2023 – Present
Ph.D., Computational Neuroscience
Boston University · Graduate Program in Neuroscience
Matthew P. Szuromi
Doctoral Student · Theoretical & Computational Neuroscience
Boston, MA [email protected] mattszuromi.com
Research Statement
Seeking to apply deep tools from mathematics and theoretical physics to hard problems in computational and theoretical neuroscience — translating rigorous, first-principles thinking into practical tools and theories.
Education
2016 – 2021
B.S., Pure Mathematics & Physics
University of Michigan · College of Literature, Science & the Arts
Publications
Modelling dysfunction-specific interventions for seizure termination in epilepsy
Dynamotypes for Dummies: A Toolbox, Atlas, and Tutorial for Simulating a Comprehensive Range of Realistic Synthetic Seizures
Optimization of ictal aborting stimulation using the dynamotype taxonomy
Research Experience
2024 – Present
Ocker Lab — Boston University
Advisor: Dr. Gabriel Ocker
- Building a symbolic–numerical pipeline for automated Feynman-diagram evaluation in nonlinear stochastic differential equations, including a novel proof of a generic loop-diagram counting algorithm
- Developing a statistical field theory for cell-type-specific stochastic spiking neural networks with correlated inputs
- Modelling thalamo-cortical network development with spike-timing-dependent plasticity during critical periods for primary sensory cortices
- Studying stochastic state transitions in spiking networks with spike resets
2023 – Present
BRAIN Dynamics Lab — Johns Hopkins University
Advisors: Dr. Catherine Chu & Dr. Mark Kramer
- Developing a machine-learning approach to detect hippocampal sharp-wave ripples in human stereo-EEG recordings
- Studying the relation between hippocampal sharp-wave ripple activity and memory consolidation in humans
2023 – Present
Independent Research
- Developing low-dimensional neural mass models as practical analytic tools for studying multi-state nonlinear systems, with application to epilepsy and neural dysfunction
2024
Han Lab — Boston University
Advisor: Dr. Xue Han
- Developed an improved action-potential detector for in-vivo voltage-imaging datasets
2020 – 2023
Stacey Lab — University of Michigan
Advisor: Dr. William Stacey
- Developed theoretical understanding of inconsistent efficacy of stimulatory treatments in epilepsy
- Developed a model of seizure dynamics to understand E/I imbalance at seizure onset
2019 – 2020
Levin Lab — University of Michigan
Advisor: Dr. Daniel Levin
- Determined properties of photomultiplier tubes for use in particle detectors
Skills
Mathematical
- Nonlinear dynamics
- Bifurcation theory, timescale separation, coarse-graining
- Stochastic dynamics
- SDEs, point processes, statistical field theory (path-integral methods)
- Graph theory
- Graph-enumeration algorithms
- Analysis
- Real & functional analysis (Volterra-kernel representations of local learning rules)
- Statistics & data
- Time-series analysis, spectral estimation, GLMs, HMMs
Computational
- Programming
- Python, MATLAB, Mathematica, C++, LaTeX, Linux/Unix
- Libraries
- NumPy, SciPy, SymPy, Matplotlib, PyTorch
- Packages
- NEST, bdtoolbox, MatCont, FOOOF, CERN ROOT
Presentations & Lectures
2025
Controlling and Probing Seizure Dynamotypes
Minisymposium · Computational Neuroscience (CNS 2025)
2023
Optimization of Ictal Aborting Stimulation
Minisymposium · SIAM Dynamical Systems (DS23)
2022
Optimization of Ictal Aborting Stimulation
Selected talk · International Conference for Technology and Analysis of Seizures (ICTALS)
2022
Optimization of Ictal Aborting Stimulation
Invited talk · University of Michigan Neural Networks Journal Club
Teaching
2022
Private Tutor
- Taught Linear Algebra (proofs & applications) one-on-one online; developed practice problems
2019
Tutor — ACE Tutor Zone, Ann Arbor, MI
- Taught Calculus I/II, elementary mechanics, AP Calculus & Physics, and SAT/ACT Math to individuals and small groups
2018 – 2019
Undergraduate Tutor — University of Michigan
- Tutored Calculus I/II and elementary mechanics one-on-one
References
Available upon request.