Abstract
An implantable device for controlling electrical conditions of body tissue. A feedback sense electrode and a compensation electrode are positioned proximal to the tissue to make electrical contact with the tissue. A feedback amplifier is referenced to ground, and takes as an input a feedback signal from the feedback sense electrode. The output of the feedback amplifier is connected to the compensation electrode. The feedback amplifier thus drives the neural tissue via the compensation electrode in a feedback arrangement which seeks to drive the feedback signal to ground, or other desired electrical value.
Claims
1. A method for controlling electrical conditions of tissue, the method comprising: providing a plurality of electrodes including at least one nominal feedback sense electrode and at least one nominal compensation electrode, the electrodes being positioned proximal to the tissue and being in electrical contact with the tissue; connecting a feedback signal from the feedback sense electrode to an input of a feedback amplifier, and referencing the amplifier to a desired electrical value; and connecting an output of the feedback amplifier to the compensation electrode such that the feedback amplifier drives the tissue via the compensation electrode in a feedback arrangement which seeks to drive the feedback signal to the desired electrical value.
2. The method of claim 1 wherein the desired electrical value is electrical ground referenced to a patient ground electrode distal from the feedback sense electrode.
3. A method for measuring a neural response to a stimulus, the method comprising: providing a plurality of electrodes including at least one nominal stimulus electrode, at least one nominal measurement electrode, at least one nominal feedback sense electrode and at least one nominal compensation electrode, the electrodes being positioned proximal to neural tissue and being in electrical contact with the tissue; applying an electrical stimulus to the neural tissue from the stimulus electrode; connecting a feedback signal from the feedback sense electrode to an input of a feedback amplifier, and referencing the amplifier to a desired electrical value; connecting an output of the feedback amplifier to the compensation electrode such that the feedback amplifier drives the neural tissue via the compensation electrode in a feedback arrangement which seeks to drive the neural tissue to the desired electrical value; and obtaining a measurement of a neural response from the measurement electrode.
4. The method of claim 3 wherein the desired electrical value is electrical ground referenced to a patient ground electrode distal from the feedback sense electrode.
5. The method of claim 3, further comprising disconnecting the feedback amplifier during application of a neural stimulus.
6. The method of claim 5 wherein the feedback amplifier is disconnected by disconnecting the feedback sense electrode from the feedback amplifier.
7. The method of claim 5 wherein the feedback amplifier is disconnected by disconnecting an output of the feedback amplifier from the compensation electrode.
8. The method of claim 3 wherein, during the entire period of stimulation, the feedback amplifier operates and is in connection with the feedback sense electrode and compensation electrode.
9. The method of claim 3 wherein the feedback sense electrode and the measurement electrode are located outside the dipole formed by the stimulus electrode and the compensating electrode.
10. An implantable device for controlling electrical conditions of tissue, the device comprising: a plurality of electrodes including at least one nominal feedback sense electrode and at least one nominal compensation electrode, the electrodes being configured to be positioned proximal to the tissue to make electrical contact with the tissue; a feedback amplifier configured to be referenced to a desired electrical value and to take as an input a feedback signal from the feedback sense electrode, an output of the feedback amplifier being connected to the compensation electrode such that the feedback amplifier is configured to drive the neural tissue via the compensation electrode in a feedback arrangement which seeks to drive the feedback signal to the desired electrical value.
