Patient condition display

• US 7,031,857 B2
• Filed: 05/29/2002
• Issued: 04/18/2006
• Est. Priority Date: 05/31/2001
• Status: Expired due to Term

• Current Assignee: Oxford University Innovation Limited (University of Oxford)
• Original Assignee: Isis Innovation Limited
• Inventors: Tarassenko, Lionel, Townsend, Neil William
• Primary Examiner(s): Hoff, Marc S.
• Assistant Examiner(s): Charioui, Mohamed
• Application Number:
  US10/479,224
• Publication Number:
  US 20040148140A1
• Time in Patent Office: 1,420 Days
• Field of Search: 702/16, 702/19, 702/32, 702/57, 702/67, 702/70, 702/71, 702/123, 702/124, 702/146, 702/158, 702/180, 702/183, 702/189, 702/193, 706/15, 706/16, 706/25, 600/300, 600/301, 600/532
• US Class Current: 702/67
• CPC Class Codes: A61B 5/0205 Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
• CPC Class Codes: A61B 5/7267 involving training the classification device
• CPC Class Codes: A61B 5/742 using visual displays A61B5/339 takes precedence
• CPC Class Codes: A61B 5/7445 Display arrangements, e.g. multiple display units
• CPC Class Codes: G16H 30/40 for processing medical images, e.g. editing
• CPC Class Codes: G16H 40/63 for local operation
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients

Abstract

Data from a plurality of sensors representing a patient'"'"'s condition, including measurement signals and also secondary parameters derived from the measurement signals, are displayed in a simple way by calculating a novelty index constituting a one-dimensional visualization space. The novelty index is based on the distance of the current data point in a multi-dimensional measurement space, whose coordinates are defined by the values of the measurement signals and secondary parameters, from a predefined normal point. This may be achieved by using a suitably trained artificial neural network to sum the distance between the current data point in the measurement space and a plurality of prototype points representing normality.

Claims

1. A method of displaying a graphical representation of a patient'"'"'s condition as measured by n parameters, where n>

3, comprising mapping data points represented by said n parameters from an n-dimensional measurement space into an m-dimensional visualization space, where m<

n, using a dimensionality reduction mapping, and displaying the visualization space and the data points mapped into it, and wherein the displaying is adapted to the display of dynamically changing values of said parameters by means of the mapping being carried out by a trained artificial neural network, wherein said mapping maps each data point on receipt onto a novelty index constituting a one dimensional visualization space and the value of said novelty index is displayed against time, and wherein the novelty index of each point is based on its distance in the n-dimensional measurement space from a predefined set of points representing a normal condition for the patient.

2. A method according to claim 1, wherein the n parameters are selected from the group consisting of:

a respiration measurement, an oxygen saturation measurement, blood pressure, skin temperature, S-T segment elevation/depression, heart rate, heart rate variability and respiration rate.

3. A method according to claim 2, wherein the respiration measurement is an impedance pneumography measurement.

4. A method according to claim 2, wherein the oxygen saturation measurement is a pulse oximetry measurement.

5. A method according to claim 1, further comprising normalizing the parameters prior to said mapping.

6. A method according to claim 5, wherein the normalization is by a zero mean, unit variance transformation.

7. A method according to claim 5, wherein the normalization is by an empirical transformation.

8. A method according to claim 1, wherein the dimensionality-reduction is a distance-preserving mapping.

9. A method according to claim 8, wherein the distance preserving mapping minimizes the difference in distance between pairs of points in the n-dimensional measurement space and the corresponding pairs of points in the m-dimensional visualization space.

10. A method according to claim 9, wherein the distance-preserving mapping is Sammon'"'"'s mapping.

11. A method apparatus according to claim 1, wherein training data for the artificial neural network is provided by pre-clustering in the n-dimensional measurement space data points each comprising a set of said parameters, deriving the centers of the clusters, and using the centers of the clusters as the training data.

12. A method according to claim 11, wherein the data points are pre-clustered using the K-means method.

13. A method according to claim 1, wherein the artificial neural network is a Radial Basis Function (RBF) neural network.

14. A method according to claim 1, wherein the artificial neural network is trained with data comprising a set of said parameters from the patient.

15. A method according to claim 1, wherein the artificial neural network is trained with data comprising a plurality of sets of said parameters from a group of patients.

16. A method according to claim 1, further comprising generating an alarm responsive to said value of the novelty index exceeding a threshold.

17. A method according to claim 1, wherein the novelty index of each point is calculated by summing the distance in the n-dimensional measurement space between it and each of a set of prototype points representing the normal condition for the patient.

