Harvard Catalyst Profiles

My research focuses on MRI (and to some extent PET) technology development and translation to neuro-imaging to help better understand the human brain. I have several areas of specializations:

(i) Modeling and optimization of radio-frequency (RF) and gradient MR sub-systems. My postdoctoral work focused on modeling of RF coils, especially parallel transmission (pTx) coils whereby nuclear spins are excited using several transmit channels as opposed to a single one (transmit phased array). I also developed methodology for the design of pTx RF pulses at low and large flip-angle with explicit control for the specific absorption rate, which is the main safety limitation in high duty-cycle MRI sequences. This work is relevant at 3 Tesla, 7 Tesla and will be crucial to the progress of MRI at 11.7 Tesla and higher. Recently, I have been working together with Prof. Lawrence Wald and Mathias Davids on a modeling and design methodology of gradient coils with an intrinsically lower propensity to create peripheral nerve stimulation (PNS). PNS is the main limitation of MRI acquisition speed in fast imaging sequences such as echo planar imaging and turbo spin echo, therefore such new gradient technology has the potential to dramatically improve speed and data acquisition rate in many clinical and research (fMRI, diffusion imaging) applications.

(ii) Neuro-imaging of patients with deep brain stimulation (DBS) implants. Generally speaking, I am interested in the combined use of neurostimulation strategies and functional imaging methods such as fMRI. Neurostimulation using DBS allows pin-point excitation of the human brain in a very controlled manner and, when combined with fMRI, offers unique brain mapping possibilities. DBS is in itself a fascinating therapy that is poorly understood and holds promise for the treatment of many psychiatric disorders. I use fMRI to study the mechanisms of action of DBS in patients with Parkinson disease and depression. In addition, I study the safety aspects of using DBS in the MR RF environment using electromagnetic modeling.

(iii) Iterative reconstruction, quantitative corrections and physics of positron emission tomography (PET). My interest in this area stems from my PhD work on fully 3D, list-mode reconstruction of PET data with corrections for scatter coincidences and non-rigid respiratory and cardiac motion. A central innovation of my work was to incorporate the information carried by the energy of the detected PET photons into the scatter correction process using a statistical model of the coincidence energy blurring incorporated in the likelihood function of the list-mode PET data. Another innovation was the use of tagged-MRI to track complex non-rigid motions due to respiration and cardiac beating and the incorporation of these motion vector fields in the 3D list-mode PET reconstruction.