





Iain P. Bruce, PhD
Ph.D. MU 

Iain P. Bruce, Ph.D.
Ph.D. in Computational Sciences
Department of Mathematics, Statistics, and Computer Science
Marquette University
Research Assistant in Functional Magnetic Resonance Image Analysis Lab
357 Cudahy Hall
Marquette University
Milwaukee, WI 532011881
Phone: 4142886347
Email: iain.bruce{at}marquette.edu
Curriculum Vitae 
Education
2014  present Postdoctoral Fellow. Brain Image and Analysis Center, Duke University, Durham, NC
2009  2014 Ph.D. in Computational Sciences, Marquette University, Milwaukee, WI
 Advisor: Dr. Daniel B. Rowe
 Dissertation: Determination of Correlations Induced by the SENSE and GRAPPA pMRI Models with an Application to MRI RF Coil Design.
2009  2011 M.S. in Computational Sciences, Marquette University, Milwaukee, WI
 Advisor: Dr. Daniel B. Rowe
 Thesis: A Statistical Examination of the SENSE Reconstruction via an Isomorphism Representation.
2007  2008 M.Sc. in Advanced Mechanical Engineering (with Merit), Imperial College London, UK
 Advisor: Dr. Andy Hayes
 Thesis: Aircraft Wake Development
20032007 B.S. in Physics (with Honors), Principia College, Elsah, IL
20032007 B.S. in Mathematics (with Honors), Principia College, Elsah, IL
Experience
Teaching Experience
Department of Mathematics, Statistics, and Computer Science, Marquette University, Milwaukee, WI.
Instructor:
• 01/2014 – 05/2014: Modern Elementary Statistics
• 01/2013 – 05/2013: Modern Elementary Statistics
Teaching Assistant:
• 08/2013 – 12/2013: Modern Elementary Statistics & Theory of Probability
• 08/2012 – 12/2012: Modern Elementary Statistics & Theory of Probability
• 08/2011 – 12/2011: Modern Elementary Statistics
• 08/2010 – 05/2011: Modern Elementary Statistics
Research Experience
Department of Mathematics, Statistics, and Computer Science, Marquette University, Milwaukee, WI.
• 06/2013 – 08/2013: Computational Sciences Summer Research Program (CSSRP)
• 06/2012 – 08/2012: Research Assistant, Marquette Regular Research Grant (Dr. Daniel B. Rowe)
• 01/2012 – 05/2012: Research Assistant
• 06/2011 – 08/2011: Computational Sciences Summer Research Program (CSSRP)
• 06/2010 – 08/2010: Computational Sciences Summer Research Program (CSSRP)
• 09/2009 – 05/2010: Research Assistant
Honors / Affiliations
 International Society for Magnetic Resonance in Medicine, 2012Present
 American Statistical Association, 20102014
 Sigma Pi Sigma – The Physics Honors Society, inducted in 2006
Research
 Quantifying the statistical implications of the SENSE and GRAPPA parallel MRI reconstruction models in fcMRI data.
An MRI scanner’s inability to instantly acquire the spatial frequency spectrum of an object places constraints on both the spatial and temporal resolution achievable in data acquired for functional connectivity MRI (fcMRI) studies. To accelerate data acquisition, parallel MRI models such as SENSE and GRAPPA use multiple receiver coils placed around the object to acquire subsampled arrays of spatial frequencies. This is done by omitting rows of frequency measurements in the phase encoding direction, and results in inverse Fourier reconstructed coil images that appear to be folded over on themselves. Parallel MRI models exploit the overlap of coil magnetic field sensitivity profiles to unfold the folded coil images through spatial localization. In the SENSE model, the folded coil images are unfolded and combined into a single image through a complexvalued weighted least squares estimation, while the GRAPPA model performs an interpolation of missing spatial frequency values prior to the inverse Fourier reconstruction. A previously unexplored consequence of the unfolding process performed by either model is that an artificial correlation of no biological origin is induced between the regions of the object that were previously folded on one another. To observe these correlations, I have developed a linear isomorphism that represents each model in terms of a series of matrix operators. With the models represented in this fashion, a mathematically equivalent reconstruction can be performed on the acquired data, and the degree to which the mean and covariance of the acquired data is changed can be quantified both precisely and directly. As the estimation of correlations between various brain regions is the mechanism for determining functional connectivity in the brain, I have verified that the artificial correlations induced by the SENSE and GRAPPA models can corrupt the conclusions drawn in fcMRI studies, where regions of the brain can appear to be either correlated or uncorrelated when they are not. My current research therefore aims to develop new parallel MRI models that can reap the benefits of accelerated acquisition schemes without suffering from the statistical consequences that the reconstruction models impose.
