Adaptive interference canceler for EEG movement and eye artifacts

• US 5,513,649 A
• Filed: 03/22/1994
• Issued: 05/07/1996
• Est. Priority Date: 03/22/1994
• Status: Expired due to Term

• Current Assignee: SAM Technology, Inc.
• Original Assignee: SAM Technology, Inc.
• Inventors: Leong, Harrison, Du, Weixiu, Gevins, Alan S.
• Primary Examiner(s): Sykes, Angela D.
• Application Number:
  US08/216,256
• Time in Patent Office: 777 Days
• Field of Search: 128/731-3, 128/715, 128/687, 128/774, 128/782, 128/745
• US Class Current: 600/544
• CPC Class Codes: A61B 5/369 Electroencephalography [EEG] devices for psychotechnics A61B5/16
• CPC Class Codes: A61B 5/398 Electrooculography [EOG], e.g. detecting nystagmus; Electroretinography [ERG]

Abstract

In an EEG (electroencephalogram) system to detect brain waves from a subject, the adverse effect on the EEG recording of artifact due to head, body and eye movements is reduced. A head and body movement reference signal can be provided by an accelerometer, motion detector or, alternatively, spatial average of EEG channels. Eye motion sensors are used as a reference for eye movement artifacts. A composite reference signal consists of the head, body and eye movement reference signals. The contaminated EEG is the primary input to the adaptive movement and eye artifact canceler, in which an adaptive filter estimates the contaminants in the measured EEG data, and then subtracts them from the primary signal to obtain the corrected EEG data.

Claims

1. An EEG (electroencephalograph) system comprising:

(a) a plurality of EEG electrode means adapted to be removably connected to the scalp of a subject to detect the subject'"'"'s brain waves;

(b) amplification means connected to the EEG electrode means to amplify the detected brain waves and produce an amplified signal output for each EEG electrode means;

(c) a three-axis translation accelerometer adapted to be removably positioned on the head of the subject to detect the subject'"'"'s head movement, the accelerometer producing at least one signal output indicative of the head movements;

(d) an interference canceler connected to the outputs of the amplification means and the accelerometer signal output, the canceler comprising a filter having an output and an adder connected to the output of the filter and the amplification means, the adder producing corrected and usable EEG signals.

2. An EEG system as in claim 1 and further comprising an eye movement sensor means adapted to be removably connected proximate an eye of the subject to generate an eye movement reference signal which is an input to the filter.

3. An EEG system as in claim 1 wherein the filter uses an algorithm embodied in a software program executed on a digital computer system or embodied in a digital hardware circuit.

4. An EEG system as in claim 3 wherein the filter is an adaptive filter having a coefficient matrix which is adjusted with time according to a Least Mean Squares (LMS) algorithm.

5. An EEG system as in claim 4 wherein the filter is an adaptive filter having a filter coefficient matrix which is adjusted with time according to an Exponentially Weighted Recursive Least Squares (EWRLS) algorithm or a Recursive Least Squares (RLS) algorithm.

6. An EEG system as in claim 3 wherein the filter uses a block adaptive filter algorithm selected from the group of autoregressive with external input (ARX), autoregressive and moving average with external input (ARMAX), Wiener filters, Kalman filters, Linear Predictors, autoregressive and moving average (ARMA) and Box-Jenkins model.

7. An EEG system as in claim 1 wherein the filter is an analog hardware circuit.

8. An EEG system as in claim 1 wherein the filter is a non-adaptive filter.

9. An EEG (electroencephalograph) system comprising:

(a) a plurality of EEG electrode means adapted to be removably connected to the scalp of a subject to detect the subject'"'"'s brain waves;

(b) amplification means connected to the EEG electrode means to amplify the detected brain waves and produce an amplified signal output for each EEG electrode means;

(c) signal averaging means to combine and average at least some of the amplified signal outputs to generate a spatial average EEG signal as a head and body reference signal; and

(d) an interference canceler connected to the outputs of the amplification means and the signal averaging means, the canceler comprising a filter having an output, and an adder connected to the output of the filter and the amplification means, the adder producing corrected and usable EEG signals.

10. An EEG system as in claim 9 and further including an accelerometer or movement detector adapted to be removably positioned on the head of the subject to detect the subject'"'"'s head movements, the accelerometer or movement detector generating at least one signal indicative of head movements, which signal is used to determine the starting and ending of the head movements, and means to operate the canceler only during said determined head movements.

12. An EEG system as in claim 9 wherein the filter is an adaptive filter.

13. An EEG system as in claim 12 wherein the adaptive filter comprises a filter coefficient matrix which is adjusted with time according to a Least Mean Squares (LMS) algorithm.

14. An EEG system as in claim 12 wherein the adaptive filter comprises a filter coefficient matrix which is adjusted with time according to an Exponentially Weighted Recursive Least Squares (EWRLS) algorithm or Recursive Least Squares (RLS) algorithm.

15. An EEG system as in claim 9 wherein the filter is an algorithm embodied in a software program executed on a digital computer system or embodied in a digital hardware circuit.

