Automated detection of sleep and waking states

• US 8,073,534 B2
• Filed: 05/09/2006
• Issued: 12/06/2011
• Est. Priority Date: 05/10/2005
• Status: Active Grant

• Current Assignee: Salk Institute for Biological Studies
• Original Assignee: Salk Institute for Biological Studies
• Inventors: Low, Philip
• Primary Examiner(s): MALLARI, PATRICIA C
• Application Number:
  US11/431,425
• Publication Number:
  US 20070016095A1
• Time in Patent Office: 2,037 Days
• Field of Search: 600/544, 600/545
• US Class Current: 600/544
• CPC Class Codes: A61B 5/374 Detecting the frequency distribution of signals, e.g. detecting delta, theta, alpha, beta or gamma waves
• CPC Class Codes: A61B 5/4812 Detecting sleep stages or cycles
• CPC Class Codes: A61B 5/4815 Sleep quality
• CPC Class Codes: A61B 5/7203 for noise prevention, reduction or removal
• CPC Class Codes: A61B 5/726 using Wavelet transforms
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: A61B 5/7267 involving training the classification device
• CPC Class Codes: G06F 2218/12 Classification; Matching
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

Determining low power frequency range information from spectral data. Raw signal data can be adjusted to increase dynamic range for power within low power frequency ranges as compared to higher-power frequency ranges to determine adjusted source data valuable for acquiring low power frequency range information. Low power frequency range information can be used in the analysis of a variety of raw signal data. For example, low power frequency range information within electroencephalography data for a subject from a period of sleep can be used to determine sleep states. Similarly, automated full-frequency spectral electroencephalography signal analysis can be useful for customized analysis including assessing sleep quality, detecting pathological conditions, and determining the effect of medication on sleep states.

Claims

1. A method, comprising:

obtaining data indicative of brainwave activity;

normalizing at least one frequency range of said data to change a power level of the data in said at least one frequency range relative to data in another frequency range, to form normalized data indicative of brainwave activity;

a second normalizing of the data, to form double normalized data, wherein said second normalizing comprises normalizing frequencies across time; and

analyzing said double normalized data indicative of brainwave activity utilizing a computing device to provide at least one parameter indicative of sleep state from said analyzing.

2. A method as in claim 1, wherein said first and second normalizing each use Z scoring for the normalizing.

3. A method as in claim 1, further comprising defining a discrimination function which represents characteristics of the said double normalized data for a plurality of different sleep states, and using said discrimination function to determine a sleep state from said double normalized data.

4. A method as in claim 3, wherein said discrimination function is a function that is in terms of frequencies which are present in specified ranges and not present in specified other ranges, to define a sleep state.

5. A method as in claim 1, further comprising characterizing a preferred frequency as a frequency which has the highest normalized value in any specified time, and analyzing the preferred frequency to determine said at least one parameter.

6. A method as in claim 5, further comprising defining a discrimination function as a function of preferred frequency, where a discrimination function defines a sleep state in terms of frequencies which are present, and frequencies which are not present.

7. A method as in claim 1, wherein analyzing said double normalized data indicative of brainwave activity further comprises analyzing a fragmentation of the double normalized data.

8. A method for determining sleep states in a subject over a period of time comprising:

receiving brain wave data for the subject over the period of time, wherein the brain wave data exhibits lower dynamic range for power in at least one low power first frequency range in a frequency spectrum as compared to a second frequency range in the frequency spectrum;

segmenting the brain wave data into one or more epochs;

weighting frequency power of the one or more epochs across time utilizing a computing device, wherein the weighting comprises increasing the dynamic range for power within the low power frequency range of the frequency spectrum as compared to the second frequency range, thereby generating one or more frequency weighted epochs; and

providing classified sleep states of the subject based on the one or more frequency weighted epochs.

9. The method as in claim 8 wherein classifying sleep states in the subject comprises:

clustering the one or more frequency weighted epochs; and

assigning sleep state designations to the one or more frequency weighted epochs according to the clustering; and

presenting the sleep state designations as indicative of sleep states in the subject for the period of time represented by the one or more frequency weighted epochs.

10. The method of claim 8 wherein clustering the one or more frequency weighted epochs comprises k-means clustering.

11. The method of claim 8 further comprising pretreating the brain wave data with component analysis.

12. The method of claim 8 wherein classifying sleep states in the subject comprises applying independent component analysis to the one or more frequency weighted epochs.

13. The method of claim 8 wherein classifying sleep states further comprises incorporating manually determined sleep states.

14. The method of claim 8 wherein assigning sleep state designations to the one or more frequency weighted epochs comprises:

determining a slow wave sleep designation from a non-slow wave sleep designation based at least on low frequency information; and

determining a rapid eye movement sleep designation from a non-rapid eye movement sleep designation based at least on high frequency information.

15. The method of claim 14 further comprising assigning a slow wave sleep designation to an epoch that has significant weighted power at low frequencies.

16. The method of claim 14 further comprising assigning a rapid eye movement sleep designation to an epoch with significant weighted power at high frequency.

17. The method of claim 8 wherein assigning sleep state designations to the one or more frequency weighted epochs further comprises applying a smoothing window to the one or more weighted epochs, wherein the smoothing can comprise averaging sleep state designations across the one or more weighted epochs.

18. The method of claim 8 further comprising presenting one or more frequency weighted epochs as canonical spectra representative of the sleep state in the subject for the period of time represented by the one or more epochs having similar sleep state designations.

19. The method of claim 18 further comprising analyzing the canonical spectra with independent component analysis to establish sleep state classification confidence.

20. The method of claim 8 further comprising presenting sleep statistics for the subject according to the sleep state designations of the one or more frequency weighted epochs.

21. The method as in claim 8, further comprising second weighting power to normalize the data according to a second dimension, prior to said classifying to form doubly normalized data.

22. The method as in claim 21, wherein said second weighting comprises normalizing at least one frequency across time.

23. A method as in claim 21, wherein said weighting and said second weighting each use Z scoring to carry out normalizing.

24. A method as in claim 21, further comprising defining a discrimination function which represents characteristics of the double normalized data for a plurality of different sleep states, and using said discrimination function to determine a sleep state from said double normalized data.

25. A method as in claim 24, wherein said discrimination function is a function that is in terms of frequencies which are present in specified ranges and not present in specified other ranges, to define a sleep state.

26. A method as in claim 21, further comprising characterizing a preferred frequency as a frequency which has the highest normalized value in any specified time, and analyzing the preferred frequency to determine said at least one parameter.

27. A method as in claim 26, further comprising defining a discrimination function as a function of said preferred frequency, where a discrimination function defines a sleep state in terms of frequencies which are present, and frequencies which are not present.

28. A method as in claim 21, further comprising analyzing a spectral fragmentation of the doubly normalized data, and using the spectral fragmentation as part of said analyzing.

29. A method as in claim 21, further comprising analyzing a temporal fragmentation of the doubly normalized data, and using the temporal fragmentation as part of said analyzing.