Methods and apparatus for identifying spectral peaks in neuronal spiking representation of a signal

• US 9,147,157 B2
• Filed: 11/06/2012
• Issued: 09/29/2015
• Est. Priority Date: 11/06/2012
• Status: Expired due to Fees

• Current Assignee: Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: Campos, Michael
• Primary Examiner(s): Gaffin, Jeffrey A
• Assistant Examiner(s): Smith, Paulinho E
• Application Number:
  US13/669,521
• Publication Number:
  US 20140129496A1
• Time in Patent Office: 1,057 Days
• US Class Current: 1/1
• CPC Class Codes: G06N 3/049 Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs

Abstract

Certain aspects of the present disclosure provide methods and apparatus for identifying spectral peaks in a neuronal spiking representation of a signal, such as an auditory signal. One example method generally includes receiving a signal; filtering the signal into a plurality of channels using a plurality of filters having different frequency passbands; sending the filtered signal in each of the channels to a first type of spiking neuron model; sending the filtered signal in each of the channels to a second type of spiking neuron model; and identifying one or more spectral peaks in the signal based on a first output of the first type of spiking neuron model and a second output of the second type of spiking neuron model for each of the channels.

Claims

1. A method for neural processing, comprising:

receiving a signal;

filtering the signal into a plurality of channels using a plurality of filters having different frequency passbands;

sending the filtered signal in each of the channels to a first type of spiking neuron model;

sending the filtered signal in each of the channels to a second type of spiking neuron model; and

identifying one or more spectral peaks in the signal based on a first output of the first type of spiking neuron model and a second output of the second type of spiking neuron model for each of the channels, wherein each of the spectral peaks comprises a relative concentration of power in a frequency spectrum of the signal.

2. The method of claim 1, wherein identifying the spectral peaks comprises:

sending the first output of the first type to a third type of spiking neuron model for each of the channels; and

sending the second output of the second type to a fourth type of spiking neuron model for each of the channels, wherein the third type differs from the fourth type in at least one characteristic.

3. The method of claim 2, wherein the third type of spiking neuron model represents a chopper cell in a cochlear nucleus.

4. The method of claim 3, wherein the fourth type of spiking neuron model represents a wideband inhibitor (WBI) cell in the cochlear nucleus.

5. The method of claim 2, wherein the at least one characteristic comprises at least one of a spiking threshold or a bandwidth.

6. The method of claim 2, wherein the third type has a lower spiking threshold and is more narrowly tuned than the fourth type of spiking neuron model.

7. The method of claim 2, wherein the fourth type inhibits the third type of spiking neuron model for each of the channels.

8. The method of claim 2, wherein identifying the spectral peaks comprises using feed-forward inhibition of the third type of spiking neuron model by the fourth type.

9. The method of claim 1, wherein the signal comprises an electrical representation of an audio signal.

10. The method of claim 9, wherein the plurality of channels span a hearing range of frequencies.

11. The method of claim 1, wherein the first type of spiking neuron model has a smaller dynamic range with respect to intensity and greater phase-locking capability than the second type of spiking neuron model.

12. The method of claim 1, wherein the first type of spiking neuron model is specialized for encoding temporal information and wherein the second type of spiking neuron model is specialized for encoding intensity information.

13. The method of claim 1, wherein the first type of spiking neuron model has a smaller fan-out than the second type of spiking neuron model.

14. The method of claim 1, wherein the first type of spiking neuron model represents a high spontaneous rate (HSR) neuron of an auditory nerve and the second type of spiking neuron model represents a low spontaneous rate (LSR) neuron of the auditory nerve.

