Claims
- 1. A method for measuring glucose present in a biological system, said method comprising:transdermally extracting a sample comprising glucose from the biological system using a sampling system that is in operative contact with a skin or mucosal surface of said biological system; obtaining a raw signal from the extracted glucose, wherein said raw signal is specifically related to glucose amount or concentration in the biological system; performing a calibration which correlates the raw signal with a measurement value indicative of the amount or concentration of glucose present in the biological system at the time of extraction; repeating said transdermal extracting and said obtaining a raw signal to obtain a series of two or more measurement values at selected time intervals; and predicting a measurement value based on the series of measurement values using a Mixtures of Experts algorithm comprising the following BG=w1BG1+w2BG2+w3BG3 (15) wherein (BG) is blood glucose, BGi is the blood glucose predicted by Expert i; wi is a weighting value, and the individual Experts BGi are further defined by the expression shown as Equations 16, 17, and 18 BG1=p1(timec)+q1(active)+r1(signal)+s1(BG|cp)+t1 (16) BG2=p2(timec)+q2(active)+r2(signal)+s2(BG|cp)+t2 (17) BG3=p3(timec)+q3(active)+r3(signal)+s3(BG|cp)+t3 (18) wherein, BGi is the blood glucose predicted by Expert i; parameters include, timec (elapsed time from a calibration of said sampling system), active (active signal), signal (calibrated signal), and BG/cp (blood glucose value at a calibration point); pi, qi, ri, and si are coefficients; and ti is a constant; and further where the weighting value, wi, is defined by the formulas shown as Equations 19, 20, and 21 w1=ⅇd1ⅇd1+ⅇd2+ⅇd3(19)w2=ⅇd2ⅇd1+ⅇd2+ⅇd3(20)w3=ⅇd3ⅇd1+ⅇd2+ⅇd3(21) where e refers to the exponential function and di is a parameter set used to determine the weights wi, given by Equations 19, 20, and 21, and d1=τ1(timec)+β1(active)+gamma1(signal)+δ1(BG|cp)+ε1 (22) d2=τ2(timec)+β2(active)+gamma2(signal)+δ2(BG|cp)+ε2 (23)d3=τ3(timec)+β3(active)+gamma3(signal)+δ3(BG|cp)+ε3 (24) where τi, βi, gammai and δi are coefficients, and where εi is a constant.
- 2. The method of claim 1, wherein the sample is extracted from the biological system into a collection reservoir to obtain an amount or concentration of the glucose in said reservoir.
- 3. The method of claim 2, wherein the collection reservoir is in contact with the skin or mucosal surface of the biological system and the sample is extracted using an iontophoretic current applied to said skin or mucosal surface.
- 4. The method of claim 3, wherein the collection reservoir comprises an enzyme that reacts with the extracted glucose to produce an electrochemically detectable signal.
- 5. The method of claim 4, wherein the enzyme is glucose oxidase.
- 6. The method of claim 1, which includes further parameters for measurement values, said parameters selected from the group consisting of temperature, ionophoretic voltage, and skin conductivity.
- 7. A method for measuring glucose in a subject, said method comprising:obtaining a raw signal from a sensing apparatus, wherein said raw signal is specifically related to an amount or concentration of glucose in the subject; performing a calibration which correlates the raw signal with a measurement value indicative of the glucose amount or concentration in the subject; repeating said obtaining of raw signal to obtain a series of measurement values at selected time intervals; and predicting a measurement value using a Mixtures of Experts algorithm comprising the following, BG=w1BG1+w2BG2+w3BG3 (15) wherein (BG) is blood glucose, BGi is the blood glucose predicted by Expert i; wi is a weighting value, and the individual Experts BGi are further defined by the expression shown as Equations 16, 17, and 18 BG1=p1(timec)+q1(active)+r1(signal)+s1(BG|cp)+t1 (16) BG2=p2(timec)+q2(active)+r2(signal)+s2(BG|cp)+t2 (17) BG3=p3(timec)+q3(active)+r3(signal)+s3(BG|cp)+t3 (18) wherein, BGi is the blood glucose predicted by Expert i; parameters include, timec (elapsed time from a calibration of said sampling system), active (active signal), signal (calibrated signal), and BG/cp (blood glucose value at a calibration point); pi, qi, ri, and si are coefficients; and ti is a constant; and further where the weighting value, wi, is defined by the formulas shown as Equations 19, 20, and 21 w1=ⅇd1ⅇd1+ⅇd2+ⅇd3(19)w2=ⅇd2ⅇd1+ⅇd2+ⅇd3(20)w3=ⅇd3ⅇd1+ⅇd2+ⅇd3(21) where e refers to the exponential function and di is a parameter set used to determine the weights wi, given by Eguations 19, 20, and 21, and d1=τ1(timec)+β1(active)+gamma1(signal)+δ1(BG|cp)+ε1 (22) d2=τ2(timec)+β2(active)+gamma2(signal)+δ2(BG|cp)+ε2 (23) d3=τ3(timec)+β3(active)+gamma3(signal)+δ3(BG|cp)+ε3 (24) where τi, βi, gammai and δi are coefficients, and where εi is a constant.
