Data compression method for implantable medical devices

• US 20060094968A1
• Filed: 11/01/2004
• Published: 05/04/2006
• Est. Priority Date: 11/01/2004
• Status: Active Grant

• Current Assignee: Medtronic Incorporated (Medtronic Plc)
• Original Assignee: Medtronic Incorporated (Medtronic Plc)
• Inventors: Drew, Touby A.
• Granted Patent: US 7,288,066 B2
• Time in Patent Office: Days
• US Class Current: 600/509
• CPC Class Codes: A61B 5/369 Electroencephalography [EEG] devices for psychotechnics A61B5/16
• CPC Class Codes: A61B 5/37 Intracranial electroencephalography [IC-EEG], e.g. electrocorticography [ECoG]
• CPC Class Codes: A61B 5/372 Analysis of electroencephalograms
• CPC Class Codes: A61B 5/4094 Diagnosing or monitoring seizure diseases, e.g. epilepsy
• CPC Class Codes: A61B 5/7232 involving compression of the physiological signal, e.g. to extend the signal recording period
• CPC Class Codes: H03M 7/30 Compression speech analysis-synthesis for redundancy reduction G10L19/00; for image communication H04N; Expansion; Suppression of unnecessary data, e.g. redundancy reduction

Abstract

A method of data compression of physiological signals for use in implantable medical devices (IMDs) in which temporal information in the input data points is retained in the compressed sample points with a potential loss of amplitude precision. The data compression method includes distributing the stream of data samples or data points of the physiologic waveform into respective major data bins within the data range of the samples. Each major data bin can then be sub-divided into minor data bins. Compressed data points associated with the data samples are assigned values indicating which major data bin the original data point falls into, or which minor data bin the data point falls into if the data point immediately follows a data point in the same major data bin. Higher levels of revision beyond the major and minor bins may also be used to achieve greater levels of amplitude precision.

Claims

1. A method of performing data compression of a physiologic waveform, comprising:

obtaining a stream of data points within a data range, the data points corresponding to successive data samples of the physiologic waveform;

distributing the data range into major data bins;

determining within which major data bin each data point falls;

determining whether each data point in the stream of data points falls within the same major data bin as an immediately preceding data point of such data point;

retaining a compressed data point for each data point, the compressed data point having a revision level component and a division value, the revision level component indicating a revision condition, and the division component indicating the data bin which the associated data point falls within;

for each data point falling in the major data bin different from the major data bin of the immediately preceding data point, setting the division value of the respective compressed data point equal to the major data bin of such data point and setting the revision level component of the respective compressed data point to indicate a non-revision; and

for each data point falling within the same major data bin as the immediately preceding data point, dividing the major data bin into which such data point falls into a plurality of minor data bins unless the major data bin has already been divided into minor data bins, determining which minor data bin such data point falls within, setting the division value of the respective compressed data point equal to the minor data bin of such data point and setting the revision level component of the respective compressed data point to indicate a first-level revision.

2. The method of claim 1, wherein the compressed data points consist of revision level components and division value components.

3. The method of claim 1, wherein the division values comprise digital data having a number of bits, Z.

4. The method of claim 3, wherein, for each data point falling in the major data bin different from the major data bin of the immediately preceding data point, the division value of the respective compressed data point is set equal to the major data bin of such data point by using the Z most significant bits of the digital data corresponding to such data point.

5. The method of claim 3, wherein, for each data point falling within the same major data bin as the immediately preceding data point, the division value of the respective compressed data point is set equal to the minor data bin of such data point by ignoring the Z most significant bits and storing the next Z bits of the digital data corresponding to such data point.

6. The method of claim 3, further including a method of reconstructing the physiologic waveform from the compressed data points, comprising:

creating an expanded data point of M-bit digital data for each compressed data point,

for each compressed data point having its revision level component indicating non-revision, forming the Z most significant bits of the associated expanded data point with the division value of such compressed data point, adding (M-Z) bits as the remaining less significant bits of the associated M-bit expanded data point; and

for each compressed data point having its revision level component indicating a first-level revision, forming the associated expanded data point with the division value of such compressed data point forming the next most significant Z bits of digital data after the first Z most significant data bits, adding (M−

2*Z) bits as the remaining less significant bits of the associated M-bit expanded data point, and forming the Z most significant bits of digital data missing from such expanded data point with the division value of the nearest preceding compressed data point having its revision level component indicating non-revision.

7. The method of claim 6, wherein the (M−

Z) bits added are zeroes.

8. The method of claim 6, wherein the (M−

2*Z) bits added are zeroes.

9. The method of claim 1, wherein the revision level component comprises digital data having a length of one bit.

10. The method of claim 1, wherein, for each data point falling in the major data bin different from the major data bin of the immediately preceding data point, the revision level component of the respective compressed data point is set to “

0”

to be indicative of a non-refined compressed data point.