15. The method of claim 1, further comprising outputting an indication of the spectral peaks to a display.

16. The method of claim 1, further comprising determining a frame timing from the spectral peaks.

17. An apparatus for neural processing, comprising:

a processing system configured to;

receive a signal;

filter the signal into a plurality of channels using a plurality of filters having different frequency passbands;

send the filtered signal in each of the channels to a first type of spiking neuron model;

send the filtered signal in each of the channels to a second type of spiking neuron model; and

identify one or more spectral peaks in the signal based on a first output of the first type of spiking neuron model and a second output of the second type of spiking neuron model for each of the channels.

18. The apparatus of claim 17, wherein the processing system is configured to identify the spectral peaks by:

sending the first output of the first type to a third type of spiking neuron model for each of the channels; and

sending the second output of the second type to a fourth type of spiking neuron model for each of the channels, wherein the third type differs from the fourth type in at least one characteristic.

19. The apparatus of claim 18, wherein the third type of spiking neuron model represents a chopper cell in a cochlear nucleus.

20. The apparatus of claim 19, wherein the fourth type of spiking neuron model represents a wideband inhibitor (WBI) cell in the cochlear nucleus.

21. The apparatus of claim 18, wherein the at least one characteristic comprises at least one of a spiking threshold or a bandwidth.

22. The apparatus of claim 18, wherein the third type has a lower spiking threshold and is more narrowly tuned than the fourth type of spiking neuron model.

23. The apparatus of claim 18, wherein the fourth type inhibits the third type of spiking neuron model for each of the channels.

24. The apparatus of claim 18, wherein the processing system is configured to identify the spectral peaks by using feed-forward inhibition of the third type of spiking neuron model by the fourth type.

25. The apparatus of claim 17, wherein the signal comprises an electrical representation of an audio signal.

26. The apparatus of claim 25, wherein the plurality of channels span a hearing range of frequencies.

27. The apparatus of claim 17, wherein the first type of spiking neuron model has a smaller dynamic range with respect to intensity and greater phase-locking capability than the second type of spiking neuron model.

28. The apparatus of claim 17, wherein the first type of spiking neuron model is specialized for encoding temporal information and wherein the second type of spiking neuron model is specialized for encoding intensity information.

29. The apparatus of claim 17, wherein the first type of spiking neuron model has a smaller fan-out than the second type of spiking neuron model.

30. The apparatus of claim 17, wherein the first type of spiking neuron model represents a high spontaneous rate (HSR) neuron of an auditory nerve and the second type of spiking neuron model represents a low spontaneous rate (LSR) neuron of the auditory nerve.

31. The apparatus of claim 17, wherein the processing system is further configured to output an indication of the spectral peaks to a display.

32. The apparatus of claim 17, wherein the processing system is further configured to determine a frame timing from the spectral peaks.

33. An apparatus for neural processing, comprising:

means for receiving a signal;

means for filtering the signal into a plurality of channels using a plurality of filters having different frequency passbands;

means for sending the filtered signal in each of the channels to a first type of spiking neuron model;

means for sending the filtered signal in each of the channels to a second type of spiking neuron model; and

means for identifying one or more spectral peaks in the signal based on a first output of the first type of spiking neuron model and a second output of the second type of spiking neuron model for each of the channels.

34. The apparatus of claim 33, wherein the means for identifying the spectral peaks is configured to:

send the first output of the first type to a third type of spiking neuron model for each of the channels; and

send the second output of the second type to a fourth type of spiking neuron model for each of the channels, wherein the third type differs from the fourth type in at least one characteristic.

35. The apparatus of claim 34, wherein the third type of spiking neuron model represents a chopper cell in a cochlear nucleus.

36. The apparatus of claim 35, wherein the fourth type of spiking neuron model represents a wideband inhibitor (WBI) cell in the cochlear nucleus.

37. The apparatus of claim 34, wherein the at least one characteristic comprises at least one of a spiking threshold or a bandwidth.

38. The apparatus of claim 34, wherein the third type has a lower spiking threshold and is more narrowly tuned than the fourth type of spiking neuron model.

39. The apparatus of claim 34, wherein the fourth type inhibits the third type of spiking neuron model for each of the channels.