11. The device of claim 10 further configured for measuring a neural response to a stimulus, and further comprising: one or more nominal stimulus electrodes; one or more nominal sense electrodes; a stimulus source for providing a stimulus to be delivered from the one or more stimulus electrodes to neural tissue; measurement circuitry for amplifying a neural signal sensed at the one or more sense electrodes; and a control unit configured to apply an electrical stimulus to the neural tissue from the stimulus electrode and obtain a measurement of a neural response from the measurement electrode.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
[0045] An example of the invention will now be described with reference to the accompanying drawings, in which:
[0046] FIG. 1 illustrates an implantable device suitable for implementing the present invention;
[0047] FIG. 2 illustrates currents and voltages which can contribute to SCP measurements;
[0048] FIG. 3 illustrates the equivalent circuit of a typical system for applying a neural stimulus and attempting to measure a neural response;
[0049] FIG. 4 is an equivalent circuit modelling the tissue/electrode interface and electrode loading;
[0050] FIG. 5a illustrates a virtual ground system configuration, with double-ended measurement, in accordance with one embodiment of the invention, and FIG. 5b illustrates a virtual ground system configuration, with single-ended measurement, in accordance with another embodiment of the invention; FIG. 5c is a model of the embodiment of FIG. 5b; FIG. 5d is an equivalent circuit of the model of FIG. 5c;
[0051] FIG. 6 is a system schematic illustrating multiplexing of virtual ground functionalities across multiple electrodes;
[0052] FIG. 7a illustrates an equivalent circuit of the virtual GND embodiment of FIG. 6;
[0053] FIG. 8a illustrates the operation of the embodiment of FIG. 6 when the virtual ground function is disabled;
[0054] FIGS. 8b and 8c show the operation of the circuit of FIG. 6 when the virtual ground function is activated, when experimentally demonstrated on a bench using a saline bath;
[0055] FIG. 9a illustrates the operation of the embodiment of FIG. 6 when the virtual ground function is disabled, with amplifier blanking; FIG. 9b illustrates the step response of the embodiment of FIG. 6 when the virtual ground function is enabled, with amplifier blanking;
[0056] FIG. 10a illustrates the performance of the embodiment of FIG. 6 when the virtual ground function is disabled, with amplifier blanking and sinusoid injection; FIG. 10b illustrates the step response of the embodiment of FIG. 6 when the virtual ground function is enabled, with amplifier blanking and sinusoid injection;
[0057] FIG. 11a plots the measurements from an electrode array in response to a stimulus delivered by the array to a sheep dorsal column, while FIG. 11b is a superimposed plot of similar data, demonstrating timing of respective signal features;
[0058] FIG. 12 illustrates an alternative embodiment intended for ASIC implementation;
[0059] FIG. 13 illustrates another embodiment intended for ASIC implementation; and
[0060] FIG. 14 illustrates yet another embodiment intended for ASIC implementation.
Description of the Preferred Embodiments
[0061] FIG. 1 illustrates an implantable device 100 suitable for implementing the present invention. Device 100 comprises an implanted control unit 110, which controls application of a sequence of neural stimuli. In this embodiment the unit 110 is also configured to control a measurement process for obtaining a measurement of a neural response evoked by a single stimulus delivered by one or more of the electrodes 122. Device 100 further comprises an electrode array 120 consisting of a three by eight array of electrodes 122, each of which may be selectively used as the stimulus electrode, sense electrode, compensation electrode or sense electrode.
[0062] FIG. 2 shows the currents and voltages that contribute to spinal cord potential (SCP) measurements in a typical system of the type shown in FIG. 3. These signals include the stimulus current 202 applied by two stimulus electrodes, which is a charge-balanced biphasic pulse to avoid net charge transfer to or from the tissue and to provide low artefact. Alternative embodiments may instead use three electrodes to apply a tripolar charge balanced stimulus for example where a central electrode. In the case of spinal cord stimulation, the stimulus currents 202 used to provide paraesthesia and pain relief typically consist of pulses in the range of 3-30 mA amplitude, with pulse width typically in the range of 100-400 μs, or alternatively may be paraesthesia-free such as neuro or escalator style stimuli. The stimuli can comprise monophasic or biphasic pulses.
[0063] The stimulus 202 induces a voltage on adjacent electrodes, referred to as stimulus crosstalk 204. Where the stimuli 202 are SCP stimuli they typically induce a voltage 204 in the range of about 1-5 V on a SCP sense electrode.
