Method for generating a net analyte signal calibration model and uses thereof
A method for generating a net analyte signal calibration model for use in detecting and/or quantifying the amount of an analyte in a test subject. The net analyte signal can be generated by providing a set of in vivo infrared spectra for a test subject during a period in which an analyte concentration is essentially constant; calculating an optimal subspace of spectra that at least substantially describes all non-analyte dependent spectral variance in the in vivo spectra; providing a pure component infrared spectrum for the analyte; and calculating a net analyte signal spectrum from a data set comprising the optimal subspace spectra and the pure analyte spectrum. The net analyte signal calibration model can be used, for example, in measuring the concentration of analyte in a test subject, and/or for evaluating the analytical significance of an in vivo multivariate calibration model.
1. A method for generating a net analyte signal calibration model for use in detecting an analyte in a test subject, comprising:
a) providing a set of in vivo infrared spectra for a test subject during a period in which an analyte concentration is essentially constant;
b) calculating an optimal subspace of spectra that at least substantially describes all non-analyte dependent spectral variance in the in vivo spectra of step a);
c) providing a pure component infrared spectrum for the analyte; and
d) calculating a net analyte signal spectrum from a data set comprising the optimal subspace spectra of b) and the pure analyte spectrum of c), wherein the net analyte signal spectrum identifies one or more in vivo spectral features specific to the analyte.
2. The method of claim 1 , wherein the spectra are absorption spectra.
3. The method of claim 2 , wherein the absorption spectra are near infrared absorption spectra in the range of from approximately 4000 cm −1 to approximately 5000 cm −1 .
4. The method of claim 2 , wherein the absorption spectra are near infrared absorption spectra in the range of from approximately 5500 cm −1 to approximately 6500 cm −1 .
5. The method of claim 1 , wherein the spectra are reflectance spectra.
6. The method of claim 1 , wherein the spectra are single-beam spectra.
7. The method of claim 1 , wherein the infrared spectra are absorption spectra in the mid infrared spectral range.
8. The method of claim 7 , wherein the absorption spectra are in the range of from approximately 1200 cm −1 to approximately 900 cm −1 .
9. The method of claim 1 , wherein the analyte is a physiological chemical.
10. The method of claim 9 , wherein the analyte is glucose, urea, lactate, triglyceride, total protein, cholesterol, or ethanol.
11. The method of claim 9 , wherein the physiological chemical comprises at least one C—H, N—H, or O—H molecular bond.
12. The method of claim 9 , wherein the analyte is glucose.
13. The method of claim 1 , wherein the test subject is a living organism.
14. The method of claim 13 , wherein the test subject is a plant.
15. The method of claim 13 , wherein the test subject is an animal.
16. The method of claim 15 , wherein the animal is non-mammalian.
17. The method of claim 15 , wherein the animal is mammalian.
18. The method of claim 13 , wherein the test subject is a microbial species.
19. The method of claim 13 , wherein the test subject is a human.
20. The method of claim 1 , wherein step b) comprises a principle component analysis.
21. The method of claim 1 , further comprising reporting the net analyte signal calibration spectrum on a display device.
22. The method of claim 1 , further comprising storing the net analyte signal spectrum on a recordable medium.
23. A method for non-invasively measuring the concentration of an analyte in a test subject, comprising:
(a) identifying a test subject in need of having an analyte concentration measured;
(b) generating an in vivo net analyte signal calibration model by:
i) providing a set of in vivo infrared spectra for a test subject during a period in which an analyte concentration is essentially constant;
ii) calculating an optimal subspace of spectra that at least substantially describes all non-analyte dependent spectral variance in the in vivo spectra of step i);
iii) providing a pure component infrared spectrum for the analyte; and
iv) calculating the net analyte signal spectrum from a data set comprising the optimal subspace spectra of ii) and the pure analyte spectrum of iii), wherein the net analyte signal spectrum identifies one or more in vivo spectral features specific to the analyte;
(c) providing an in vivo infrared spectrum of the test subject; and
(d) calculating a predicted concentration of the analyte in the test subject from a data set comprising the net analyte signal calibration model and the in vivo infrared spectrum of the test subject.
24. The method of claim 23 , wherein the analyte is a physiological chemical.
25. The method of claim 24 , wherein the analyte is glucose, urea, lactate, triglyceride, total protein, cholesterol, or ethanol.
26. The method of claim 25 , wherein the analyte is glucose.
27. The method of claim 23 , wherein the test subject is any living organism.
28. The method of claim 27 , wherein the test subject is a plant.
29. The method of claim 27 , wherein the test subject is a mammal.
30. The method of claim 27 , wherein the test subject is a human.
31. The method of claim 23 , further comprising reporting the predicted concentration on a display device.
32. The method of claim 23 , further comprising storing the predicted concentration on a recordable medium.