EEG signal analysis system

• US 4,610,259 A
• Filed: 08/31/1983
• Issued: 09/09/1986
• Est. Priority Date: 08/31/1983
• Status: Expired due to Term

• Current Assignee: Aequitron Medical, Inc. (Medtronic Plc)
• Original Assignee: CNS Incorporated
• Inventors: Cohen, Daniel E., Strobl, Frederick T.
• Primary Examiner(s): Valentine, Donald R.
• Application Number:
  US06/527,955
• Time in Patent Office: 1,105 Days
• Field of Search: 128/731
• US Class Current: 600/544
• CPC Class Codes: A61B 5/374 Detecting the frequency distribution of signals, e.g. detecting delta, theta, alpha, beta or gamma waves

Abstract

An electroencephalograph (EEG) signal analyzer allows the examination of the changes in EEG cerebral activity at various sites in response to an event (e.g. a stimulus or task). The analyzer includes one or more signal processing modules which periodically sample the EEG signal from each of the sites, convert the sampled signals to digital sample values, and store those values. Digitized waveforms based on the stored digital sample values and having a length equal to or greater than the period of the lowest frequency of interest are transformed from the time to the frequency domain. For each of a plurality of epochs, a frequency spectrum is produced having frequency content which has a content uniquely due to the digital sample values from that epoch. A weighted means frequency value for each site during each epoch is derived from the corresponding frequency spectrum. Based upon the weighted mean frequency values, output signals are provided to a display and a printer to produce graphical representations of cerebral activity with time at the various sites in response to the event.

Claims

1. A method of processing an EEG signal to provide an indication of cerebral activity in response to an event, the method comprising:

sampling the EEG signal to produce digital sample values representative of amplitude of the EEG signal as a function of time;

providing a plurality of digitized waveforms which are based at least in part on the digital sample values, which have a first length, and each of which includes fill values surrounding the digital sample values derived from one of a plurality of epochs having different time relationships to the event, each epoch having a second length which is less than the first length;

transforming each digitized waveform from a time domain to a frequency domain to produce a frequency spectrum corresponding to one of the plurality of epochs, each frequency spectrum having frequency components uniquely attributable to the digital sample values derived from its corresponding epoch;

deriving a frequency value for each epoch based upon the frequency spectrum corresponding to that epoch; and

providing an output indicative of cerebral activity based upon the frequency values.

2. The method of claim 1 wherein providing an output indicative of cerebral activity comprises:

comparing the frequency value for each epoch with a reference frequency value to produce a frequency difference value for each epoch; and

providing the output as a function of the frequency difference values.

3. The method of claim 2 wherein the reference frequency value is a calculated frequency value from a frequency spectrum based upon a digitized waveform which began and ended prior to the event.

4. The method of claim 2 wherein the reference frequency value is an average mean frequency, and wherein the method comprises:

averaging the frequency values for the plurality of epochs to produce the reference frequency value.

5. The method of claim 2 wherein providing the output as a function of the frequency difference values comprise:

comparing a maximum and a minimum calculated frequency value from among the frequency values for the plurality of epochs to produce a range value;

deriving an adjusted frequency value for each epoch based upon a ratio of the frequency difference value for that epoch and the range value; and

providing the output as a function of the adjusted frequency values.

6. The method of claim 2 wherein providing the output as a function of the frequency difference values comprises:

providing a plurality of graphical representations of a human subject'"'"'s head, each graphic representation corresponding to one of the plurality of epochs; and

providing an indicium associated with each graphical representation which is representative of a region from which the EEG signal is derived, the indicium having a characteristic which is a function of the frequency difference value for the epoch corresponding to the graphical representation.

7. The method of claim 6 wherein the characteristic of the indicia is a function of sign of the frequency difference value.

8. The method of claim 7 wherein the characteristic which is a function of the sign of the frequency difference value is a color of the indicium, and wherein the indicium has a first color representative of a positive frequency difference value and a second color representative of a negative frequency difference value.

9. The method of claim 6 wherein the characteristic of the indicium is a function of magnitude of the frequency difference value.

10. The method of claim 9 wherein providing the output further comprises:

comparing a maximum and a minimum calculated frequency value from among the frequency values for the plurality of epochs to produce a range value;

deriving an adjusted frequency value for each epoch based upon a ratio of the frequency difference value for that epoch and the range value; and

wherein the indicium has an area which is a function of magnitude of the adjusted frequency value.