18. Apparatus for displaying a graphical representation of a patient'"'"'s condition as measured by n parameters, where n>

3, comprising a processor which maps data points represented by said n parameters from an n-dimensional measurement space into an m-dimensional visualization space, where m<

n, using a dimensionality reduction mapping, and a display which displays the visualization space and the data points mapped into it, and which is adapted to the display of dynamically changing values of said parameters by means of the mapping being carried out by a trained artificial neural network, wherein said processor maps each data point on receipt onto a novelty index constituting a one dimensional visualization space and said display displays the value of said novelty index against time, and wherein the novelty index of each point is based on its distance in the n-dimensional measurement space from a predefined set of points representing a normal condition for the patient.

19. Apparatus according to claim 18, wherein the n parameters are selected from the group consisting of:

a respiration measurement, an oxygen saturation measurement, blood pressure, skin temperature, S-T segment elevation/depression, heart rate, heart rate variability and respiration rate.

20. Apparatus according to claim 19, wherein the respiration measurement is an impedance pneumography measurement.

21. Apparatus according to claim 19, wherein the oxygen saturation measurement is a pulse oximetry measurement.

22. Apparatus according to claim 18, wherein said processor normalizes the parameters prior to said mapping.

23. Apparatus according to claim 22, wherein the processor normalizes the parameters using a zero mean, unit variance transformation.

24. Apparatus according to claim 22, wherein the processor normalizes the parameters using an empirical transformation.

25. Apparatus according to claim 18, wherein the dimensionality reduction is a distance-preserving mapping.

26. Apparatus according to claim 25, wherein the distance preserving mapping minimizes the difference in the distance between pairs of points in the n-dimensional measurement space and the corresponding pairs of points in the m-dimensional visualization space.

27. Apparatus according to claim 26, wherein the distance-preserving mapping is Sammon'"'"'s mapping.

28. Apparatus according to claim 18, wherein the artificial neural network is a Radial Basis Function (RBF) neural network.

29. Apparatus according to claim 18, wherein the artificial neural network is trained with data comprising a plurality of sets of said parameters from the patient.

30. Apparatus according to claim 18, wherein the artificial neural network is trained with data comprising a plurality of sets of said parameters from a group of patients.

31. Apparatus according to claim 18, further comprising an alarm responsive to said value of the novelty index exceeding a threshold.

32. Apparatus according to claim 18, wherein the novelty index of each point is calculated by summing the distance in the n-dimensional measurement space between it and each of a set of prototype points representing the normal condition for the patient.

33. A patient condition monitor system comprising:

sensors for monitoring a condition of a patient; and

apparatus for displaying a graphical representation of a patient'"'"'s condition as measured by n parameters based on data from the sensors, where n>

3, the apparatus comprising a processor which maps data points represented by said n parameters from an n-dimensional measurement space into an m-dimensional visualization space, where m<

n, using a dimensionality reduction mapping, and a display which displays the visualization space and the data points mapped into it, and which is adapted to the display of dynamically changing values of said parameters by means of the mapping being carried out by a trained artificial neural network, wherein said processor maps each data point on receipt onto a novelty index constituting a one dimensional visualization space and said display displays the value of said novelty index against time, and wherein the novelty index of each point is based on its distance in the n-dimensional measurement space from a predefined set of points representing a normal condition for the patient.

34. A computer program tangibly embodied on a computer readable storage medium and comprising instructions which, when executed by a computer system, perform a method of displaying a graphical representation of a patient'"'"'s condition as measured by n parameters, where n>

3, the method comprising mapping data points represented by said n parameters from an n-dimensional measurement space into an m-dimensional visualization space, where m<

n, using a dimensionality reduction mapping, and displaying the visualization space and the data points mapped into it, and wherein the displaying is adapted to the display of dynamically changing values of said parameters by means of the mapping being carried out by a trained artificial neural network, wherein said mapping maps each data point on receipt onto a novelty index constituting a one dimensional visualization space and the value of said novelty index is displayed against time, and wherein the novelty index of each point is based on its distance in the n-dimensional measurement space from a predefined set of points representing a normal condition for the patient.

35. A computer readable storage medium on which a computer program is tangibly embodied, the computer program comprising instructions which, when executed by a computer system, perform a method of displaying a graphical representation of a patient'"'"'s condition as measured by n parameters, where n>

3, the method comprising mapping data points represented by said n parameters from an n-dimensional measurement space into an m-dimensional visualization space, where m<

n, using a dimensionality reduction mapping, and displaying the visualization space and the data points mapped into it, and wherein the displaying is adapted to the display of dynamically changing values of said parameters by means of the mapping being carried out by a trained artificial neural network, wherein said mapping maps each data point on receipt onto a novelty index constituting a one dimensional visualization space and the value of said novelty index is displayed against time, and wherein the novelty index of each point is based on its distance in the n-dimensional measurement space from a predefined set of points representing a normal condition for the patient.