Modeling Cerebral Hemodynamics Using BOLD Magnetic Resonance Imaging and its Application in Mild Cognitive Impairment

Objective This study develops a procedure and related analytical methods for deriving indices of cerebral hemodynamics in the magnetic resonance imaging (MRI) setting using resting state recordings of systemic blood pressure, pulse rate, and end-tidal CO 2 synchronized with the MRI image acquisitions of blood oxygenation level dependent (BOLD) data as a measure of cerebral perfusion. Methods We employed the concept of Principal Dynamic Modes (PDM) to model the effect of three determinants of cerebral perfusion: mean arterial blood pressure (MABP), end-tidal CO 2 (PETCO 2 ), and pulse rate (PR). The relation between these signals and the BOLD signal were used respectively to quantify cerebral autoregulation (CA), CO 2 vasoreactivity (CVR), and pulse rate reactivity (PRR). Results Hemodynamic indices were obtained from 129 participants with normal cognition (NC) and mild cognitive impairment (MCI). CA was reduced in MCI compared to NC in the parietal lobe, CVR was reduced in MCI in the occipital and temporal lobes, and PRR was reduced in the frontal, parietal, occipital and temporal lobes. Reduced CVR and PRR were associated with worse cognitive scores including memory and executive function. Conclusion Employed acquisition and analysis of MRI hemodynamic identified cerebral hemodynamic alterations in MCI, related to PR and ETCO2 changes. Significance This modeling approach may offer a novel way to clinically assess cerebral hemodynamics during MRI.