AUTHOR=Kannurpatti Sridhar S., Biswal Bharat B.

TITLE=Prediction of Task-Related BOLD fMRI with Amplitude Signatures of Resting-State fMRI

JOURNAL=Frontiers in Systems Neuroscience

VOLUME=Volume 6 - 2012

YEAR=2012

URL=https://www.frontiersin.org/journals/systems-neuroscience/articles/10.3389/fnsys.2012.00007

DOI=10.3389/fnsys.2012.00007

ISSN=1662-5137

ABSTRACT=We explored task-activated fMRI (T-fMRI) signals and their relationship to the resting state fMRI (R-fMRI) signals based on the hypothesis that they arise from a common hemodynamic substrate. In the first group of twelve healthy human subjects, BOLD signal changes in response to a motor task, was regressed with the vascular sensitivity signals represented by the BOLD signal change during the performance of a hypercapnic breath hold (BH) task. Motor task versus resting-state fluctuation of amplitude (RSFA) relationship was also determined. Within any subject, a significant linear correlation was observed between motor task and BH across voxels. Averaged across the whole brain, the subject-wise correlation between the motor task and BH showed a similar linear relationship. In a similar manner, a significant linear correlation was observed between motor task and RSFA both across voxels and subjects. We term the linear dependence between motor task and RSFA as rest-task (R-T) relationship, evident in both the low and high frequencies of RSFA. Using the R-T relationship determined from the first group of 12 healthy subjects, we predicted T-fMRI responses using the low and high frequency R-fMRI (RSFA) signals in a second cohort of 7 healthy subjects. Both low and high frequency RSFA could predict the magnitude of T-fMRI responses from each subject within an error limit of 25 and 5% respectively. Also the high frequency RSFA was a better predictor of task-induced response than low frequency RSFA across voxels within a subject. The better predictive power of high frequency RSFA at the voxel and subject levels stemmed from its lower between voxel and between subject variability compared to low frequency RSFA.