WO1997006728A1

WO1997006728A1 - Improved recording configuration for the recording of electro-physiological signals (and non-physiological physical measures)

Info

Publication number: WO1997006728A1
Application number: PCT/DK1996/000348
Authority: WO
Inventors: Ronald Raymond Riso; Christian Pflaum
Original assignee: Ronald Raymond Riso; Christian Pflaum
Priority date: 1995-08-16
Filing date: 1996-08-16
Publication date: 1997-02-27
Also published as: AU6655896A

Abstract

The invention concerns a method of interference rejection in electrical recording systems, which rejects the voltage drop IDMV caused by the interference sources Vinterf at Ztl and Zt2, but amplifies the signals uncorrupted. A recording system according to the invention consists of the recording inputs (E1, E2, E3, and E4, ... En, respectively) and an output (6). An input stage (7) and a computing stage (9) provide a transfer function Vout = f(V1, V2, V3, and V4, ... Vn, respectively), which exhibits the above-mentioned interference rejection properties.

Description

Improved Recording Configuration for the Recording of Electro-Physiological Signals (and Non-Physiological Physical Measures)

Introduction

Usually electro-physiological signals such as Electromy- ography (EMG) and Electroneurography (ENG) are recorded with a single instrumentation amplifier [40] (Fig. 9), which provides a sufficient rejection of common mode voltages (CMV) of interfering electromagnetic sources (V interf. ) , such as that derived from the power mains. In¬ terfering currents (I interf. ) , which cause voltage-drops along the tissue impedances Ztl, Zt2 and Zt3 (Fig. 9), are picked up by the conventional bipolar recording meth¬ ods as difference mode voltage and are thus amplified by the full gain of the amplifier. This type of interfer¬ ence, which we will refer to as Interference Difference Mode Voltage (IDMV) (Fig. 9), has been considered to be not suppressible. However, the recorded bio-electric sig¬ nal is frequently disturbed by a considerable interfer¬ ence difference mode voltage from the power supply, power rectifiers, electrical stimulation systems (FES-systems) and other bio-electric sources.

Our invention, the tripolar configuration, solves this problem of interfering currents by using (in the simplest version) two instrumentation amplifiers [42,43] (Fig. 9, which are connected to three electrodes [2,3,4] (Fig. 9. These three electrodes, (instead of two which are usually employed) , are located with equal distance along the ac¬ tive bio-electric structure such as a muscle or nerve. In the first amplification stage one instrumentation ampli¬ fier [42] is connected to one outer El [2] and the middle electrode E2 [3] and the other instrumentation amplifier [43] is connected to the remaining outer E3 [4] and the middle electrode E2 [3]. The signals from these two bipo¬ lar amplifiers are then subtracted from each other by a subtraction stage [44]. The IDMV is a voltage drop caused by longitudinal cur¬ rents following within the tissue or other structures be¬ neath the pickup electrodes [2,3,4]. Ideally, assuming that the voltage drop (IDMV) along the tissue impedance between the first and the middle electrode is equal to the voltage drop (IDMV) along the tissue impedance be¬ tween the middle electrode and the third electrode, each of the first stage amplifiers will have an identical out- put and the resulting interference will be canceled out by the subtraction stage [44]. However, because the volt¬ age drops along the tissue impedances are never exactly equal, we must supply the subtraction stage [44] with an adjustable weighting network, which can be used to mini^ mize the interference by adjusting the contribution of each input channel to the resulting output of the second stage amplifier.

The properties of the bio-electric signals on the other hand are totally different from that of the Interference Difference Mode Voltage (IDMV) so that they do not cancel out in the second amplification stage. The bio-electric signal, may be considered as a propagating voltage wave which moves along the active bio-electric structure (i.e. muscle or nerve) . The relatively slow movement of the bio-electric potentials (3-5m/s for muscles and 30-70m/s for nerves) beneath the recording electrodes causes a phase-shift between the potentials seen at the two ampli¬ fier input channels. If the distance between the record- ing electrodes is sufficient large to get a substantial phase shift between the inputs to the two amplifiers, the result of the subtracting stage will be a bio-electric signal, and this will have an amplitude which is in the range of a conventional single bipolar amplifier.

