Skip to main content
SpringerLink
Log in

Estimation and model selection in constrained deconvolution

  • Published:
Annals of Biomedical Engineering Aims and scope Submit manuscript

Abstract

We analyze in detail the estimation problem associated with the following problem. Givenn noisy measurements (y i ,i=1, ...,n) of the response of a system to an input (A(t) wheret indicates time), obtain an estimate ofA(t) given a knownK(t) (the unit impulse response function of the system) under the model:

$$y_i = \int_0^{t_i } {A(s)K(t_i - s)ds + \varepsilon _i } $$

where ε1,...ε n are independent identically distributed random variables with mean zero and common finite variance. In the solution to the problem, the unknown function is represented by a spline function, and the problem is recast in terms of (inequality constrained) linear regression. The main issues addressed are: (a) the comparison of different nonparametric regression methods in this context, and (b) how to do model selection, i.e., given a (finite) set of candidate spline functions, select the (possibly unique) best one using some (statistically based) selection criteria. Different spline candidate sets, and different asymptotic and resampling-based statistical selection criteria are compared by means of simulations. Due to the particular nature of the estimation problem, modifications to the criteria are suggested. Applications to simulated and real pharmacokinetics data are reported.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

References

  1. Akaike, H. A new look at the statistical model identification. IEEE Trans. Automat. Contr. 19:716–723; 1974.

    Google Scholar 

  2. Allen, D. The relationship between variable selection and data augmentation and a method for prediction. Technometrics 16:125–127; 1974.

    Google Scholar 

  3. Anderson, G.; Anderson, P. The Unix C shell field guide. Englewood Cliffs, NJ: Prentice Hall; 1986.

    Google Scholar 

  4. Anscombe, F.J. Topics in the investigation of linear relations fitted by least squares (with discussion). J. Roy. Satist. Soc. Ser. B. (29 1–52) 29:1–59; 1967.

    Google Scholar 

  5. Bellman, R.; Jacquez, J.A.; Kalaba, R. Some mathematical aspects of chemotherapy: I. One-organ models. Bull. Math. Biophys. 22:181–198; 1960.

    Google Scholar 

  6. Benowitz, N.L. Personal communication; 1990.

  7. Bickel, P. J.; Doksum, K. A. Mathematical statistics. Oakland, CA; Holden-Day, Inc.; 1977.

    Google Scholar 

  8. Bischoff, K. Some fundamental considerations of the application of pharmacokinetics to cancer therapy. Cancer Chemother. Rep. 59:777–793; 1975.

    CAS  PubMed  Google Scholar 

  9. Breiman, L.; Peters, P. Comparing automatic smoothers (A public service enterprise). Technical Report No. 161. Department of Statistics. University of California at Berkeley, Berkeley, CA; June 1988; revised July 1990.

    Google Scholar 

  10. Chambers, J.M.; Cleveland, W.S.; Kleiner, B.; Tukey, P.A. Graphical methods for data analysis. Wadsworth, Belmont, CA; 1983.

    Google Scholar 

  11. Cobelli, C.; Mari, A.; Del-Prato, S.; De-Kreutzenberg, S.; Nosadini, R.; Jensen, I. Reconstructing the rate of appearance of subcutaneous insulin by deconvolution. Am. J. Physiol. 253 (Regulatory Integrative Comp. Physiol. 16):E584-E590; 1987.

    CAS  PubMed  Google Scholar 

  12. Craven, P.; Wahba, G. Smoothing noisy data with spline functions. Numer. Math. 31:377–403; 1979.

    Google Scholar 

  13. DeBoor, C. A practical guide to splines. New York: Springer-Verlag; 1978.

    Google Scholar 

  14. Ebling, W.F.; Mills-Williams, L.; Harapat, S.R.; Stanski, D.R. A high-resolution liquid chromatographic method for determining thiopental concentrations in twelve rat tissues: Application to physiologic modeling of disposition of barbiturate. J. Chromatography 490:339–353; 1989.