40. The apparatus of claim 34, wherein the means for identifying the spectral peaks is configured to use feed-forward inhibition of the third type of spiking neuron model by the fourth type.

41. The apparatus of claim 33, wherein the signal comprises an electrical representation of an audio signal.

42. The apparatus of claim 41, wherein the plurality of channels span a hearing range of frequencies.

43. The apparatus of claim 33, wherein the first type of spiking neuron model has a smaller dynamic range with respect to intensity and greater phase-locking capability than the second type of spiking neuron model.

44. The apparatus of claim 33, wherein the first type of spiking neuron model is specialized for encoding temporal information and wherein the second type of spiking neuron model is specialized for encoding intensity information.

45. The apparatus of claim 33, wherein the first type of spiking neuron model has a smaller fan-out than the second type of spiking neuron model.

46. The apparatus of claim 33, wherein the first type of spiking neuron model represents a high spontaneous rate (HSR) neuron of an auditory nerve and the second type of spiking neuron model represents a low spontaneous rate (LSR) neuron of the auditory nerve.

47. The apparatus of claim 33, further comprising means for outputting an indication of the spectral peaks to a display.

48. The apparatus of claim 33, further comprising means for determining a frame timing from the spectral peaks.

49. A computer program product for neural processing, comprising a non-transitory computer-readable medium comprising instructions executable to:

receive a signal;

filter the signal into a plurality of channels using a plurality of filters having different frequency passbands;

send the filtered signal in each of the channels to a first type of spiking neuron model;

send the filtered signal in each of the channels to a second type of spiking neuron model; and

identify one or more spectral peaks in the signal based on a first output of the first type of spiking neuron model and a second output of the second type of spiking neuron model for each of the channels.

50. The computer program product of claim 49, wherein identifying the spectral peaks comprises:

sending the first output of the first type to a third type of spiking neuron model for each of the channels; and

sending the second output of the second type to a fourth type of spiking neuron model for each of the channels, wherein the third type differs from the fourth type in at least one characteristic.

51. The computer program product of claim 50, wherein the third type of spiking neuron model represents a chopper cell in a cochlear nucleus.

52. The computer program product of claim 51, wherein the fourth type of spiking neuron model represents a wideband inhibitor (WBI) cell in the cochlear nucleus.

53. The computer program product of claim 50, wherein the at least one characteristic comprises at least one of a spiking threshold or a bandwidth.

54. The computer program product of claim 50, wherein the third type has a lower spiking threshold and is more narrowly tuned than the fourth type of spiking neuron model.

55. The computer program product of claim 50, wherein the fourth type inhibits the third type of spiking neuron model for each of the channels.

56. The computer program product of claim 50, wherein identifying the spectral peaks comprises using feed-forward inhibition of the third type of spiking neuron model by the fourth type.

57. The computer program product of claim 49, wherein the signal comprises an electrical representation of an audio signal.

58. The computer program product of claim 57, wherein the plurality of channels span a hearing range of frequencies.

59. The computer program product of claim 49, wherein the first type of spiking neuron model has a smaller dynamic range with respect to intensity and greater phase-locking capability than the second type of spiking neuron model.

60. The computer program product of claim 49, wherein the first type of spiking neuron model is specialized for encoding temporal information and wherein the second type of spiking neuron model is specialized for encoding intensity information.

61. The computer program product of claim 49, wherein the first type of spiking neuron model has a smaller fan-out than the second type of spiking neuron model.

62. The computer program product of claim 49, wherein the first type of spiking neuron model represents a high spontaneous rate (HSR) neuron of an auditory nerve and the second type of spiking neuron model represents a low spontaneous rate (LSR) neuron of the auditory nerve.

63. The computer program product of claim 49, further comprising instructions executable to output an indication of the spectral peaks to a display.

64. The computer program product of claim 49, further comprising instructions executable to determine a frame timing from the spectral peaks.