[0064] The stimulus 202 also induces electrode artefact. The mechanism of artefact production can be considered as follows. The stimulus crosstalk can be modelled as a voltage, with an equivalent output impedance. In a human spinal cord, this impedance is typically around 500 ohms per electrode, but will be larger or smaller in different applications. This resistance has little effect in the circuit, but is included for completeness. The stimulus crosstalk drives the measurement amplifiers through the electrode/tissue interface. This interface is shown in FIG. 4 as a set of series capacitance/resistance pairs, modelling a component referred to in the literature as a “Warburg element”. The RC pairs model the complex diffusion behaviour at the electrode surface, and have time constants from micro-seconds to seconds. The cables from the electrode to the amplifier add capacitance which loads the electrode, along with the resistive input impedance of the amplifier itself. Typical loading would be 200 pF of capacitance and 1 megohms of resistance. Following this is an ideal amplifier in this equivalent circuit of FIG. 4.
[0065] The electrode artefact is the response of the electrode/tissue interface, when driven by the stimulus crosstalk and loaded by the capacitance and resistance at the amplifier input. It can be observed, either with a circuit simulator or in a laboratory. It can also be observed that the sign of the artefact is opposite for capacitive and resistive loading. Electrical artefact usually also arises from the behaviour of the amplifier circuitry in response to these particular circumstances.
[0066] It is possible to reduce artefact by reducing the loading on the electrode, however in practical situations there are limits to how low this capacitance can be made. Increasing the electrode surface area also decreases artefact but again in practical situations there will be limits to the electrode size. Artefact can also be reduced by adding resistance or capacitance to the amplifier input relying on the opposite sign of the artefact produced by these terms. However, this only works to a limited extent, and changing the size of the electrode changes the size of the required compensation components which makes it difficult to make a general purpose amplifier that can be connected to a range of electrodes. One can also reduce artefact by reducing the size of the stimulus crosstalk, and this is the aim of the virtual ground circuit embodiment of this invention shown in FIG. 5, which relates to evoking and measuring a neural response.
[0067] Referring again to FIGS. 2 and 3, an appropriate electrical stimulus 202 will induce nerves to fire, and thereby produces an evoked neural response 206. In the spinal cord, the neural response 206 can have two major components: a fast response lasting ˜2 ms and a slow response lasting ˜15 ms. The slow response only appears at stimulation amplitudes which are larger than the minimum stimulus required to elicit a fast response. Many therapeutic stimuli paradigms seek to evoke fast responses only, and to avoid evoking any slow response. Thus, the neural response of interest for neural response measurements concludes within about 2 ms. The amplitude of the evoked response seen by epidural electrodes is typically no more than hundreds of microvolts, but in some clinical situations can be only tens of microvolts.
[0068] In practical implementation a measurement amplifier used to measure the evoked response does not have infinite bandwidth, and will normally have infinite impulse response filter poles, and so the stimulus crosstalk 204 will produce an output 208 during the evoked response 206, this output being referred to as electrical artefact.
[0069] Electrical artefact can be in the hundreds of millivolts as compared to a SCP of interest in the tens of microvolts. Electrical artefact can however be somewhat reduced by suitable choice of a high-pass filter pole frequency.
[0070] The measurement amplifier output 210 will therefore contain the sum of these various contributions 202-208. Separating the evoked response of interest (206) from the artefacts 204 and 208 is a significant technical challenge. For example, to resolve a 10 μV SCP with 1 μV resolution, and have at the input a 5V stimulus, requires an amplifier with a dynamic range of 134 dB. As the response can overlap the stimulus this represents a difficult challenge of amplifier design.
[0071] FIG. 5a illustrates a neural stimulus and response system providing differential neural measurements and using a shared electrode for measurement and for feedback sense. Alternative embodiments could use two separate electrodes for measurements and feedback sense, respectively.
[0072] FIG. 5b illustrates a configuration of device 100 for controlling electrical conditions of neural tissue in accordance with another embodiment of the present invention, providing single ended neural measurements. In the configuration of FIG. 5b the device has a current source 502 which drives current into tissue via stimulus electrode 504 in order to stimulate the neural tissue and evoke a neural response. A feedback sense electrode 506, compensation electrode 508 and measurement electrode 510 are also provided. The electrodes 504-510 are positioned proximal to neural tissue to make electrical contact with the tissue. A feedback amplifier 512 is referenced to ground 514 and takes as an input a feedback signal from the feedback sense electrode 506. An output of the feedback amplifier 512 is connected to the compensation electrode 508 such that the feedback amplifier 512 is configured to drive the tissue via the compensation electrode 508 in a feedback arrangement which seeks to drive the feedback signal to ground. This mechanism will thus operate to quash stimulus artefact at the tissue electrode interface, improving measurement conditions for neural response measurement circuitry 516.
