Magnetic Resonance Imaging of Human Brain Anatomy and Function

National Institute Of Neurological Disorders And Stroke

Investigators

Linked publications, trials & patents

Abstract

In a first project, we have continued investigating spontaneous brain activity across a broad spectrum of arousal states occurring during natural sleep. During our pilot study, we developed a better understanding of the relationship between arousal state and autonomic physiology, and the strong effect the latter has on the fMRI signal. Notably, by directly comparing the incidence of EEG K-complexes with fMRI and autonomic signals, we found that during sleep, episodic sympathetic constriction of the cerebral vasculature substantially affects blood flow and the fMRI signal. These findings will be published shortly. We plan to further investigate how relevant this mechanism is for the awake state, during which most fMRI experiments are performed. This mechanism may also affect interpretation of fMRI studies of pathologies that affect the sympathetic nervous system. The eventual goal is to develop a comprehensive model that would allow accounting for the confounding influence of autonomic physiology on the fMRI signal across the full range of arousal states. We have started to use deep learning approaches to identify relationships between the various autonomous markers (hear rate, respiration, blood pressure, peripheral vascular tone) and the fMRI signal. Based on the initial findings from the pilot sleep study, we slightly adjusted the design of our sleep experiment to include an additional measure of peripheral vascular tone from photoplethysmography, and insert a one week recovery between the first (adaptation) an second night of fMRI recording. The amended protocol was approved late spring 2019 and we so far had 6 successful experiments out of 7 attempted. In a second project, we investigated the possibility of improving fMRI spatial resolution by manipulating the underlying image contrast. Currently, this limit at 7T is about 1mm isotropic, which is close to allowing distinguishing between cortical layers. Current state of the art methods are based on blood volume contrast and have shown laminar contrast in several brain areas. Since perfusion changes with brain activity strongly exceed blood volume changes, we evaluated the utility of a previously developed perfusion MRI technique for high resolution fMRI. Employing a dual spin inversion strategy, this techniques eliminates all extravascular signal, improving the detection of any vascular effects associated with brain activity. Unexpectedly, activation related perfusion changes were too small to reliably detect activity at resolutions below 1.5 mm, making this approach unsuitable for laminar-resolution fMRI. We are currently investigating the source of the apparent discrepancy in sensitivity between blood volume and perfusion based fMRI techniques. In a third project, we investigated the mechanism by which myelin affects MRI contrast through the so-called mechanism of magnetization transfer (MT). MT has been used extensively in MRI for the study of CNS in a variety of pathologies but the underlying contrast mechanism is incompletely understood. One of the outstanding issues in this regard is what the T1 relaxation time constant of myelin protons is, as this parameter cannot be directly measured in-vivo. Using a 2-pool model approximation for MT between myelin and water protons, we had previously determined myelin T1 at 3T and 7T MRI field strength and found a strong field dependence. To confirm and further investigate this dependence, we have now performed additional measurements at 1.5T and even 0.5T, the latter having recently become available on a new NHLBI research system. We established that myelin T1 increases close to linearly with B1, over the range 0.5-7T consistent with our previous result and supporting the 2-pool model approach. This finding will help quantifying results from MT MRI and help optimize contrast for MRI at fields higher than 7T.