Analysis and development of strategies for magnetic resonance functional neuroimaging

Abstract

This thesis focuses on strategies for oxygenation-sensitive MRI that are designed to measure alterations in brain physiology induced by changes in neuronal activity.

In a first study, it was shown that the temporal evolution of BOLD MRI signal changes in response to both brief and sustained visual activation are equally well visible in FLASH and EPI recordings. These results demonstrated that previously reported inconsistencies with regard to sustained visual activation cannot be explained by technical differences and confirmed independence of brain mapping studies from MRI acquisition physics.

In a second step, series of echo planar images were acquired to determine the temporal response profiles to repetitive visual activation (5 Hz reversing black and white checkerboard vs. darkness). Analysis of the real-time data resulted in a strong positive BOLD MRI response (3-6%) as well as a marked post-stimulus undershoot (1-2%, duration 60-90 s) for all paradigms. Insufficient recovery periods between stimuli gave rise to a "wrap around" effect when analysing time-locked averages from multiple activation cycles. The procedure may artificially generate an "initial dip" in pertinent response curves. Similar findings were obtained for subsecond visual stimulation (checkerboard stimulation vs. darkness). The use of a 0.2 s stimulus and a 90 s recovery period revealed an initial latency phase (2 s, no signal change), a positive BOLD response, and a post-stimulus undershoot lasting for about 50-60 s.

The temporal and spatial aspects of MRI responses to repetitive brief (1 s) visual stimuli (checkerboard vs. darkness) were addressed as a function of inter-stimulus intervals in order to substantiate the utility of event- related protocols. A main finding was a reduction of the functional contrast as given by the peak-to-peak difference between conditions when inter-stimulus intervals become shorter than the response time between stimulus onset and maximum signal strength. The spatial extent of brain activation remained unaffected as long as the underlying responses were temporally resolved.

The effect of stimulus quality on the physiological response characteristics of BOLD MRI signals were studied with use of paradigms comprising checkerboard vs. darkness or vs. gray light as well as gray light vs. darkness. For all paradigms tested, the resulting time courses exhibited similar physiological responses. These findings support the assumption of a uniform response function for the evaluation of event-related paradigms.

Finally, a baseline study addressed the temporal and spatial characteristics of positive and negative BOLD MRI responses to visual activation and deactivation, respectively. Pertinent paradigms used reversed stimulus presentations with gray light and checkerboard as distinct functional states. For sustained stimulation (>= 60 s) the two conditions resulted in markedly different steady state BOLD MRI signal strengths. The transient responses to brief stimulation (< 18 s) differed in so far as positive BOLD responses and negative undershoot effects are temporally separated by about 10 s for activation processes, whereas negative BOLD responses and undershoot contributions overlap for deactivation processes. Apart from differences in stimulus features (e.g., motion) the used activation and deactivation protocols revealed similar maps of neuronal activity changes.