This K99/R00 application supports additional research training in computational mathematics and computer vision which will enable Dr. Madhusudhanan Balasubramanian-the applicant, to become an independent multidisciplinary investigator in computational ophthalmology. Specifically, in the K99 training phase of this grant, Dr. Balasubramanian will train at UC San Diego under the direction of Linda Zangwill PhD, an established glaucoma clinical researcher in the Department of Ophthalmology, as well as a team of co- mentors, including, Dr. Michael Holst from the Department of Mathematics and co-director for the Center for Computational Mathematics, and co-director of the Comptutational Science, Mathematics and Engineering and Dr. David Kriegman from Computer Science and Engineering. Training will be conducted via formal coursework, hands-on lab training, mentored research, progress review by an advisory committee, visiting collaborating researchers and regular attendance at seminars and workshops. The subsequent R00 independent research phase involves applying Dr. Balasubramanian's newly acquired computational techniques to the difficult task of identifying glaucomatous change over time from optical images of the optic nerve head and retinal nerve fiber layer. A documented presence of progressive optic neuropathy is the best gold standard currently available for glaucoma diagnosis. Confocal Scanning Laser Ophthalmoscope (CSLO) and Spectral Domain Optical Coherence Tomography (SD-OCT) are two of the optical imaging instruments available for monitoring the optic nerve head health in glaucoma diagnosis and management. Currently, several statistical and computational techniques are available for detecting localized glaucomatous changes from the CSLO exams. SD-OCT is a new generation ophthalmic imaging instrument based on the principle of optical interferometry. In contrast to the CSLO technology, SDOCT can resolve retinal layers from the internal limiting membrane (ILM) through the Bruch's membrane and can capture the 3-D architecture of the optic nerve head at a very high resolution. These high-resolution, high-dimensional volume scans introduce a new level of data complexity not seen in glaucoma progression analysis before and therefore, powerful (high-performance) computational techniques are required to fully utilize the high precision retinal measurements for glaucoma diagnosis. The central focus of this application in the K99 mentored phase of the application will be in 1) developing computational and statistical techniques for detecting structural glaucomatous changes in various retinal layers from the SDOCT scans, and 2) developing a new avenue of research in glaucoma management where in strain in retinal layers will be estimated non-invasively to characterize glaucomatous progression. In the R00 independent phase, the specific aims focus on developing 1) statistical and computational techniques for detecting volumetric glaucomatous change over time using 3-D SD-OCT volume scans and 2) a computational framework to estimate full-field 3-D volumetric strain from the standard SD-OCT scans.
Detecting the onset and progression of glaucomatous changes in the eye is central to glaucoma diagnosis and management. This multidisciplinary project focuses on developing powerful (high-performance) computational, mathematical, and statistical techniques for detecting volumetric glaucomatous changes from the Confocal Scanning Laser Ophthalmoscopy (CSLO) scans and volumetric Spectral Domain Optical Coherence Tomography (SD-OCT) scans of the optic nerve head. In addition, the Principal Investigator of this proposal will receive extensive training in the areas of computational mathematics and computer vision to augment and strengthen his multidisciplinary expertise essential to execute the proposed specific aims.
|Yousefi, Siamak; Balasubramanian, Madhusudhanan; Goldbaum, Michael H et al. (2016) Unsupervised Gaussian Mixture-Model With Expectation Maximization for Detecting Glaucomatous Progression in Standard Automated Perimetry Visual Fields. Transl Vis Sci Technol 5:2|
|Balasubramanian, Madhusudhanan; Arias-Castro, Ery; Medeiros, Felipe A et al. (2014) Detecting glaucoma progression from localized rates of retinal changes in parametric and nonparametric statistical framework with type I error control. Invest Ophthalmol Vis Sci 55:1684-95|
|Belghith, Akram; Balasubramanian, Madhusudhanan; Bowd, Christopher et al. (2014) A unified framework for glaucoma progression detection using Heidelberg Retina Tomograph images. Comput Med Imaging Graph 38:411-20|
|Bowd, Christopher; Weinreb, Robert N; Balasubramanian, Madhusudhanan et al. (2014) Glaucomatous patterns in Frequency Doubling Technology (FDT) perimetry data identified by unsupervised machine learning classifiers. PLoS One 9:e85941|
|Belghith, Akram; Bowd, Christopher; Medeiros, Felipe A et al. (2014) Glaucoma progression detection using nonlocal Markov random field prior. J Med Imaging (Bellingham) 1:034504|
|Yousefi, Siamak; Goldbaum, Michael H; Balasubramanian, Madhusudhanan et al. (2014) Learning from data: recognizing glaucomatous defect patterns and detecting progression from visual field measurements. IEEE Trans Biomed Eng 61:2112-24|
|Yousefi, Siamak; Goldbaum, Michael H; Balasubramanian, Madhusudhanan et al. (2014) Glaucoma progression detection using structural retinal nerve fiber layer measurements and functional visual field points. IEEE Trans Biomed Eng 61:1143-54|
|Balasubramanian, Madhusudhanan; Kriegman, David J; Bowd, Christopher et al. (2012) Localized glaucomatous change detection within the proper orthogonal decomposition framework. Invest Ophthalmol Vis Sci 53:3615-28|
|Bowd, Christopher; Lee, Intae; Goldbaum, Michael H et al. (2012) Predicting glaucomatous progression in glaucoma suspect eyes using relevance vector machine classifiers for combined structural and functional measurements. Invest Ophthalmol Vis Sci 53:2382-9|