Neural measurement

Abstract

Measuring a neural response to a stimulus comprises applying an electrical stimulus, then imposing a delay during which the stimulus electrodes are open circuited. During the delay, a neural response signal present at sense electrodes is measured with a measurement amplifier, while ensuring that an impedance between the sense electrodes is sufficiently large that a voltage arising on the sense electrode tissue interface in response to the stimulus is constrained to a level which permits assessment of the neural response voltage seen at the sense electrode. For example the input impedance to the measurement amplifier (Z.sub.IN) can be $Z_{\mathrm{IN}}>Z_C\frac{\left(V_{S1}-V_{S2}\right)}{V_E},$
where Z.sub.C is the sense electrode(s) constant phase element impedance, V.sub.s1V.sub.s2 is the differential voltage arising on the sense electrode tissue interface, and V.sub.E is the neural response voltage seen at the sense electrode.

Claims

1. A method for measuring a neural response to a stimulus, the method comprising: applying an electrical stimulus from stimulus electrodes to neural tissue; imposing a delay during which the stimulus electrodes are open circuited; and during the delay, measuring a neural response signal present at sense electrodes with a measurement amplifier, while ensuring that an impedance between the sense electrodes is sufficiently large that a voltage arising on the sense electrode tissue interface in response to the stimulus is constrained to a level which permits assessment of the neural response voltage seen at the sense electrode, wherein an input impedance to the measurement amplifier (Z.sub.IN) is limited by:
Z.sub.IN>AZ.sub.C(V.sub.s1V.sub.s2)/V.sub.E where A is a scalar provided to give sufficient margin of V.sub.E over (V.sub.s1V.sub.s2), Z.sub.C is the constant phase element impedance of the or each sense electrode, V.sub.s1V.sub.s2 is the differential voltage arising on the sense electrode tissue interface in response to the stimulus, and V.sub.E is the neural response voltage seen at the sense electrode.

2. The method of claim 1 wherein A=1.

3. The method of claim 1 wherein A is greater than 0.067.

4. The method of claim 3 wherein A is greater than 0.5.

5. The method of claim 3 wherein A is greater than 1.

6. The method of claim 3 wherein A is greater than 2.

7. The method of claim 1, further comprising providing a sense electrode capacitor in series between the sense electrode and the measurement amplifier, the sense electrode capacitor being chosen to have a capacitance which maintains a desired Z.sub.IN, for a selected duration of the electrical stimulus.

8. The method of claim 1 further comprising obtaining neural measurements repeatedly over time and monitoring for changes in the neural response to a given stimulus.

9. The method of claim 8 further comprising providing feedback control of a therapy delivered to the patient.

10. An implantable device for measuring a neural response to a stimulus, the device comprising: a plurality of electrodes including one or more nominal stimulus electrodes and one or more nominal sense electrodes; a stimulus source for providing a stimulus to be delivered from the one or more stimulus electrodes to neural tissue in order to evoke a neural response; a measurement amplifier for amplifying a neural response signal sensed at the one or more sense electrodes, wherein an impedance between the sense electrodes is sufficiently large that a voltage arising on the sense electrode tissue interface in response to the stimulus is constrained to a level which permits assessment of the neural response voltage seen at the sense electrode, wherein an input impedance to the measurement amplifier (Z.sub.IN) is limited by:
Z.sub.IN>AZ.sub.C(V.sub.s1V.sub.s2)/V.sub.E where A is a scalar provided to give sufficient margin of V.sub.E over (V.sub.s1V.sub.s2), Z.sub.C is the constant phase element impedance of the or each sense electrode, V.sub.s1V.sub.s2 is the differential voltage arising on the sense electrode tissue interface in response to the stimulus, and V.sub.E is the neural response voltage seen at the sense electrode; and a control unit configured to control application of a stimulus to the neural tissue and measurement of an evoked neural response, the control unit configured to apply an electrical stimulus from the stimulus electrodes to neural tissue, the control unit further configured to impose a delay during which the stimulus electrodes are open circuited, and the control unit further configured to, during the delay, measure a neural response signal present at the sense electrodes with the measurement amplifier.

11. The device of claim 10 wherein A=1.

12. The device of claim 10 wherein A is greater than 0.067.

13. The device of claim 12 wherein A is greater than 0.5.

14. The device of claim 12 wherein A is greater than 1.

15. The device of claim 12 wherein A is greater than 2.

16. The device of claim 10, further comprising a sense electrode capacitor in series between the or each sense electrode and the measurement amplifier, the or each sense electrode capacitor having a capacitance which maintains a desired Z.sub.IN, for a selected duration of the electrical stimulus.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

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

(2) FIG. 1 illustrates a prior art approach to neural response measurement;

(3) FIG. 2 illustrates a neural response measurement system in accordance with one embodiment of the present invention;

(4) FIG. 3 illustrates an embodiment of the invention utilising electrode capacitors;

(5) FIG. 4 is another illustration of the embodiment of FIG. 3, showing the stimulus electrode shorting arrangement;

(6) FIG. 5 is a simplified model of the driving circuitry of an implantable device and the surrounding tissue;

