POSTURE STATE DETECTION

• US 20100280579A1
• Filed: 01/26/2010
• Published: 11/04/2010
• Est. Priority Date: 04/30/2009
• Status: Abandoned Application

• Current Assignee: Medtronic Incorporated (Medtronic Plc)
• Original Assignee: Medtronic Incorporated (Medtronic Plc)
• Inventors: Denison, Timothy J., Carlson, David L., Shoeb, Ali H., Linde, David E.
• Application Number:
  US12/694,053
• Publication Number:
  US 20100280579A1
• Time in Patent Office: Days
• US Class Current: 607/62
• CPC Class Codes: A61N 1/36082 Cognitive or psychiatric applications, e.g. dementia or Alzheimer's disease
• CPC Class Codes: G06F 18/241 relating to the classification model, e.g. parametric or non-parametric approaches
• CPC Class Codes: G06F 2221/2101 Auditing as a secondary aspect
• CPC Class Codes: G06F 2221/2105 Dual mode as a secondary aspect
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 20/10 using kernel methods, e.g. support vector machines [SVM]
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 50/30 for calculating health indices; for individual health risk assessment
• CPC Class Codes: G16H 50/50 for simulation or modelling of medical disorders

Abstract

A patient state is detected with at least one classification boundary generated by a supervised machine learning technique, such as a support vector machine. The patient state can be, for example, a patient posture state. In some examples, the patient state detection is used to at least one of control the delivery of therapy to a patient, to generate a patient notification, to initiate data recording, or to evaluate a patient condition. In addition, an evaluation metric can be determined based on a feature vector, which is determined based on characteristics of a patient parameter signal, and the classification boundary. Example evaluation metrics can be based on a distance between at least one feature vector and the classification boundary and/or a trajectory of a plurality of feature vectors relative to the classification boundary over time.

Claims

1. A method comprising:

receiving a signal indicative of a patient parameter;

receiving information identifying an occurrence of a posture state;

determining at least a first value of a characteristic of the signal that is indicative of the patient being in the posture state and at least a second value of the characteristic of the signal that is indicative of the patient not being in the posture state, wherein the first and second values are different; and

applying a supervised machine learning technique to define a classification boundary based on the first and second values of the characteristics of the signal, wherein a medical device utilizes the classification boundary to classify a subsequently sensed signal of the patient as indicative of the posture state.

2. The method of claim 1, wherein the supervised machine learning technique comprises at least one of a genetic algorithm or an artificial neural network.

3. The method of claim 2, wherein the artificial neural network comprises at least one of a support vector machine or a Bayesian classifier technique.

4. The method of claim 1, wherein the patient parameter comprises a least one of patient motion or activity, heart rate, respiratory rate, electrodermal activity, thermal activity or muscle activity.

5. The method of claim 1, wherein the characteristic of the signal comprises at least one of an amplitude value or a frequency domain characteristic of the signal.

6. The method of claim 1, wherein receiving information identifying the occurrence of the posture state comprises receiving input from a user indicating the occurrence of the posture state.

7. The method of claim 1, wherein the posture state comprises at least one of an upright posture state, an upright and active posture state, or a lying down posture state.

8. The method of claim 1, wherein the characteristic comprises a first characteristic, the method further comprising determining a third value of a second characteristic of the signal that is indicative of the patient being in the posture state, and determining a fourth value of the second characteristic of the signal that is indicative of the patient not being in the posture state, wherein applying a supervised machine learning technique to define a classification boundary comprises applying the supervised machine learning technique to define the classification boundary based on the first and second values of the first characteristic and the third and fourth values of the second characteristic.

9. The method of claim 1, wherein receiving the signal indicative of the patient posture comprises receiving the signal from an accelerometer, and wherein determining the first value of the characteristic of the signal that is indicative of the patient being in the posture state comprises determining respective values of the characteristic for at least two portions of the signal indicative of movement along respective axes of the accelerometer and wherein determining the at least the second value of the characteristic of the signal that is indicative of the patient not being in the posture state comprises determining respective values of the characteristic for at least two portions of the signal indicative of movement along respective axes of the accelerometer.

10. A system comprising:

a sensing module that generates a signal indicative of a patient parameter;

a processor that receives the signal indicative of the patient parameter, receives information identifying an occurrence of a posture state, determines at least a first value of a characteristic of the signal that is indicative of the patient being in the posture state and at least a second value of the characteristic of the signal that is indicative of the patient not being in the posture state, wherein the first and second values are different, and applies a supervised machine learning technique to define a classification boundary based on the first and second values of the characteristic of the signal; and

a medical device that utilizes the classification boundary to classify a subsequently sensed signal of the patient as indicative of the posture state.

11. The system of claim 10, wherein the medical device comprises the processor.

12. The system of claim 10, further comprising a medical device programmer comprising the processor.

13. The system of claim 10, wherein the supervised machine learning technique comprises at least one of a genetic algorithm or an artificial neural network.

14. The system of claim 13, wherein the artificial neural network comprises at least one of a support vector machine or a Bayesian classifier technique.

15. The system of claim 10, wherein the patient parameter comprises a least one of patient motion or activity, heart rate, respiratory rate, electrodermal activity, thermal activity or muscle activity.