These amplifier qualities of the invention, the ability to reject IDMV, while at the same time to provide a sub¬ stantial bio-electric signal output, have been demon¬ strated for the recording of EMG (electromyography) in experiments with human subjects and for ENG (electroneurography) in a series of experiments with ani¬ mal models. In addition, we have demonstrated an im¬ proved S/N ratio (signal to noise ratio) by comparing our new tripolar recording configuration with the conven¬ tional bipolar configuration. The benefits from our con¬ figuration derive from the fact that the amplifier noise from the two first stage amplifiers increases only by a factor of the square root of two (1.4), while the tripo- lar configuration amplifies the potential peaks below the middle electrode by a factor of 2.

This measurement principle is also useful for applica¬ tions beyond the biomedical sector. It could be applied to any kind of recording problem, where external sources of interference cause potential differences between the recording sites, while potential peaks of the signal are traveling relatively slow underneath the pick up site. An example would be in measuring perturbation in pressure within a length of pipe in which there is also a fluid flow.

The underlying principles of this new recording method are described below on a more universal level, in order to increase the scope of application for this new record¬ ing method.

State of the Art

Usually a simple electrical recording system consists of several system inputs and a system output. Two basic electrical recording systems are the monopolar and the bipolar amplifier.

The monopolar amplifier compares the electrical potential at the input (VI) with a reference potential (Vref) and amplifies the difference by the factor G. V_mono

- Vref)

The bipolar amplifier compares the potential of its two inputs (VI and V3) and amplifies the difference by the factor G.

V_bipolar

- V3)

Interference Rejection

Figure 1 (derived from Fig. 9) shows a simplified elec¬ trical model, which illustrates the influence of an in¬ terference source (V interf.) and of the bio-electric signal sources (V signall, V signal2, V signal3) on the potentials (VI, V2, V3) underneath the recording elec¬ trodes (E1,E2,E3) [2,3,4]. The source of interference (V interf.) causes an interference current (I interf. ) , which flows through the ( tissue) impedances Zt(c)l, Zt(c)2, Ztl, Zt2, Zt3. The bioelectric signals are repre- sented by voltage sources (V signall, V signal2, V sig¬ nals) connected to the (tissue) impedances Ztl, Zt2 and Zt3.

Interference free recording can not be realized by a mo- nopolar amplifier at all and only partially by a bipolar amplifier, because the sources of interference cause a voltage drop along the (tissue) impedances Ztl, Zt2 and Zt3, which effects the recording output substantially.

Equations 3 and 4 describe the influence of interference source on monopolar and bipolar amplifiers.

V_mmo = G ^■ [vsignall + CMV + ^-j ( 3 )

^v ' _bbiip_poollaarr - = iG-J ''^[ Vr ysiiggnnuaih i -- V Y osiiggnιιua,/ J3 T + — — ++ IIDDMMV\^r \ ( 4 ) VV + V3' CMV = Vref . . . . C.ommon-Mode-Voltage (5a)

IDMV = VV - V3' . . . . interference-Difference-Mode-

Voltage (5b)

While the monopolar amplifier neither rejects CMV nor IDMV, the bipolar amplifier is only able to reject CMV by the factor 1/CMR (common mode rejection), and the IDMV is fully amplified.

Noise Behavior

Figure 2 illustrates the noise behavior of a bipolar am¬ plifier. The model comprises noise sources (Vnoisel, Vnoise2) and signal sources (Vsignall and Vsignal2).

Assuming V_noi_sel=V_noi_se2=V noise , ^vsignall-^vsignal2^=v signal and V2=0 the S/N ratio ( signal to noise ratio) is about

S ,, . , y G (Vsignal+ Vsign l) - Vsignal

— (bipolar) = ' = — . ( 6 )

N G ■ J Vnoise² + Vnoise² Vnoise

The New Method (Invention)

The aim of the our invention is to reject not only the common mode voltage (CMV) but also to suppress interfer- ence difference mode voltage (IDMV) and to improve the signal -to-noise (S/N) ratio.

