Alex
D. Leow, MD Ph.D
I am currently with Neuropsychiatric hospital at UCLA (psychiatry is my
practice). However, my main ongoing research activities are through Laboratory
of NeuroImaging.
My research
My research mainly consists of two branches: application and theory. Though practicing medicine, I do
consider myself more of a theoretically/mathematically inclined person. On the
theory front, my main research interests, at this moment, are the following:
·
The statistical analyses of group deformation fields as
well as their applications to brain imaging in neuropsychiatric diseases such
as, but not limited to, Alzheimer's disease, schizophrenia, autism, and bipolar
disorder (please refer to list of publications for recent papers).
·
The development of novel non-linear registration
frameworks incorporating mathematical techniques in continuum mechanics,
differential geometry, as well as information theory.
·
The quantification of diffusion tensor images, both
within subject and across groups. This would help visualize the configuration
of white matter myelination, both longitudinally and cross-sectionally, leading
to better understandings of the human brain.
In short and lay person term, I am trying to figure out how one we judge
different non-linear registration methods. Mostly recently, I have been
developing a novel framework, based on information theory, which would provide
a statistically correct framework for analyzing Jacobian maps ( Unbiased
Large-Deformation Non-linear Image Registration; refer to IPMI
05, the MICCAI workshop paper below, and 07 IEEE TMI paper. For a short
description of this idea, refer to UCLA CAM report 06-64 that can be found here
http://www.math.ucla.edu/applied/cam/).
Another new field that I have worked on is diffusion tensor imaging (DTI), q-ball imaging, and HARDI, an exciting field. In particular, I have looked into multi-tensor
estimation, and developed new methodology -the computation of Tensor
Distribution Function-
to directly estimate the particle displacement function. Orientation diffusion
tensor, and dominant fiber tract directions can be easily derived from TDF
through analytic relations (see http://www.math.ucla.edu/applied/cam/
for a technical report on the concept of Tensor Distribution Function).
I have been a regular reviewer, for IEEE Transactions on Medical
Imaging, IEEE Transactions on Image Processing, NeuroImage, as well as other
journals. I also served as a reviewer for various conferences/workshops
including the 2007 ISBI (International Symposium on Biomedical Imaging), and
the MICCAI 2007 workshop on statistical registration: pair-wise and group-wise
alignment and atlas formation (http://www.miccai2007.org/workshops/)
My Main collaborators
I work extensively with Prof. Paul Thompson, as well as some of the most
amazing graduate students/research assistants in our group: Agatha Lee,
Allen Lu, Andrea Klunder, and Kiralee Hayashi (I am eternally grateful to these people). I
consider them the nicest, smartest, most mature, and most
efficient people in
the world. As far as mathematical theory/algorithm design goes, I also work
with Dr. Ming-Chang Chiang, who I think is the hardest-working person I've ever met.
Most recently, I have also mentored Igor Yanovsky, at the Mathematics department of
UCLA, on fine-tuning algorithm implementation as well as testing of the
proposed Unbiased registration framework. The goal is to vigorously establish
and test theories leading to unbiased statistics when making inference of
Jacobian operators resulting from deformation maps. For some of our work on the
Unbiased Large Deformation Registration, please see his website here: http://www.math.ucla.edu/~yanovsky/.
Here are some images from our most recent work, showing that
ULDIR outperforms regular fluid registration, even on a visual level. The
images in this figure are part of the Alzheimer’s Disease Neuroimaging
Initiative, where normal elderly subjects were scanned twice within a two-week
time period. As a result, these images are used for calibration purposes (i.e.,
no structural changes are present)
Click here if you can’t
see the image
This figure examines the inverse consistency of two different deformation models in tensor-based morphometry. Products of Jacobian maps generated using fluid registration and ULDIR for forward direction (time 1 to time 2) and backward direction (time 2 to time 1) are shown. Both the fluid and ULDIR registrations are driven by L2 intensity matching functional. Jacobian maps are superimposed on the deformed volumes. Rows depict slices in axial, sagittal, and coronal planes. For the ULDIR method, the visualization of the product of Jacobian maps is less noisy, with values closer to 1, showing better inverse consistency.
Hobbies
PIANO- a hobby shared by both
Paul and I. Although the similarity ends here!! Paul likes piano music with
extreme technical difficulties (Rachmaninoff,
some Chopin, and Liszt), while I prefer intimate music, like some other Chopin
(mutually exclusive!!), Ravel, Mozart (mainly piano concertos).