16. The system of claim 10, further comprising a programmer, wherein the processor receives information identifying the occurrence of the posture state from a user via the programmer.

17. A method comprising:

receiving a signal indicative of patient parameter;

applying a classification algorithm determined based on a supervised machine learning technique to classify a patient posture state based on the signal, wherein the classification algorithm defines a classification boundary; and

controlling therapy delivery to the patient based on the determined patient posture state.

18. The method of claim 17, wherein the supervised machine learning technique comprises at least one of a genetic algorithm or an artificial neural network.

19. The method of claim 18, wherein the artificial neural network comprises at least one of a support vector machine-based or a Bayesian classifier.

20. The method of claim 17, wherein the patient parameter comprises a least one of patient motion or activity, heart rate, respiratory rate, electrodermal activity, thermal activity or muscle activity.

21. The method of claim 17, wherein applying the classification algorithm to classify the patient posture state based on the signal comprises:

determining a feature vector based on at least a first value of a first characteristic of the signal and a second value of a second characteristic of the signal, wherein the supervised machine learning technique determined the classification boundary based on training data including values of the first and second characteristics;

determining a side of the classification boundary on which the feature vector lies; and

determining the patient posture state associated with the side of the classification boundary on which the feature vector lies.

22. The method of claim 17, wherein controlling therapy delivery to the patient based on the determined patient posture state comprises at least one of activating therapy delivery, deactivating therapy delivery or adjusting one or more therapy parameters.

23. A system comprising:

a therapy module that delivers therapy to a patient;

a sensor that generates a signal indicative of patient posture; and

a processor that applies a classification algorithm determined based on a supervised machine learning technique to classify a patient posture state based on the signal and controls the therapy module based on the determined patient posture state.

24. The system of claim 23, wherein the supervised machine learning technique comprises at least one of a genetic algorithm or an artificial neural network.

25. The system of claim 24, wherein the artificial neural network comprises at least one of a support vector machine or a Bayesian classifier.

26. The system of claim 23, wherein the patient parameter comprises a least one of patient motion or activity, heart rate, respiratory rate, electrodermal activity, thermal activity or muscle activity.

27. The system of claim 23, wherein the processor applies the classification algorithm to classify the patient posture state based on the signal by at least:

determining a feature vector based on at least a first value of a first characteristic of the signal and a second value of a second characteristic of the signal, wherein the supervised machine learning technique determined the classification boundary based on training data including values of the first and second characteristics;

determining a side of the classification boundary on which the feature vector lies; and

determining the patient posture state associated with the side of the classification boundary on which the feature vector lies.

28. The system of claim 23, wherein the processor controls the therapy module by at least one of activating therapy delivery by the therapy module based on the determined patient posture state, deactivating therapy delivery by the therapy module based on the determined patient posture state or adjusting at least one parameter with which the therapy module delivers therapy to the patient based on the determined patient posture state.

29. A system comprising:

means for receiving a signal indicative of a patient posture;

means for receiving information identifying an occurrence of a posture state;

means for determining at least a first value of a characteristic of the signal that is indicative of the patient being in the posture state and at least a second value of the characteristic of the signal that is indicative of the patient not being in the posture state, wherein the first and second values are different; and

means for applying a supervised machine learning technique to define a classification boundary based on the first and second values of the characteristics of the signal, wherein a medical device utilizes the classification boundary to classify a subsequently sensed signal of the patient as indicative of the posture state.

30. The system of claim 29, wherein the supervised machine learning technique comprises a support vector machine.

31. A system comprising:

means for receiving a signal indicative of patient parameter;

means for applying a classification algorithm determined based on a supervised machine learning technique to classify a patient posture state based on the signal, wherein the classification algorithm defines a classification boundary; and

means for controlling therapy delivery to the patient based on the determined patient posture state.

32. The system of claim 31, wherein the means for applying a classification algorithm to classify a patient posture state based on the signal comprises:

means for determining a feature vector based on at least a first value of a first characteristic of the signal and a second value of a second characteristic of the signal, wherein the supervised machine learning technique determined the classification boundary based on training data including values of the first and second characteristics; and

means for determining a side of the classification boundary on which the feature vector lies, wherein the means for applying a classification algorithm to classify a patient posture state determines the patient posture state corresponding to the side of the classification boundary on which the feature vector lies.

33. A computer-readable medium comprising instructions that cause a programmable processor to:

receive a signal indicative of a patient posture;

receive information identifying an occurrence of a posture state;

determine at least a first value of a characteristic of the signal that is indicative of the patient being in the posture state and at least a second value of the characteristic of the signal that is indicative of the patient not being in the posture state, wherein the first and second values are different; and

apply a supervised machine learning technique to define a classification boundary based on the first and second values of the characteristics of the signal, wherein a medical device utilizes the classification boundary to classify a subsequently sensed signal of the patient as indicative of the posture state.

34. A computer-readable medium comprising instructions that cause a programmable processor to:

receive a signal indicative of patient parameter;

apply a classification algorithm determined based on a supervised machine learning technique to classify a patient posture state based on the signal, wherein the classification algorithm defines a classification boundary; and

control therapy delivery to the patient based on the determined patient posture state.