View in EDS

Estimating the effective degrees of freedom in univariate multiple regression analysis

Saved in:
Bibliographic Details
Title: Estimating the effective degrees of freedom in univariate multiple regression analysis
Authors: Kruggel, F. kruggel@cns.mpg.de, Pélégrini-Issac, M., Benali, H.
Source: Medical Image Analysis. Mar2002, Vol. 6 Issue 1, p63. 13p.
Subjects: Magnetic resonance imaging, Diagnostic imaging
Abstract: The general linear model provides the most widely applied statistical framework for analyzing functional MRI (fMRI) data. With the increasing temporal resolution of recent scanning protocols, and more elaborate data preprocessing schemes, data independency is no longer a valid assumption. In this paper, we revise the statistical background of the general linear model in the presence of temporal autocorrelations. First, when detecting the activation signal, we explicitly account for the temporal autocorrelation structure, which yields a generalized F-test and the associated corrected (or effective) degrees of freedom (DOF). The proposed approach is data driven and thus independent of any specific preprocessing method. Then, for event-related protocols, we propose a new model for the temporal autocorrelations (“damped oscillator” model) and compare this model to another, previously used in the field (first-order autoregressive model, or AR(1) model). In the case of long fMRI time series, an efficient approximation for the number of effective DOF is provided for both models. Finally, the validity of our approach is assessed using simulated and real fMRI data and is compared with more conventional methods. [Copyright &y& Elsevier]
Copyright of Medical Image Analysis is the property of Elsevier B.V. and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
Description
Abstract:The general linear model provides the most widely applied statistical framework for analyzing functional MRI (fMRI) data. With the increasing temporal resolution of recent scanning protocols, and more elaborate data preprocessing schemes, data independency is no longer a valid assumption. In this paper, we revise the statistical background of the general linear model in the presence of temporal autocorrelations. First, when detecting the activation signal, we explicitly account for the temporal autocorrelation structure, which yields a generalized <F>F</F>-test and the associated corrected (or effective) degrees of freedom (DOF). The proposed approach is data driven and thus independent of any specific preprocessing method. Then, for event-related protocols, we propose a new model for the temporal autocorrelations (“damped oscillator” model) and compare this model to another, previously used in the field (first-order autoregressive model, or AR(1) model). In the case of long fMRI time series, an efficient approximation for the number of effective DOF is provided for both models. Finally, the validity of our approach is assessed using simulated and real fMRI data and is compared with more conventional methods. [Copyright &y& Elsevier]
ISSN:13618415
DOI:10.1016/S1361-8415(01)00052-4