Interference Rejection

In order to reject IDMV, the invention utilizes a minimum of three recording electrodes [2,3,4] and a reference electrode [5], which are connected to the inputs of the recording system [1] (Fig. 1) inputs to pick up the bio- electric signals. Because the system uses three elec¬ trodes instead of two (as is more commonly used), we re¬ fer to the invention as a tripolar amplifier.

Figure 1 depicts the general influences of an interfer¬ ence source (V interf.) on the electrode potentials VI , V2, V3 and V ref., which relate to the signal recorded below the electrodes [2,3,4,5].

The new measurement system (invention), consisting of an input stage [7] and a computing stage [8] applies a spe¬ cific transfer function to the input signals VI, V2 and V3, which allows a significant reduction of IDMV without reducing the amplitude of the bio-electric signal. This IDMV-rejection only works if the proper coefficients of the transfer function are chosen. In the following pages the conditions for the optimal transfer function coeffi¬ cients are derived from theoretical considerations.

The following mathematical model derives the conditions, that the transfer function has to fulfill in order to re¬ ject IDMV, even when Ztl and Zt2 are assumed not to be equal. The potentials VI', V2^' and V3^' are determined by the voltage divider Ztl and Zt2, if the interference cur- rent I in_ter_ferenc_e flows through the impedances Ztl and Zt2. The complex voltage divider £tl and Zt2 is described by equations 7a, 7b and 8.

Z..=.Z.~ (7a) Z = Z, .^ (7b) Z_t = Z_t, +Z_n (8)

K is a complex mismatch factor, that characterizes the relationship between Ztl and Zt2. The potential V2^' can be calculated from the potentials VI^' and V3^' and the factor K .

VT = vv -^—≡- + v - -≡- (9)

2 2 The potentials VI, V2 and V3 found at the inputs of the tripolar amplifier (El, E2 and E3) are determined by the potentials VI^', V2^', V3^' and the signal sources Vεignall, Vεignal2 and Vsignal3.

V\ = VY+Vsignαl\ (10a)

V2 = V2' + Vsignαl2 (10b)

V3 = V3' +Vsigιιαl3 (10c)

The transfer function V_tri_polar= f (VI, V2, V3) of the tripolar amplifier can be described as illustrated in equation 11. HI, H2 and H3 represent the complex gain- and weighting- factors of the transfer function.

V_tripolαr=H\-V\ + H2-V2 + H3-V3 (11)

Equation 12 is derived from equation 11 by substituting the potentials VI, V2 and V3 with equations 10a , 10b, 10c and 9.

V_poi_αr = K> ' Vsignah + H2 ^■ VsigιιaI2 + H3 ^• VsignaB +

(12) +V\ ' ■ ( HI + H2 • — £) + V3 ' ■ f H3 + H2 • U^

From equation 12 the conditional equations 13a and 13b are derived.

Hl= -H2- \-K (13a)

H3= -H2-^ (13b)

The relationship between the coefficients Ηl, Η2 and H3 of a realization of an adjustable tripolar configuration are determined by the adjustment factor k, as illustrated in equation 14a and 14b. l - k

Hl = -H2 - (14a)

\ + k

H3 = -H2 - (14b)

If the conditional equations are exactly fulfilled by the tripolar amplifier (which implies that k=K) , the IDMV is suppressed completely, while the EMG signal is amplified as can be seen from equation 15.

V_tnpolar(a,_ustMe) = HI • Vsignall + H2 • Vsigιιal2 + H3 • Vsigιιal3 + 0 • IDMV =

\ - k \ + k

- -H2 • I — =^• • Vsignalλ - Vsignal2 + — = • VsignaI3 + - IDMV

( 15 )

In practice, two types of tripolar amplifiers can be de¬ signed, whereas the simple type allows only an amplitude adjustment (of the trimming factor k ) and the more ad- vanced type permits a phase and amplitude adjustment (of the factor k ). Even with the simple type (just ampli¬ tude adjustment) more than 35 dB IDMV rejection can be achieved.

