The Physiological Basis of BOLD Functional MRI /

Functional magnetic resonance imaging (fMRI) is an exciting technology, and researchers have used it to give us beautiful maps of the brain performing a multitude of tasks. A non-invasive tool that can be safely applied in humans, fMRI also has the potential to go beyond a simple mapping tool by providing insight into the physiology underlying neural activity, particularly the cerebral metabolic rate of oxygen (CMRO₂), which is an indicator of the underlying neural activity. The blood oxygenation level dependent (BOLD) signal is the main fMRI signal, but a major concern with BOLD is the inability to quantitatively relate it to underlying physiological changes. This is because the BOLD signal depends not only on oxygen metabolism changes but also on cerebral blood flow (CBF) and the baseline state of the brain. However by performing simultaneous measurements of CBF and BOLD in response to a stimulus and also in response to inhaled CO₂ for calibration, quantitative measurements of CMRO₂ can be made using a simple mathematical model to relate the data. One measure commonly reported is the ratio of the CBF change to the CMRO₂ change, commonly known as the CBF- CMRO₂ coupling parameter, n. In this work, I explore the physiological basis of the BOLD signal in situ by developing a detailed biophysical model of the BOLD signal and optimizing the simple mathematical model used for data analysis. I also present a new heuristic model of the BOLD signal that suggests a new, straightforward "ratio" method for the analysis of combined BOLD and CBF measurements. I applied the optimized simple model to study the effects of caffeine on the physiological response of the brain to a visual stimulus and found that caffeine increases the absolute CMRO₂ change to the same stimulus by 61% suggesting increased neuronal excitability. Using the detailed biophysical method, I also demonstrate that the ratio method accurately determines when CBF-CMRO₂ coupling in response to a stimulus differs between two stimuli using only the measured BOLD and CBF signals. Finally I demonstrate a new method for BOLD calibration that is more reliable than the inhaled CO₂ metho