Processing for multi-channel signals

• US 8,849,390 B2
• Filed: 12/29/2009
• Issued: 09/30/2014
• Est. Priority Date: 12/29/2008
• Status: Active Grant

• Current Assignee: Cyberonics, Inc
• Original Assignee: Cyberonics, Inc
• Inventors: Echauz, Javier Ramn, Snyder, David E., Leyde, Kent W.
• Primary Examiner(s): BERHANU, ETSUB D
• Application Number:
  US12/649,098
• Publication Number:
  US 20100168604A1
• Time in Patent Office: 1,736 Days
• Field of Search: 600/544, 600/545
• US Class Current: 600/544
• CPC Class Codes: A61B 5/369 Electroencephalography [EEG] devices for psychotechnics A61B5/16
• CPC Class Codes: A61B 5/4094 Diagnosing or monitoring seizure diseases, e.g. epilepsy

Abstract

Method and apparatus for improved processing for multi-channel signals. In an exemplary embodiment, an anomaly metric is computed for a multi-channel signal over a time window. The magnitude of the anomaly metric may be used to determine whether an anomaly is present in the multi-channel signal over the time window. In an exemplary embodiment, the anomaly metric may be a condition number associated with the singular values of the multi-channel signal over the time window, as further adjusted by the number of channels to produce a data condition number. Applications of the anomaly metric computation include the scrubbing of signal archives for epileptic seizure detection/prediction/counter-prediction algorithm training, pre-processing of multi-channel signals for real-time monitoring of bio-systems, and boot-up and/or adaptive self-checking of such systems during normal operation.

Claims

1. A method for detecting anomalies in a multi-channel signal, the method comprising:

sampling the multi-channel signal over a time window, said multi-channel signal representing physiological signals derived from outputs of a plurality of electrodes positioned to acquire physiological signals from a patient, wherein each channel of the multi-channel signal corresponds to a different one of the plurality of electrodes;

computing with a processor an anomaly metric for the multi-channel signal over the time window, wherein said computing an anomaly metric comprises;

computing a condition number of the multi-channel signal over the time window; and

adiusting the condition number based on a parameter of the multi-channel signal to generate a data condition number (DCN); and

identifying the presence of an anomaly based on a magnitude of the anomaly metric, wherein said identifying the presence of an anomaly comprises comparing the magnitude of the DCN to at least one threshold.

2. The method of claim 1, further comprising:

wirelessly receiving the sampled outputs at a wireless receiver unit.

3. The method of claim 2, further comprising:

identifying the anomaly during a boot-up phase of a computation device used to process the received sampled outputs.

4. The method of claim 1, further comprising:

updating an anomaly data service over time, the anomaly data service indicating at least one threshold for comparing the magnitude of the anomaly metric to identify the presence of an anomaly;

wherein said identifying the presence of an anomaly comprises comparing the anomaly metric to the at least one threshold indicated by the anomaly data service.

5. The method of claim 4, further comprising:

transmitting information wirelessly from the updated anomaly data service to a computation device used to process the received sampled outputs.

6. The method of claim 4, wherein said updating the anomaly data service over time comprises:

sampling multi-channel signals from a plurality of seizure detection systems; and

generating the at least one threshold indicated by the anomaly data service based on the sampling of the multi-channel signals.

7. The method of claim 1, wherein the plurality of electrodes are configured to acquire neurological signals from the patient to derive the multi-channel signal.

8. The method of claim 7, wherein the plurality of electrodes are configured to be positioned on or implanted in the patient'"'"'s brain.

9. The method of claim 7, further comprising:

processing the received sampled outputs using a seizure detection algorithm; and

based on the identified presence of an anomaly, modifying the processing of the received sampled outputs using the seizure detection algorithm.

10. The method of claim 9, the modifying the processing comprising omitting from the processing portions of the sampled outputs corresponding to identified anomalies.

11. The method of claim 10, the processing comprising adjusting parameters of an adaptive seizure detection algorithm, the method further comprising detecting a seizure based on the received sampled outputs.

12. The method of claim 7, further comprising:

processing the received sampled outputs using a seizure prediction algorithm; and

based on the identified presence of an anomaly, modifying the processing of the received sampled outputs using the seizure prediction algorithm.

13. The method of claim 7, further comprising:

processing the received sampled outputs using a seizure likelihood computation algorithm; and

based on the identified presence of an anomaly, modifying the processing of the received sampled outputs using the seizure likelihood computation algorithm.

14. The method of claim 7, further comprising:

generating a DCN time series corresponding to a plurality of time windows;

generating an anomaly log based on the DCN time series;

merging anomalies in the anomaly log separated by less than a minimum separation to generate a modified anomaly log;

identifying segments of the multi-channel signal corresponding to anomalies in the modified anomaly log; and

outputting time-expanded versions of the identified segments to a record.

15. The method of claim 1, wherein said computing the condition number comprises:

generating a matrix A[k] for a time window k, each column of the matrix A[k] comprising sampled values of a single channel of the multi-channels signal over a time window k;

computing a maximum singular value of the matrix A[k];

computing a minimum singular value of the matrix A[k]; and

dividing the maximum singular value by the minimum singular value to generate the condition number.

16. The method of claim 1, wherein said computing the condition number comprises:

generating a square matrix A[k] for a time window k, each column of the matrix A[k] comprising sampled values of a single channel of the multi-channels signal over a time window k;

deriving a norm of the matrix A[k];

deriving the norm of the matrix A-1[k]; and

dividing the norm of the matrix A[k] by the norm of the matrix A-1[k] to generate the condition number.

