fMRI: Acquisition & Optimization

Thursday 15 May 2014

Bertram J. Wilm¹, Lars Kasper¹, Yolanda Duerst¹, Benjamin E. Dietrich¹, Simon Gross¹, Thomas Schmid¹, David O. Brunner¹, Christoph Barmet^1,2, and Klaas P. Pruessmann^1
1Institute for Biomedical Engineering, ETH Zurich and University of Zurich, Zurich, Zurich, Switzerland, ²Skope Magnetic Resonance Technologies, Zurich, Switzerland

In fMRI time series temporal SNR is reduced by field drifts as well as physiological noise such as stemming from breathing or limb motion. To address this problem we present the use of a real-time field control system to compensate for field fluctuations independently of the employed sequence without the need to alter image reconstruction or additional pre-processing.

Tae Kim¹, Tiejun Zhao², Yoojin Lee¹, Piotr Starewicz³, Hoby Hetherington¹, and Jullie Pan^1
1Radiology, University of Pittsburgh, Pittsburgh, PA, United States, ²Siemens Medical Solution USA, INC., Siemens MediCare USA, Pittsburgh, PA, United States, ³Resonance Research Inc., Billerica, MA, United States

High degree/order shimming was applied to increase field homogeneity to reduce susceptibility-induced signal loss in gradient-echo EPI using a shim insert coil at 7T. Use of the shim insert improved the overall homogeneity across the entire brain by 30% in comparison to conventional 1st&2nd degree/order shimming. For challenging brain regions such as the anterior temporal and frontal lobes use of the shim insert increased the temporal-signal-to-noise ratios by 111.2%. Our study demonstrates that the use of higher order/degree shims improves GE-EPI BOLD signal at high field.

Ariane Fillmer¹, Milan Scheidegger^1,2, Signe Johanna Vannesjo¹, Matteo Pavan¹, Klaas Paul Pruessmann¹, and Anke Henning^1,3
1Institute for Biomedical Engineering, University and ETH Zurich, Zurich, ZH, Switzerland, ²Clinic of Affective Disorders and General Psychiatry, University Hospital of Psychiatry Zurich, Zurich, Switzerland, ³Max Planck Institute for Biological Cybernetics, Tuebingen, Germany

EPI, the “work horse” of fMRI, is prone to artifacts, induced by B0 inhomogeneities. Therefore, sophisticated B0 shimming techniques are required. A promising, but technically demanding approach for B0 shimming is dynamic shim updating (DSU). The application of DSU to fMRI is, however, only feasible, if an accurate pre-emphasis calibration for eddy current correction (ECC) is implemented. This work presents the first report of resting-state fMRI data, acquired with a 3rd-order dynamically updated B0 shim along with a dynamic F0 determination.

Tim van Mourik¹, Peter J Koopmans¹, and David G Norris^1
1Donders Centre for Cognitive Neuroimaging, Nijmegen, Netherlands

We propose a registration method that corrects local distortion in functional scans, using the brain surfaces as constructed by FreeSurfer. An iterative procedure of Boundary Based Registration (Greve & Fischl 2009) is used in a multiscale approach to register parts of the brain, first at the whole-brain level, then at smaller levels. The specificity increases at each stage and accurately and robustly corrects for local mismatches between the structural and the functional volumes. This non-linear registration greatly improves the spatial accuracy, which allows for laminar analysis.

Barbara Dymerska¹, Benedikt Poser², Wolfgang Bogner¹, Eelke Visser³, Pedro Cardoso¹, Markus Barth^4,5, Siegfried Trattnig¹, and Simon Robinson^1
1Departement of Biomedical Imaging and Image-guided Therapy, Medical University of Vienna, Vienna, Vienna, Austria, ²Department of Psychology and Neuroscience, Cognitive Neuroscience, Maastricht University, Maastricht, Netherlands, ³Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, Oxford, United Kingdom, ⁴Erwin L. Hahn Institute for Magnetic Resonance Imaging, Essen, Germany, ⁵Donders Institute for Brain, Cognition and Behaviour, Donders Centre for Cognitive Neuroimaging, Radboud University Nijmegen, Nijmegen, Netherlands

Echo Planar Imaging suffers from geometric distortions caused by local magnetic field inhomogeneities. We here present a new dynamic distortion correction (DDC) based on an EPI sequence, in which the echo time is alternated between two values in the odd and even time points allowing calculation of the field maps between adjacent time points. Proposed method showed more accurate distortion correction than static approach. EPI images were successfully unwarped at 7T using DDC despite head motion. Additionally, the precision of the activation localization in fMRI measurements with volunteers performing chin and motion tasks was improved.

