Richard B. Buxton - US grants

In the last few years functional magnetic resonance imaging (fMRI) has become a powerful and widely used tool for investigating the working human brain. The small changes in the MR signal due to the Blood Oxygenation Dependent (BOLD) effect that accompany local changes in brain metabolism can be used to map patterns of brain activation during performance of a variety of sensory, motor and cognitive tasks. But despite the widespread use of fMRI techniques, the basic physiological and biophysical mechanisms underlying the observed signal changes are still poorly understood. The broad goal of the proposed work is to answer two basic questions: 1) What are the physiological changes accompanying human brain activation?, and 2) How can we quantitatively interpret observed MR effects in terms of physiological changes? This project brings together two lines of research that have developed in our laboratory over the last few years: 1) Theoretical mathematical modeling of the physiological changes occurring during activation and the quantitative translation of these changes into MR signal changes; and 2) Development and evaluation of MRI experimental techniques for quantitative perfusion measurements. Based on the theoretical modeling, we have framed this project around two central hypotheses: 1) The changes in cerebral oxygen metabolism (CMRO2) and cerebral blood flow (CBF) are tightly coupled during brain activation, but in a nonlinear fashion requiring large changes in CBF to support small changes in CMRO2 because of limited O2 extraction from the capillary; and 2) The temporal profile of signal changes observed in fMRI experiments is highly sensitive to the relative time courses for blood flow and blood volume changes during activation. These hypotheses will be tested using recently developed MRI techniques for measurement of perfusion and blood volume changes in combination with the conventional fMRI signal sensitive to blood oxygenation. Experiments will measure changes in these physiological variables during performance of four types of stimulation (sensorimotor, visual, auditory, and cognitive). The three sets of experiments to be performed are: 1) variable stimulus amplitude to vary the physiological response; 2. high temporal resolution measurements of the time course of BOLD signal, blood flow and blood volume changes; and 3) sustained activation to test the continued coupling of CBF and CMRO2. In addition, the theoretical models will be further developed to include viscoelastic properties of blood vessels. The end result of this work will be an experimental and theoretical characterization of the physiological changes accompanying human brain activation.

[unreadable] DESCRIPTION (provided by applicant): Functional magnetic resonance imaging (fMRI) is an essential part of many basic neuroscience studies, providing a powerful tool for mapping patterns of activation in the working human brain. However, these techniques have been slow to move into the clinical realm, due both to a lack of sensitivity of current techniques and also to a lack of understanding of the nature and variability of the local hemodynamic response and the associated Blood Oxygenation Level Dependent (BOLD) signal response to activation. [unreadable] [unreadable] Because the hemodynamic response (HR) is the key physiological process in the brain linking neural activity with blood flow and energy metabolism, a detailed characterization of the HR has the potential to provide a powerful probe for early assessment of disease and for investigations of the pathophysiology involved. There are three problems blocking progress in the development of these clinical applications: 1) The BOLD effect is highly variable even in healthy subjects, making it difficult to detect meaningful differences between individual subjects, 2) Nonlinearities in the BOLD response introduce a strong dependence on the experimental design used, and 3) The interpretation of alterations in the BOLD response observed in a disease population is intrinsically ambiguous because the BOLD effect depends on combined changes in cerebral blood flow (CBF), cerebral metabolic rate of oxygen (CMRO2) and cerebral blood volume (CBV). [unreadable] [unreadable] The basis for this proposal is the hypothesis that simultaneous measurement of the CBF response and the BOLD response using arterial spin labeling (ASL) techniques, when analyzed with a dynamic model for the HR, will greatly expand the specificity and reliability of HR measurements. To validate the feasibility of this approach we will: 1) verify that the proposed model is adequate for describing the CBF and BOLD responses by testing whether the model parameters calculated with one experimental design allow accurate prediction of the responses with a different experimental design; 2) determine the reproducibility of the HR parameter measurements in repeated studies on the same subject, and assess the variability of the parameters in young healthy adults; 3) determine the degree to which variations in resting CBF, CMRO2 changes and CBV changes contribute to the variability of the BOLD response, and 4) determine whether there is a significant variation in the HR parameters with healthy development and aging.

DESCRIPTION (provided by applicant): The overall long-term goal of these studies is to understand the mechanisms by which pulmonary blood flow is controlled, and how that control contributes to gas exchange defects or optimization in health and disease. The specific goal is to use a novel MRI technique to quantify the spatial and temporal dynamics of blood flow in the normal human lung. Temporal heterogeneity in a number of physiological systems has been found to be a mark of healthy function, yet little is known about the temporal dynamics of blood flow in the human lung because the appropriate tools for measuring temporal heterogeneity have not been available. Recently we developed a noninvasive MRI technique that provides quantitative measurements of pulmonary blood flow with a spatial resolution of <1 cm3 and a temporal resolution of ~10 s in the human lung, permitting us to examine spatial-temporal heterogeneity in the human lung for the first time. The Specific Aims are designed to systematically explore the normal spectrum of spatial-temporal heterogeneity, testing: 1) the effects of altered inspired gas (O2 and CO2) in healthy subjects; 2) the effects of exercise; 3) the effects of ageing; and 4) the effects of altered O2 and CO2 in the lungs of subjects susceptible to high altitude pulmonary edema. This will be the first systematic, quantitative study of the spatial and temporal dynamics of pulmonary blood flow in human subjects, and will lay a foundation for applying these methods in the early detection and characterization of disease. PUBLIC HEALTH RELEVANCE: Oxygenation of the blood in the lungs depends on a close matching of ventilation and blood flow: fresh gas and blood flow need to be at the same place and at the same time. We have developed a novel imaging technique for measuring the distribution of blood flow in the human lung not only spatially, but also over time, a measurement that has not been possible before. In this work we will explore the normal dynamics of blood flow as a foundation for applying these methods to identify, and to better understand, the underlying mechanisms of disease.