In October of 2006, my dearest friend Dr. Clayton Bullock MD PhD (a great violinist
and harpsichord-ist) and I had formally formed a group, the Duo
Doctor Duo, and have
since given several recitals. We are currently working very hard on our
repertoire, which has included music by Handel, Mozart, Brahms, various violin
transcriptions, as well as piano solo by Chopin, Bach, Rachmaninoff, Ravel. We
recently did the first of the 2007 concert series at the
A big goal in my life is to own one grand piano from the following brands: Hamburg
Steinway (the new NY
Steinway’s really not so good), Fazioli (Ferarri of pianos) (for people
who never heard of it, link is http://www.fazioli.com/,
Bluthner
(another romantic sounding instrument, link here http://www.bluthnerpiano.com/ ), Mason
and Hamlin (the old
company actually kind of went out of bzness, but the new ones are reported to
be made using the same specs. Link here http://www.masonhamlin.com/
. Mason and Hamlin probably, to my knowledge, makes pianos with the most rich
overtone, reminding me of the good old days, turn-of-the-century-ish. I highly
recommend any piano lover to give it a try.), and
TENNIS- I used to be a very
very avid tennis player.
Contact me at alexfeuillet@gmail.com if you have any questions or comments
about my research. Thanks for visiting my homepage.
Publications
(submitted papers, preprints,
or technical reports not included):
1. W. Liao,
K. Lange, M. Bergsneider and S.C. Huang (2002) Optimal design in dynamic PET
data acquisition: A new approach using simulated annealing and component-wise
metropolis updating. IEEE Transactions on Nuclear Science, 49(5) pp. 2291-2296.
2. A.D. Leow,
C.L. Yu, S.J. Lee, S.C. Huang, R. Nicolson, K.M. Hayashi, H. Protas, A.W. Toga,
and P.M. Thompson (2004) Brain Structural Mapping using a Novel Hybrid
Implicit/Explicit Framework based on the Level-Set Method. NeuroImage 24(3) pp.
910-927
3.N. Hattori, S.C. Huang, H.M. Wu, W. Liao, T. Glenn, P. Vespa, M. Phelps, D. Hovda, M.
Bergsneider (2004) Acute changes in regional cerebral FDG kinetics in patients
with traumatic brain injury [In press]. Journal of Nuclear Medicine, 2004.
4. E.F.Lee, R. Jacobs, I. Dinov, A.D. Leow, A.W. Toga (2005) standard atlas space for C57
BL/6J neonatal mouse brain. Anatomy and Embryology NOV, 210(4): 245-63, 2005
5.Leow AD,
Klunder AD, Jack CR, Toga AW, Dale AM, Bernstein MA, Britson PJ, Gunter JL,
Ward CP, Whitwell JL, Borowski B, Fleisher A, Fox NC, Harvey D, Kornak J,
Schuff N, Studholme C, Alexander GE, Weiner MW, Thompson PM; For the ADNI
Preparatory Phase Study (2005). Longitudinal Stability of MRI for Mapping Brain
Change using Tensor-Based Morphometry, NeuroImage 2006; 31:627-40
6.AD Lee, AD Leow, A Lu, A. Reiss, S. Hall, M-C. Chiang, A. Toga, P. Thompson (2006) 3D
pattern of brain abnormalities in fragile X syndrome visualized using
tensor-based morphometry. Neuroimage (in press)
7 AD Leow,
M-C. Chiang, I. Yanovsky, AD Lee, A. Klunder, J. Becker, S. Davis, A. Toga, P.
Thompson (2006) Statistical properties of Jacobian maps and the realization of
unbiased large-deformation non-linear image registration, to appear in IEEE
Transactions on Medical Imaging.
8. A.D. Leow,
S.C. Huang, A. Geng, J. Becker, S. Davis, A. Toga, and P. Thompson (2005) Inverse
consistent mapping in 3D deformable image registration: its construction and
statistical properties, IPMI 2005, G.E. Christensen and M. Sonka (Eds.):
Lecture Notes in Computer Science 3565, pp. 493-503, 2005
9. W. Liao,
L. Vese, S.C. Huang, M. Bergsneider, and S. Osher (2003) Computational anatomy
and implicit object representation: a level set approach, Lecture Notes in
Computer Science, vol. 2717, pp. 40-49.