Jürgen Finsterbusch^1,2
1Department of Systems Neuroscience, University Medical Center Hamburg-Eppendorf, Hamburg, Germany, ²Neuroimage Nord, Hamburg-Kiel-Lübeck, Germany

Recently, an approach has been presented to acquire different regions distributed in the brain in a single projection image with a high temporal resolution. To distinguish the signals of the different target regions unambiguously, their projections must not overlap which constrains the positions of the regions and the orientation of the projection plane. Here, the approach is extended by gradient blips in the slice direction in order to shift the projections of the different target regions relative to each other (similar to blipped-CAIPIRINHA). Thus, the positions of the target regions and the image plane can be chosen more flexibly.

Rüdiger Stirnberg¹ and Tony Stöcker^1
1German Center for Neurodegenerative Diseases (DZNE), Bonn, Germany

This work studies the temporal SNR characteristics of conventional 2D-EPI and segmented 3D-EPI acquisitions under realistic conditions for standard and advanced applications. Therefore, different whole-brain protocols optimized for BOLD fMRI at normal and high spatial resolutions are prepared. Time series data are acquired at 3 and 7 Tesla. As expected, segmented 3D-EPI achieves higher tSNR than 2D-EPI at high spatial resolutions (thermal noise dominated regime). Beyond that, however, a general signal sensitivity advantage due to higher imaging acceleration capabilities (larger data rate, hence increased robustness of fMRI data) renders 3D-EPI useful even at typical, rather coarse resolutions.

Hao Sun¹, Jeffrey A. Fessler¹, Douglas C. Noll², and Jon-Fredrik Nielsen^2
1Electrical Engineering and Computer Science, the University of Michigan, Ann Arbor, MI, United States, ²Biomedical Engineering, the University of Michigan, Ann Arbor, MI, United States

Most functional brain MR imaging uses T2*-weighted gradient-echo sequences with single-shot readout (BOLD fMRI), providing high activation contrast but suffering from off-resonance-induced image artifacts (signal drop, distortions or blurring). Steady-state fMRI based on balanced steady-state free precession (bSSFP) uses segmented readouts and can produce excellent image quality, but is susceptible to dark “banding” artifacts, and generally has lower functional contrast than BOLD. Small-tip fast recovery (STFR) imaging is a recently proposed steady-state imaging sequence that has similar image intensity to bSSFP, but with reduced signal variations (banding) due to resonance offsets. STFR relies on a tailored “tip-up,” or “fast recovery,” RF pulse to align the spins with the longitudinal axis after each data readout segment. Using Monte Carlo Bloch simulation and preliminary in vivo experiments, it has been demonstrated that STFR can produce detectable fMRI signal. Here we investigate the spoiled STFR fMRI sequence in more detail by: (1) performing a quantitative comparison between simulation and in vivo experiments, and (2) estimating test–retest reliability of STFR (and BOLD) functional maps.

Ute Goerke¹, M. Elias Kersten², Siraj Bachani², David A. Bereiter², and Donald R. Nixdorf^2
1CMRR/Radiology, University of Minnesota, Minneapolis, Minnesota, United States, ²School of Dentistry, University of Minnesota, Minnesota, United States

In previous studies, it has been demonstrated that the emerging ultrafast imaging sequence RASER is beneficial for fMRI at ultrahigh magnetic field strength. In this paper, we show the potential of RASER for mapping neural correlates of pain processing in the human brain stem with high spatial and temporal resolution. In particular, 3D-RASER captures the BOLD response with higher temporal resolution compared to multi-slice technique. Activation maps were generated using the post-processing technique, SSBA. Robust activation was found in trigeminal nucleus (V2), the chief sensory nucleus and the periaqueductal grey induced by a moderately painful prick stimulus of the gingiva.

Andreas Fehlner¹, Sebastian Hirsch¹, Jing Guo¹, Jürgen Braun², and Ingolf Sack^1
1Department of Radiology, Charité - Universitätsmedizin Berlin, Berlin, Germany, ²Department of Medical Informatics, Charité - Universitätsmedizin Berlin, Berlin, Germany

This study is on the human brain's mechanical response to functional activation by fast single shot MR elastography (MRE). In a total of 57 volunteers we observed a significant reduction of cerebral viscoelasticity in the order of 2.5% due to visual stimulation. The sensitivity of functional MRE was higher at very low vibration frequencies of 25 and 30 Hz compared to 40 and 50 Hz indicating the involvement of poroelastic effects. In contrast to activity patterns revealed by BOLD fMRI, the viscoelastic response to brain function appears to be a global phenomenon which may arise from activity-induced alteration of micro vascular pressure and effective tissue pressure.