DESCRIPTION (provided by applicant): Our overall goal is to establish the basis for a new experimental paradigm for functional magnetic resonance imaging (fMRI) that makes possible the determination of fluctuating brain activity patterns during performance of complex tasks, at rest, or in response to a drug, in quantitative units of absolute cerebral blood flow (CBF). Conventional functional magnetic resonance imaging (fMRI) is based on detection of blood oxygenation level dependent (BOLD) signal modulations. The BOLD signal is a sensitive indicator of underlying physiological changes, but BOLD-fMRI applications are currently limited because the magnitude of the BOLD signal does not provide a reliable quantitative measure of a physiologically meaningful quantity. Arterial spin labeling (ASL) methods provide quantitative measurements of CBF, a well-defined physiological variable. However, sensitive measurement of CBF dynamics remains challenging because of the low signal to noise ratio of the ASL measurement. The key idea of this proposal is a new method to take simultaneous measurements of ASL and BOLD time series, and with an appropriate model of the BOLD response, treat these signals as both being generated from the same underlying time series of CBF fluctuations. The combined data are used to estimate the CBF fluctuations without knowing anything about the underlying drivers of those fluctuations. The proposed new methodology rests on two assumptions: 1) the CBF/CMRO2 coupling ratio for a local region remains constant during the measurement period; and 2) there are no systematic fluctuations of the BOLD signal that are unrelated to CBF fluctuations. Neither assumption is strictly true, so the high risk hal of this proposal is the open question of whether these effects are sufficiently small or can be adequately corrected for the methodology to be robust. We propose to test the feasibility of this method by: Measuring simultaneous ASL and BOLD responses to visual stimuli in healthy human subjects with an experimental paradigm designed to challenge the basic assumptions of the methodology, including variable CBF/CMRO2 coupling, dynamic transitions and BOLD transients (Aim 1); and developing two new analysis techniques to improve the accuracy of the method, one to adapt a recent independent components analysis (ICA) method to use our multi-echo acquisition to identify and remove artifact components in the measured BOLD signals, and the second to improve estimation of the model parameters and deal with a time varying CBF/CMRO2 coupling ratio with a Bayesian approach. The assessment of systematic errors, and the development of robust analysis tools for minimizing their effect, will establish a basis fo widespread application of the new method. This will substantially broaden the possible applications of fMRI, including measurement of brain activity during complex behavior, and quantitative assessments of the effects of development, disease, or drug administration on both the baseline physiological state and stimulus-evoked responses.

DESCRIPTION (provided by applicant): Our overall goal is to establish the basis for a new experimental paradigm for functional magnetic resonance imaging (fMRI) that makes possible quantitative measurement of the dynamics of the cerebral metabolic rate of oxygen metabolism (CMRO2) noninvasively in the human brain. Functional MRI methods based on blood oxygenation level dependent (BOLD) signal changes clearly have the potential to provide a window on CMRO2 dynamics, using simultaneous measurement of both the BOLD response to activation and the cerebral blood flow (CBF) response with a spiral dual-echo arterial spin labeling (ASL) technique. We and others have combined these tools in calibrated-BOLD studies to quantify changes in CMRO2, but these studies have focused on sustained changes in an approximate steady-state. The primary obstacle to extending these methods to measuring full CMRO2 dynamics is a physiological question: Do the dynamics of venous cerebral blood volume (CBVV) strongly differ from the dynamics of CBF? The key variable needed to estimate the dynamics of CMRO2 is the dynamics of the venous hemoglobin saturation, and the basic problem is that the BOLD effect depends primarily on changes in total deoxyhemoglobin, and thus also on the dynamics of venous blood volume. Dynamic measurements of CBF and BOLD signals provide sufficient information to estimate CMRO2 dynamics only if CBVV follows CBF. A primary example of this fundamental ambiguity of the BOLD signal is a long-standing issue in fMRI: is the post-stimulus undershoot of the BOLD signal a neural, vascular or metabolic effect? Despite considerable effort by many groups, there is still no clear answer, and the possibility of a dissociation of venous blood volume changes from CBF changes due to different dynamic time constants currently stands in the way of developing reliable tools for measuring CMRO2 dynamics. The motivation for this high risk/high gain proposal is that our recent studies of the effect of hyperoxia on the BOLD signal suggest a novel approach for addressing this primary physiological question, with a method that is specifically sensitive to CBVV. In addition, current models for the BOLD response and for analyzing the ASL experiment are essentially steady-state models, and these need to be expanded to include full dynamics. We will address these two basic limitations to measuring CMRO2 dynamics with two Aims. Aim 1: Extend our current modeling framework to include dynamics as well as potentially confounding physiologically variables, and use this to develop a Bayesian framework for estimating CMRO2 dynamics. Aim 2: Using the post- stimulus undershoot as a test case, use the hyperoxia approach to measure the dynamics of CBVV in human primary visual cortex in response to visual stimuli with varying duration and intensity. The endpoint will be a novel assessment of the dynamics of CBVV that will establish the feasibility of measuring the dynamics of CMRO2 for future applications in health and disease.