(7) FIG. 6 is an illustrative equivalent circuit of the constant phase element at each electrode-tissue interface;

(8) FIG. 7 is a plot produced by a simulation of the model of FIG. 5, showing the artifact arising after a stimulus in the presence of various values of amplifier input impedance, both capacitive and resistive;

(9) FIG. 8 shows experimental data points, and simulation curves, of artefact arising from a stimulus when the amplifier input resistance and capacitance are varied;

(10) FIG. 9 shows the RMS artifact contribution from resistance and capacitance respectively;

(11) FIG. 10 shows artefact variation with resistance and capacitance; and

(12) FIG. 11 shows RMS artefact variation with resistance and capacitance.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

(13) FIG. 2 illustrates a neural response measurement system in accordance with one embodiment of the present invention. Two sense electrodes each having a constant phase element (CPE) impedance of Z.sub.C are used to detect a neural response signal Ve arising in neural tissue of an implant recipient. A stimulus applied by stimulus electrodes of the implant (shown in FIG. 4) gives rise to the neural response, but also causes stimulus voltages V.sub.s1 and V.sub.s2 to be present on the sense electrodes. An input impedance of Z.sub.in is present at each input of the differential measurement amplifier.

(14) The input impedance required in this embodiment of the invention is determined by noting that noise input is comparable to stimulation voltage, and that the goal is for the stimulus to induce a voltage (V.sub.s1V.sub.s2) on the CPE of the sense electrodes which is less than the evoked response V.sub.E. Consequently the desired input impedance is given by:

(15) $Z_{\mathrm{IN}}>Z_C\frac{\left(V_{S1}-V_{S2}\right)}{V_E}$

(16) In one embodiment, being a spinal cord stimulator (SCS) having electrodes with an area of 14 mm.sup.2, Z.sub.c=20, (V.sub.s1V.sub.s2)1V, V.sub.e=50 uV, so that the above equation dictates that the minimum value of Z.sub.in is 400 k. To give a sufficient margin of V.sub.e over artefact, a more desirable value of Z.sub.in is larger, perhaps in the range 1-2 M. In alternative embodiments such as a cochlear implant with electrode area of about 0.1 mm.sup.2, being a fraction of the area of an SCS electrode, the minimum required amplifier input impedance is many times higher; 8 M or for sufficient margin more preferably 20 M, illustrating the difficulties of the resistance values chosen in FIG. 1.

(17) FIG. 3 shows an embodiment of the present invention utilising an ASIC amplifier having a very high value of Zin. Electrode capacitors are provided to block DC insertion to the tissue, the electrode capacitors having a value of C.sub.in=5 pF. Since the ASIC amplifier of FIG. 3 automatically settles to zero during off periods there is no need for resistance to be added at the amplifier input.

(18) FIG. 4 is another illustration of the embodiment of FIG. 2. Electrode capacitors are provided on all electrodes to block DC. The electrode capacitors can store their own charge which in turn can produce uncontrolled current on switch-on. Accordingly, the control module closes the switches to equilibrate the stimulus electrodes prior to each stimulus. The switches are closed only in short bursts so that the equilibration current does not rise to a level which is perceivable by the implant recipient. Similar embodiments may be provided having additional resistance and/or capacitance on the inputs of the measurement amplifier, so long as the input impedance obeys the equation above.

(19) The importance of including the constant phase element model of the electrode-to-tissue interface in FIG. 2 for example arises from a simplified model of the driving circuitry and saline as shown in FIG. 5. The circuit consists of the spreading resistance, being a mesh of resistors that model the current through the bulk saline; the constant phase elements (CPE) where the saline meets the electrode metal; an excitation source having an output impedance including some stray capacitance; loading on each electrode and a ground connection. The saline bath has a bulk voltage point sBath. The saline bath is used to mimic tissue. In FIG. 5 a single-ended measurement can be made between electrodes e1 and e2, and a differential measurement can be made between e2 and e3.

(20) An equivalent circuit of a CPE is shown in FIG. 6. It consists of a set of series RC networks connected in parallel. To adequately model a saline bath, the CPE might have 20-30 RC pairs, but the simplified version of FIG. 6 is shown for understanding. The RC pairs have time constants that change exponentially, in this case by a factor of sqrt(10), however the notable fact is that the time constants of each RC pair are different from all other RC pairs in the CPE. Following a stimulus, the output voltage of a CPE will change over time as charge redistributes between the capacitors, even though no net current is flowing in or out. This property is shared by a single parallel RC network, although a CPE has no R value that can be found at DC.

(21) Unlike an RC network that shows a response characteristic of the circuit, the response of a CPE is dominated by the RC networks that have a similar time constant to that of the length of the stimulation. For example a SCS may have a stimulus pulse width in the range of 100-500 s. This result is important for defining the apparent conductance of a capacitor as discussed below.