    CAS  Google Scholar 

  15. Efron, B. The jackknife, the bootstrap and other resampling planes. Philadelphia: Society for Industrial and Applied Mathematics; 1982.

    Google Scholar 

  16. Friedman, J.H.; Silverman, B.W. Flexible parsimonious smoothing and additive modeling. Technometrics 31:3–39; 1989.

    Google Scholar 

  17. Friedman, J.H.; Tibshirani, R. The monotone smoothing of scatterplots. Technometrics 26:243–250; 1984.

    Google Scholar 

  18. Gill, P.E.; Hammarling, S.J.; Murray, W.; Saunders, M.A.; Wright, M.H. LSSOL. Report SOL 86-1: Users guide for LSSOL (Version 1.0). Stanford, CA: Department of Operations Research, Stanford University; 1986.

    Google Scholar 

  19. Gillespie, W.R. CVPEX: Programs for the convolution of piecewise polynomial and polyexponential functions (submitted for publication).

  20. Huffel, S.V.; Vandewalle, J.; Roo, M.C.; Willems, J.L. Reliable and efficient deconvolution technique based on total linear least squares for calculating the renal retention function. Med. Biol. Eng. Comput. 25:26–33; 1987.

    PubMed  Google Scholar 

  21. Hunt, B.R. The inverse problem of radiography. Mathematical Biosciences 8:161–179; 1970.

    Article  Google Scholar 

  22. Jacquez, J.A.; Bellman, R.; Kalaba, R. Some mathematical aspects of chemotherapy: II. The distribution of drug in the body. Bull. Math. Biophys. 22:309–322; 1960.

    CAS  Google Scholar 

  23. Jansson, P.A.; Hunt, R.H.; Plyler, E.K. Resolution enhancement of spectra. J. Opt. Soc. Am. 60:596–599; 1970.

    CAS  Google Scholar 

  24. Kelly, C.; Rice, J. Monotone smoothing with application toddose-response curves and the assessment of synergism. Biometrics 46:1071–1085; 1990.

    CAS  PubMed  Google Scholar 

  25. Langenbucher, F. Numerical convolution/deconvolution as a tool for correlatingin vitro within vivo drug availability. Pharm. Ind. 44:1166–1172; 1982.

    CAS  Google Scholar 

  26. Luenberger, D.G. Optimization by vector space methods. New York: John Wiley & Sons; 1969.

    Google Scholar 

  27. Lyche, T.; Morken, K. Knot removal for parametric B-spline curves and surfaces. Computer Aided Geometric Design 4:217–230; 1987.

    Article  Google Scholar 

  28. Mendelsohn, J.; Rice, J. Deconvolution of microfluorometric histograms with B-splines. J. American Statistical Association 77:748–753; 1985.

    Google Scholar 

  29. O'Sullivan, F. A statistical perspective on ill-posed inverse problems. Statistical Sciences 1:502–527; 1986.

    Google Scholar 

  30. O'Sullivan, F.; O'Sullivan, J. Deconvolution of episodic hormone data: an analysis of the role of season on the onset of puberty in cows. Biometrics 44:339–353; 1988.

    PubMed  Google Scholar 

  31. Provencher, S.W.; Vogel, R.H. Regularization techniques for inverse problems in molecular biology. In: Numerical treatment of inverse problems in differential and integral equations. Deuflhard, P.; Hairer, E. ed. Birkhauser; 1983.

  32. QPROG, Houston, TX: IMSL.

  33. Ramsay, J.O. Monotone regression splines in action. Statistical Science 3:425–461; 1988.

    Google Scholar 

  34. Rice, J.A. Choice of smoothing parameter in deconvolution problems. Contemporary Mathematics 59:137–151; 1986.

    Google Scholar 

  35. Smolen, V.F. Theoretical and computational basis for drug bioavailability determinations using pharmacological data. I. General considerations and procedures. J. Pharmacokinet. Biopharm. 4:337–353; 1976.

