hide
Free keywords:
-
Abstract:
Introduction: The blood oxygenation level-dependent (BOLD) signal using fMRI is currently the most popular imaging method to study brain function non-invasively. However, it is not fully understood quantitavely how the differ-
ent water (proton) pools inside blood vessels and/or in tissue contribute to the
total MRI signal both intrinsically and as a function of blood oxygenation and
volume and how these effects change with magnetic field or spin preparation.
In this study, through simulations, we quantitatively assess the various BOLD
signal contributions by proposing a ge
neral model for the BOLD signal for
field strengths up to 16.4T for both gradient-echo (GRE) and spin-echo (SE).
Methods:
For each of the contributions, we provide analytical formulas for: a)
intrinsic intra-vascular (IV) and extra-vascular (EV) relaxation rate (with no
deoxygenated Hemoglobin (deoxy-Hb)) derived from published experimen-
tal data; b) deoxy-Hb dependence, derived for EV BOLD signal from Monte-
carlo Simulations and for IV BOLD signal, again from experimental values.
This BOLD signal model was used to investigate the various contributions to
the BOLD signal assuming oxygenation and blood volume values typical for
micro- and macrovasculature.
Results:
The results indicate, most notably, that for SE: a) the IV BOLD sig-
nal (Figure 1) does not disappear for high field strengths but rather shifts to
blood vessels with high blood oxygenation; b) diffusion weighting at low field
strengths increases micro-vasculature weighting; c) using a TE larger than the
T2 of tissue also enhances micro-vasculat
ure weighting, though compromising
signal-to-noise ratio for sp
atial specificity; d) surprisingly, the highest micro-
vasculature weighting is achieved for field strength between 4 and 7T. For GRE,
there is no field strength for which micro-vasculature BOLD signal is larger
than macro-vasculature BOLD signal.
Discussion:
A general model of the BOLD signal up to 16.4T for GRE and SE
was provided. For both IV and EV BOLD signal (Figure 1 and 2), analytical
expressions for intrinsic relaxation rates and as a function blood oxygenation
and volume were derived from experimental and computer simulation data. It
was found that IV and EV signal contributions vary with field strength, echo time and MRI sequence used and these imaging parameters can be optimized
to yield high micro-vasculature weighting. The results have consequences for
assessing spatial specificity of fMRI and for the exact formulation of some
standard fMRI techniques currently relying solely on theoretical estimates of
EV BOLD signal, e.g. calibrated
BOLD signal and vessel size imaging.