(22) Following a stimulus, there are three mechanisms or sources of artifact that can be identified in the circuit of FIG. 5. For each of these mechanisms, the load and current source impedances are considered infinite unless otherwise noted: The voltage on the CPE on electrode 1 changes. This can be seen in a single ended measurement e2-e1, or on the stimulating electrode e1. This is not seen in the differential measurement as this voltage is common mode between e2 and e3. If the current source output impedance is finite, the change in the electrode 1 CPE voltage causes a current to flow through the spreading resistance. This appears differentially on electrodes e2 and e3. This only occurs due to the mesh nature of the spreading resistance; if modelled by a star resistor or a single string of resistors this will not be observed. If the input impedance of either sense amplifier is finite, then during stimulus current will flow into this load. This will then settle.

(23) The ability of the model of FIG. 5 to predict the voltage on e4 was experimentally tested. All stimulation used 4 mA 400 us biphasic pulses. These were used to give rise to an artifact large enough to resolve above noise, and with a voltage on the electrodes that could be digitized without anomaly. This stimulation level delivers 1.6 uC per stimulus, which is in the upper end of the range of charge required for comfort level stimulation in a SCS. Measurements were averaged over 99 iterations. As artefact can take many different profiles of either polarity, a single artefact measure was defined as being the integral of the V.Math.t product of the signal, after resetting the DC value to a baseline.

(24) In addition to experimental verification a simulation of FIG. 5 was conducted. FIG. 7 shows a simulation output showing the artifact over a selected range after the stimulus, in which the y-axis indicates RMS voltagetime, and the x-axis indicates admittance, with admittance of capacitances being calculated as Y=C.Math.t, where t is the stimulus pulse width. Input impedance on the amplifier was selected to be either 330 pF, 1000 pF, 3300 pF, 330 k, and 100 k, giving rise to respective artefact waveforms 702, 704, 706, 708, 710. It is notable that capacitance and resistance give rise to artefact of opposite polarity. Although these are simple waveforms, in practice there can be several sources of artifact with different time-constants so that the actual artefact seen can be more complex than the simple monotic decreasing curves shown.

(25) FIG. 8 shows both experimental data points and simulation curves, where the load resistance and capacitance are varied. The conductance of the capacitors, being their value divided by the length of each phase of the biphasic pulse, is a measure that has the same slope of artefact as for a resistor, and is thus preferred to using the entire length of the stimulus in FIGS. 8 to 11. The simulated line and the experimentally obtained data point groups having a positive slope in FIG. 8 show the effect of adding resistance, while the simulated line and the experimentally obtained data point groups having a negative slope show the effect of adding capacitance to the amplifier input impedance. The slopes of the capacitive and resistive lines are very similar for all electrodes, and closely match that of the simulation, indicating that the model of FIG. 5 is largely correct. The electrodes have different y-intercepts. Electrode 1 (the r1.txt data points) has a peak artifact of 700 uV when a resistive load is reduced, which is a very large artefact and would certainly obscure a neural response signal of around 10 uV. In the absence of loading, artifact can be positive or negative. The y-intercept offsets are outside the control of the electronics, and must be handled by techniques such as filtering.

(26) While the plot of FIG. 8 validates the simulation model, it also shows that there is a missing element that causes artifact in the absence of loading and causes the y-intercept offsets. The y-intercept offsets vary from one electrode to the next, and is perhaps the result of metallic contamination on each electrode surface creating a small galvanic cell and asymmetric behaviour for the phases of the biphasic pulse.

(27) FIG. 9 shows the RMS contribution to simulated artefact from resistance and capacitance respectively.

(28) FIG. 10 shows artefact variation when both resistance and capacitance are progressively changed.

(29) FIG. 11 shows artefact variation with resistance and capacitance using the above described RMS method.

(30) In FIGS. 10 and 11, the curve dips then rises, consistent with FIG. 8. As expected, due to the DC offset, the RMS method obscures the fundamental accuracy of the model.

(31) From the simulation model, using the above described baseline definition of artifact and a 400 us pulse width, the sensitivity of artefact to resistance is 4.110.sup.2 V.Math.s per mho, and the sensitivity of artefact to capacitance is 2.8510.sup.2 Vs per mho. Thus for a load of R, and where the artifact is over a 1 ms interval, then the voltage is
V(r,t)=4.110.sup.2/(Rt)

(32) So for example, for an amplifier input resistance of 100 K and a 1 ms artefact interval:
V(100 k,1 ms)=400 uV

(33) Further, for a capacitive load, and where the artifact is over a 1 ms interval, then the voltage is:
V(C,t)=7.1410.sup.1C/t

(34) So for example for a 1000 pF load, artifact over 1 ms, artifact=71.4 uV.

(35) Using this artefact calculation method, the following table shows the artifact contributions of various stray impedances which might be present in a typical SCS.

(36) TABLE-US-00001 Artifact Contribution Stray Impedance Value for 1 ms in uV Cable 350p 25 input impedance 50k 820 Star load 270k 152 Output impedance of current source 135k 304 Reference inputs to amplifier 83.3K 492

(37) As can be seen in the above table, appropriate adjustment and control of such impedances present in the neural measurement system can allow considerable sources of artefact to be reduced and ease the task measuring a neural signal of the order of 10 uV.

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