Imaging and Analysis Strategies for Rapid MRI


University of California, Berkeley

Monday, July 18th, 2011. 13:30 @EE-314

Abstract: Magnetic resonance imaging (MRI) is a powerful modality that depicts the morphology and function of biological tissues noninvasively. Recent hardware developments have opened a venue for reducing the relatively long scan times that typically constrain the spatiotemporal resolution of MRI. As a result, there has been growing interest in rapid imaging and analysis methods that take full advantage of these advances.

In the first part of the talk, I will present fast imaging strategies that employ state-of-the-art steady-state pulse sequences and efficient parallel-imaging/compressed-sensing reconstructions. These strategies can achieve substantial improvements in image quality for numerous applications such as cardiac imaging, positive-contrast cellular imaging, and high-resolution peripheral angiography.

In the second part, I will describe sophisticated post-processing algorithms devised for rapid event-related fMRI. These modern statistical tools are used to build quantitative models that can accurately predict the brain’s response during natural vision. In addition to advancing our understanding of the visual system, these models can also be utilized for reliably decoding brain activity, i.e., mind reading.

BIO: Tolga Çukur graduated from Bilkent University in 2003 with a B.S. degree in Electrical Engineering. He continued his studies at Stanford University under the supervision of Prof. Dwight G. Nishimura, and received his Ph.D. degree in Electrical Engineering in 2009. He is currently a postdoctoral fellow in Prof. Jack L. Gallant’s lab in the Helen Wills Neuroscience Institute at the University of California, Berkeley.

His research interests include rapid data acquisition, image reconstruction, and statistical analysis strategies for magnetic resonance imaging (MRI), with a broad range of applications including angiographic, cardiac, cellular, and functional imaging. His current work focuses on building quantitative models of the human visual system during natural stimulation, using functional MRI measurements.

MRI Image Reconstruction Algorithms:
Compressed Sensing & Quantitative Susceptibility Mapping
Berkin Bilgic
PhD Candidate, Magnetic Resonance Imaging Group
Massachusetts Institute of Technology
Friday, June 10, 2011
At UMRAM, Cyberpark ,Block C, 2nd Floor on13:30 pm
Nonlinear reconstruction techniques that involve regularization gained substantial popularity in medical imaging community recently. This popularity is based on the demonstrated ability of these algorithms to reconstruct data sampled below the Nyquist rate, thus reducing scan times or making certain imaging applications more practical.
This talk presents a Bayesian Compressed Sensing algorithm that exploits similarities between MRI images obtained at different contrast settings and reconstructs them jointly from measurements below the Nyquist rate. The joint inference problem is formulated in a hierarchical Bayesian setting, wherein the variance of image gradients across contrasts for a single voxel is represented with a single hyperparameter. All of the images from the same anatomical region, but with different contrast properties, contribute to the estimation of the hyperparameters, and once they are found, the Fourier space data belonging to each image are used independently to infer the image gradients. Examples demonstrate improved reconstruction quality (up to a factor of 4 in root-mean-square error) compared to the state of the art compressed sensing algorithm.
An immediate application of regularized reconstruction is in Quantitative Susceptibility Mapping (QSM). Iron concentration is strongly correlated with the tissue susceptibility, hence QSM has useful applications such as quantifying tissue iron deposition in vivo or estimating vessel oxygenation. We introduce a QSM algorithm, calledL1-QSM, which estimates the tissue susceptibility based on MRI signal phase. This algorithm solves for the underlying susceptibility distribution that gives rise to the observed signal phase by placing sparsity inducing priors on the susceptibility map in spatial gradient domain. For validation, L1-QSM was tested on a dataset collected from young and elderly subjects, and compared well with publishedpostmortem measurements. Results indicate that the elderly group have significantly more iron in striatal and brain stem ROIs than the young group, offering an explanation to slowness in motor tasks with advancing age.
Short bio:
Berkin received his B.S. degrees in Electrical and Electronics Engineering and Physics from Bogazici University in 2008, and his S.M. degree in Electrical Engineering and Computer Science from MIT in 2010. He joined the MRI group in the Research Laboratory of Electronics in February 2010, and is currently pursuing his Ph.D. degree.His main research interests include developing image reconstruction algorithms to speed up data acquisition, magnetic susceptibility mapping and artifact reduction in spectroscopic imaging.

Bilkent Laboratory and International School students visited UMRAM with their Psychology teach to learn more about investigations on how brain works.

In the 16th annual meeting of the Turkish Magnetic Resonance Society, UMRAM researchers presented 11 scientific studies. Presenters were Oktay Algın, Aslıhan Örs, Davut İbrahim Mahçiçek, Mehmet Can Kerse, Ergin Atalar (as a replacement of Emre Kopanoğlu), Volkan Açıkel, Ali Çağlar Özen, Yiğitcan Eryaman, Esra Abacı, Taner Demir and Özgür Yılmaz. Murat Tümer of Bogazici University presented the catheter design which was developed using UMRAM resources. In addition UMRAM ve Boğaziçi University Life Science Institute opened a stand during the meeting.

Researchers from Yale, SUNY, Acibadem, Cerrahpasa, Hacettepe, and Bilkent Universities have collaborated to find the source of a genetic disorder that causes malformation in brain development. The study was published in Nature Genetics.


Ergin Atalar became Fellow of International Society of Magnetic Resonance in Medicine (ISMRM). Dr. Atalar's fellowship has been announced in the upcoming ISMRM meeting at Montreal in May 7, 2011. Dr. Atalar named Fellow for his contributions to the fields of Interventional MRI and Safety of Magnetic Resonance Imaging.

Manuscript entitled "Reverse Polarized Inductively Coupling to Transmit and Receive RF Coil Arrays" by Haydar Celik, Ibrahim Davut Mahcicek and Ergin Atalar have been accepted for publication in Magnetic Resonance in Medicine.