- 8. The method of claim 7, wherein the sensing apparatus is a near-IR spectrometer.
- 9. The method of claim 7, wherein the sensing apparatus comprises a biosensor having an electrochemical sensing element.
- 10. The method of claim 7, wherein before said obtaining of a raw signal, a sample comprising glucose is extracted from the subject into a collection reservoir to obtain an amount or concentration of the glucose in said reservoir.
- 11. The method of claim 10, wherein the collection reservoir is in contact with the skin or mucosal surface of the biological system and the sample is transdermally extracted.
- 12. The method of claim 11, wherein said sample is extracted using an iontophoretic current applied to said skin or mucosal surface.
- 13. The method of claim 12, wherein the collection reservoir comprises an enzyme that reacts with the extracted glucose to produce an electrochemically detectable signal.
- 14. The method of claim 13, wherein the enzyme is glucose oxidase.
- 15. A monitoring system for measuring glucose present in a biological system, said monitoring system comprising, in operative combination:a sampling device for extracting a sample comprising glucose from the biological system, wherein said sampling device extracts the sample across a skin or mucosal surface of said biological system; a sensing device in operative contact with the sample extracted by the sampling device, wherein said sensing device obtains a raw signal from the extracted glucose and said raw signal is specifically related to an amount or concentration of glucose present in the biological system; and one or more microprocessors in operative communication with the sampling device and the sensing device, wherein said microprocessors (i) control the sampling device and the sensing device to obtain a series of raw signals at selected time intervals during a measurement period, (ii) correlate the raw signals with measurement values indicative of the amount or concentration of glucose present in the biological system, and (iii) predict a measurement value using a Mixtures of Experts algorithm comprising the following, BG=w1BG1+w2BG2+w3BG3 (15) wherein (BG) is blood glucose, BGi is the blood glucose predicted by Expert i; wi is a weighting value, and the individual Experts BGi are further defined by the expression shown as Equations 16, 17, and 18 BG1=p1(timec)+q1(active)+r1(signal)+s1(BG|cp)+t1 (16) BG2=p2(timec)+q2(active)+r2(signal)+s2(BG|cp)+t2 (17) BG3=p3(timec)+q3(active)+r3(signal)+s3(BG|cp)+t3 (18) wherein, BGi is the blood glucose predicted by Expert i; parameters include, timec (elapsed time from a calibration of said sampling system), active (active signal), signal (calibrated signal), and BG/cp (blood glucose value at a calibration point); pi, qi, ri, and si are coefficients; and ti is a constant; and further where the weighting value, wi, is defined by the formulas shown as Equations 19, 20, and 21 w1=ⅇd1ⅇd1+ⅇd2+ⅇd3(19)w2=ⅇd2ⅇd1+ⅇd2+ⅇd3(20)w3=ⅇd3ⅇd1+ⅇd2+ⅇd3(21) where e refers to the exponential function and di is a parameter set used to determine the weights wi, given by Equations 19, 20, and 21, and d1=τ1(timec)+β1(active)+gamma1(signal)+δ1(BG|cp)+ε1 (22) d2=τ2(timec)+β2(active)+gamma2(signal)+δ2(BG|cp)+ε2 (23) d3=τ3(timec)+β3(active)+gamma3(signal)+δ3(BG|cp)+ε3 (24) where τi, βi, gammai and δi are coefficients, and where εi is a constant.
- 16. The monitoring system of claim 15, wherein the sampling device includes one or more collection reservoirs for containing the extracted sample.
- 17. The monitoring system of claim 16, wherein the sampling device uses an iontophoretic current to extract the sample from the biological system.
- 18. The monitoring system of claim 17, wherein the collection reservoir comprises an enzyme that reacts with the extracted glucose to produce an electrochemically detectable signal.
- 19. The monitoring system of claim 18, wherein the enzyme is glucose oxidase.
- 20. A monitoring system for measuring glucose in a subject, said system comprising, in operative combination:a sensing device in operative contact with a sample extracted from the subject, said sample comprising glucose, wherein said sensing device obtains a raw signal specifically related to an amount or concentration of glucose in the subject; and one or more microprocessors in operative communication with the sensing device, wherein said microprocessors (i) control the sensing device to obtain a series of raw signals at selected time intervals, (ii) correlate the raw signals with measurement values indicative of the amount or concentration of glucose present in the subject, and (iii) predict a measurement value using a Mixtures of Experts algorithm comprising the following, BG=w1BG1+w2BG2+w3BG3 (15) wherein (BG) is blood glucose, BGi is the blood glucose predicted by Expert i; wi is a weighting value, and the individual Experts BGi are further defined by the expression shown as Equations 16, 17, and 18 BG1=p1(timec)+q1(active)+r1(signal)+s1(BG|cp)+t1 (16) BG2=p2(timec)+q2(active)+r2(signal)+s2(BG|cp)+t2 (17) BG3=p3(timec)+q3(active)+r3(signal)+s3(BG|cp)+t3 (18) wherein, BGi is the blood glucose predicted by Expert i; parameters include, timec (elapsed time from a calibration of said sampling system), active (active signal), signal (calibrated signal), and BG/cp (blood glucose value at a calibration point); pi, qi, ri, and si are coefficients; and ti is a constant; and further where the weighting value, wi, is defined by the formulas shown as Equations 19, 20, and 21 w1=ⅇd1ⅇd1+ⅇd2+ⅇd3(19)w2=ⅇd2ⅇd1+ⅇd2+ⅇd3(20)w3=ⅇd3ⅇd1+ⅇd2+ⅇd3(21) where e refers to the exponential function and di is a parameter set used to determine the weights wi, given by Equations 19, 20, and 21, and d1=τ1(timec)+β1(active)+gamma1(signal)+δ1(BG|cp)+ε1 (22) d2=τ2(timec)+β2(active)+gamma2(signal)+δ2(BG|cp)+ε2 (23) d3=τ3(timec)+β3(active)+gamma3(signal)+δ3(BG|cp)+ε3 (24) where τi, βi, gammai and δi are coefficients, and where εi is a constant.
