Pushing the Bounds of fMRI Resolution
Click on to view the abstract pdf and click on to view the video presentation.
Thursday May 12th
Room 512A-G  16:00 - 18:00 Moderators: Nan-Kuei Chen and Seong-Gi Kim

16:00 631.   Ultra-fast fMRI of human visual cortex using echo-shifted magnetic resonance inverse imaging 
Wei-Tang Chang1, Thomas Witzel2, Kevin Wen-Kai Tsai1, Wen-Jui Kuo3, and Fa-Hsuan Lin1
1Institute of Biomedical Engineering, National Taiwan University, Taipei, Taiwan, 2Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, MA, United States, 3Institute of Neuroscience, National Yang-Ming University, Taipei, Taiwan

To ensure a good BOLD contrast with TE = 34 ms at 3T, the scanning time of magnetic resonance inverse imaging (InI) can be further shortened to 25 ms per volume by using the echo shifting technique. Here we demonstrated the feasibility of using echo-shifted InI (esInI) with an unprecedented temporal resolution (40 Hz / volume) to detect hemodynamic responses of human visual cortex. Validated by EPI, esInI was found with good spatial localization using the minimum norm estimation (MNE) for volumetric reconstruction.

16:12 632.   Dynamic magnetic resonance multi-projection inverse imaging (mInI) with isotropic spatial resolution 
Kevin Wen-Kai Tsai1, Aapo Nummenmaa2, Thomas Witzel2,3, Wei-Tang Chang4, Wei-Jui Kuo5, and Fa-Hsuan Lin1,2
1Institute of Biomedical Engineering, National Taiwan University, Taipei, Taiwan, 2A. A. Martinos Center, Massachusetts General Hospital, Charlestown, MA, United States,3Harvard-MIT Divisions of Health Sciences and Technique, Charlestown, MA, United States, 4Institute of Biomedical Engineering, National Taiwan University, Taipei, Taiwan, Taiwan, 5Cognitive Neuropsychology Laboratory, National Yang-Ming University, Taipei, Taiwan, Taiwan

MR inverse imaging (InI) can achieve 100 ms temporal resolution with whole-brain coverage using highly parallel detection at the cost of compromised spatial resolution due to solving ill-posed inverse problems in image reconstruction. Here we propose the multi-projection InI (mInI) method. Using projection data in three orthogonal directions from all RF channels separately can achieve a high temporal resolution without compromising spatial resolution. Results of using mInI to study the hemodynamic responses in human visuomotor system from 12 participants using a 32-channel head coil array at 3T can achieve 100 ms temporal resolution and 4 mm3 isotropic spatial resolution.

16:24 633.   Single-Shot Whole Brain Echo Volume Imaging for Temporally Resolved Physiological Signals in fMRI 
Thomas Witzel1,2, Jonathan R Polimeni1, FaHsuan Lin1,3, Aapo Nummenmaa1, and Lawrence L Wald1,2
1A.A. Martinos Center MGH Department of Radiology, Harvard Medical School, Boston, MA, United States, 2Harvard-MIT Division of Health Sciences and Technology, Cambridge, MA, United States, 3Biomedical Engineering, National Taiwan University, Taipei, Taiwan

We developed a 64x64x56 matrix, whole-head 3T single-shot EVI sequences for fMRI. We image the whole brain at 8 frames per second and 3.4mm isotropic resolution with the kz phase encoding direction accelerated by a total of 21 fold to reduce distortions in this “slow” phase encoding direction 21 fold. We analyze the spectral composition of BOLD and physiological signals to show that full Nyquist the cardiac frequency is helpful in mitigating their effect on the estimation of the fMRI activation.

16:36 634.   Tracking Dynamic Resting-State Networks with High Temporal Resolution fMRI 
Hsu-Lei Lee1, Benjamin Zahneisen1, Thimo Grotz1, Pierre LeVan1, and Jürgen Hennig1
1Medical Physics, University Medical Center Freiburg, Freiburg, Germany

Most resting-state network analysis methods assume the networks are stationary during the course of the scan session, usually ranged from 5 to 15 minutes. However it might not be the case. We propose to use a highly under-sampled fast acquisition scheme (shell trajectory) to record images at a 10 Hz sampling rate and use sliding window to track the dynamic change in those coherent networks with a temporal resolution of 20 seconds.

16:48 635.   Multiplexed Echo Planar Imaging with Sub-second Whole Brain FMRI and Fast Diffusion Imaging 
David A Feinberg1,2, Steen Moeller3, Stephen Smith4, Edward Auerbach3, Kamil Ugurbil3, and Essa Yacoub3
1Advanced MRI Technologies, Sebastopol, CA, United States, 2University of California, Berkeley and San Francisco, CA, United States, 3Center for Magnetic Resonance Research, University of Minnesota, 4FMRIB, Oxford University

A multiplexed-EPI (M-EPI) pulse sequence is presented that combines temporal multiplexing (m) utilizing simultaneous echo refocused (SIR) EPI and spatial multiplexing (n) with multibanded RF pulses (MB) to achieve m x n images in an EPI echo train. This resulted in significant increases in temporal resolution for whole brain fMRI evaluated at 0.8s and 0.4s TR in resting state and in substantial reductions in scan time 8.5m in HARDI neuronal fibertracks using 256 b-values. Multiplexed EPI can be used to acquire higher spatial resolutions, increase diffusion encoding, or to investigate temporal dynamics of the fMRI response.