We refer to the special case of k=0 as equally balanced tripolar configuration. The IDMV rejection is degraded as illustrated in equation 16, if an equally balanced tripolar amplifier without adjustment facility is em¬ ployed and the tissue impedances remain mismatched with the factor K.

Assuming k=0 and Η1=Η3=-Η2/2=G

^^ = G • 2 • Vsignal2 - (Vsignall + VsignaB) + IDMV • ) ( 16 )

Noise Behavior The S/N behavior for the tripolar amplifier is character¬ ized in Figure 3. This model comprises noise sources (Vnoisel, Vnoise2, Vnoise3), signal sources (Vsignall, Vsignal2, Vsignal3 ) and a recording system with three inputs, instead of two. For an easier mathematical treat¬ ment, the mismatch factor K is assumed to be negligible. Assuming further that Vnoisel=Vnoise2=Vnoise3=Vnoise; Vsignall= Vsignal2 = Vsignal3= Vsignal ; K=0 =>H1=- H2/2=H3=H and V2=0 the S/N ratio is determined by equa- tion 17.

S ,. . y H - (2 Vsignal + 2 Vsignal) 2 Vsignal

— (invention) = ^~ ■ = ( 17 )

N H • V2 • Vnoise ² + 2 • Vnoise ² Vnoise

By comparing the S/N ratio of the invention with that calculated for the conventional bipolar amplifier (see equation 6) , the improvement becomes obvious (factor

Realization of the Invention

As previously mentioned, the invention can be realized in many ways. But the basic idea remains the same. A tripo¬ lar recording system [1] (Fig. 1) with an appropriate transfer function rejects the voltage drop along (tissue) impedance substantially. Instead of describing the inven¬ tion detailed for all possible realizations a more gen¬ eral description is chosen.

Basically the system (Fig. 1) consists of an input stage [7] and a computing stage [8] with the system output [6] . The input stage provides sufficient impedance transforma¬ tion, amplification and CMRR (common mode rejection ra¬ tio) in order to elevate the recorded signal into the de¬ sired amplitude range, without being corrupted by imped¬ ance changes of the recording electrodes or high common mode voltages (CMV) . The overall gain of the system has to be chosen such that the maximum IDMV can not drive the amplifiers into saturation. The bandwidth has to be cho- sen wider than the spectra of the original signal, in order to avoid phase-shift differences between the input amplifiers, which could cause additional interference to the system.

The computation stage [8] provides the actual function of the invention, by supplying the appropriate coefficients (H1,H2,H3) for the transfer function. The implementation can be accomplished by using an A/D Converter and per- forming the calculation of the transfer function with a digital network or computer instead of employing an ana¬ log network. In addition, the computation stage [8] can be realized either as a weighting adder/subtraction stage (Fig. 6) or a weighting averaging stage (Fig. 7).

The simplest realization is to employ two conventional instrumentation amplifiers [16,17] (Fig. 5 ), which are connected with alternate polarization to El [2], E2[3] (VI- V2) and E2 [3], E3[4] (V3-V2), respectively, fol- lowed by a weighting averaging stage (Fig. 7).

A version using two conventional instrumentation amplifi¬ ers [16,17] (Fig. 5) having the same polarization and a weighting subtraction stage (Fig. 6) is another possible solution.

The weighting factors of the adder/subtraction stage (Fig. 6) are determined by the impedances Zl, Z2, Z3, Z4 and Z5 [25,26,27,28,29] and the operational amplifier [34]. In terms of adjustment for high IDMV rejection, Zl [25] and Z3 [27] play a crucial roll, since they deter¬ mine the trimming factor k. In the simplest case Zl and Z3 are realized as a potentiometer Pl [35] substituting these impedances as easily adjusted resistors.

The weighting factors of the averaging stage are deter¬ mined by the two impedances Zl and Z2 [30,31], which are easily realized as a potentiometer Pl [32] as well. The instrumentation amplifier [33] determines the gain of the subtraction stage.

