Welcome to my corner of the internet! I am a second year PhD Physics student at Yale University interested in the neural circuits of vision in Drosophila. My interests stemmed from first studying the biophysics of neuron membranes. My scope changed to a more macroscopic scale when I began to study the whole brain as a complex system. Having dabbled at the scales of single-cell and whole brain, I found a happy balance at the circuit level where I try to uncover how networks of neurons perform algorithms. Inspired by the field of computer vision, I dedicated my PhD to studying the neural circuits of vision.
Supervisor: Damon Clark
Supervisor: Maged Goubran
· Collaboration with Li Ye (Scripps Research) and Ailey Crow (Stanford) to apply modern
deep-learning techniques for single-axon tracing in light microscopy for multimodal image registration with
diffusion MRI and connectivity analysis
· Developed and implemented pipelines for data preparation, labelling, training, and inference on local and
remote Compute Canada high-performance computing hardware
· Tested different deep-learning architectures including U-net and flood filling networks
Volunteer position conducted remotely due to COVID-19
Supervisor: Siew Ann Cheong and Jörn Davidsen
· Initiated collaborations between the Complexity Institute and the Complexity Science Group studying systematic risks of financial portfolio overlap and criticality in the stock market
· Developed and implemented numerical simulations modelling financial portfolios on a bipartite network with
varying degrees of overlap and unique trading dynamics to study cascades of portfolio re-balancing
· Analyzed simulated and real-world data from a statistical physics point of view
Funded by the University of Calgary in lieu of the Mitacs Globalink Research Award due to COVID-19. This work resulted in an oral presentation at a national conference and is continued towards an Honours Thesis.
Supervisor: Jörn Davidsen
· Demonstrated criticality on various real biological neural networks while undergoing the branching process
with spontaneous activity
· Implemented numerical simulations and applied statistical data analysis and visualization to classify the
resulting dynamics into specific universality classes
· Implemented high-performance computers to parse diffusion MRI data for generating networks with up to 14
million nodes.
This work resulted in 2 presentations at regional conferences and was funded by the Natural Sciences and
Engineering Research Council of Canada Undergraduate Student Research Award
Supervisor: Helge Ewers and Jia Hui Li
· Developed a new method combining directed manipulation of membrane proteins by fluorescent magnetic
nanoparticles with super-resolution microscopy to investigate nanoscale biological structures
· Demonstrated that neuronal actin forms evenly spaced ring-like structures following a periodicity of ∼180-190
nm along axonal shafts using light microscopy
· Developed analysis code as well as a deep-learning pipeline using MASK-RCNN to identify cellular structures
within images
This work resulted in 1 manuscript in Nature Communications, 1 conference proceedings paper, as well
as 4 presentations at international, national, and local conferences. This work was funded by the Deutscher
Akademischer Austauschdienst and Mitacs
Supervisor: Pina Collarusso and Rima-Marie Wazen
· Developed pipelines for computational super-resolution microscopy image reconstruction and analysis via
traditional and deep-learning methods
· Demonstrated that human telomeres form t-loop like structures at a nanometer scale using light microscopy
This work resulted in 1 presentation at a local symposium and was funded by Canada Summer Jobs.
Part of Yale's Integrated Graduate Program in Physical and Engineering Biology program
Thesis: Modelling overlapping portfolio networks and re-balancing cascades in the stock market
Expected First Class Honours
· Statistical Mechanics I/II (A+/A+)
· Computational Physics I/II/III (B+/A/TBA)
· Electromagnetic Theory I/II/III (A+/A/A)
· Mathematical Methods in Physics (A)
· Solid State Physics (A+)
· Nonlinear Dynamics and Chaos (TBA)
4.08/4.30 GPA
Lots of fun
Overlapping portfolios create hidden systematic risks which are not addressed in Modern Portfolio Theory. We are working on addressing this risk through a dynamic complex network perspective. (Image from Delpini et al. 2019)
View intro presentationThe brain is an extremely complex system. We do not know exactly how it functions, but expirements have shown evidence of critical behaviour similar to that of a continuous phase transition. To study this computationally, we apply the branching process with spontaneous activity on real biological neural networks obtained via diffusion MRI. Preliminary results show that criticality with spontaneous activity is characterized by a transition from directed percolation to undirected percolation universality classes due to the merging of neuronal avalanches. (Image from Tagliazucchi et al. 2012)
View Project PosterBiological neural networks of the brain and their alterations are a heavily researched topic in the field neurological disorders. Multimodal Image Registration And Connectivity anaLysis (MIRACL) is an automated pipeline that combines brain maps from various methods such as microscopy and diffusion MRI. To improve on the MIRACL pipeline, we are working on axon tracing in 3D CLARITY light microscopy images using modern deep-learning methods. (Goubran et al. 2019)
View Paper by Dr. GoubranMembrane diffusion barriers are required in maintaining cellular identity. For example, the axon and dendrites of a neuron perform drastically different tasks and thus will contain different biochemicals. Therefore, a diffusion barrier must exist between the two to maintain their distinct identities. Loss of this identity has been shown to correspond with ailments such as bipolar disorder. To study these barriers, we developed a novel method using fluorescent magnetic nanoparticles. (Li et al. 2019)
View paper in Nature Communications