10. S.C. Huang, W. Liao,
11. W. Liao,
H. Protas, M. Bergsneider, L. Vese, S.C. Huang, and S. Osher (2003) A new
framework of object warping: a semi-Lagrangian level set approach. International
Conference on Image Processing, Sep. 2003.
12. A.D. Leow, P. Thompson, H. Protas, and S.C. Huang (2004) Brain warping with
implicit representation. International Symposium on Biomedical Imaging
ISBI2004.
13. A.D. Leow, M.C. Chiang, P. Thompson, and S.C. Huang (2004) Linear and non-linear
geometric object matching with implicit representations. International
Conference on Pattern Recognition ICPR04.
14. A.D. Leow, M.C. Chiang, S.C. Huang, A.W. Toga, and P.M. Thompson, Realizing
unbiased deformation: a theoretical consideration. In Mathematical foundations
of computational anatomy workshop, MICCAI 2006.
15.N. Lepore, A.D. Leow, P.M. Thompson (2006) Landmark Matching on the
Sphere Using Distance Functions, IEEE International Symposium on Biomedical
Imaging (ISBI2006), April 6-9 2006.
16. I Yanovsky, MC
Chiang, P Thompson, A Klunder, J
Becker, S Davis, A Toga, A Leow (20007) QUANTIFYING
DEFORMATION USING INFORMATION THEORY: THE LOG-UNBIASED NONLINEAR REGISTRATION.
ISBI 2007.
17. Igor Yanovsky, Paul Thompson,
18. Igor Yanovsky, Paul Thompson, Stanley
Osher, Luminita Vese, Alex Leow, Multiphase Segmentation of Deformation using
Logarithmic Priors, IEEE Computer Society Workshop on Image Registration and
Fusion, 2007.
19. Igor Yanovsky,
20. P.M. Thompson, E.R.
Sowell, N. Gogtay, J.N. Giedd, C.N. Vidal, K.M. Hayashi, A.D. Leow, R. Nicolson, J.L. Rapoport, A.W. Toga (2005).
Structural MRI and Brain Development, in: Glabus M [ed.], International Review
of Neurobiology (IRN), 2005.
21 AD Leow, I. Yanovsky (2007) MICCAI workshop
paper (details coming soon)
Abstracts/Posters
Liao WH, Thompson PM, Hayashi KM, Bearden C,
Nicoletti MA, Monkul SE, Brambilla P, Sassi RB, Mallinger AG, Soares JC (2005).
Lithium Effects on Human Brain Structure Mapped Using Longitudinal MRI, 35th Annual Meeting of the Society for Neuroscience,
Liao WH, Thompson PM, Hayashi KM, Bearden C,
Nicoletti MA, Monkul SE, Brambilla P, Sassi RB, Mallinger AG, Soares JC (2005).
Lithium Effects on Human Brain Structure Mapped Using Longitudinal MRI,
A.D. Leow, A.D. Lee, M.C. Chiang, R.A.
Dutton, K.M. Hayashi, S.C. Huang, J.T. Becker, S.W. Davis, A.W. Toga, P.M.
Thompson (2005) Mapping Brain Atrophy in A Single Case of Semantic
Dementia Using Serial MRI with Inverse-Consistent Non-Rigid Registration. Human
Brain Mapping 2005.
A.D. Leow, S.C. Huang, J. Becker, and P.
Thompson (2005) Study of tissue atrophy involvement in semantic dementia using
elastic mapping. Brian 05’.
W.H. Liao, K. Lange, S.C. Huang (2001) The
Use of MM algorithm for regularization of parametric images in PET. BrainPET
2001.
W.H. Liao, L. Kashida, L. Vese, M.
Bergsneider, and S.C. Huang (2002) A variational approach for noise removal of
parametric images in tracer kinetic modeling. Neuroimage, 16(3), 2002, S68-S69 Part 2 Suppl. S.
A.D. Leow, H. Protas, S.C. Huang,
and P.M. Thompson (2004) Evaluation of a landmark curve matching technique
in brain mapping based on the level set method. Human Brain Mapping, 2004.
A.D. Leow, A. Lu, A.D. Lee, N.
Gogtay, J.L. Rapoport, A.W. Toga, and P.M. Thompson (2006) Tensor-Based Morphometry Reveals Significant Growth
Pattern Abnormalities in Childhood-Onset Schizophrenia. To be presented in
Society of biological Psychiatry,