11. The method of claim 2 wherein each of a plurality of EEG signals derived from different sites are processed according to the method for each of the plurality of epochs to produce a frequency value and a frequency difference value for each site for each epoch.

12. The method of claim 11 wherein providing an output indicative of cerebral activity further comprises:

comparing a maximum and a minimum calculated frequency value from among the frequency values for the plurality of epochs to produce a range value;

deriving an adjusted frequency value for each epoch based upon a ratio of the frequency difference value for that epoch and the range value;

comparing the adjusted frequency values of the sites for each epoch;

providing an output representative of relative amounts of change in frequency at the different sites corresponding to the EEG signals for each of the epochs, based upon the comparing of adjusted frequency values.

13. The method of claim 1 and further comprising:

multiplying the digitized waveforms by a window function prior to transforming.

14. A method of processing a time-varying analog biological signal to provide a frequency value for each of a plurality of epochs which have different time relationships to an event, the method comprising:

digitizing the biological signal to produce digital sample values representative of amplitude of the biological signal as a function of time;

providing for each epoch, a first digitized waveform based at least in part upon the digital sample values derived from the epoch and in part upon fill values, the first digitized waveform having a length which is longer than the epoch;

producing a frequency spectrum for each epoch based at least in part upon the first digitized waveform containing digital sample values derived from that epoch, the frequency spectrum having a frequency content which is uniquely attributable to the digital sample values derived from the epoch; and

deriving a frequency value for each epoch based upon the frequency spectrum corresponding to that epoch.

15. A method of processing a time-varying analog biological signal to provide an indication of biological activity in response to an event, the method comprising:

digitizing the biological signal to produce digital sample values representative of amplitude of the biological signal as a function of time;

providing a first non-averaged digitized waveform of a first length, the first digitized waveform including digital sample values derived from an epoch which has a time duration which is less than the first length and which has a time relationship to the event;

providing a second non-averaged digitized waveform of a second length equal to the first length and partially overlapping in time with the first digitized waveform, but which does not include digital sample values derived from the epoch;

transforming the first digitized waveform to a first frequency spectrum;

transforming the second digitized waveform to a second frequency spectrum;

producing a difference frequency spectrum representative of a difference between the first frequency spectrum and the second frequency spectrum;

deriving a frequency value for the epoch based upon the difference frequency spectrum; and

providing an output indicative of biological activity during the epoch based upon the frequency value for that epoch.

16. A method of processing an EEG signal, the method comprising:

sampling the EEG signal during a time interval having a time relationship to an event to produce digital sample values representative of amplitude of the EEG signal as a function of time;

producing, for each of a plurality of epochs having different time relationships to the event, a frequency spectrum based upon a digitized waveform which covers a time period longer than the epoch and which has a frequency content which is uniquely attributable to the digital sample values derived from that epoch;

deriving, for each of the plurality of epochs, a frequency value based upon the frequency spectrum corresponding to that epoch; and

providing an output as a function of the frequency values for the plurality of epochs.

17. A method of processing an EEG signal to provide an indication of cerebral activity in response to an event, the method comprising:

providing a plurality of tests which include;

sampling the EEG signal to produce digital sample values representative of amplitude of the EEG signal as a function of time;

providing first and second non-averaged digitized waveforms which are based at least in part upon the digital sample values and which represent first and second time intervals having essentially equal lengths but different time relationships to the event, wherein the first and second time intervals are staggered and are partially overlapping to define an epoch which has a time duration which is less than a the lengths of the first and second time intervals and which has a time relationship to the stimulus so that only one of the first and second time intervals includes digital sample values corresponding to the epoch;

transforming the first and second digitized waveforms from a time domain to a frequency domain to produce a first frequency spectrum and a second frequency spectrum representative of amplitude of the EEG signal at selected frequencies during the first and second time intervals, respectively;

subtracting the second frequency spectrum from the first frequency spectrum to produce a difference frequency spectrum which is a function of amplitude of the EEG signal at selected frequencies during the epoch;

deriving a frequency value for the epoch based upon the difference frequency spectrum;

averaging the frequency values for the epoch obtained from the plurality of tests to produce an averaged frequency value; and

providing an indication of cerebral activity as a function of the averaged frequency value.

18. The method of claim 17 wherein the plurality of tests are performed for each of a plurality of different epochs, and wherein providing an indication of cerebral activity comprises:

comparing the averaged frequency value for each epoch with a reference frequency value to produce a frequency difference value for each epoch; and

providing an output as a function of the frequency difference values for the plurality of epochs.

