Constraining the general linear model for sensible hemodynamic response function waveforms

We propose a method to do constrained parameter estimation and inference from neuroimaging data using general linear model (GLM). Constrained approach precludes unrealistic hemodynamic response function (HRF) estimates to appear at the outcome of the GLM analysis. The permissible ranges of waveform...

Full description

Saved in:
Bibliographic Details
Published in:Medical & biological engineering & computing 2008-08, Vol.46 (8), p.779-787
Main Authors: Çiftçi, Koray, Sankur, Bülent, Kahya, Yasemin P., Akın, Ata
Format: Article
Language:English
Subjects:
Adult
Bayes Theorem
Bayesian analysis
Biomedical and Life Sciences
Biomedical Engineering and Bioengineering
Biomedicine
Computer Applications
Female
Hemodynamics
Human Physiology
Humans
Imaging
Infrared imaging systems
Linear Models
Male
Medical imaging
Neuroimaging
Original Article
Parameter estimation
Prefrontal Cortex - physiology
Radiology
Signal Processing, Computer-Assisted
Spectroscopy, Near-Infrared - methods
Spectrum analysis
Studies
Time series
Waveform analysis
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We propose a method to do constrained parameter estimation and inference from neuroimaging data using general linear model (GLM). Constrained approach precludes unrealistic hemodynamic response function (HRF) estimates to appear at the outcome of the GLM analysis. The permissible ranges of waveform parameters were determined from the study of a repertoire of plausible waveforms. These parameter intervals played the role of prior distributions in the subsequent Bayesian analysis of the GLM, and Gibbs sampling was used to derive posterior distributions. The method was applied to artificial null data and near infrared spectroscopy (NIRS) data. The results show that constraining the GLM eliminates unrealistic HRF waveforms and decreases false activations, without affecting the inference for “realistic” activations, which satisfy the constraints.
ISSN:0140-0118
1741-0444
DOI:10.1007/s11517-008-0347-6