DESCRIPTION (provided by applicant): Our overall goal is to establish a new methodology for quantifying changes in brain state in single subjects, a paradigm for clinical applications in assessing the effects of a drug or for following the progression of disease and the response to therapy. The method builds on several developments during the previous period of support related to quantitative fMRI methods to measure cerebral blood flow (CBF) and the cerebral metabolic rate of oxygen (CMRO2): the development of an alternative approach to calibration in the calibrated BOLD method that does not require inhalation of special gas mixtures; a series of studies showing that the BOLD response alone is relatively insensitive for detecting a change in brain state that leads to a change in the evoked physiological response to a standard stimulus; and a particular example study of caffeine effects showing that quantitative fMRI methods are able to detect substantial changes in both baseline and activation responses of CBF and CMRO2 that were essentially undetected by BOLD alone. Based on this work we propose that a set of four metrics, reflecting baseline CBF and CMRO2 and their responses to a standard stimulus (analogous to a 'stress test'), can be acquired noninvasively and without requiring special gas mixtures in ~20 min. The goal of the project is to establish a basis for clinical applications of this approach by testing the ability of these metrics to track brain state changes in individual subjects. Aim 1 will test that the new approach to calibration based on measuring the relaxation rate R2' in the baseline state gives essentially the same required information as the standard hypercapnia challenge for quantifying the CMRO2 response to a stimulus. Aim 2 will test the reproducibility of the four metrics in a healthy population in both back to back test and in studies ~1 month apart. Aim 3 is a blinded test case of the ability of these metrics to detect changes in brain state due to caffeine on an individual basis. This work will establish the sensitivity and reliability of these metrics for following brain state changes in individual subjecs as a basis for future focused clinical applications in assessing drug effects (e.g., identifying responders, or quantifying the effect on disease), for determining the progression of disease, or for assessing the effect of therapy.

Functional magnetic resonance imaging (FMRI) based on the blood oxygen-level dependent (BOLD) signal has revolutionized cognitive neuroscience. However, the FMRI BOLD signal is an indirect measure of neural activity dependent on fluctuations in the oxygenation level of blood that is neither a direct measure of neural activity nor the hemodynamics responding to that activity. There is accumulating evidence that the indirect nature of the BOLD signal can provide inaccurate estimates of neural activity when comparing groups or individuals with differences in neurovascular coupling. This means that the BOLD signal may provide inaccurate estimates of neural activation in contrasts for typical and atypical development, psychiatric and neurological disorders, and response to psychoactive or pharmacological agents. The BOLD signal may systematically under- or over-estimate neural activity at rest or in response to task or challenge factors in one group or the other. To increase the validity of FMRI, it will be critical to understand how the additional factors modulating the BOLD signal vary across contrasts. This project seeks to meet this challenge by applying a newly developed quantitative FMRI (QFMRI) technology that, unlike other methods of QFMRI, uses standard MRI approaches that do not require the inhalation of special gases (e.g., CO2) and thus are capable of being used in a variety of settings and populations just as traditional BOLD FMRI is used today. Here, we will evaluate the advantages of QFMRI across typical development, an area that has strong evidence for inaccurate estimates of neural activity based on the BOLD signal. This project will use a dual-echo arterial spin labeling (ASL) to obtain baseline cerebral blood flow (CBF) and CBF and BOLD changes during performance on four perceptual processing tasks (visual, auditory, face, and object processing) in three developmental groups: children (8-9 years), adolescents (14-15 years) and young adults (20-30 years). In addition, we will obtain a measure of the cerebral metabolism rate of oxygen (CMRO2) in the baseline state using the newly developed GESSE technique for measuring R2'? (scaling parameter M). In Aim 1, we will use our direct measure of M to define neural activity in our tasks more precisely across development. This will allow us to characterize the validity of the BOLD method in developmental studies. In Aim 2, we will investigate the relationship of our QFMRI estimates of the physiological measures of neural activity with structural brain measures including cortical thickness, area, volume, and whole brain volume. The findings from this study of typical development will have broad implications in all areas of basic and clinical neuroscience, and may provide metrics allowing researchers to adjust prior findings of BOLD signal developmental trajectories to correct for potential underestimates of neural activity in children and adolescents based on the BOLD signal alone.