How big of a change are you talking about? Hemoglobin is one of many
physiological parameters that might influence resting state results. I
would guess that it would tend to have a global rather than a spatially
specific effect, and could influence the vulnerability of resting state
data to global physiological noise (e.g. from variations in blood pCO2).
It might also influence the amplitude of resting state fluctuations.
Including it as a covariate of no interest is the most conservative way to
approach this confound, however if your medication typically causes a drop
in hemoglobin, you will also remove some of your no medication/medication
effect of interest by doing this.
Peace,
Matt.
On 7/4/16, 9:54 AM, "FSL - FMRIB's Software Library on behalf of YW Yoon"
<[log in to unmask] on behalf of [log in to unmask]> wrote:
>Dear FSL users,
>
>I have a question regarding the effect of hemoglobin level in BOLD signal.
>We are trying to investigate the resting-state fMRI change during the
>pre- and post-medication.
>
>The problem is that after the medication, few number of participants has
>shown hemoglobin level change.
>
>According to the previous studies, (Yang, Front Neurosci. 2015 doi:
>10.3389/fnins.2014.00452) (Levin, Magn Reson Imaging. 2001
>Oct;19(8):1055-62) hemoglobin level influences BOLD signal.
>However, we want to see the pure effect of medication on the
>resting-state brain network.
>
>I planning to put the hemoglobin level as covariate during the 2nd level
>anaylsis to overcome this problem.
>
>Is there anyone who struggled with the similar problem?
>If you have, how did you deal with it?
>I tried to find a paper which has solved the problem, but I couldn't.
>Are there any publication which overcome this problem?
>
>I appreciate your help in advance.
>
>Best,
>YW
|