Brain Fractal Blood-Oxygen Level Dependent (BOLD) Signals: The Effect of MRI Acquisition Parameters on Temporal Fractal Dimension (FD) Stability

Fractal analysis of the temporal patterns in functional magnetic resonance imaging (fMRI) brain data has recently gained momentum in assessing diseased and pharmacokinetically altered brain. However, there is no agreed standard method of acquisition and analysis. Therefore, we examined the fractal dimension (FD) of the brain's blood oxygen level dependent (BOLD) signal over time, with varying time points, MRI k-space filling techniques, repetition times (TR), and scan lengths. Furthermore, fractal dimension was determined with two different approaches: relative dispersion (FD_RD), and power spectrum (FD_PS) techniques. At a reduced number of time points, FD_RD is resistant to differing TR. FD_PS is time-point invariant at a low TR (250 ms). With constant scan times, FD_PS is the least variable. Our results show that some constraints to FD measurement in the brain require a certain amount of consistency in technique when comparing multiple scans. But within these constraints, FD can provide a reliable method of examining the brain's resting state BOLD signal.