    CAS  PubMed  Google Scholar 

  36. S-PLUS, Seattle, WA: Statistical Sciences Inc.; 1989.

  37. Stephenson, J.J. Theory and transport in linear biological systems: I. Fundamental integral equation. Bull. Math. Biophys. 22:1–17; 1960.

    Google Scholar 

  38. Stephenson, J.J. Theory and transport in linear biological systems: II. Multiflux problems. Bull. Math. Biophys. 22:113–138; 1960.

    Google Scholar 

  39. Stone, C.J. Additive regression and other nonparametric models. Annals Statistics 13:689–706; 1985.

    Google Scholar 

  40. Vajda, S.; Godfrey, K.R.; Valko, P. Numerical deconvolution using system identification methods. J. Pharmacokin. Biopharm. 16:85–107; 1988.

    Article  CAS  Google Scholar 

  41. Verotta, D. An inequality-constrained least-squares deconvolution method. J. Pharmacokin. Biopharm. 19:269–289; 1989.

    Google Scholar 

  42. Verotta, D. Bootstrap confidence bands in constrained linear regression. Chapter 3. Dept. Of Biostatistics, University of California at Berkeley; 1992. Available from: University Microfilms International. Ph D Dissertation.

  43. Verotta, D.; Beal, S.; Sheiner, L.B. Semiparametric approach to pharmacokinetic-pharmacodynamic data. Am. J. Physiol. 256(Regulatory Integrative Comp. Physiol. 25):R1005-R1010; 1989.

    CAS  PubMed  Google Scholar 

  44. Verotta, D.; Sheiner, L.B.; Ebling, W.F.; Stanski, D. A semiparametric approach to physiological flow models. J. Pharmacokin. Biopharm. 17:463–491; 1989.

    CAS  Google Scholar 

  45. Villalobos, M.; Wahba, G. Inequality-constrained multivariate smoothing splines with applications to the estimation of posterior probabilities. J. American Statistical Association 82:239–248; 1987.

    Google Scholar 

  46. Volterra, V. Theory of functionals and of integral and integro-differential equations, New York: Dover; 1959.

    Google Scholar 

  47. Wahba, G. Ill posed problems: Numerical and statistical methods for midly, moderately and severely ill posed problems with noisy data: Technical Report No. 595. Madison, WI: Department of Statistics, University of Wisconsin; 1980.

    Google Scholar 

  48. Wahba, G. Constrained regularization for ill posed linear operator equations, with applications in meteorology and medicine. In: Statistical decision theory and related topics III. Vol. 2. Gupta, S.S.; Berger, J.O., eds. Academic Press; 1982: pp. 383–418.

  49. Wahba, G. Spline models for observational data. Philadelphia: Society for Industrial and Applied Mathematics; 1990.

    Google Scholar 

  50. Wiener, N. Nonlinear problems in random theory, Cambridge, MA: MIT Press; 1958.

    Google Scholar 

Download references

Author information

Authors and Affiliations

  1. Department of Pharmacy and Pharmaceutical Chemistry, University of California San Francisco, San Francisco, CA

    Davide Verotta

  2. Department of Epidemiology and Biostatistics, University of California San Francisco, San Francisco, CA

    Davide Verotta

  3. Department of Biostatistics, University of California at Berkeley, Berkeley, CA

    Davide Verotta

Authors
  1. Davide Verotta
    View author publications

    You can also search for this author in PubMed Google Scholar

Rights and permissions

Reprints and permissions

About this article

Cite this article

Verotta, D. Estimation and model selection in constrained deconvolution. Ann Biomed Eng 21, 605–620 (1993). https://doi.org/10.1007/BF02368641

Download citation

  • Issue Date:

  • DOI: https://doi.org/10.1007/BF02368641

Keywords

Search

Navigation