Robust fitting for neuroreceptor mapping

Document Type

Article

Publication Date

3-15-2009

Abstract

Among many other uses, positron emission tomography (PET) can be used in studies to estimate the density of a neuroreceptor at each location throughout the brain by measuring the concentration of a radiotracer over time and modeling its kinetics. There are a variety of kinetic models in common usage and these typically rely on nonlinear least-squares (LS) algorithms for parameter estimation. However, PET data often contain artifacts (such as uncorrected head motion) and so the assumptions on which the LS methods are based may be violated. Quantile regression (QR) provides a robust alternative to LS methods and has been used successfully in many applications. We consider fitting various kinetic models to PET data using QR and study the relative performance of the methods via simulation. A data adaptive method for choosing between LS and QR is proposed and the performance of this method is also studied. Copyright © 2008 John Wiley & Sons, Ltd.

Identifier

65549117518 (Scopus)

Publication Title

Statistics in Medicine

External Full Text Location

https://doi.org/10.1002/sim.3510

e-ISSN

10970258

ISSN

02776715

PubMed ID

19109810

First Page

1004

Last Page

1016

Issue

6

Volume

28

Grant

P50MH062185

Fund Ref

National Institute of Mental Health

This document is currently not available here.

Share

COinS