A method for determining venous contribution to BOLD contrast sensory activation

Hall, DA, Gonçalves, MS, Smith, S, Jezzard, P, Haggard, MP and Kornak, J, 2002. A method for determining venous contribution to BOLD contrast sensory activation. Magnetic Resonance Imaging, 20 (10), pp. 695-706.

Preview

Text 193471_1851 Hall PostPrint.pdf Download (889kB) | Preview

Abstract

While BOLD contrast reflects haemodynamic changes within capillaries serving neural tissue, it also has a venous component. Studies that have determined the relation of large blood vessels to the activation map indicate that veins are the source of the largest response, and the most delayed in time. It would be informative if the location of these large veins could be extracted from the properties of the functional responses, since vessels are not visible in BOLD contrast images. The present study describes a method for investigating whether measures taken from the functional response can reliably predict vein location, or at least be useful in down-weighting the venous contribution to the activation response, and illustrates this method using data from one subject. We combined fMRI at 3 Tesla with high-resolution anatomical imaging and MR venography to test whether the intrinsic properties of activation time courses corresponded to tissue type. Measures were taken from a gamma fit to the functional response. Mean magnitude showed a significant effect of tissue type (P<0.001) where CSF > veins ≈ grey matter > white matter. Mean delays displayed the same ranking across tissue types (P<0.001), except that veins > grey matter. However, measures for all tissue types were distributed across an overlapping range. A logistic regression model correctly discriminated 72% of the veins from grey matter in the absence of independent information of macroscopic vessels (ROC=0.72). Whilst tissue classification was not perfect for this subject, weighting the T contrast by the predicted probabilities materially reduced the venous component to the activation map.

Item Type:	Journal article
Publication Title:	Magnetic Resonance Imaging
Creators:	Hall, D.A., Gonçalves, M.S., Smith, S., Jezzard, P., Haggard, M.P. and Kornak, J.
Publisher:	Elsevier (not including Cell Press)
Date:	2002
Volume:	20
Number:	10
Identifiers:	Number Type 10.1016/S0730-725X(02)00607-0 DOI
Divisions:	Schools > School of Social Sciences
Record created by:	EPrints Services
Date Added:	09 Oct 2015 09:52
Last Modified:	23 Aug 2016 09:06
URI:	https://irep.ntu.ac.uk/id/eprint/4152

Actions (login required)

Edit View

Statistics

Views

Downloads