High Temporal Resolution Functional Magnetic Resonance Imaging at 7 Tesla

Speed has always been a critical consideration in magnetic resonance imaging (MRI), promising dramatic gains in imaging speed, a reduction in motion and susceptibility artifacts, and ultimately increased throughput for clinical studies. This thesis presents several new approaches to increase temporal resolution of a three dimensional (3D) echo planar imaging (EPI) sequence and explores the benefits of using high temporal resolution for both event-related brain activation and resting state networks fMRI studies. In the first approach, a new k-space traversal strategy for segmented 3D EPI is introduced which we termed as 3D reduced excitation EPI (3D RE-EPI). In this approach, two partitions are encoded per radio frequency (RF) excitation, effectively reducing the number of excitations used to acquire a 3D EPI dataset by a factor of two. With this strategy, whole-brain images with a nominal voxel size of 2mmisotropic could be acquired with a temporal resolution under half a second, using traditional parallel imaging acceleration upto 4x in the partition-encoding direction and using novel data acquisition speed-up of 2x. With 8x data acquisition speed-up in the partition-encoding direction, 3D RE-EPI produced acceptable image quality without introduction of noticeable additional artifacts. Due to increased temporal signal-to-noise ratio (tSNR) and better characterization of physiological fluctuations, the new strategy allowed detection of more resting state networks compared to standardmulti-slice two dimensional (2D) EPI and segmented 3D EPI. The approach above suffered from geometry factor (g-factor related) signal-to-noise ratio (SNR) losses, while using very high parallel imaging acceleration factor. To achieve substantial increase in the temporal resolution whilemaintaining low g-factors when high parallel imaging acceleration factors are used, the combination of segmented 3D EPI using 2D controlled aliasing with generalized autocalibrating partially parallel acquisitions (GRAPPA) was implemented in the next approach termed 3D-EPI-CAIPI. FunctionalMRI data with whole-brain coverage, a voxel size of 2 mm isotropic and a temporal resolution of 371 ms was acquired with acceptable image quality. 10-fold parallel imaging accelerated 3D-EPI-CAIPI data was shown to lower g-factor losses by as much as 10% with respect to segmented 3D EPI at 7 Tesla. Additional resting state networks were detected using 3D-EPI-CAIPI compared to a comparable standard 2D EPI acquisition. This was attributed to the improved statistics due to an increased number of volumes acquired in a given duration and because of the improved characterization of physiological signal fluctuations. Functional MRI with 400 ms temporal resolution allowed the detection of time-to-peak variations in temporal, occipital and frontal cortices haemodynamic responses due to multisensory facilitation of the order of ~200 ms. To conclude, this thesis introduces several novel techniques to increase temporal resolution of fMRI. Besides building theoretical foundations for these techniques, their applications for studying human brain functional connectivity, event-related brain activation and the ability to characterize physiological signal fluctuations at ultra-high field strength (7 Tesla) are evaluated.