19. A system for processing an analog EEG signal, the system comprising:

means for producing a first non-averaged digitized waveform representative of amplitude of the EEG signal as a function of time during a first time interval;

means for producing a second non-averaged digitized waveform representative of amplitude of the EEG signal as a function of time during a second time interval, wherein the first and second intervals are essentially equal in length and staggered in time to define an overlapping portion when both intervals are occurring and a nonoverlapping portion in which only one of the intervals is occuring, the nonoverlapping portion having a time duration which is less than the length of the intervals;

means for transforming the first digitized waveform to a first frequency spectrum representative of amplitude of the EEG signal as a function of frequency of the EEG signal during the first time interval;

means for transforming the second digitized waveform to a second frequency spectrum representative of amplitude of the EEG signal as a function of frequency of the EEG signal during the second time interval;

means for subtracting the second frequency spectrum from the first frequency spectrum to produce a difference frequency spectrum representative of amplitude of the EEG signal as a function of frequency of the EEG signal during the nonoverlapping portion; and

means for providing an indication of frequency response during the nonoverlapping portion based upon the difference frequency spectrum.

20. A system for processing EEG signals derived from a plurality of sites to provide an indication of cerebral activity as a function of time in response to a stimulus, the system comprising:

means for providing a stimulus during each of a plurality of tests;

analog-to-digital converter means for sampling each EEG signal;

storage means for storing digital sample values produced by the analog-to-digital converter means for each EEG signal, the digital sample values being representative of amplitude of the corresponding EEG signal as a function of time;

means for providing first and second non-averaged digitized waveforms, for each EEG signal and each test, which are based at least in part on the digital sample values and represent a first and a second time interval of essentially equal length, respectively, which define one of a plurality of epochs having different time relationships to the stimulus and a duration which is less than the lengths of the first and second time intervals, the first and second time intervals being staggered and partially overlapping and wherein the epoch is defined by presence of only one of the first and second time intervals;

means for transforming the first and second digitized waveforms for each EEG signal from a time domain to a frequency domain to produce a first and a second frequency spectrum, respectively;

means for subtracting the second frequency spectrum from the first frequency spectrum for each EEG signal to produce a difference frequency spectrum;

means for deriving a frequency value for each EEG signal during each of the plurality of epochs based upon a corresponding difference frequency spectrum; and

means for providing an indication of cerebral activity at each of a plurality of sites from which the EEG signals are derived as a function of the frequency values.

21. A system for processing a time-varying analog biological signal to provide a frequency value for each of a plurality of epochs which have different time relationships to an event, the system comprising:

means for digitizing the biological signal to produce digital sample values representative of amplitude of the biological signal as a function of time;

means for providing, for each epoch, a non-averaged digitized waveform based at least in part upon the digital sample values from the epoch, the digitized waveform having a length which is greater than a length of the epoch;

means for producing a frequency spectrum for each epoch as a function of the digitized waveform for that epoch, the frequency spectrum having a frequency content which is uniquely attributable to the digital sample values from the epoch; and

means for deriving the frequency value for each epoch based upon the frequency spectrum corresponding to that epoch.

22. A system for processing an EEG signal, the method comprising:

means for sampling the EEG signal during a time interval having a time relationship to an event to produce digital sample values representative of amplitude of the EEG signal as a function of time;

means for producing, for each of a plurality of epochs having different time relationships to the event, a frequency spectrum based upon a digitized waveform which is longer than the epoch and which has a frequency content which is uniquely attributable to the digital sample values corresponding to that epoch;

means for deriving, for each of the plurality of epochs, a frequency value based upon the frequency spectrum corresponding to that epoch; and

means for providing an output as a function of the frequency values for the plurality of epochs.

23. A method of processing an EEG signal to provide an indication of cerebral activity during an epoch, the method comprising:

digitizing the EEG signal to produce digital sample values representative of amplitude of the EEG signal as a function of time, including at least one digital sample value representative of amplitude of the EEG signal during the epoch;

forming a digitized waveform having those digital sample values which are representative of amplitude of the EEG signal during the epoch located at its center and having fill values surrounding those digital sample values;

transforming the digitized waveform from a time domain to a frequency domain to produce a frequency spectrum; and

deriving a frequency value for the epoch based upon the frequency spectrum.

24. The method of claim 23, wherein forming a digitized waveform comprises:

selecting a plurality of digital sample values associated with the epoch;

multiplying the digital sample values by a window function; and

forming a digitized waveform in which the digital sample values, as multiplied by the window function, are surrounded at each end by a plurality of fill values.