Method for real-time detection and identification of neuroelectric signals

• US 4,603,703 A
• Filed: 04/13/1984
• Issued: 08/05/1986
• Est. Priority Date: 04/13/1984
• Status: Expired due to Fees

• Current Assignee: Board of Trustees of the Leland Stanford Junior University (Stanford Management Co.)
• Original Assignee: Board of Trustees of the Leland Stanford Junior University (Stanford Management Co.)
• Inventors: McGill, Kevin C., Dorfman, Leslie J.
• Primary Examiner(s): Kaplan, G. L.
• Application Number:
  US06/599,889
• Time in Patent Office: 844 Days
• Field of Search: 364/417, 364/415, 128/731, 128/733
• US Class Current: 600/544
• CPC Class Codes: A61B 5/35 by template matching
• CPC Class Codes: A61B 5/7239 using differentiation including higher order derivatives
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00

Abstract

Neuroelectric signals are detected and processed by digital filtering to detect, classify and determine the relative proportions of the different types of signals eminating from muscle, nerve or brain tissue.

Claims

1. A method of real time analysis of neuroelectrophysiological signals for detecting and identifying action potentials comprising the steps of digitally filtering the signals using a low-pass differentiator and automatically detecting and identifying the recurring spikes in the filtered signal.

2. A method as claimed in claim 1 wherein said identifying step includes the step of issuing a trigger pulse on one of a plurality of output lines to indicate the identity of the detected action potential.

3. A method as claimed in claim 1 wherein the step of digitally filtering the neuroelectrophysiological signal compares the filtered neuroelectrophysiological signal to an assigned threshold value, the spikes representing action potentials exceeding the threshold.

4. A method as claimed in claim 3 wherein the detecting step includes dividing the action potential signal into intervals, wherein each interval contains one of said spikes, and assigning a signature to each signal interval.

5. A method as claimed in claim 4 wherein said signal intervals are disjoint and do not overlap.

6. A method as claimed in claim 4 including the step of comparing each signal signature to a plurality of stored signatures stored as templates, and indicating the closest match between signal and template signature.

7. A method as claimed in claim 6 including the step of adapting the template indicated as the closest match to the signal signature to improve the template'"'"'s signal to noise ratio and to track slow changes in signature configuration.

8. A method as claimed in claim 7 including calculating the error differential between the signal signature and the template indicated as the closest match, comparing the differential error to a stored threshold error value, and inhibiting the output if the error differential exceeds the stored error threshold level.

9. A method as claimed in claim 8 further comprising the steps of replacing the template having a stored history of the fewest previous matches with the signal signature in response to the indication of error differential exceeding the threshold value.

10. A method as claimed in claim 8 further comprising the steps of replacing the template having a stored history of the fewest previous matches with the signal signature in response to the indication of error differential exceeding the threshold value.

11. A method as claimed in claim 6 wherein each of said signal signatures and template signatures comprises an equal number of words, said process further comprising the step of aligning said spike signature for comparison with the template signatures by centering the word representing the peak value of the neuroelectrophysiological signal in said interval.

12. A method as claimed in claim 11 including the steps of initially sampling the neuroelectrophysiological signals at a rate fast enough to locate the peaks with high temporal resolution, and then decimating the samples to obtain said signal signature, keeping only the theoretically minimum number of samples needed to characterize the signal signature.

13. A method as claimed in claim 11 wherein said process further includes the step of simultaneously comparing the signal signature with a plurality of said template signatures.

14. A method as claimed in claim 13 wherein said digitally filtering step includes a low pass differentiating of the action potential signal utilizing a formula having only additive terms in the algorithm.

15. A method as claimed in claim 14 wherein said digital filtering step comprises developing a first order differential of the signal according to the formula

16. A method as claimed in claim 15 wherein said digital filtering step comprises developing a second order differential of the signal according to the formula

17. A method as claimed in claim 13 including the step of adapting the template indicated as the closest match to the signal signature to reflect the actual signature.

18. A method as claimed in claim 17 including calculating the error differential between the signal signature and the template indicated as the closest match, comparing the differential error to a stored threshold error value, and inhibiting the output if the error differential exceeds the stored error threshold level.

19. A method as claimed in claim 17 wherein each of said signal signatures and template signatures comprises an equal number of words, said process further comprising the steps of aligning said spike signature for comparison with the template signature by centering the word representing the peak value of the signal in the signal signature.

20. A method as claimed in claim 19 wherein said process further includes the steps of simultaneously comparing the signal signature with a plurality of said template signatures.

21. A method of real time analysis of neuroelectrophysiological signals for detecting and identifying action potentials comprising the steps of

digitally filtering the signal, and

detecting and identifying the recurring spikes in the filtered signal by

(a) detecting as a spike each interval containing a peak whose amplitude exceeds a known threshold,

(b) matching each detected spike with one of a plurality of stored spike templates, and

(c) indicating the result of said matching step.