target neural structure a pulse train comprising a plurality of pulses together with the determined compensation current or voltage so as to reduce a stimulation artefact of the pulse train; anddetecting neural activity of the target neural structure after delivering the pulse train.

22. The method of claim 21 wherein the resultant waveform and the response of the target neural structure to the pulse train are detected by an amplifier and wherein the compensation charge, current or voltage is configured to reduce or avoid saturation of the amplifier.

23. The method of claim 21 or 22 wherein the at least one calibration pulse comprises a first calibration pulse and a second calibration pulse which generate first and second resultant waveforms, wherein a charge delivered by the second calibration pulse is greater than the charge delivered by the first calibration pulse.

24. The method of any one of claims 21 to 23 wherein the one or more waveform characteristics extracted from the resultant waveform include a plurality of impedance measurements over time based on the at least one calibration pulse and resultant waveform.

25. The method of any one of claims 21 to 24 wherein determining a compensation current or voltage comprises utilising the extracted waveform characteristics of the resultant waveform in a predetermined model of the one or more electrodes which takes the extracted waveform characteristics as an input and outputs a recommended compensation charge, current or voltage based on the extracted waveform characteristics.

26. The method of claim 25, wherein method uses the predetermined model to:determine an effective surface area of an electrode based on the detected waveform characteristics; andcalculate a compensation current or voltage based on the determined effective area and one or more of the detected waveform characteristics.

27. The method of claim 25 or 26 wherein the predetermined model is configured to determine a compensation current or voltage which reduces saturation of an amplifier in a pre-determined period of time after the pulse train has ended.

28. The method of any one of claims 21 to 27 wherein the compensation charge, current or voltage is a pulse or a phase of a pulse of the pulse train.

29. The method of any one of claims 21 to 28 wherein the calibration pulse (or each calibration pulse) is a current pulse and the resultant waveform (or each resultant waveform) is a voltage waveform.

30. The method of any one of claims 21 to 29 wherein the compensation current is delivered prior to a final pulse of the pulse train, or is delivered at least in part during the final pulse of the pulse train but prior to a final phase of the final pulse.

31. The method of any one of claims 21 to 30 wherein the pulse train has a frequency of at least 5Hz.

32. The method of claim 21 wherein the pulse train has the features of the pulse train described in any one of claims 1 to 16.

33. The method of any one of claims 21 to 32, wherein the detected neural activity is an evoked neural response of the target neural structure to the pulse train, or spontaneous neural activity of the target neural structure.

34. An apparatus comprising:a signal generator;a measurement circuit; anda processing unit configured to control the signal generator and the measurement circuit to perform the method of any one of claims 21 to 33.

35. A non-transitory computer readable storage medium storing instructions which are executable by a processor to perform the method of any one of claims 21 to 33.

Images