17. The method of claim 16, wherein the norm comprises an L1-norm.

18. The method of claim 16, wherein the norm comprises an L-infinity norm.

19. The method of claim 1, wherein said adjusting the condition number based on the parameter of the multi-channel signal to generate the DCN comprises:

computing

20. The method of claim 1, wherein the parameter used to adjust the condition number comprises at least one of window size, signal measurement bandwidth, and electrode montage.

21. An apparatus for processing a multi-channel signal, the apparatus comprising:

an anomaly metric computation module configured to compute with a computer an anomaly metric for the multi-channel signal over a time window, said multi-channel signal representing physiological signals derived from outputs of a plurality of electrodes positioned to acquire physiological signals from a patient, wherein each channel of the multi-channel signal corresponds to a different one of the plurality of electrodes, wherein the anomaly metric computation module is further configured to;

compute the anomaly metric by computing a condition number of the multi-channel signal over the time window; and

adjust the condition number by a parameter of the multi-channel signal to generate a data condition number (DCN); and

an anomaly identification module configured to identify the presence of an anomaly in the multi-channel signal based on a magnitude of the anomaly metric, wherein the anomaly identification module is further configured to identify the presence of the anomaly by comparing the magnitude of the DCN to at least one threshold.

22. The apparatus of claim 21, wherein the plurality of electrodes are configured to acquire neurological signals from the patient to derive the multi-channel signal.

23. The apparatus of claim 22, wherein the plurality of electrodes are configured to be positioned on or implanted in the patient'"'"'s brain.

24. The apparatus of claim 21, wherein the anomaly metric computation module and anomaly identification module are positioned in a wireless receiver unit, the wireless receiver unit wirelessly receiving the multi-channel signal from the plurality of electrodes.

25. The apparatus of claim 24, wherein the anomaly identification module is further configured to identify an anomaly during a boot-up phase of a computation device used to process the received multi-channel signal.

26. The apparatus of claim 21, further comprising:

an anomaly data service configured to be updated over time, the anomaly data service configured to indicate at least one threshold for comparing the magnitude of the anomaly metric to identify the presence of an anomaly;

wherein the anomaly identification module is configured to compare the magnitude of the anomaly metric to the at least one threshold indicated by the anomaly data service.

27. The apparatus of claim 26, further comprising a wireless receiver for receiving information from the updated anomaly data service.

28. The apparatus of claim 21, further comprising an algorithm training module configured to train an algorithm to detect seizures.

29. The apparatus of claim 21, further comprising an algorithm training module configured to train an algorithm to predict seizures.

30. The apparatus of claim 21, further comprising an algorithm training module configured to train an algorithm to identify conditions under which a patient is susceptible to seizure.

31. The apparatus of claim 21, wherein the plurality of electrodes are configured to acquire neurological signals from the patient, the apparatus further configured to:

generate a DCN time series corresponding to a plurality of time windows;

generate an anomaly log based on the DCN time series;

merge anomalies in the anomaly log separated by less than a minimum separation to generate a modified anomaly log;

identify segments of the multi-channel signal corresponding to anomalies in the modified anomaly log; and

output time-expanded versions of the identified segments to a record.

32. The apparatus of claim 21, wherein the anomaly metric computation module is further configured to:

generate a matrix A[k] for a time window k, each column of the matrix A[k] comprising sampled values of a single channel of the multi-channels signal over a time window k;

compute a maximum singular value of the matrix A[k];

compute a minimum singular value of the matrix A[k]; and

divide the maximum singular value by the minimum singular value to generate the condition number.

33. The apparatus of claim 21, wherein the anomaly metric computation module is further configured to:

generate a square matrix A[k] for a time window k, each column of the matrix A[k] comprising sampled values of a single channel of the multi-channels signal over a time window k;

derive a norm of the matrix A[k];

derive the norm of the matrix A-1[k]; and

divide the norm of the matrix A[k] by the norm of the matrix A-1[k] to generate the condition number.

34. The apparatus of claim 21, wherein the anomaly metric computation module is further configured to compute the DCN as:

35. An apparatus for processing a multi-channel signal, the apparatus comprising:

means for sampling the multi-channel signal over a time window;

means for computing with a computer an anomaly metric for the multi-channel signal over the time window; and

means for identifying the presence of an anomaly based on the magnitude of the anomaly metric.

36. The apparatus of claim 35, wherein the means for computing the anomaly metric is further configured to:

compute the anomaly metric by computing a condition number of the multi-channel signal over the time window;

adjust the condition number by a parameter of the multi-channel signal to generate a data condition number (DCN); and

identify the presence of the an anomaly by comparing the magnitude of the DCN to at least one threshold.

37. The apparatus of claim 36, further comprising a plurality of sensors configured to acquire physiological signals from the patient to derive the multi-channel signal.

38. The apparatus of claim 37, wherein the plurality of sensors comprises a plurality of electrodes configured to acquire neurological signals from the patient to derive the multi-channel signal.

39. The apparatus of claim 36, further comprising:

means for generating a DCN time series corresponding to a plurality of time windows;

means for generating an anomaly log based on the DCN time series;

means for merging anomalies in the anomaly log separated by less than a minimum separation to generate a modified anomaly log;

means for identifying segments of the multi-channel signal corresponding to anomalies in the modified anomaly log; and

means for outputting time-expanded versions of the identified segments to a record.

40. The apparatus of claim 36, wherein the means for computing the anomaly metric is further configured to compute the DCN as:

41. The apparatus of claim 37, wherein said plurality of sensors comprises a plurality of electrodes.

42. The apparatus of claim 35, further comprising an algorithm training module configured to train an algorithm to detect seizures.

43. The apparatus of claim 35, further comprising an algorithm training module configured to train an algorithm to predict seizures.

44. The apparatus of claim 35, further comprising an algorithm training module configured to train an algorithm to identify conditions under which a patient is susceptible to seizure.