 Aircraft Wake Development: A theoretical analysis of the three dimensional vortex sheet
Publications
Peer  Reviewed Journal Publications
 Karaman, MM, Bruce, IP, Rowe, DB. Incorporating relaxivities to more accurately reconstruct MR images. Magn. Reson. Imaging, 33(4):374384, 2015.
 Bruce IP, Rowe DB.Quantifying the Statistical Impact of GRAPPA in fcMRI Data with a RealValued Isomorphism.IEEE Transactions on Medical Imaging 33(2): 495503, (2014).
 Karaman, MM, Bruce, IP, Rowe, DB. A Statistical fMRI Model for Differential T2* Contrast Incorporating T1 and T2* of Grey Matter.Magnetic Resonance Imaging 32(1): 927, (2014).
 Bruce IP, Karaman MM, Rowe DB. The SENSEIsomorphism Theoretical Image Voxel Estimation (SENSEITIVE) Model for Reconstruction and Observing Statistical Properties of Reconstruction Operators. Magnetic Resonance Imaging 30(8): 11431166 (2012).
 Bruce IP, Karaman MM, Rowe DB. A Statistical Examination of SENSE Image Reconstruction via an Isomorphism Representation. Magnetic Resonance Imaging 29(9): 12671287 (2011).
Abstracts
 Bruce IP, Muftuler LT, Rowe DB. Spatial Normalization Can Morph RF Coils into Brain Region Optimized Geometries for fcMRI Studies. Proc. Intl. Soc. Magn. Reson. Med. 22:4144, (2014).
 Bruce IP, Muftuler LT, Rowe DB. SENSE Induced Correlations are used to Optimize RF Coil Design for Specific fcMRI Studies. Proc. Intl. Soc. Mag. Reson. Med. 21:2733, (2013).
 Bruce IP, Karaman, MM, Rowe DB. Artificial Correlations Induced by SENSE and GRAPPA Corrupt fcMRI Conclusions. Proc. Intl. Soc. Mag. Reson. Med. 21:2229, (2013).
 Karaman, MM, Bruce, IP, Rowe, DB. Incorporation of Grey Matter T1 and T2* Improves Brain Activation Statistics in fMRI. Proc. Intl. Soc. Mag. Reson. Med. 21:2285, (2013).
 Rowe DB, Bruce IP: Processing Induced Voxel Correlation in SENSE FMRI Via the AMMUST Framework. Proc. Second Biennial International Conference on Resting State Connectivity, Medical College of Wisconsin, Milwaukee, Wisconsin, F052, 2010. http://www.restingstate.com/
Conferences and Invited Talks
 Bruce, IP: Using the Gfactor and SENSE Induced Correlations for Optimizing fcMRI Study Specific RF Coil Design. Workshop on Medical Image Analysis, University of Wisconsin, Madison, WI, 2013. http://brainimaging.waisman.wisc.edu/
 Bruce IP. Observing Spatial Correlations Induced by the SENSE and GRAPPA Parallel MRI Image Reconstruction Models Using an Isomorphic Framework. Workshop on Brain Image Analysis, University of Wisconsin, Madison, Wisconsin, 2012. http://brainimaging.waisman.wisc.edu/
 Bruce, IP: A Statistical Investigation of the ComplexValued GRAPPA Parallel MRI Reconstruction Model through a RealValued Isomorphic Representation. Department of Mathematics, Statistics and Computer Science, Marquette University, Milwaukee, WI, 2011.
 Rowe DB, Bruce IP: Processing Induced Voxel Correlation in SENSE FMRI Via the AMMUST Framework. Proc. Second Biennial International Conference on Resting State Connectivity, Medical College of Wisconsin, Milwaukee, Wisconsin, 2010. http://www.restingstate.com/
 Bruce, IP: An Adjustment to SENSE Image Reconstruction through a SENSE Isomorphism Theoretical Image Voxel Estimation (SENSEITIVE) Model: Statistical Properties and Implications. Department of Mathematics, Statistics and Computer Science, Marquette University, Milwaukee, WI, 2010.
 Bruce IP: The SENSEIsomorphism Theoretical Image Voxel Estimation (SENSEITIVE) Model for Reconstruction and Observing Statistical Properties of Reconstruction Operators. Department of Biophysics, Medical College of Wisconsin, Milwaukee, WI, 2010.