- 21. The monitoring system of claim 20, which includes further parameters for raw signals, said parameters selected from the group consisting of temperature, ionophoretic voltage, and skin conductivity.
- 22. The monitoring system of claim 20, wherein the sensing device comprises a biosensor having an electrochemical sensing element.
- 23. The monitoring system of claim 20, wherein the sensing device comprises a near-IR spectrometer.
CROSS-REFERENCE TO RELATED APPLICATIONS
This application is a continuation of U.S. patent application Ser. No. 09/405,976, filed Sep. 27, 1999, now U.S. Pat. No. 6,326,160, which is a continuation-in-part of U.S. patent application Ser. No. 09/241,929, filed Feb. 1, 1999, now abandoned, which is a continuation-in-part of U.S. patent application Ser. No. 09/198,039, filed Nov. 23, 1998, now abandoned, which is a continuation-in-part of U.S. patent application Ser. No. 09/163,856, filed Sep. 30, 1998, now U.S. Pat. No. 6,180,416; priority is claimed to these applications under 35 U.S.C. §120, and these applications are herein incorporated by reference in their entireties.
US Referenced Citations (10)
Foreign Referenced Citations (13)
Number |
Date |
Country |
WO 9502357 |
Jan 1995 |
WO |
WO 9600109 |
Jan 1996 |
WO |
WO 9600110 |
Jan 1996 |
WO |
WO 9635370 |
Nov 1996 |
WO |
WO 9702811 |
Jan 1997 |
WO |
WO 9710356 |
Mar 1997 |
WO |
WO 9710499 |
Mar 1997 |
WO |
WO 9724059 |
Jul 1997 |
WO |
WO 9842252 |
Oct 1998 |
WO |
WO 9958050 |
Nov 1999 |
WO |
WO 9958051 |
Nov 1999 |
WO |
WO 9958190 |
Nov 1999 |
WO |
WO 9958973 |
Nov 1999 |
WO |
Non-Patent Literature Citations (9)
Entry |
Alcock et al., “Continuous Analyte Monitoring to Aid Clinical Practice,” I.E.E.E. Engineering in Medicine & Biology 13:(3):319-325 (1994). |
Barlow, T.W., “Feed-Forward Neural Networks for Secondary Structure Prediction,” Journal of Molecular Graphics 13:175-183 (1995). |
Ghahramani, Z. And Wolpert, D.M., “Modular Decomposition in Visuomotor Learning,” Nature 386(6623):392-395 (1997). |
Hamilton, J.D. and Susmel, Raul, “Autoregressive Conditional Heteroskedasticity and Changes in Regime,” Journal of Econometrics 64:307-333 (1994). |
Hu et al., “Customized ECG Beat Classifier Using Mixture of Experts,” Proceedings of the 1995 IEEE Workshop pp. 459-464 (1995). |
Jordan et al., “Hierarchical Mixtures of Experts and the EM Algorithm,” Neural Comput. 6(2):181-214 (1994). |
Waterhouse, Steven Richard, “Classification and Regression Using Mixtures of Experts,” Jesus College, Cambridge and Department of Engineering, University of Cambridge, Oct. 1997. |
Waterhouse et al., in: D.S. Touretzky (Ed.) “Bayesian Methods for Mixtures of Experts,” Advances in Neural Information Processing Systems MIT Press. Cambridge MA. 8:351-357 (1996). |
Weigend et al., “Nonlinear Gated Experts for Time Series: Discovering Regimes and Avoiding Overfitting,” International Journal of Neural Systems 6(4):373-399 (1995). |
Continuations (1)
|
Number |
Date |
Country |
Parent |
09/405976 |
Sep 1999 |
US |
Child |
09/911341 |
|
US |
Continuation in Parts (3)
|
Number |
Date |
Country |
Parent |
09/241929 |
Feb 1999 |
US |
Child |
09/405976 |
|
US |
Parent |
09/198039 |
Nov 1998 |
US |
Child |
09/241929 |
|
US |
Parent |
09/163856 |
Sep 1998 |
US |
Child |
09/198039 |
|
US |