Method and system for limiting interference in electroencephalographic signals

• US 7,809,433 B2
• Filed: 08/09/2005
• Issued: 10/05/2010
• Est. Priority Date: 08/09/2005
• Status: Expired due to Fees

• Current Assignee: Adidas AG
• Original Assignee: Adidas AG
• Inventors: Keenan, Barry
• Primary Examiner(s): Layno; Carl H
• Assistant Examiner(s): Behringer; Luther G
• Application Number:
  US11/200,674
• Publication Number:
  US 20070038382A1
• Time in Patent Office: 1,883 Days
• Field of Search: 381/94.1, 455/296, 600/323, 600/509, 600/510, 600/515, 600/544, 600/546, 607/45
• US Class Current: 600/544
• CPC Class Codes: A61B 5/7203 for noise prevention, reduction or removal
• CPC Class Codes: A61B 5/7257 using Fourier transforms
• CPC Class Codes: A61B 5/726 using Wavelet transforms
• CPC Class Codes: G06F 2218/04 Denoising

Abstract

This invention provides methods and systems for removing or limiting interference and artifacts often found in EEG signals and/or EOG signals and/or other physiological signals, and produces output signals that can be of greater use for analytic or diagnostic purposes, such as sleep studies. Interference and artifacts are removed by applying a plurality of filtering methods targeted to the particular time-frequency characteristics of the anticipated interference.

Claims

1. A method for processing a composite physiological signal having at least one physiological signal of interest and at least one unwanted spurious signal, the composite physiological signal having a first duration, the method comprising:

determining a variable spurious signal threshold, said spurious signal threshold being determined as a function of signal spectral energy;

or eliminating at least one narrow-band spurious signal in said physiological signal throughout said signal first duration that meets or exceeds said spurious signal threshold by performing a plurality of Fourier transforms, each of said Fourier transforms being performed on at least one of a plurality of overlapping temporal windows of said physiological signal, and filtering Fourier coefficients by reducing said Fourier coefficients in narrow bands of frequencies, whereby a window-to-window temporal variation of said Fourier coefficients to be filtered is distinguishably less than a window-to-window temporal variation of non-filtered Fourier coefficients;

recreating an intermediate signal throughout substantially all of said physiological signal first duration by inverse Fourier transforming said filtered Fourier coefficients;

eliminating at least one intermittently-occurring spurious signal in said intermediate signal by performing at least one wavelet analysis on said intermediate signal, said intermittently-occurring spurious signal including a plurality of intermittently-occurring wavelet coefficients, and filtering said intermittently-occurring wavelet coefficients by reducing or zeroing intermittently-occurring wavelet coefficients that are distinguishably greater than the remaining intermittently-occurring wavelet coefficients in said intermittently-occurring spurious signal; and

recreating a second physiological signal in filtered form by performing a plurality or inverse wavelet analyses on said filtered intermittently-occurring wavelet coefficients.

2. The method of claim 1 wherein Fourier coefficients are found to have a temporal variation substantially less than the temporal variation of said physiological signal if the standard deviation of said coefficients is less than a threshold.

3. The method of claim 2 wherein said threshold depends on a standard deviation of said Fourier coefficients of interest.

4. The method of claim 1 wherein said filtering Fourier coefficients further comprises

determining the zero-crossing bandwidth of Fourier coefficients;

and

limiting the amplitudes of Fourier coefficients within said zero-crossing bandwidth.

5. The method of claim 1 wherein said filtering wavelet coefficients further comprises applying a threshold function to said wavelet coefficients.

6. The method of claims 5 wherein said threshold function sets to zero wavelet coefficients having amplitudes exceeding a threshold value.

7. The method of claim 1 further comprising

performing at least one inverse Fourier transform of said filtered Fourier coefficients;

performing at least one inverse wavelet transform of said filtered wavelet coefficients; and

.

outputting the signal resulting from said inverse Fourier transform and said inverse wavelet transform.

8. The method of claim 1, wherein a total temporal duration of all occurrences of said intermittently-occurring spurious signals is less than or equal to approximately one-fifth of said total temporal duration of a signal window being processed.

9. The method of claim 1 wherein said intermittently-occurring spurious signals comprise at least one occurrence of a QRS wave.