[0073] As shown in FIG. 5c, the adjacent stimulus, compensation, feedback sense and measurement electrodes in contact with resistive tissue can be modelled as respective contacts each connected to a tissue rail by a respective resistance R1, R2, R3 and R4.
[0074] An equivalent circuit of FIG. 5c is shown in FIG. 5d. Stimulus and stimulus artefact occurring upon the stimulus electrode creates a current I through R1. Feedback amplifier 512 operates to maintain zero current at each amplifier input, and also operates to maintain the voltage at each input to be identical. Therefore in the configuration of FIGS. 5a-d, the voltage at each amplifier input is zero, because the positive input is referenced to ground. Moreover, current through R3 is forced to zero, being the same as the input current to the amplifier 312. This ensures that there is no voltage differential across R3, and that the tissue node must therefore be forced to ground, in this model. This effect is referred to herein as providing a virtual ground.
[0075] The voltage caused by the current stimulus travels at the speed of light in the tissue medium, whereas an evoked action potential in the neural tissue travels at around 60 m/s. When the feedback sense electrode is subject to (or senses) the evoked response it will cancel the stimulus crosstalk in the tissue, but due to the larger propagation delay, the voltages produced by the evoked response at different electrodes (such as the measurement electrodes) will differ, and can be recorded. It will simply be the voltage that would otherwise by recorded as the difference between the measurement electrode and the feedback sense electrode. Alternatively, the sense electrode can be placed elsewhere in the tissue further from the stimulus electrode(s), and substantially no cancellation of the evoked response will then occur, although the electrode will be subject to other electrical signals in the body from muscles, and other nerve bundles. This might be the situation when the sense electrode is on the body of an implant, with the stimulating electrodes on a lead.
[0076] FIG. 6 is a system schematic illustrating multiplexing of virtual ground functionalities across multiple electrodes, applicable to the embodiment of FIG. 5a or FIG. 5b. This shows the buffer at the output of the reference MUX. A set of electrodes is connected to a current source and a set of amplifiers. The “current source MUX” allows the stimulus current to be directed to any electrode. The “Ground MUX” allows any electrode to be chosen as the second of the pair of stimulating electrodes. Switches “Std” and “VG” allow the circuit to selectively provide conventional stimulation, or stimulation according to this invention. A third multiplexor selects the electrode to be used as the reference point (feedback sense electrode). Once the electrode configuration has been chosen, the circuit operates according to FIGS. 5a-d. Each of the N electrodes of the array of FIG. 6 may thus, at any given time, nominally serve as any one of the stimulus, sense or measurement electrode.
[0077] The circuit of FIG. 6 may alternatively be modified such that the reference voltage passed to the virtual ground feedback amplifier is a combination of the voltages on the measurement electrodes, e.g. the average of two or more electrode voltages.
[0078] In principle, the virtual ground circuit of the embodiment of FIG. 5b provides for the stimulus electrode 504 to be driven by a current source. An op-amp circuit 512 and compensating electrode 508 provides a feedback loop holding the tissue voltage, as measured at a sense electrode 510, to 0 V. In an ideal situation, the voltage on the compensating electrode 508 is identical in amplitude but of opposite polarity to the voltage on the stimulating electrode 504. This ideally leaves the potential on a measurement electrode 510 unchanged by the stimulation and thus significantly improves conditions for measurement of an evoked response with reduced artefact.
[0079] Referring to FIG. 6, it is noted that the components making up the virtual ground circuit are spread throughout the device 100, having components in the reference multiplexer (Ref. MUX). As shown in FIG. 6, an electrode is wired to the inverting input multiplexer. This includes a buffer to drive the capacitance on the inputs of the array of amplifiers, and the virtual ground amplifier. These components, their parasitics, and the capacitance of the wiring on the circuit boards from which the system is made must be considered in order to design a stable circuit.