Thirdly, the invention can be realized by using three mo- nopolar amplifiers [12,13,14] as shown in Figure 4 (the amplifiers reference inputs (15) are connected together), followed by a computing network (either adder/subtraction stage or averaging stage), which realizes the appropriate transfer-function. The monopolar amplifiers are realized as non-inverting amplifiers as seen from Figure 8. While R2 [20] and Rl [18] determine the gain in the pass-band. The capacitor Cl [19] rejects the DC-components of the signal, to prevent the amplifiers from saturating, due to DC and low frequency components arising from the electro- chemical interface between the pick-up electrodes and the tissue.

Depending on the recording task, the invention can be de¬ signed as fixed, pre-adjusted or adjustable during use. The latter can be realized either by a circuit, which al¬ lows the adjustment by hand or by an automatic adjustment circuit, which minimizes the interference continuously. The adjustable version provides the best IDMV rejection. The fixed adjustment version provides a lower rejection ability since the "tuning" may not be ideal during use. However, even the fixed adjustment tripolar amplifier will usually perform better than a conventional bipolar amplifier.

Claims

Patent Claims:

1. A method for interference rejection in electrical re¬ cording systems, which amplifies, weights and subtracts signals from each other (difference amplifier), in which the signals of at least three inputs [2,3,4] are ampli¬ fied, weighted, added and subtracted, respectively and finally presented to the output [6], in order to imple¬ ment a specific transfer-function, which is able to re- ject interference difference mode voltage (IDMV) partly or fully, so that the signal is amplified uncorrupted, while the interference is attenuated or totally cancelled out and/or the S/N ratio (signal to noise ratio) is im¬ proved.

2. The method of claim 1, in which the electrical poten¬ tials V1,V2,V3 and Vref are fed directly to the inputs (E1,E2,E3) [2,3,4] and the reference input (Eref) [5] of the recording system [1] or any other physical measures are converted by a transducer and/or a converter to the input potentials (V1,V2,V3 and Vref) and then fed to the system^'s inputs.

3. The method of claim 1, in which the electrical record- ing system incorporates a transfer-function

Vout=f(Vl, V2, V3)= =H1. V1+H2. V2+H3. V3, that provides the desired interference rejection and amplifies the signals properly.

4. The method of claim 1, in which the complex coeffi¬ cients (H1,H2,H3) of the transfer- function are either ad¬ justable by hand or automatically (automatic interference minimization ) during use or are fixed, or pre-adjusted.

5. The method version of claim 1, in which the electrical recording system incorporates a transfer-function, that can be described by the equation Vout= HI . V1+H2. V2+H3. V3 and whose coefficients fulfil the conditional equations _{m =} „_H2. K and H3= -H2-^

— — 2 - ^~ 2

(the complex factor K describes the mismatch of the im¬ pedances Ztl and Zt2) exactly or approximately and so the interference difference mode voltage caused by any inter¬ ference source can be partially or fully rejected.

6. The devices of claim 1, in which the electrical re¬ cording system [1] consists of an input stage [7] and a computing stage [8], which provide a transfer function, which is able to partly or fully reject interference dif¬ ference mode voltage.

7. The device realization of claim 6, in which the ad- just ent of the coefficients (Η1,Η2,Η3 and H4,..,Hn) of the transfer-function is done only in the input stage or only in the computing stage or in the input stage as well as in the computing stage.

8. The device realization of claim 6, in which the imped¬ ance transformation, amplification and common mode rejec¬ tion (CMR) are provided by the input stage, which can consist either of three monopolar amplifiers [12,13,14] (Fig. 4) with a common reference potential (15) or of two bipolar amplifiers (instrumentation amplifier) [16,17] (Fig. 5) or other devices, which are able to pro¬ vide impedance transformation, amplification and common mode rejection (CMR) .

9. The device realization of claim 6, in which the com¬ puting stage [8] is realized either by an A/D converter followed by a digital computing network (i.e. signal processor, microcontroller ... ) or by an analogue comput¬ ing network , which performs as a weighting ad- der/subtraction stage (Fig. 6) or as a weighting avera¬ ger (Fig. 7) or more generally as a device, which pro¬ vides by using the outputs of the input stage a transfer function for the whole system, which is able to reject interference difference mode voltage partly or fully.