17:00 636.   Resting-state correlations between depths within columns of voxels radial to the cortical surface 
Jonathan Rizzo Polimeni1, Kyoko Fujimoto1, Boris Keil1, Douglas N. Greve1, Bruce Fischl1,2, and Lawrence L. Wald1,3
1A. A. Martinos Center for Biomedical Imaging, Department of Radiology, Harvard Medical School, Massachusetts General Hospital, Charlestown, MA, United States,2Computer Science and AI Lab (CSAIL), Massachusetts Institute of Technology, Cambridge, MA, United States, 3Harvard-MIT Division of Health Sciences and Technology, Massachusetts Institute of Technology, Cambridge, MA, United States

Here we provide characterization of the radial independence of the fluctuations at each depth below a given cortical location. This radial correlation length of resting state fluctuations was measured using gradient echo and spin echo EPI. The radial correlation length is strongly affected by the removal of confounding physiological signals in a counter-intuitive way--namely the global signals regressed out do not seem to produce an artifactual high-correlation between layers, but rather mask the presence of correlations between different depths. Similarly, the spin echo acquisition, which is also expected to remove relatively global fluctuations also uncovers additional underlying radial correlations.

17:12 637.   High Resolution fMRI of the Functionally-defined Fusiform Face Area using 7T 
Rankin Williams McGugin1, Christopher Gatenby2,3, and Isabel Gauthier1
1Psychology, Vanderbilt University, Nashville, TN, United States, 2Radiology & Radiological Sciences, Vanderbilt University Medical Center, Nashville, TN, United States,3Radiology, University of Washington, Seattle, Washington, United States

We imaged the inferior temporal lobe in humans at 7T, using built-in realtime fMRI analysis in combination with the 3D FFE-SENSE sequence to allow for bilateral imaging of the Fusiform Face Area (FFA). We demonstrate robust and reliable functional heterogeneity of the fine-grain neural architecture of bilateral FFA, such that clusters of voxels maximally responsive to non-face categories (irrespective of hemisphere) are spatially interdigitated amongst regions showing maximal response to faces. We demonstrate that a functionally-defined region that appears to be homogenously face-selective at standard resolution (~3mm isotropic) is, in fact, functionally heterogeneous when explored at HR.

17:24 638.   Tonotopic Mapping in Inferior Colliculus using bSSFP fMRI and Sweeping Frequency Auditory Stimulation 
Matthew Man Hin Cheung1,2, Joe S. Cheng1,2, Iris Y Zhou1,2, Kevin C. Chan1,2, Condon Lau1,2, and Ed X. Wu1,2
1Laboratory of Biomedical Imaging and Signal Processing, The University of Hong Kong, Pokfulam, Hong Kong SAR, China, People's Republic of, 2Department of Electrical and Electronic Engineering, The University of Hong Kong, Pokfulam, Hong Kong SAR, China, People's Republic of

The inferior colliculus (IC) is an important auditory relay center for ascending pathways and the central nucleus of IC receives input from lower auditory centers and projects to the medial geniculate body in a strictly tonotopic manner. In this study, a novel approach was proposed to map the tonotopic organization in IC with fMRI by using (1) periodic frequency-modulated auditory stimulus with linearly sweeping frequency; and (2) distortion-free balanced steady state free precession (bSSFP) sequence. Tonotopic mapping in IC by fMRI was here demonstrated successfully for the first time in animals or humans.

17:36 639.   Cortical depth dependent temporal dynamics of the BOLD response in the human brain 
Jeroen Cornelis Willem Siero1,2, Natalia Petridou1,2, Johannes Marinus Hoogduin1,2, Peter R Luijten2, and Nick F Ramsey1
1Rudolf Magnus Institute, University Medical Center Utrecht, Utrecht, Netherlands, 2Radiology, University Medical Center Utrecht, Utrecht, Netherlands

The BOLD specificity to the cortical vascular architecture was investigated in the human brain, at 7T: temporal dynamics (delay and width) of the BOLD response were characterized across cortical depth in the primary motor and visual cortices. Faster and narrower (<4s) BOLD responses corresponded to the deeper gray matter, and these temporal properties increased in an orderly manner toward the cortical surface. A pooling time of oxygenated blood from the deeper cortex to the pial surface was estimated <~1s. These findings matched the known vascular organization across cortical depth, and may lead to new insights in neurovascular coupling in humans.

17:48 640.   Ipsilateral FMRI Response in Primary Somatosensory Cortex (Area 3b) of Awake Marmosets 
Junjie V Liu1, Matthew Huberty1, and Afonso C Silva1
1NINDS, National Institutes of Health, Bethesda, MD, United States

We conducted systematic measurements of BOLD fMRI responses to unilateral electrical stimulation, in both the contralateral and the ipsilateral primary somatosensory cortex (SI) of awake non-human primates (marmosets). Taking advantage of T1 mapping co-registered with fMRI, we found many interesting differences between the responses in low-T1 and high-T1 regions within area 3b of SI, specifically regarding to their relationships with stimulus frequency, duration and intensity. Because T1 is correlated with density of thalamic inputs, with our results the ipsilateral SI responses driven mainly by corticocortical neural inputs can be distinguished from the contralateral SI responses driven mainly by thalamic inputs.