Graham Smith

Writer & Computational Neuroscientist

Education

Ph.D. in Computational Neuroscience

The University of Chicago • 2016–2022

GPA: 3.98/4.0; Pritzker Fellow in Neuroscience, 2016-17 (best performing first year in program; $100k). Dissertation: Modeling and Characterizing Population-Scale Neural Activity, with Applications to Epilepsy

S.B. in Mathematics

The University of Chicago • 2011–2015

GPA: 3.63/4.0; National Merit Scholar; Dean's List, 2011-2014; General Honors

Research Experience

Ph.D Candidate

The University of Chicago • 2016–2022

  • Dissertation: Modeling and Characterizing Population-Scale Neural Activity, with Applications to Epilepsy
  • Developed unique characterization of network activity (triple correlation)
  • Detected seizures in newborn infants using triple correlation (publication pending)
  • Proved necessity of non-monotonic population input response in focal seizures (publication in review)
  • Synthesized ongoing study results for half-hour talks to 70+ faculty/students annually
  • Published first-author editorial on deep learning in clinical seizure detection

Neural Networks Research, Ph.D Rotations

The University of Chicago • 2015–2016

  • Integrated structural learning into spiking neural networks
  • Programmed recurrent neural network to capture high-order statistics of natural visual world

Research Assistant

The University of Chicago • 2012–2015

  • Trained a monkey to control a robotic arm via brain implant
  • Analyzed neural timeseries in MATLAB/Python (first-author conference paper)
  • Systematized organization of terabytes of data for collaborative access

Publications

  • Smith, G.A., Cowan, J.D., Palmer, S.E., & van Drongelen, W. Failure of inhibition provides transition from population-scale computation to focal seizure. (in review).
  • Smith, G.A., Henry, J., van Drongelen, W. Detecting heterogeneous seizures in newborn infants using triple correlation. (in prep).
  • Deshpande, S.S.*, Smith, G.A.*, & van Drongelen, W. Third-order motifs... Sci Rep 13, 238 (2023).
  • Smith, G., Lee, S., & van Drongelen, W. Can we use machines to interpret EEG? Clinical Neurophysiology (2018).
  • Smith, G., Balasubramanian, K., et al. Pharmacologically Induced Laminar Oscillations... EMBC, IEEE (2015).

Teaching Experience

Tutor

Aurelian Learning (owner) • 2022–present

  • 750+ hours private tutoring (high school through professionals)
  • Programming, mathematics, technical writing
  • Consistent 5.0/5.0 rating (75+ students)

Teaching Assistant

The University of Chicago • 2016–2020

  • Computational Approaches to Cognitive Neuroscience
  • qBio quantitative bootcamp (2016–2020, Marine Biological Lab)
  • Mathematical Modeling of Large-Scale Brain Dynamics
  • Topics in Data Analysis in Biomedical Research: Big Data
  • Multimedia Programming as Interdisciplinary Art
  • Led and redesigned programming reproducibility tutorials, moved to Zoom/Discord

Tutor and Tech

Computer Science Instructional Laboratories • 2013–2015

  • Helped students/faculty with Linux/Apple OS & applications
  • Taught UNIX, LaTeX, MATLAB, R minicourses
  • Designed/improved scheduling program
  • Used fabmin/Ansible to manage lab computers and configuration
  • Transitioned lab between buildings

Presentations

  • Unique Characterization of Spatiotemporal Neuronal Network Activity – University of Twente (2022)
  • Triple Correlation Uniquely Characterizes Neural Activity – UChicago (2022)
  • Focal Seizures from Failure of Inhibition – UChicago (2021)
  • Wave Propagation in Wilson-Cowan Model – UChicago (2020)
  • Traveling Waves (various talks and retreats, 2017–2019)
  • Analysis of Neural Field Dynamics – UChicago (2018)
  • Multiple talks on neural modeling, structural plasticity, and generative models, 2015–2018

Leadership & Teamwork

Treasurer

Dean's Council of Biological Sciences Division • 2017–2019

Accounted for $50k annual budget across 20 programs

Appointee

University-wide Standing Committee on Disruptive Conduct • 2018–2019

Selected to sit on committee hearing disciplinary cases

Singer

CMAC: UChicago Glee Club • 2016–2022

Three concerts and tour annually

Singer

Chicago Chorale • 2018–2022

Three concerts annually; tour to Spain (2022); Classical repertoire: Mozart, Bach, Rautavaara, and contemporary works

President

CMAC Glee Club • 2016–2018

Fostered positive culture/events (50% membership growth); Raised $15k donation; organized two cross-country tours

Programming

Languages: Julia, Python, Mathematica, C, Java, R, MATLAB, Haskell

Paradigms: Multiple dispatch, functional, OOP, array, parallel

Packages: NumPy, MATLAB, SciPy, Tensorflow, signal processing, differential equations, optimization, unit testing (PyTest)

Operating Systems: Ubuntu (2009–present), Arch (sporadically 2011–2019; consistently 2020–present)

Interests

Singing, running, climbing, reading & writing fantasy, guitar, tennis, tea