16. An EEG system as in claim 9 wherein the filter is an analog hardware circuit.

17. An EEG system as in claim 9 wherein the filter comprises a block adaptive filter algorithm selected from the group of autoregressive with external input (ARX), autoregressive and moving average with external input (ARMAX), Wiener filters, Kalman filters, Linear Predictors, autoregressive and moving average (ARMA) and Box-Jenkins model.

18. An EEG system as in claim 9 and further comprising an eye movement sensor means adapted to be removably connected proximate an eye of the subject to generate an eye movement reference signal which is an input to the filter.

19. An EEG system as in claim 9 wherein the filter is a non-adaptive filter.

11. An EEG system as in claim 11 wherein an accelerometer is positioned on the subject'"'"'s head and the accelerometer is a three-axis translational accelerometer.

20. An EEG (electroencephalograph) method comprising the steps of:

(a) connecting a plurality of EEG electrodes to the scalp of a subject to detect the subject'"'"'s brain waves;

(b) amplifying the subject'"'"'s brain waves using amplification means connected to the EEG electrodes to produce an amplified signal output for each EEG electrode;

(c) connecting a three-axis translation on the head of the subject to detect the subject'"'"'s head movements, the accelerator producing at least one reference signal output indicative of the head movements;

(d) using an interference canceler to produce corrected EEG signals, the canceler being connected to the output of the amplification means and to the reference signal, the canceler comprising a filter having an input connected to the output of the reference signal and having an output, and an adder connected to the output of the filter and the amplification means, the adder producing corrected and usable EEG signals.

21. An EEG method as in claim 20 wherein the filter is an adaptive filter executed in an algorithm embodied in a software program executed on a digital computer system or embodied in a digital hardware circuit.

22. An EEG method as in claim 21 wherein the adaptive filter comprises a filter coefficient matrix which is adjusted with time according to a Least Means Squares (LMS) algorithm.

23. An EEG method as in claim 21 wherein the adaptive filter comprises a filter coefficient matrix which is adjusted with time according to an Exponentially Weighted Recursive Least Squares (EWRLS) algorithm or a Recursive Least Squares (LRS) algorithm.

24. An EEG method as in claim 20 wherein the filter comprises a block adaptive filter algorithm selected from the group of autoregressive with external input (ARX), autoregressive and moving average with external input (ARMAX), Wiener filters, Kalman filters, Linear Predictors, autoregressive and moving average (ARMA) and Box-Jenkins model.

25. An EEG method as in claim 20 and further including the step of generating an eye movement reference signal as an input to the filter, the eye movement reference signal being generated from an eye movement sensor removably connected proximate an eye of the subject.

26. An EEG method as in claim 20 wherein the filter is a non-adaptive filter.

27. An EEG (electroencephalograph) method comprising the steps of:

(a) connecting a plurality of EEG electrodes to the scalp of a subject to detect the subject'"'"'s brain waves;

(b) analyzing the subject'"'"'s brain waves using amplification means connected to the EEG electrodes to produce an amplified signal output for each EEG electrode;

(c) combining and averaging at least some of the amplified signal outputs and thereby generating a spatial average EEG signal as a head and body reference signal; and

(d) using an interference canceler to produce corrected EEG signals, the canceler being connected to the output of the amplification means and to the reference signal, the canceler comprising a filter having an input connected to the reference signal and having an output, and an adder connected to the output of the filter and the amplification means, the adder producing corrected and usable EEG signals.

28. An EEG method as in claim 27 and further comprising removably connecting an accelerometer or movement detector on the head of the subject to detect head movements, the accelerometer or movement detector generating at least one signal indicative of head movements, using the head movement signal to determine the starting and ending of head movements, and operating the interference canceler only during said determined head movements.

29. An EEG method as in claim 28 wherein an accelerometer is connected on the head of the subject and the accelerometer is a three-axis translational accelerometer.

30. An EEG method as in claim 27 wherein the filter is an adaptive filter which is an algorithm embodied in a software program executed on a digital computer system or embodied in a digital hardware circuit.

31. An EEG method as in claim 27 wherein the filter is an analog hardware circuit.

32. An EEG method as in claim 27 wherein the filter is an adaptive filter having a filter coefficient matrix which is adjusted with time according to a Least Mean Squares (LMS) algorithm.

33. An EEG method as in claim 27 wherein the filter is an adaptive filter having a filter coefficient matrix which is adjusted with time according to an Exponentially Weighted Recursive Least Squares (EWRLS) algorithm or a Recursive Least Squares (RLS) algorithm.

34. An EEG method as in claim 27 wherein the filter uses a block adaptive filter algorithm selected from the group of autoregressive with external input (ARX), autoregressive and moving average with external input (ARMAX), Wiener filters, Kalman filters, Linear Predictors, autoregressive and moving average (ARMA) and Box-Jenkins model.

35. An EEG method as in claim 27 wherein the filter is a non-adaptive filter.