10. The method of claim 9

wherein said filtering wavelet coefficients further comprises applying a threshold function to said wavelet coefficients;

and

wherein said threshold value is between a first expected amplitude of wavelet coefficients not in a vicinity of any QRS wave occurrence and a second expected amplitude of distinguishable wave coefficients in the vicinity of at least one QRS wave occurrence.

11. The method of claim 10 wherein vicinities of QRS wave occurrences are determined by a QRS detection method applied to a physiological signal.

12. The method of claim 10 wherein said first and/or second expected amplitudes are determined in dependence on the means and standard deviations of wavelet coefficients.

13. The method of claim 1 wherein said physiological signals comprise one or more of electroencephalographic (EEG) signals, and/or electrooculographic (EOG) signals, and/or electromyographic (EMG) signals.

14. The method of claim 1 wherein said windows are chosen to overlap so that all portions of a physiological signal are within one or more of said windows.

15. The method of claim 1, wherein at least said Fourier coefficients are filtered by being set to zero.

16. The method of claim 1, wherein first Fourier coefficients in a narrow frequency band present in all windows have window-to-window variations that are distinguishably less than the window-to-window variations of second Fourier coefficients if the standard deviation of said first Fourier coefficients is less than a first threshold.

17. The method of claim 1 wherein a wavelet coefficient is reduced or zeroed only when approximately coincident with a QRS complex in a concurrent ECG signal.

27. A computer system for processing a physiological signal of interest also comprising unwanted spurious signals, the system comprising a non-transitory computer readable medium comprising encoded instructions for causing a processor to perform the steps of claim 1.

18. A method for processing a physiological signal of interest having a certain duration and comprising unwanted spurious signals, the method comprising:

applying to said physiological signal one or more filters for producing a plurality of transform coefficients that are responsive to signal components present in sub-domains of a time/frequency domain plane having a plurality of different shapes, said filters comprising one or more windowed Fourier transformations responsive to signal components in sub-domains having a shape with a relative temporal duration substantially greater than a relative frequency bandwidth, and also comprising one or more wavelet transformations responsive to signal components in sub-domains having a limited relative temporal duration, and performing

a first filtering operation that reduces or eliminates spurious signals by reducing or zeroing first Fourier transform coefficients in one or more narrow bands with a substantially constant frequency that have temporal variations distinguishably less than the temporal variations of second Fourier transform coefficients, and a second filtering operation that reduces or eliminates spurious signals by reducing or zeroing third wavelet transform coefficients with a limited relative temporal duration and with amplitudes that are distinguishably greater than the amplitudes of fourth wavelet transform coefficients; and

performing a plurality of inverse transformations to recreate a signal of interest in filtered form.

19. The method claim 18 wherein applying a filter further comprises determining whether signal components present in a time/frequency sub-domain arise from spurious signals, and limiting said signal components that arise from spurious signals.

20. The method of claim 19 wherein limiting said signal components further comprises setting an amplitude of said signal components to zero.

21. The method of claim 19 wherein limiting said signal components further comprises applying a threshold function to said signal components.

22. The method of claim 19 wherein a signal component present in a time/frequency sub-domain is determined to arise from spurious signals if said signal component has a temporal variation in said sub-domain that is substantially less than the temporal variation of said physiological signal of interest.

23. The method of claim 19 wherein a signal component present in a time/frequency sub-domain is determined to arise from spurious signals if said signal component has characteristics that are statistically distinguishable from said physiological signal of interest.

24. The method of claim 18 wherein applying a filter for filtering signal components present in a time/frequency sub-domain having a shape with a relative temporal duration substantially greater than a relative frequency bandwidth further comprises performing a windowed Fourier transform.

25. The method of claim 18 wherein applying said filters further comprises applying said filters in a temporal sequence.

26. The method of claim 18 wherein said physiological signals comprise one or more of electroencephalographic (EEG) signals, and/or electrooculographic (EOG) signals, and/or electromyographic (EMG) signals, and/or electrocardiographic (ECG) signals.

28. A computer system for processing a physiological signal of interest said physiological signal including unwanted spurious signals, said system comprising a processor and non-transitory computer readable memory having encoded instructions for causing said processor to perform the steps of claim 18.