Quantifying Cardiac, Respiratory, and Low Frequency Components of CSF Motion From fMRI Inflow Effects.

purposeCerebrospinal fluid (CSF) flow oscillations have emerged as a potentially important marker related to brain clearance, but their acquisition often relies on specialized imaging MRI sequences. The purpose of this work was to enable quantitative assessment of CSF flow associated with cardiac, respiratory, and low-frequency cycles using widely available functional magnetic resonance imaging (fMRI) acquisitions.

methodsA method was developed to translate fMRI-derived CSF inflow signals into quantitative flow rates. This approach modeled the spin-history of an oscillating ensemble of molecules. Validation was performed using phantom experiments with cardiac-, respiratory-, and low-frequency-like oscillatory flow. The method was further applied to resting-state data from 48 older adults (68-82 years, 19 women) to characterize CSF flow at the foramen magnum.

resultsPhantom experiments demonstrated excellent correlations between estimated and true velocities for cardiac- and respiratory-like frequencies (r = 0.94 and 0.97, respectively) and moderate correlation for the low-frequency-like oscillation (r = 0.58). In the population cohort, median CSF stroke volumes were 0.77 [0.57, 1.09] mL for the cardiac cycle, 0.38 [0.26, 0.88] mL for the respiratory cycle, and 0.26 [0.14, 0.39] mL for the low-frequency cycle.

conclusionThe proposed spin-history modeling method enabled quantitative estimation of CSF flow components using a conventional fMRI dataset and showed that the cardiac cycle dominates CSF motion at the foramen magnum.