[0080] FIG. 7a illustrates the actual circuit used for the amplifier in the reference multiplexer (“Ref. MUX”) of FIG. 6, in the preferred embodiment of the virtual ground circuit. The reference signal selected by the MUX is buffered before it is passed to the amplifier negative inputs, and the virtual ground circuit. The buffer uses a current feedback amplifier, because this amplifier is inside the virtual ground feedback loop and this amplifier introduces less phase shift than a voltage feedback device. This has been used both for experimental verification in a saline bath (from which the attached figures were obtained) and in a human subject.
[0081] As shown in FIG. 7b, the virtual ground circuit of this embodiment includes an inverting amplifier and a high-speed buffer. The op-amp alone does not have the current sourcing capability to drive the current available from the current sources, which can deliver up to 50 mA. The 470 p capacitor provides dominant pole compensation to the loop. The switches are paralleled to provide low impedance paths, and match the switch configuration of FIG. 6.
[0082] The FETs Q902 and Q903 in FIG. 7b, from the input to the supplies, provide electrostatic discharge (ESD) protection. The capacitor C900 and pulldown resistor R901 set the DC bias point for the loop.
[0083] FIG. 8a illustrates the problem of stimulus crosstalk when the virtual ground function is disabled. Trace 801 is from Electrode 1 (stimulus electrode), trace 804 is from Electrode 2 (ground electrode), trace 802 is from Electrode 4 (first measurement electrode), and trace 803 is from Electrode 5 (second measurement electrode).
[0084] FIGS. 8b and 8c show the behaviour of the circuit of FIG. 6 when the virtual ground function is activated, when experimentally implemented on a bench using an actual saline bath. FIGS. 8b and 8c show the response to a 10 mA biphasic pulse stimulus in 1/10 PBS (phosphate buffered saline). As can be seen, now the stimulus and feedback electrodes (811 and 814) swing in opposite directions, while the measurement electrodes (812 and 813) mostly stay at ground. The ringing at the measurement electrodes is the response of the feedback loop to the very high slew-rate current source edges, but represents significantly reduced artefact as compared to FIG. 8a.
[0085] FIG. 9a shows artefact with virtual ground disabled, and in particular shows the behaviour of the circuit of FIG. 6 when the virtual ground function is deactivated, when experimentally implemented on a bench using an actual saline bath. In FIG. 9a, the stimulus occurred and concluded at some time prior to t=0.6 ms, at which time the amplifiers were blanked. The amplifiers were then unblanked at t=0.6 ms. FIG. 9 shows the measured voltages on two measurement electrodes, with a peak artefact of around 400 uV.
[0086] FIG. 9b shows the amplifier outputs with virtual ground enabled, but otherwise identical stimulation as produced FIG. 9a. As can be seen, when using virtual ground in the circuit of FIG. 6, artefact is about 100 uV, or smaller by about 75%.
[0087] FIG. 10a shows the amplifier output for the same experiment as FIG. 9a, but with a 50 uV pp sinusoidal signal injected in series with electrode 4, to give an idea of how an evoked response would superimpose upon the artefact. The sinusoidal signal cannot be easily seen before about 1.5 ms, i.e. the first 1 ms of measurement time is obscured by the artefact.
[0088] FIG. 10b is for the same experiment as FIG. 10a but with virtual ground enabled. Here, the sinusoidal signal can be seen distinct from the artefact from around 0.75 ms, or about 750 us (80%) earlier than for FIG. 10a relative to the unblanking time.
[0089] FIG. 11a shows the evoked response in a sheep dorsal column. In particular, FIG. 11a plots the measurements obtained simultaneously from 22 electrodes of a 24 electrode array in response to a stimulus delivered by two adjacent electrodes positioned centrally in the array. As can be seen, evoked responses propagate simultaneously both caudally and rostrally from the central stimulus site. The current required to evoke such a response in a sheep is much lower than in humans, and the evoked response signals are higher, so artefact is less of a problem. In other regards the sheep signals are similar to the human case. In FIG. 11a the amplifiers are unblanked at approximately 0.75 msec and the response finishes within another 0.75 ms. FIG. 11b is a superimposed plot of similar data, demonstrating timing of respective signal features when measuring on multiple electrodes at increasing distance from the stimulus site. FIGS. 11a and 11b illustrate the importance of reducing artefact during the period immediately after stimulation.