11. The method of claim 1, wherein the major data bins are not uniformly sized.

12. The method of claim 1, wherein the minor data bins are not uniformly distributed within the major data bins.

13. The method of claim 1, wherein the number of major data bins equals the number of minor data bins.

14. The method of claim 1, wherein the physiologic waveform is an electroencephalogram (EEG) signal.

15. The method of claim 1, further including a method of smoothing over anomalous cross-over points, comprising:

identifying cross-over points as data points crossing over into the major data bin adjacent to the major data bin of two or more of the nearest preceding data points,

reconstructing the data points associated with such cross-over points to fall numerically between the data points immediately preceding and immediately following such cross-over points.

16. A method of performing data compression of a physiologic waveform, comprising:

obtaining a stream of data points within a data range, the data points corresponding to successive data samples of the physiologic waveform;

defining the data range into data bins, each data bin at least once further dividing itself into successively narrower data bins, each successively narrower division of the data bins defining a respective revision level, the data range divided into data bins at an initial revision level, and a last narrower division of the data bins defining a narrowest revision level;

creating a compressed data point for each data point, the compressed data point having a revision level component and a division value;

proceeding one level at a time from the initial revision level towards the narrowest revision level, comparing each data point to its immediately preceding data point until reaching the first of the level where the compared data points fall in different data bins, the narrowest revision level, or one level beyond the revision level of the compressed data point associated with the immediately preceding data point, the first such level reached being defined as the selected revision level, and setting the revision level component of the compressed data point corresponding to such data point to the selected revision level and setting the division value of such compressed data point to the data bin at the selected revision level that such data point falls within.

17. The method of claim 16, wherein the data range is divided into major data bins, and each data bin is only once further divided into narrower data bins defined as minor data bins, the one narrower division of the data defining a second revision level.

18. The method of claim 16, wherein the physiologic waveform is an electroencephalogram (EEG) signal.

19. The method of claim 16, wherein each data point comprises digital data having a number of bits, M, and each compressed data point comprises digital data having a number of bits, X.

20. The method of claim 16, wherein X is less than M.

21. The method of claim 16, wherein the revision level components comprises digital data having a length of two bits.

22. The method of claim 16, wherein the division value comprises digital data having a number of bits, Z.

23. The method of claim 22, wherein the number of data bins at the initial revision level equals 2^Z.

24. The method of claim 16, wherein the number of data bins in each level is the same.

25. The method of claim 16, wherein the data bins are not uniformly sized or distributed.

26. The method of claim 16, wherein the revision level of the compressed data point associated with a first data point in the stream of data points is set to the initial revision level.

27. The method of claim 16, wherein the division value of the compressed data point associated with a first data point in the stream of data points is set to the data bin in the initial revision level that the first data points falls within.

28. The method of claim 16, wherein one or more of the compressed sample points identifies events of interest by having its revision level component and division value equal the revision level and division value of the compressed sample point associated with the immediately preceding data point.

29. A system for performing data compression of a physiologic waveform, comprising:

means for obtaining a stream of data points within a data range, the data points corresponding to successive data samples of the physiologic waveform;

means for defining the data range into data bins, each data bin at least once further dividing itself into successively narrower data bins, each successively narrower division of the data bins defining a respective revision level, the data range divided into data bins at an initial revision level, and a last narrower division of the data bins defining a narrowest revision level;

means for creating a compressed data point for each data point, the compressed data point having a revision level component and a division value;

proceeding one level at a time from the initial revision level towards the narrowest revision level, means for comparing each data point to its immediately preceding data point until reaching the first of the level where the compared data points fall in different data bins, the narrowest revision level, or one level beyond the revision level of the compressed data point associated with the immediately preceding data point, the first such level reached being defined as the selected revision level, and means for setting the revision level component of the compressed data point corresponding to such data point to the selected revision level and setting the division value of such compressed data point to the data bin at the selected revision level that such data point falls within.

30. The system of claim 29, wherein the system is located in an implantable medical device (IMD).

31. The system of claim 29, wherein the means for comparing each data point to its immediately preceding data point includes digital logic circuitry.

32. The system of claim 30, wherein the IMD includes a hermetically sealed housing, a battery disposed within the housing, and electronic circuitry including a microprocessor and waveform memory.

33. The system of claim 32, wherein the electronic circuitry possesses the means for comparing each data point to its immediately preceding data point.

34. A method of performing data compression of a physiologic waveform, comprising:

obtaining a stream of data points within a data range, the data points corresponding to successive data samples of the physiologic waveform;

distributing the data range into major data bins;

subdividing each major data bin into minor data bins;

determining within which major data bin each data point falls;

retaining a compressed data point associated with each data point;

for successive data points falling in the same major data bin, assigning the compressed data point associated with the second of the successive data points a value equal to the major data bin within which the second of the successive data points falls; and

for successive data points falling in different major data bins, assigning the compressed data point associated with the second of the successive data points a value equal to the minor data bin within which the second of the successive data points falls.