[0090] In a first mode of operation in accordance with some embodiments of the invention, at the end of stimulation the stimulation electrodes are both disconnected. The bath (or subject) is floating at this point, as there is no connection between the bath and the circuit ground. Since the amplifiers are all differential, taking the difference between the reference electrode and the other epidural electrodes will compensate for any change in voltage. This mode of operation reflects the logic that other choices of which electrode to ground seem likely to worsen artefact: connecting a stimulation electrode will cause the bath potential to change as the electrode voltage settles; connecting an epidural electrode to GND might put a transient on it which would be seen on all the channels.
[0091] In a second mode of operation of other embodiments of the invention, the VG circuit remains active after the stimulation, which makes the bioelectrical situation quite different. The voltage on the compensation electrode will change as the electrode potentials settle, but the VG loop will compensate for this so it will not affect the bath potential. At the same time, the VG circuit can hold the bath at a fixed voltage—GND. The VG circuit will attempt to keep the epidural space at a static voltage, namely GND.
[0092] In another embodiment, the present invention is implemented in an application-specific integrated circuit (ASIC). The primary difference that is encountered in an ASIC implementation is that whereas most PCB amplifiers and components are intended for split supply operation, most ASIC designs, especially one intended for implantable operation, will operate from a single supply. Also, in an ASIC, the desire to produce a low-cost design is increased, as an ASIC implementation would be preferable for commercial exploitation.
[0093] FIG. 12 shows a design intended for ASIC operation. The design uses separate current sources to provide anodic and cathodic current. These are connected to an electrode array. They also electrodes to be arbitrarily connected to the output and input of the virtual ground amplifier. The point “Vgref” provides the bias point for the amplifier; this would typically be half the power supply. In this case the power supply is called “VDDH”, indicating it is a high-voltage supply suitable for tissue stimulation.
[0094] The switched connections directly to VDDH and GND allow stimulation modes that do not use the virtual ground amplifier. In the design of FIG. 5a, the virtual ground amplifier provides the entire current for the second of the two stimulating electrodes. To create the ASIC implementation some changes were required. In the PCB design of FIG. 5a, the amplifier output provides the entire opposite current to that of the output of the current driver. This requires an amplifier with a considerable output drive, for example if the current source can drive 50 mA, so must the amplifier. An amplifier with this current drive, stable in a feedback loop, and with the required bandwidth can be difficult to obtain, although they are available.
[0095] Thus a problem in an ASIC implementation is to provide the virtual ground amplifier with sufficient current capability to balance the current source; this takes considerable silicon area which incurs cost. Noting that both positive and negative current sources are available in the ASIC, the present embodiment thus uses the circuit of FIG. 13, suitable for integrated circuit implementation. This implementation requires the use of matched positive and negative current drivers. It also operates from a single supply, simplifying the system implementation. The feedback loop holds the tissue voltage at a stable voltage—the midpoint of the two supplies. The amplifier only has to source or sink a current equal to the mismatch between the two amplifiers, denoted dI in FIG. 13. Since the current source has a high output impedance, the load seen by the amplifier is unchanged as compared to the case of FIG. 5a where the amplifier provides all the drive, however the amplifier no longer has to provide high current. In a case where the current sources match to 10%, a 0.1 reduction in capability is achieved.
[0096] FIG. 14 illustrates yet another embodiment, in which the amplifier is connected as a unity gain buffer. It is noted that this proposal may be integrated into the design of Australian Provisional Patent Application No. 2012904838.
[0097] It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. For example while application of the method to neural stimulation is described, it is to be appreciated that the techniques described in this patent apply in other situations involving measurement of a voltage within tissue during or after stimulation.
[0098] The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.
